Article

On January 1, 2021, the Current Procedural Terminology (CPT) evaluation and management (E/M) reporting rules changed dramatically, with “bullet counting” no longer necessary and the coding level now based on either the new medical decision making (MDM) table or time spent on all activities relating to the care of the patient on the day of the encounter.¹ This is described in the CPT Professional Edition 2023, a book every practitioner should review annually.² In particular, every provider should read and reread pages 1 to 14—and beyond if you provide services beyond standard office visits. These changes were made with the intent to simplify the process of documentation and allow a provider to spend more time with patients, though there is still a paucity of data related to whether the new system achieves these aims.

The general rule of reporting work with CPT codes can be simply stated—“Document what you did, do what you documented, and report that which is medically necessary” (David McCafferey, MD, personal communication)—and you should never have any difficulty with audits. Unfortunately, the new system does not let an auditor, who typically lacks a medical degree, audit effectively unless they have a clear understanding of diseases and their stages. Many medical societies, including the American Medical Association³ and American Academy of Dermatology,⁴ have provided education that focuses on how to report a given vignette, but specific examples of documentation with commentary are uncommon.

To make your documentation more likely to pass audits, explicitly link parts of your documentation to CPT MDM descriptors. We offer scenarios and tips. In part 1 of this series, we discuss how to approach the “spot check,” a commonly encountered chief concern (CC) within dermatology.

Scenario 1: A Funny-Looking New Spot

A 34-year-old presents with a new spot on the left cheek that seems to be growing and changing shape rapidly. You examine the patient and discuss treatment options. The documentation reads as follows:

• CC: New spot on left cheek that seems to be growing and changing shape rapidly.

• History: No family history of skin cancer; concerned about scarring, no blood thinner.

• Examination: Irregular tan to brown to black 8-mm macule. No lymphadenopathy.

• Impression: rule out melanoma.

• Plan: Consent, biopsy via shave technique. Lidocaine hydrochloride 1% with epinephrine, 1 cc, prepare and drape, hemostasis obtained, ointment and bandage applied, and care instructions provided.

As was the case before 2021, you still need a CC, along with a medically (and medicolegally) appropriate history and physical examination. A diagnostic impression and treatment plan also should be included.

In this situation, reporting is straightforward. There is no separate E/M visit; only the CPT code 11102 for tangential biopsy is reported. An International Classification of Diseases, Tenth Revision code of D48.5 (neoplasm of uncertain behavior of skin) will be included.

Why no E/M code? This is because the biopsy includes preservice and postservice time and work that would be double reported with the E/M. Remember that the preservice work would include any history and physical examination related to the area to be biopsied.

Specifically, preservice work includes:

Inspect and palpate lesion to assess surface size, subcutaneous depth and extension, and whether fixed to underlying structures. Select the most representative and appropriate site to obtain specimen. Examine draining lymph node basins. Discuss need for skin biopsy and biopsy technique options. Describe the tangential biopsy procedure method and expected result and the potential for inconclusive pathology result. Review procedural risks, including bleeding, pain, edema, infection, delayed healing, scarring, and hyper- or hypopigmentation.⁵

Postservice work includes:

Instruct patient and family on postoperative wound care and dressing changes, as well as problems such as bleeding or pain and restrictions on activities, and follow-up care. Provide prescriptions for pain and antibiotics as necessary. Advise patient and family when results will be available and how they will be communicated. The pathology request form is filled out and signed by the physician. Complete medical record and communicate procedure/results to referring physician as appropriate.⁵

The Takeaway—Procedure codes include preservice and postservice work. If additional work for the procedure is not documented beyond that, an E/M cannot be included in the encounter.

Scenario 2: What If We Don’t Biopsy?

A 34-year-old presents with a new spot on the left cheek that seems to be growing and changing shape rapidly. You examine the patient and discuss treatment options. The documentation reads as follows:

• CC: New spot on left cheek that seems to be growing and changing shape rapidly.

• History: No family history of skin cancer; concerned about scarring, no blood thinner.

• Examination: Irregular tan to brown to black 8-mm macule. No lymphadenopathy.

• Impression: rule out melanoma.

• Plan: Review risk, benefits, and alternative options. Schedule biopsy. Discuss unique risk factor of sebaceous peau d’orange skin more prone to contour defects after biopsy.

When determining the coding level for this scenario by MDM, 3 components must be considered: number and complexity of problems addressed at the encounter (column 1), amount and/or complexity of data to be reviewed and analyzed (column 2), and risk of complications and/or morbidity or mortality of patient management (column 3).¹ There are no data that are reviewed, so the auditor will assume minimal data to be reviewed and/or analyzed (level 2, row 2 in the MDM table). However, there may be a lot of variation in how an auditor would address the number and complexity of problems (level 1). Consider that you must explicitly state what you are thinking, as an auditor may not know melanoma is a life-threatening diagnosis. From the perspective of the auditor, could this be a:

• Self-limited or minor problem (level 2, or minimal problem in the MDM table)?¹

• Stable chronic illness (level 3, or low-level problem)?¹

• Undiagnosed new problem with uncertain prognosis (level 4, or moderate level problem)?¹

• Acute illness with systemic symptoms (level 4, or moderate level problem)?¹

• Acute or chronic illness or injury that poses a threat to life or bodily function (level 5, or high-level problem)?¹

• All of the above?

Similarly, there may be variation in how the risk (column 3) would be interpreted in this scenario. The treatment gives no guidance, so the auditor may assume this has a minimal risk of morbidity (level 2) or possibly a low risk of morbidity from additional diagnostic testing or treatment (level 3), as opposed to a moderate risk of morbidity (level 4).¹The Takeaway—In the auditor’s mind, this could be a straightforward (CPT codes 99202/99212) or lowlevel (99203/99213) visit as opposed to a moderate-level (99204/99214) visit. From the above documentation, an auditor would not be able to tell what you are thinking, and you can be assured they will not look further into the diagnosis or treatment to learn. That is not their job. So, let us clarify by explicitly stating what you are thinking in the context of the MDM grid.

Modified Scenario 2: A Funny-Looking New Spot With MDM Descriptors to Guide an Auditor

Below are modifications to the documentation for scenario 2 to guide an auditor:

• CC: New spot on left cheek that seems to be growing and changing shape rapidly.

• History: No family history of skin cancer; concerned about scarring, no blood thinner.

• Examination: Irregular tan to brown to black 8-mm macule. No lymphadenopathy.

• Impression: rule out melanoma (undiagnosed new problem with uncertain prognosis).

• Plan: Discuss risks, benefits, and alternatives, including biopsy (decision regarding minor surgery with identified patient or procedure risk factors) vs a noninvasive gene expression profiling melanoma rule-out test. Patient prefers the latter.

In this scenario, the level of MDM is much more clearly documented (as bolded above).

The number and complexity of problems would be an undiagnosed new problem with uncertain prognosis, which would be moderate complexity (column 1, level 4).¹ There are no data that are reviewed or analyzed, which would be straightforward (column 2, level 2). For risk, the discussion of the biopsy as part of the diagnostic choices should include discussion of possible scarring, bleeding, pain, and infection, which would be considered best described as a decision regarding minor surgery with identified patient or procedure risk factors, which would make this of moderate complexity (column 3, level 4).¹

Importantly, even if the procedure is not chosen as the final treatment plan, the discussion regarding the surgery, including the risks, benefits, and alternatives, can still count toward this category in the MDM table. Therefore, in this scenario with the updated and clarified documentation, this would be reported as CPT code 99204 for a new patient, while an established patient would be 99214.

Scenario 1 Revisited: A Funny-Looking New Spot

Below is scenario 1 with enhanced documentation, now applied to our procedure-only visit.

• CC: New spot on left cheek that seems to be growing and changing shape rapidly.

• History: No family history of skin cancer; concerned about scarring, no blood thinner.

• Examination: Irregular tan to brown to black 8-mm macule. No lymphadenopathy.

• Impression: rule out melanoma (undiagnosed new problem with uncertain prognosis).

• Plan: Discuss risks, benefits, and alternatives, including biopsy (decision regarding minor surgery with identified patient or procedure risk factors) vs a noninvasive 2 gene expression profiling melanoma rule-out test. Patient wants biopsy. Consent, biopsy via shave technique. Lidocaine hydrochloride 1% with epinephrine, 1 cc, prepare and drape, hemostasis obtained, ointment and bandage applied, and care instructions provided.

This documentation would only allow reporting the biopsy as in Scenario 1, as the decision to perform a 0- or 10-day global procedure is bundled with the procedure if performed on the same date of service.

Final Thoughts

Spot checks are commonly encountered dermatologic visits. With the updated E/M guidelines, clarifying and streamlining your documentation is crucial. In particular, utilizing language that clearly defines number and complexity of problems, amount and/or complexity of data to be reviewed and analyzed, and appropriate risk stratification is crucial to ensuring appropriate reimbursement and minimizing your pain with audits.

On January 1, 2021, the Current Procedural Terminology (CPT) evaluation and management (E/M) reporting rules changed dramatically, with “bullet counting” no longer necessary and the coding level now based on either the new medical decision making (MDM) table or time spent on all activities relating to the care of the patient on the day of the encounter.¹ This is described in the CPT Professional Edition 2023, a book every practitioner should review annually.² In particular, every provider should read and reread pages 1 to 14—and beyond if you provide services beyond standard office visits. These changes were made with the intent to simplify the process of documentation and allow a provider to spend more time with patients, though there is still a paucity of data related to whether the new system achieves these aims.

The general rule of reporting work with CPT codes can be simply stated—“Document what you did, do what you documented, and report that which is medically necessary” (David McCafferey, MD, personal communication)—and you should never have any difficulty with audits. Unfortunately, the new system does not let an auditor, who typically lacks a medical degree, audit effectively unless they have a clear understanding of diseases and their stages. Many medical societies, including the American Medical Association³ and American Academy of Dermatology,⁴ have provided education that focuses on how to report a given vignette, but specific examples of documentation with commentary are uncommon.

To make your documentation more likely to pass audits, explicitly link parts of your documentation to CPT MDM descriptors. We offer scenarios and tips. In part 1 of this series, we discuss how to approach the “spot check,” a commonly encountered chief concern (CC) within dermatology.

Scenario 1: A Funny-Looking New Spot

A 34-year-old presents with a new spot on the left cheek that seems to be growing and changing shape rapidly. You examine the patient and discuss treatment options. The documentation reads as follows:

• CC: New spot on left cheek that seems to be growing and changing shape rapidly.

• History: No family history of skin cancer; concerned about scarring, no blood thinner.

• Examination: Irregular tan to brown to black 8-mm macule. No lymphadenopathy.

• Impression: rule out melanoma.

• Plan: Consent, biopsy via shave technique. Lidocaine hydrochloride 1% with epinephrine, 1 cc, prepare and drape, hemostasis obtained, ointment and bandage applied, and care instructions provided.

As was the case before 2021, you still need a CC, along with a medically (and medicolegally) appropriate history and physical examination. A diagnostic impression and treatment plan also should be included.

In this situation, reporting is straightforward. There is no separate E/M visit; only the CPT code 11102 for tangential biopsy is reported. An International Classification of Diseases, Tenth Revision code of D48.5 (neoplasm of uncertain behavior of skin) will be included.

Why no E/M code? This is because the biopsy includes preservice and postservice time and work that would be double reported with the E/M. Remember that the preservice work would include any history and physical examination related to the area to be biopsied.

Specifically, preservice work includes:

Inspect and palpate lesion to assess surface size, subcutaneous depth and extension, and whether fixed to underlying structures. Select the most representative and appropriate site to obtain specimen. Examine draining lymph node basins. Discuss need for skin biopsy and biopsy technique options. Describe the tangential biopsy procedure method and expected result and the potential for inconclusive pathology result. Review procedural risks, including bleeding, pain, edema, infection, delayed healing, scarring, and hyper- or hypopigmentation.⁵

Postservice work includes:

Instruct patient and family on postoperative wound care and dressing changes, as well as problems such as bleeding or pain and restrictions on activities, and follow-up care. Provide prescriptions for pain and antibiotics as necessary. Advise patient and family when results will be available and how they will be communicated. The pathology request form is filled out and signed by the physician. Complete medical record and communicate procedure/results to referring physician as appropriate.⁵

The Takeaway—Procedure codes include preservice and postservice work. If additional work for the procedure is not documented beyond that, an E/M cannot be included in the encounter.

Scenario 2: What If We Don’t Biopsy?

A 34-year-old presents with a new spot on the left cheek that seems to be growing and changing shape rapidly. You examine the patient and discuss treatment options. The documentation reads as follows:

• CC: New spot on left cheek that seems to be growing and changing shape rapidly.

• History: No family history of skin cancer; concerned about scarring, no blood thinner.

• Examination: Irregular tan to brown to black 8-mm macule. No lymphadenopathy.

• Impression: rule out melanoma.

• Plan: Review risk, benefits, and alternative options. Schedule biopsy. Discuss unique risk factor of sebaceous peau d’orange skin more prone to contour defects after biopsy.

When determining the coding level for this scenario by MDM, 3 components must be considered: number and complexity of problems addressed at the encounter (column 1), amount and/or complexity of data to be reviewed and analyzed (column 2), and risk of complications and/or morbidity or mortality of patient management (column 3).¹ There are no data that are reviewed, so the auditor will assume minimal data to be reviewed and/or analyzed (level 2, row 2 in the MDM table). However, there may be a lot of variation in how an auditor would address the number and complexity of problems (level 1). Consider that you must explicitly state what you are thinking, as an auditor may not know melanoma is a life-threatening diagnosis. From the perspective of the auditor, could this be a:

• Self-limited or minor problem (level 2, or minimal problem in the MDM table)?¹

• Stable chronic illness (level 3, or low-level problem)?¹

• Undiagnosed new problem with uncertain prognosis (level 4, or moderate level problem)?¹

• Acute illness with systemic symptoms (level 4, or moderate level problem)?¹

• Acute or chronic illness or injury that poses a threat to life or bodily function (level 5, or high-level problem)?¹

• All of the above?

Similarly, there may be variation in how the risk (column 3) would be interpreted in this scenario. The treatment gives no guidance, so the auditor may assume this has a minimal risk of morbidity (level 2) or possibly a low risk of morbidity from additional diagnostic testing or treatment (level 3), as opposed to a moderate risk of morbidity (level 4).¹The Takeaway—In the auditor’s mind, this could be a straightforward (CPT codes 99202/99212) or lowlevel (99203/99213) visit as opposed to a moderate-level (99204/99214) visit. From the above documentation, an auditor would not be able to tell what you are thinking, and you can be assured they will not look further into the diagnosis or treatment to learn. That is not their job. So, let us clarify by explicitly stating what you are thinking in the context of the MDM grid.

Modified Scenario 2: A Funny-Looking New Spot With MDM Descriptors to Guide an Auditor

Below are modifications to the documentation for scenario 2 to guide an auditor:

• CC: New spot on left cheek that seems to be growing and changing shape rapidly.

• History: No family history of skin cancer; concerned about scarring, no blood thinner.

• Examination: Irregular tan to brown to black 8-mm macule. No lymphadenopathy.

• Impression: rule out melanoma (undiagnosed new problem with uncertain prognosis).

• Plan: Discuss risks, benefits, and alternatives, including biopsy (decision regarding minor surgery with identified patient or procedure risk factors) vs a noninvasive gene expression profiling melanoma rule-out test. Patient prefers the latter.

In this scenario, the level of MDM is much more clearly documented (as bolded above).

The number and complexity of problems would be an undiagnosed new problem with uncertain prognosis, which would be moderate complexity (column 1, level 4).¹ There are no data that are reviewed or analyzed, which would be straightforward (column 2, level 2). For risk, the discussion of the biopsy as part of the diagnostic choices should include discussion of possible scarring, bleeding, pain, and infection, which would be considered best described as a decision regarding minor surgery with identified patient or procedure risk factors, which would make this of moderate complexity (column 3, level 4).¹

Importantly, even if the procedure is not chosen as the final treatment plan, the discussion regarding the surgery, including the risks, benefits, and alternatives, can still count toward this category in the MDM table. Therefore, in this scenario with the updated and clarified documentation, this would be reported as CPT code 99204 for a new patient, while an established patient would be 99214.

Scenario 1 Revisited: A Funny-Looking New Spot

Below is scenario 1 with enhanced documentation, now applied to our procedure-only visit.

• CC: New spot on left cheek that seems to be growing and changing shape rapidly.

• History: No family history of skin cancer; concerned about scarring, no blood thinner.

• Examination: Irregular tan to brown to black 8-mm macule. No lymphadenopathy.

• Impression: rule out melanoma (undiagnosed new problem with uncertain prognosis).

• Plan: Discuss risks, benefits, and alternatives, including biopsy (decision regarding minor surgery with identified patient or procedure risk factors) vs a noninvasive 2 gene expression profiling melanoma rule-out test. Patient wants biopsy. Consent, biopsy via shave technique. Lidocaine hydrochloride 1% with epinephrine, 1 cc, prepare and drape, hemostasis obtained, ointment and bandage applied, and care instructions provided.

This documentation would only allow reporting the biopsy as in Scenario 1, as the decision to perform a 0- or 10-day global procedure is bundled with the procedure if performed on the same date of service.

Final Thoughts

Spot checks are commonly encountered dermatologic visits. With the updated E/M guidelines, clarifying and streamlining your documentation is crucial. In particular, utilizing language that clearly defines number and complexity of problems, amount and/or complexity of data to be reviewed and analyzed, and appropriate risk stratification is crucial to ensuring appropriate reimbursement and minimizing your pain with audits.

On January 1, 2021, the Current Procedural Terminology (CPT) evaluation and management (E/M) reporting rules changed dramatically, with “bullet counting” no longer necessary and the coding level now based on either the new medical decision making (MDM) table or time spent on all activities relating to the care of the patient on the day of the encounter.¹ This is described in the CPT Professional Edition 2023, a book every practitioner should review annually.² In particular, every provider should read and reread pages 1 to 14—and beyond if you provide services beyond standard office visits. These changes were made with the intent to simplify the process of documentation and allow a provider to spend more time with patients, though there is still a paucity of data related to whether the new system achieves these aims.

The general rule of reporting work with CPT codes can be simply stated—“Document what you did, do what you documented, and report that which is medically necessary” (David McCafferey, MD, personal communication)—and you should never have any difficulty with audits. Unfortunately, the new system does not let an auditor, who typically lacks a medical degree, audit effectively unless they have a clear understanding of diseases and their stages. Many medical societies, including the American Medical Association³ and American Academy of Dermatology,⁴ have provided education that focuses on how to report a given vignette, but specific examples of documentation with commentary are uncommon.

To make your documentation more likely to pass audits, explicitly link parts of your documentation to CPT MDM descriptors. We offer scenarios and tips. In part 1 of this series, we discuss how to approach the “spot check,” a commonly encountered chief concern (CC) within dermatology.

Scenario 1: A Funny-Looking New Spot

A 34-year-old presents with a new spot on the left cheek that seems to be growing and changing shape rapidly. You examine the patient and discuss treatment options. The documentation reads as follows:

• CC: New spot on left cheek that seems to be growing and changing shape rapidly.

• History: No family history of skin cancer; concerned about scarring, no blood thinner.

• Examination: Irregular tan to brown to black 8-mm macule. No lymphadenopathy.

• Impression: rule out melanoma.

• Plan: Consent, biopsy via shave technique. Lidocaine hydrochloride 1% with epinephrine, 1 cc, prepare and drape, hemostasis obtained, ointment and bandage applied, and care instructions provided.

As was the case before 2021, you still need a CC, along with a medically (and medicolegally) appropriate history and physical examination. A diagnostic impression and treatment plan also should be included.

In this situation, reporting is straightforward. There is no separate E/M visit; only the CPT code 11102 for tangential biopsy is reported. An International Classification of Diseases, Tenth Revision code of D48.5 (neoplasm of uncertain behavior of skin) will be included.

Why no E/M code? This is because the biopsy includes preservice and postservice time and work that would be double reported with the E/M. Remember that the preservice work would include any history and physical examination related to the area to be biopsied.

Specifically, preservice work includes:

Inspect and palpate lesion to assess surface size, subcutaneous depth and extension, and whether fixed to underlying structures. Select the most representative and appropriate site to obtain specimen. Examine draining lymph node basins. Discuss need for skin biopsy and biopsy technique options. Describe the tangential biopsy procedure method and expected result and the potential for inconclusive pathology result. Review procedural risks, including bleeding, pain, edema, infection, delayed healing, scarring, and hyper- or hypopigmentation.⁵

Postservice work includes:

Instruct patient and family on postoperative wound care and dressing changes, as well as problems such as bleeding or pain and restrictions on activities, and follow-up care. Provide prescriptions for pain and antibiotics as necessary. Advise patient and family when results will be available and how they will be communicated. The pathology request form is filled out and signed by the physician. Complete medical record and communicate procedure/results to referring physician as appropriate.⁵

The Takeaway—Procedure codes include preservice and postservice work. If additional work for the procedure is not documented beyond that, an E/M cannot be included in the encounter.

Scenario 2: What If We Don’t Biopsy?

A 34-year-old presents with a new spot on the left cheek that seems to be growing and changing shape rapidly. You examine the patient and discuss treatment options. The documentation reads as follows:

• CC: New spot on left cheek that seems to be growing and changing shape rapidly.

• History: No family history of skin cancer; concerned about scarring, no blood thinner.

• Examination: Irregular tan to brown to black 8-mm macule. No lymphadenopathy.

• Impression: rule out melanoma.

• Plan: Review risk, benefits, and alternative options. Schedule biopsy. Discuss unique risk factor of sebaceous peau d’orange skin more prone to contour defects after biopsy.

When determining the coding level for this scenario by MDM, 3 components must be considered: number and complexity of problems addressed at the encounter (column 1), amount and/or complexity of data to be reviewed and analyzed (column 2), and risk of complications and/or morbidity or mortality of patient management (column 3).¹ There are no data that are reviewed, so the auditor will assume minimal data to be reviewed and/or analyzed (level 2, row 2 in the MDM table). However, there may be a lot of variation in how an auditor would address the number and complexity of problems (level 1). Consider that you must explicitly state what you are thinking, as an auditor may not know melanoma is a life-threatening diagnosis. From the perspective of the auditor, could this be a:

• Self-limited or minor problem (level 2, or minimal problem in the MDM table)?¹

• Stable chronic illness (level 3, or low-level problem)?¹

• Undiagnosed new problem with uncertain prognosis (level 4, or moderate level problem)?¹

• Acute illness with systemic symptoms (level 4, or moderate level problem)?¹

• Acute or chronic illness or injury that poses a threat to life or bodily function (level 5, or high-level problem)?¹

• All of the above?

Similarly, there may be variation in how the risk (column 3) would be interpreted in this scenario. The treatment gives no guidance, so the auditor may assume this has a minimal risk of morbidity (level 2) or possibly a low risk of morbidity from additional diagnostic testing or treatment (level 3), as opposed to a moderate risk of morbidity (level 4).¹The Takeaway—In the auditor’s mind, this could be a straightforward (CPT codes 99202/99212) or lowlevel (99203/99213) visit as opposed to a moderate-level (99204/99214) visit. From the above documentation, an auditor would not be able to tell what you are thinking, and you can be assured they will not look further into the diagnosis or treatment to learn. That is not their job. So, let us clarify by explicitly stating what you are thinking in the context of the MDM grid.

Modified Scenario 2: A Funny-Looking New Spot With MDM Descriptors to Guide an Auditor

Below are modifications to the documentation for scenario 2 to guide an auditor:

• CC: New spot on left cheek that seems to be growing and changing shape rapidly.

• History: No family history of skin cancer; concerned about scarring, no blood thinner.

• Examination: Irregular tan to brown to black 8-mm macule. No lymphadenopathy.

• Impression: rule out melanoma (undiagnosed new problem with uncertain prognosis).

• Plan: Discuss risks, benefits, and alternatives, including biopsy (decision regarding minor surgery with identified patient or procedure risk factors) vs a noninvasive gene expression profiling melanoma rule-out test. Patient prefers the latter.

In this scenario, the level of MDM is much more clearly documented (as bolded above).

The number and complexity of problems would be an undiagnosed new problem with uncertain prognosis, which would be moderate complexity (column 1, level 4).¹ There are no data that are reviewed or analyzed, which would be straightforward (column 2, level 2). For risk, the discussion of the biopsy as part of the diagnostic choices should include discussion of possible scarring, bleeding, pain, and infection, which would be considered best described as a decision regarding minor surgery with identified patient or procedure risk factors, which would make this of moderate complexity (column 3, level 4).¹

Importantly, even if the procedure is not chosen as the final treatment plan, the discussion regarding the surgery, including the risks, benefits, and alternatives, can still count toward this category in the MDM table. Therefore, in this scenario with the updated and clarified documentation, this would be reported as CPT code 99204 for a new patient, while an established patient would be 99214.

Scenario 1 Revisited: A Funny-Looking New Spot

Below is scenario 1 with enhanced documentation, now applied to our procedure-only visit.

• CC: New spot on left cheek that seems to be growing and changing shape rapidly.

• History: No family history of skin cancer; concerned about scarring, no blood thinner.

• Examination: Irregular tan to brown to black 8-mm macule. No lymphadenopathy.

• Impression: rule out melanoma (undiagnosed new problem with uncertain prognosis).

• Plan: Discuss risks, benefits, and alternatives, including biopsy (decision regarding minor surgery with identified patient or procedure risk factors) vs a noninvasive 2 gene expression profiling melanoma rule-out test. Patient wants biopsy. Consent, biopsy via shave technique. Lidocaine hydrochloride 1% with epinephrine, 1 cc, prepare and drape, hemostasis obtained, ointment and bandage applied, and care instructions provided.

This documentation would only allow reporting the biopsy as in Scenario 1, as the decision to perform a 0- or 10-day global procedure is bundled with the procedure if performed on the same date of service.

Final Thoughts

Spot checks are commonly encountered dermatologic visits. With the updated E/M guidelines, clarifying and streamlining your documentation is crucial. In particular, utilizing language that clearly defines number and complexity of problems, amount and/or complexity of data to be reviewed and analyzed, and appropriate risk stratification is crucial to ensuring appropriate reimbursement and minimizing your pain with audits.

Dermatology certainly is the most visual medical specialty. In the current era of powerful electronic imaging and laboratory techniques, the skills of physical diagnosis seem to have become less important in medicine—not so in dermatology, in which the experienced clinician is able to identify many conditions by simply looking at the skin. Of course, dermatologists do heavily rely on dermatopathologists to microscopically visualize biopsies to distinguish diseases. Even as we acknowledge the dominant role of visual recognition, there is increasing progress in making clinical determinations based on molecular events. The era of genomic dermatology is here.

The Genodermatoses

There are more than 500 dermatologic conditions resulting from heritable mutational events.¹ The rarity of most of these diseases and variability in phenotypic manifestations presents considerable diagnostic challenges, typically the province of a select group of clinical pediatric dermatologists whose abilities have been developed by experience.² However, the addition of genomic analysis has now made reliable identification more accessible to a wider group of clinicians.³ The Human Genome Project was arguably the most successful health policy endeavor in human history, promoting the development of massive automated, information theory–driven applications to analyze DNA sequences.⁴ We all think of DNA analysis as the ultimate means to detect mutations by sequencing whole exomes—and in fact the entire genome of affected individuals searching for mutations—but DNA sequencing often is insufficient to detect mutations in noncoding regions of genes and to identify abnormalities of gene expression (eg, splice variants). Building on the advances in high-throughput nucleic acid sequencing and massive computerized analysis, the field has now taken a quantum leap further to sequence transcribed RNA to detect abnormalities.⁵

The techniques are straightforward: RNA is isolated and reverse transcribed to complementary DNA. The complementary DNA is amplified and then processed by high-throughput sequencers. The sequences are then identified by computer algorithms. It is possible to fully define the transcriptomes of multiple genes, even reaching the threshold of resolution of gene expression emanating from a single cell.⁶

Studying Gene Expression for Malignant Melanoma

As much as we rely on visual interpretations, we acknowledge that many conditions look very similar, whether to the naked eye or under the microscope. This is true for rare diseases but also for the rashes we routinely see. A group of investigators recently used RNA transcriptome sequencing to analyze differences between atopic dermatitis and psoriasis, permitting better differentiation of these 2 common conditions.⁷

One of the greatest challenges confronting dermatologists and their dermatopathologist partners is to distinguish malignant melanoma from benign nevi.⁸ Despite staining for a number of molecular markers, some lesions defy histopathology, such as distinguishing benign and malignant Spitz nevi; however, recent work on RNA transcriptomes suggests that gene expression may increase confidence in assessing atypical Spitz nevi.⁹ A 23-gene expression panel has yielded a sensitivity of 91.5% and a specificity of 92.5% in differentiating benign nevi from malignant melanoma.¹⁰

From the Research Laboratory to Routine Clinical Use

Undoubtedly, it is a large step from proof-of-concept studies to accepted clinical use. The ultimate achievement for a laboratory technique is to enter approved clinical use. Gene expression panels have now been approved by numerous third-party insurers to help predict future clinical evolution of biopsied melanomas. Although early in situ melanomas are eminently curable by wide excision, lesions that have more concerning characteristics (eg, depth >0.8 mm, ulceration) may progress to metastatic disease. The gratifying success of checkpoint inhibitor therapy has improved the previously dismal outlook for advanced melanomas.¹¹ Dermatologists search for clues to suggest which patients may benefit from adjuvant therapy. Sentinel lymph node biopsy (SLNB) has been a standard-of-care technique to help make this determination.¹²

It has now been demonstrated that gene expression array analysis can provide evidence complementing SLNB results or even independent of SLNB results. In extensive validation studies, a 31-gene expression panel analyzing initial melanoma biopsy specimens showed predictive value for later recurrence and development of metastatic disease.^13,14 The gene expression studies have identified patients with negative SLNBs who have gone on to develop metastatic melanomas.¹⁵ It has been suggested that gene expression panel diagnosis may reduce the need for invasive SLNBs in patients in whom the surgical procedure may involve risk.¹⁶

Looking to the Future

The progress of science is the result of many small steps building on prior work. The terms breakthrough and game changer in medicine have been popularized by the media and rarely are valid. On the contrary, sequential development of methods over many years has preceded the acclaimed successes of medical research; for example, the best-known medical breakthrough—that of Salk’s inactivated polio vaccine—was preceded by the use of an inactivated polio vaccine by Brodie and Park¹⁷ in 1935. However, it was the development of tissue culture of poliomyelitis virus by Enders et al¹⁸ that provided the methodology to Salk’s group to produce their inactivated polio vaccine.

The ability to go beyond our visual senses will be of great importance in characterizing the variability of skin diseases, especially in skin of color patients; for example, acral melanoma is perhaps the primary melanocytic malignancy in darker-skinned patients and is the target of RNA transcriptomic research.¹⁹ Progress is continuing on gene therapy for a growing number of skin conditions.^20,21 In vivo correction of abnormal genes is being attempted for a number of inherited cutaneous diseases,²² notably for disorders of skin fragility.²³ For now, we welcome the addition of genomic capabilities to the visual practice of dermatology and the capability to go beyond that which we can see with our eyes.

Dermatology certainly is the most visual medical specialty. In the current era of powerful electronic imaging and laboratory techniques, the skills of physical diagnosis seem to have become less important in medicine—not so in dermatology, in which the experienced clinician is able to identify many conditions by simply looking at the skin. Of course, dermatologists do heavily rely on dermatopathologists to microscopically visualize biopsies to distinguish diseases. Even as we acknowledge the dominant role of visual recognition, there is increasing progress in making clinical determinations based on molecular events. The era of genomic dermatology is here.

The Genodermatoses

There are more than 500 dermatologic conditions resulting from heritable mutational events.¹ The rarity of most of these diseases and variability in phenotypic manifestations presents considerable diagnostic challenges, typically the province of a select group of clinical pediatric dermatologists whose abilities have been developed by experience.² However, the addition of genomic analysis has now made reliable identification more accessible to a wider group of clinicians.³ The Human Genome Project was arguably the most successful health policy endeavor in human history, promoting the development of massive automated, information theory–driven applications to analyze DNA sequences.⁴ We all think of DNA analysis as the ultimate means to detect mutations by sequencing whole exomes—and in fact the entire genome of affected individuals searching for mutations—but DNA sequencing often is insufficient to detect mutations in noncoding regions of genes and to identify abnormalities of gene expression (eg, splice variants). Building on the advances in high-throughput nucleic acid sequencing and massive computerized analysis, the field has now taken a quantum leap further to sequence transcribed RNA to detect abnormalities.⁵

The techniques are straightforward: RNA is isolated and reverse transcribed to complementary DNA. The complementary DNA is amplified and then processed by high-throughput sequencers. The sequences are then identified by computer algorithms. It is possible to fully define the transcriptomes of multiple genes, even reaching the threshold of resolution of gene expression emanating from a single cell.⁶

Studying Gene Expression for Malignant Melanoma

As much as we rely on visual interpretations, we acknowledge that many conditions look very similar, whether to the naked eye or under the microscope. This is true for rare diseases but also for the rashes we routinely see. A group of investigators recently used RNA transcriptome sequencing to analyze differences between atopic dermatitis and psoriasis, permitting better differentiation of these 2 common conditions.⁷

One of the greatest challenges confronting dermatologists and their dermatopathologist partners is to distinguish malignant melanoma from benign nevi.⁸ Despite staining for a number of molecular markers, some lesions defy histopathology, such as distinguishing benign and malignant Spitz nevi; however, recent work on RNA transcriptomes suggests that gene expression may increase confidence in assessing atypical Spitz nevi.⁹ A 23-gene expression panel has yielded a sensitivity of 91.5% and a specificity of 92.5% in differentiating benign nevi from malignant melanoma.¹⁰

From the Research Laboratory to Routine Clinical Use

Undoubtedly, it is a large step from proof-of-concept studies to accepted clinical use. The ultimate achievement for a laboratory technique is to enter approved clinical use. Gene expression panels have now been approved by numerous third-party insurers to help predict future clinical evolution of biopsied melanomas. Although early in situ melanomas are eminently curable by wide excision, lesions that have more concerning characteristics (eg, depth >0.8 mm, ulceration) may progress to metastatic disease. The gratifying success of checkpoint inhibitor therapy has improved the previously dismal outlook for advanced melanomas.¹¹ Dermatologists search for clues to suggest which patients may benefit from adjuvant therapy. Sentinel lymph node biopsy (SLNB) has been a standard-of-care technique to help make this determination.¹²

It has now been demonstrated that gene expression array analysis can provide evidence complementing SLNB results or even independent of SLNB results. In extensive validation studies, a 31-gene expression panel analyzing initial melanoma biopsy specimens showed predictive value for later recurrence and development of metastatic disease.^13,14 The gene expression studies have identified patients with negative SLNBs who have gone on to develop metastatic melanomas.¹⁵ It has been suggested that gene expression panel diagnosis may reduce the need for invasive SLNBs in patients in whom the surgical procedure may involve risk.¹⁶

Looking to the Future

The progress of science is the result of many small steps building on prior work. The terms breakthrough and game changer in medicine have been popularized by the media and rarely are valid. On the contrary, sequential development of methods over many years has preceded the acclaimed successes of medical research; for example, the best-known medical breakthrough—that of Salk’s inactivated polio vaccine—was preceded by the use of an inactivated polio vaccine by Brodie and Park¹⁷ in 1935. However, it was the development of tissue culture of poliomyelitis virus by Enders et al¹⁸ that provided the methodology to Salk’s group to produce their inactivated polio vaccine.

The ability to go beyond our visual senses will be of great importance in characterizing the variability of skin diseases, especially in skin of color patients; for example, acral melanoma is perhaps the primary melanocytic malignancy in darker-skinned patients and is the target of RNA transcriptomic research.¹⁹ Progress is continuing on gene therapy for a growing number of skin conditions.^20,21 In vivo correction of abnormal genes is being attempted for a number of inherited cutaneous diseases,²² notably for disorders of skin fragility.²³ For now, we welcome the addition of genomic capabilities to the visual practice of dermatology and the capability to go beyond that which we can see with our eyes.

Dermatology certainly is the most visual medical specialty. In the current era of powerful electronic imaging and laboratory techniques, the skills of physical diagnosis seem to have become less important in medicine—not so in dermatology, in which the experienced clinician is able to identify many conditions by simply looking at the skin. Of course, dermatologists do heavily rely on dermatopathologists to microscopically visualize biopsies to distinguish diseases. Even as we acknowledge the dominant role of visual recognition, there is increasing progress in making clinical determinations based on molecular events. The era of genomic dermatology is here.

The Genodermatoses

There are more than 500 dermatologic conditions resulting from heritable mutational events.¹ The rarity of most of these diseases and variability in phenotypic manifestations presents considerable diagnostic challenges, typically the province of a select group of clinical pediatric dermatologists whose abilities have been developed by experience.² However, the addition of genomic analysis has now made reliable identification more accessible to a wider group of clinicians.³ The Human Genome Project was arguably the most successful health policy endeavor in human history, promoting the development of massive automated, information theory–driven applications to analyze DNA sequences.⁴ We all think of DNA analysis as the ultimate means to detect mutations by sequencing whole exomes—and in fact the entire genome of affected individuals searching for mutations—but DNA sequencing often is insufficient to detect mutations in noncoding regions of genes and to identify abnormalities of gene expression (eg, splice variants). Building on the advances in high-throughput nucleic acid sequencing and massive computerized analysis, the field has now taken a quantum leap further to sequence transcribed RNA to detect abnormalities.⁵

The techniques are straightforward: RNA is isolated and reverse transcribed to complementary DNA. The complementary DNA is amplified and then processed by high-throughput sequencers. The sequences are then identified by computer algorithms. It is possible to fully define the transcriptomes of multiple genes, even reaching the threshold of resolution of gene expression emanating from a single cell.⁶

Studying Gene Expression for Malignant Melanoma

As much as we rely on visual interpretations, we acknowledge that many conditions look very similar, whether to the naked eye or under the microscope. This is true for rare diseases but also for the rashes we routinely see. A group of investigators recently used RNA transcriptome sequencing to analyze differences between atopic dermatitis and psoriasis, permitting better differentiation of these 2 common conditions.⁷

One of the greatest challenges confronting dermatologists and their dermatopathologist partners is to distinguish malignant melanoma from benign nevi.⁸ Despite staining for a number of molecular markers, some lesions defy histopathology, such as distinguishing benign and malignant Spitz nevi; however, recent work on RNA transcriptomes suggests that gene expression may increase confidence in assessing atypical Spitz nevi.⁹ A 23-gene expression panel has yielded a sensitivity of 91.5% and a specificity of 92.5% in differentiating benign nevi from malignant melanoma.¹⁰

From the Research Laboratory to Routine Clinical Use

Undoubtedly, it is a large step from proof-of-concept studies to accepted clinical use. The ultimate achievement for a laboratory technique is to enter approved clinical use. Gene expression panels have now been approved by numerous third-party insurers to help predict future clinical evolution of biopsied melanomas. Although early in situ melanomas are eminently curable by wide excision, lesions that have more concerning characteristics (eg, depth >0.8 mm, ulceration) may progress to metastatic disease. The gratifying success of checkpoint inhibitor therapy has improved the previously dismal outlook for advanced melanomas.¹¹ Dermatologists search for clues to suggest which patients may benefit from adjuvant therapy. Sentinel lymph node biopsy (SLNB) has been a standard-of-care technique to help make this determination.¹²

It has now been demonstrated that gene expression array analysis can provide evidence complementing SLNB results or even independent of SLNB results. In extensive validation studies, a 31-gene expression panel analyzing initial melanoma biopsy specimens showed predictive value for later recurrence and development of metastatic disease.^13,14 The gene expression studies have identified patients with negative SLNBs who have gone on to develop metastatic melanomas.¹⁵ It has been suggested that gene expression panel diagnosis may reduce the need for invasive SLNBs in patients in whom the surgical procedure may involve risk.¹⁶

Looking to the Future

The progress of science is the result of many small steps building on prior work. The terms breakthrough and game changer in medicine have been popularized by the media and rarely are valid. On the contrary, sequential development of methods over many years has preceded the acclaimed successes of medical research; for example, the best-known medical breakthrough—that of Salk’s inactivated polio vaccine—was preceded by the use of an inactivated polio vaccine by Brodie and Park¹⁷ in 1935. However, it was the development of tissue culture of poliomyelitis virus by Enders et al¹⁸ that provided the methodology to Salk’s group to produce their inactivated polio vaccine.

The ability to go beyond our visual senses will be of great importance in characterizing the variability of skin diseases, especially in skin of color patients; for example, acral melanoma is perhaps the primary melanocytic malignancy in darker-skinned patients and is the target of RNA transcriptomic research.¹⁹ Progress is continuing on gene therapy for a growing number of skin conditions.^20,21 In vivo correction of abnormal genes is being attempted for a number of inherited cutaneous diseases,²² notably for disorders of skin fragility.²³ For now, we welcome the addition of genomic capabilities to the visual practice of dermatology and the capability to go beyond that which we can see with our eyes.

Practice Gap

Nail matrix and nail bed injections with triamcinolone acetonide are used to treat trachyonychia and inflammatory nail conditions, including nail psoriasis and nail lichen planus. The procedure should be quick in well-trained hands, with each nail injection taking only seconds to perform. Typically, patients have multiple nails involved, requiring at least 1 injection into the nail matrix or the nail bed (or both) in each nail at each visit. Patients often are anxious when undergoing nail injections; the nail unit is highly innervated and vascular, which can cause notable transient discomfort during the procedure^1,2 as well as postoperative pain.³

Nail injections must be repeated every 4 to 6 weeks to sustain clinical benefit and maximize outcomes, which can lead to heightened anxiety and apprehension before and during the visit. Furthermore, pain and anxiety associated with the procedure may deter patients from returning for follow-up injections, which can impact treatment adherence and clinical outcomes.

Dermatologists should implement strategies to decrease periprocedural anxiety to improve the nail injection experience. In our practice, we routinely incorporate stress-reducing techniques—music, talkesthesia, a sleep mask, cool air, ethyl chloride, and squeezing a stress ball—into the clinical workflow of the procedure. The goal of these techniques is to divert attention away from painful stimuli. Most patients, however, receive injections in both hands, making it impractical to employ some of these techniques, particularly squeezing a stress ball. We employed a unique method involving polyurethane tubing to reduce stress and anxiety during nail procedures.

The Technique

A patient was receiving treatment with intralesional triamcinolone injections to the nail matrix for trachyonychia involving all of the fingernails. He worked as an equipment and facilities manager, giving him access to polyurethane tubing, which is routinely used in the manufacture of some medical devices that require gas or liquid to operate. He found the nail injections to be painful but was motivated to proceed with treatment. He brought in a piece of polyurethane tubing to a subsequent visit to bite on during the injections (Figure) and reported considerable relief of pain.

What you were not taught in United States history class was that this method—clenching an object orally—dates to the era before the Civil War, before appropriate anesthetics and analgesics were developed, when patients and soldiers bit on a bullet or leather strap during surgical procedures.⁴ Clenching and chewing have been shown to promote relaxation and reduce acute pain and stress.⁵

Practical Implications

Polyurethane tubing can be purchased in bulk, is inexpensive ($0.30/foot on Amazon), and unlikely to damage teeth due to its flexibility. It can be cut into 6-inch pieces and given to the patient at their first nail injection appointment. The patient can then bring the tubing to subsequent appointments to use as a mastication tool during nail injections.

We instruct the patient to disinfect the dedicated piece of tubing after the initial visit and each subsequent visit by soaking it for 15 minutes in either a 3% hydrogen peroxide solution, antibacterial mouthwash, a solution of baking soda (bicarbonate of soda) and water (1 cup of water to 2 teaspoons of baking soda), or white vinegar. We instruct them to thoroughly dry the disinfected polyurethane tube and store it in a clean, reusable, resealable zipper storage bag between appointments.

In addition to reducing anxiety and pain, this method also distracts the patient and therefore promotes patient and physician safety. Patients are less likely to jump or startle during the injection, thereby reducing the risk of physically interfering with the nail surgeon or making an unanticipated advance into the surgical field.

Although frustrated patients with nail disease may need to “bite the bullet” when they accept treatment with nail injections, lessons from our patient and from United States history offer a safe and cost-effective pain management strategy. Minimizing discomfort and anxiety during the first nail injection is crucial because doing so is likely to promote adherence with follow-up injections and therefore improve clinical outcomes.

Future clinical studies should validate the clinical utility of oral mastication and clenching during nail procedures compared to other perioperative stress- and anxiety-reducing techniques.

Practice Gap

Nail matrix and nail bed injections with triamcinolone acetonide are used to treat trachyonychia and inflammatory nail conditions, including nail psoriasis and nail lichen planus. The procedure should be quick in well-trained hands, with each nail injection taking only seconds to perform. Typically, patients have multiple nails involved, requiring at least 1 injection into the nail matrix or the nail bed (or both) in each nail at each visit. Patients often are anxious when undergoing nail injections; the nail unit is highly innervated and vascular, which can cause notable transient discomfort during the procedure^1,2 as well as postoperative pain.³

Nail injections must be repeated every 4 to 6 weeks to sustain clinical benefit and maximize outcomes, which can lead to heightened anxiety and apprehension before and during the visit. Furthermore, pain and anxiety associated with the procedure may deter patients from returning for follow-up injections, which can impact treatment adherence and clinical outcomes.

Dermatologists should implement strategies to decrease periprocedural anxiety to improve the nail injection experience. In our practice, we routinely incorporate stress-reducing techniques—music, talkesthesia, a sleep mask, cool air, ethyl chloride, and squeezing a stress ball—into the clinical workflow of the procedure. The goal of these techniques is to divert attention away from painful stimuli. Most patients, however, receive injections in both hands, making it impractical to employ some of these techniques, particularly squeezing a stress ball. We employed a unique method involving polyurethane tubing to reduce stress and anxiety during nail procedures.

The Technique

A patient was receiving treatment with intralesional triamcinolone injections to the nail matrix for trachyonychia involving all of the fingernails. He worked as an equipment and facilities manager, giving him access to polyurethane tubing, which is routinely used in the manufacture of some medical devices that require gas or liquid to operate. He found the nail injections to be painful but was motivated to proceed with treatment. He brought in a piece of polyurethane tubing to a subsequent visit to bite on during the injections (Figure) and reported considerable relief of pain.

What you were not taught in United States history class was that this method—clenching an object orally—dates to the era before the Civil War, before appropriate anesthetics and analgesics were developed, when patients and soldiers bit on a bullet or leather strap during surgical procedures.⁴ Clenching and chewing have been shown to promote relaxation and reduce acute pain and stress.⁵

Practical Implications

Polyurethane tubing can be purchased in bulk, is inexpensive ($0.30/foot on Amazon), and unlikely to damage teeth due to its flexibility. It can be cut into 6-inch pieces and given to the patient at their first nail injection appointment. The patient can then bring the tubing to subsequent appointments to use as a mastication tool during nail injections.

We instruct the patient to disinfect the dedicated piece of tubing after the initial visit and each subsequent visit by soaking it for 15 minutes in either a 3% hydrogen peroxide solution, antibacterial mouthwash, a solution of baking soda (bicarbonate of soda) and water (1 cup of water to 2 teaspoons of baking soda), or white vinegar. We instruct them to thoroughly dry the disinfected polyurethane tube and store it in a clean, reusable, resealable zipper storage bag between appointments.

In addition to reducing anxiety and pain, this method also distracts the patient and therefore promotes patient and physician safety. Patients are less likely to jump or startle during the injection, thereby reducing the risk of physically interfering with the nail surgeon or making an unanticipated advance into the surgical field.

Although frustrated patients with nail disease may need to “bite the bullet” when they accept treatment with nail injections, lessons from our patient and from United States history offer a safe and cost-effective pain management strategy. Minimizing discomfort and anxiety during the first nail injection is crucial because doing so is likely to promote adherence with follow-up injections and therefore improve clinical outcomes.

Future clinical studies should validate the clinical utility of oral mastication and clenching during nail procedures compared to other perioperative stress- and anxiety-reducing techniques.

Practice Gap

Nail matrix and nail bed injections with triamcinolone acetonide are used to treat trachyonychia and inflammatory nail conditions, including nail psoriasis and nail lichen planus. The procedure should be quick in well-trained hands, with each nail injection taking only seconds to perform. Typically, patients have multiple nails involved, requiring at least 1 injection into the nail matrix or the nail bed (or both) in each nail at each visit. Patients often are anxious when undergoing nail injections; the nail unit is highly innervated and vascular, which can cause notable transient discomfort during the procedure^1,2 as well as postoperative pain.³

Nail injections must be repeated every 4 to 6 weeks to sustain clinical benefit and maximize outcomes, which can lead to heightened anxiety and apprehension before and during the visit. Furthermore, pain and anxiety associated with the procedure may deter patients from returning for follow-up injections, which can impact treatment adherence and clinical outcomes.

Dermatologists should implement strategies to decrease periprocedural anxiety to improve the nail injection experience. In our practice, we routinely incorporate stress-reducing techniques—music, talkesthesia, a sleep mask, cool air, ethyl chloride, and squeezing a stress ball—into the clinical workflow of the procedure. The goal of these techniques is to divert attention away from painful stimuli. Most patients, however, receive injections in both hands, making it impractical to employ some of these techniques, particularly squeezing a stress ball. We employed a unique method involving polyurethane tubing to reduce stress and anxiety during nail procedures.

The Technique

A patient was receiving treatment with intralesional triamcinolone injections to the nail matrix for trachyonychia involving all of the fingernails. He worked as an equipment and facilities manager, giving him access to polyurethane tubing, which is routinely used in the manufacture of some medical devices that require gas or liquid to operate. He found the nail injections to be painful but was motivated to proceed with treatment. He brought in a piece of polyurethane tubing to a subsequent visit to bite on during the injections (Figure) and reported considerable relief of pain.

What you were not taught in United States history class was that this method—clenching an object orally—dates to the era before the Civil War, before appropriate anesthetics and analgesics were developed, when patients and soldiers bit on a bullet or leather strap during surgical procedures.⁴ Clenching and chewing have been shown to promote relaxation and reduce acute pain and stress.⁵

Practical Implications

Polyurethane tubing can be purchased in bulk, is inexpensive ($0.30/foot on Amazon), and unlikely to damage teeth due to its flexibility. It can be cut into 6-inch pieces and given to the patient at their first nail injection appointment. The patient can then bring the tubing to subsequent appointments to use as a mastication tool during nail injections.

We instruct the patient to disinfect the dedicated piece of tubing after the initial visit and each subsequent visit by soaking it for 15 minutes in either a 3% hydrogen peroxide solution, antibacterial mouthwash, a solution of baking soda (bicarbonate of soda) and water (1 cup of water to 2 teaspoons of baking soda), or white vinegar. We instruct them to thoroughly dry the disinfected polyurethane tube and store it in a clean, reusable, resealable zipper storage bag between appointments.

In addition to reducing anxiety and pain, this method also distracts the patient and therefore promotes patient and physician safety. Patients are less likely to jump or startle during the injection, thereby reducing the risk of physically interfering with the nail surgeon or making an unanticipated advance into the surgical field.

Although frustrated patients with nail disease may need to “bite the bullet” when they accept treatment with nail injections, lessons from our patient and from United States history offer a safe and cost-effective pain management strategy. Minimizing discomfort and anxiety during the first nail injection is crucial because doing so is likely to promote adherence with follow-up injections and therefore improve clinical outcomes.

Future clinical studies should validate the clinical utility of oral mastication and clenching during nail procedures compared to other perioperative stress- and anxiety-reducing techniques.

The incidence of skin cancer continues to increase, and it is by far the most common malignancy in the United States. Based on the sheer incidence and prevalence of skin cancer, early detection and treatment are critical. Looking at melanoma alone, the 5-year survival rate is greater than 99% when detected early but falls to 71% when the disease reaches the lymph nodes and 32% with metastasis to distant organs.¹ Furthermore, a 2018 study found stage I melanoma patients who were treated 4 months after biopsy had a 41% increased risk of death compared with those treated within the first month.² However, many patients are not seen by a dermatologist first for examination of suspicious skin lesions and instead are referred by a general practitioner or primary care mid-level provider. Therefore, many patients experience a longer time to diagnosis or treatment, which directly correlates with survival rate.

Dermoscopy is a noninvasive diagnostic tool for skin lesions, including melanoma. Using a handheld dermoscope (or dermatoscope), a transilluminating light source magnifies skin lesions and allows for the visualization of subsurface skin structures within the epidermis, dermoepidermal junction, and papillary dermis.³ Dermoscopy has been shown to improve a dermatologist’s accuracy in diagnosing malignant melanoma vs clinical evaluation with the unaided eye.^4,5 More recently, dermoscopy has been digitized, allowing for the collection and documentation of case photographs. Dermoscopy also has expanded past the scope of dermatologists and has become increasingly useful in primary care.⁶ Among family physicians, dermoscopy also has been shown to have a higher sensitivity for melanoma detection compared to gross examination.⁷ Therefore, both the increased diagnostic performance of malignant melanoma using a dermoscope and the expanded use of dermoscopy in medical care validate the evaluation of an artificial intelligence (AI) algorithm in diagnosing malignant melanoma using dermoscopic images.

Triage (Triage Technologies Inc) is an AI application that uses a web interface and combines a pretrained convolutional neural network (CNN) with a reinforcement learning agent as a question-answering model. The CNN algorithm can classify 133 different skin diseases, 7 of which it is able to classify using dermoscopic images. This study sought to evaluate the performance of Triage’s dermoscopic classifier in identifying lesions as benign or malignant to determine whether AI could assist in the triage of skin cancer cases to shorten time to diagnosis.

Materials and Methods

The MClass-D test set from the International Skin Imaging Collaboration was assessed by both AI and practicing medical providers. The set was composed of 80 benign nevi and 20 biopsy-verified malignant melanomas. Board-certified US dermatologists (n=23), family physicians (n=7), and primary care mid-level providers (n=12)(ie, nurse practitioners, physician assistants) were asked to label the images as benign or malignant. The results from the medical providers were then compared to the performance of the AI application by looking at the sensitivity, specificity, accuracy, positive predictive value (PPV), and negative predictive value (NPV). Statistical significance was determined with a 1 sample t test run through RStudio (Posit Software, PBC), and P<.05 was considered significant.

Results

The AI application performed extremely well in differentiating between benign nevi and malignant melanomas, with a sensitivity of 80%, specificity of 95%, accuracy of 92%, PPV of 80%, and NPV of 95% (Table 1). When compared with practicing medical providers, the AI performed significantly better in almost all categories (P<.05)(Figure 1). With all medical providers combined, the AI had significantly higher accuracy, sensitivity, and specificity (P<.05). The accuracy of the individual medical providers ranged from 32% to 78%.

Compared with dermatologists, the AI was significantly more specific and accurate and demonstrated a higher PPV and NPV (P<.05). There was no significant difference between the AI and dermatologists in sensitivity or labeling the true malignant lesions as malignant. The dermatologists who participated had been practicing from 1.5 years to 44 years, with an average of 16 years of dermatologic experience. There was no correlation between years practicing and performance in determining the malignancy of lesions. Of 14 dermatologists, dermoscopy was used daily by 10 and occasionally by 3, but only 6 dermatologists had any formal training. Dermatologists who used dermoscopy averaged 11 years of use.

The AI also performed significantly better than the primary care providers, including both family physicians and mid-level providers (P<.05). With the family physicians and mid-level provider scores combined, the AI showed a statistically significantly better performance in all categories examined, including sensitivity, specificity, accuracy, PPV, and NPV (P<.05). However, when compared with family physicians alone, the AI did not demonstrate a statistically significant difference in sensitivity.

 

 

Comment

Automatic Visual Recognition Development—The AI application we studied was developed by dermatologists as a tool to assist in the screening of skin lesions suspicious for melanoma or a benign neoplasm.⁸ Developing AI applications that can reliably recognize objects in photographs has been the subject of considerable research. Notable progress in automatic visual recognition was shown in 2012 when a deep learning model won the ImageNet object recognition challenge and outperformed competing approaches by a large margin.^9,10 The ImageNet competition, which has been held annually since 2010, required participants to build a visual classification system that distinguished among 1000 object categories using 1.2 million labeled images as training data. In 2017, participants developed automated visual systems that surpassed the estimated human performance.¹¹ Given this success, the organization decided to deliver a more challenging competition involving 3D imaging—Medical ImageNet, a petabyte-scale, cloud-based, open repository project—with goals including image classification and annotation.¹²

Convolutional Neural Networks—Convolutional neural networks are computer system architectures commonly employed for making predictions from images.¹³ Convolutional neural networks are based on a set of layers of learned filters that perform convolution, a mathematical operation that reflects the relationship between the 2 functions. The main algorithm that makes the learning possible is called backpropagation, wherein an error is computed at the output and distributed backward through the neural network’s layers.¹⁴ Although CNNs and backpropagation methods have existed since 1989, recent technologic advances have allowed for deep learning–based algorithms to be widely integrated with everyday applications.¹⁵ Advances in computational power in the form of graphics processing units and parallelization, the existence of large data sets such as the ImageNet database, and the rise of software frameworks have allowed for quick prototyping and deployment of deep learning models.^16,17

Convolutional neural networks have demonstrated potential to excel at a wide range of visual tasks. In dermatology, visual recognition methods often rely on using either a pretrained CNN as a feature extractor for further classification or fine-tuning a pretrained network on dermoscopic images.^18-20 In 2017, a model was trained on 130,000 clinical images of benign and malignant skin lesions. Its performance was found to be in line with that of 21 US board-certified dermatology experts when diagnosing skin cancers from clinical images confirmed by biopsy.²¹

Triage—The AI application Triage is composed of several components contained in a web interface (Figure 2). To use the interface, the user must sign up and upload a photograph to the website. The image first passes through a gated-logic visual classifier that rejects any images that do not contain a visible skin condition. If the image contains a skin condition, the image is passed to a skin classifier that predicts the probability of the image containing 1 of 133 classes of skin conditions, 7 of which the application can diagnose with a dermoscopic image.

Image courtesy of Triage Technologies Inc and Izhaar Tejani, BA (Toronto, Ontario, Canada).

The AI application uses several techniques when training a CNN model. To address skin condition class imbalances (when more examples exist for 1 class than the others) in the training data, additional weights are applied to mistakes made on underrepresented classes, which encourages the model to better detect cases with low prevalence in the data set. Data augmentation techniques such as rotating, zooming, and flipping the training images are applied to allow the model to become more familiar with variability in the input images. Convolutional neural networks are trained using a well-known neural network optimization method called Stochastic gradient descent with momentum.²²

The final predictions are refined by a question-and-answer system that encodes dermatology knowledge and is currently under active development. Finally, the top k most probable conditions are displayed to the user, where k≤5. An initial prototype of the system was described in a published research paper in the 2019 medical imaging workshop of the Neural Information Systems conference.²³

The prototype demonstrated that combining a pretrained CNN with a reinforcement learning agent as a question-answering model increased the classification confidence and accuracy of its visual symptom checker and decreased the average number of questions asked to narrow down the differential diagnosis. The reinforcement learning approach increases the accuracy more than 20% compared with the CNN-only approach, which only uses visual information to predict the condition.²³

This application’s current visual question-answering system is trained on a diverse set of data that includes more than 20 years of clinical encounters and user-uploaded cases submitted by more than 150,000 patients and 10,000 clinicians in more than 150 countries. All crowdsourced images used for training the dermoscopy classifier are biopsy-verified images contributed by dermatologists. These data are made up of case photographs that are tagged with metadata around the patient’s age, sex, symptoms, and diagnoses. The CNN algorithm used covers 133 skin disease classes, representing 588 clinical conditions. It also can automatically detect 7 malignant, premalignant, and benign dermoscopic categories, which is the focus of this study (Table 2). Diagnoses are verified by patient response to treatment, biopsy results, and dermatologist consensus.

In addition to having improved performance, supporting more than 130 disease classes, and having a diverse data set, the application used has beat competing technologies.^20,24 The application currently is available on the internet in more than 30 countries after it received Health Canada Class I medical device approval and the CE mark in Europe.

Can AI Reliably Detect Melanoma?—In our study, of the lesions labeled benign, the higher PPV and NPV of the AI algorithm means that the lesions were more reliably true benign lesions, and the lesions labeled as malignant were more likely to be true malignant lesions. Therefore, the diagnosis given by the AI compared with the medical provider was significantly more likely to be correct. These findings demonstrate that this AI application can reliably detect malignant melanoma using dermoscopic images. However, this study was limited by the small sample size of medical providers. Further studies are necessary to assess whether the high diagnostic accuracy of the application translates to expedited referrals and a decrease in unnecessary biopsies.

Dermoscopy Training—This study looked at dermoscopic images instead of gross examination, as is often done in clinic, which draws into question the dermoscopic training dermatologists receive. The diagnostic accuracy using dermoscopic images has been shown to be higher than evaluation with the naked eye.^5,6 However, there currently is no standard for dermoscopic training in dermatology residencies, and education varies widely.²⁵ These data suggest that there may be a lack of dermoscopic training among dermatologists, which could accentuate the difference in performance between dermatologists and AI. Most primary care providers also lack formal dermoscopy training. Although dermoscopy has been shown to increase the diagnostic efficacy of primary care providers, this increase does not become apparent until the medical provider has had years of formal training in addition to clinical experience, which is not commonly provided in the medical training that primary care providers receive.^8,26

Conclusion

It is anticipated that AI will shape the future of medicine and become incorporated into daily practice.²⁷ Artificial intelligence will not replace physicians but rather assist clinicians and help to streamline medical care. Clinicians will take on the role of interpreting AI output and integrate it into patient care. With this advancement, it is important to highlight that for AI to improve the quality, efficiency, and accessibility of health care, clinicians must be equipped with the right training.^27-29

The incidence of skin cancer continues to increase, and it is by far the most common malignancy in the United States. Based on the sheer incidence and prevalence of skin cancer, early detection and treatment are critical. Looking at melanoma alone, the 5-year survival rate is greater than 99% when detected early but falls to 71% when the disease reaches the lymph nodes and 32% with metastasis to distant organs.¹ Furthermore, a 2018 study found stage I melanoma patients who were treated 4 months after biopsy had a 41% increased risk of death compared with those treated within the first month.² However, many patients are not seen by a dermatologist first for examination of suspicious skin lesions and instead are referred by a general practitioner or primary care mid-level provider. Therefore, many patients experience a longer time to diagnosis or treatment, which directly correlates with survival rate.

Dermoscopy is a noninvasive diagnostic tool for skin lesions, including melanoma. Using a handheld dermoscope (or dermatoscope), a transilluminating light source magnifies skin lesions and allows for the visualization of subsurface skin structures within the epidermis, dermoepidermal junction, and papillary dermis.³ Dermoscopy has been shown to improve a dermatologist’s accuracy in diagnosing malignant melanoma vs clinical evaluation with the unaided eye.^4,5 More recently, dermoscopy has been digitized, allowing for the collection and documentation of case photographs. Dermoscopy also has expanded past the scope of dermatologists and has become increasingly useful in primary care.⁶ Among family physicians, dermoscopy also has been shown to have a higher sensitivity for melanoma detection compared to gross examination.⁷ Therefore, both the increased diagnostic performance of malignant melanoma using a dermoscope and the expanded use of dermoscopy in medical care validate the evaluation of an artificial intelligence (AI) algorithm in diagnosing malignant melanoma using dermoscopic images.

Triage (Triage Technologies Inc) is an AI application that uses a web interface and combines a pretrained convolutional neural network (CNN) with a reinforcement learning agent as a question-answering model. The CNN algorithm can classify 133 different skin diseases, 7 of which it is able to classify using dermoscopic images. This study sought to evaluate the performance of Triage’s dermoscopic classifier in identifying lesions as benign or malignant to determine whether AI could assist in the triage of skin cancer cases to shorten time to diagnosis.

Materials and Methods

The MClass-D test set from the International Skin Imaging Collaboration was assessed by both AI and practicing medical providers. The set was composed of 80 benign nevi and 20 biopsy-verified malignant melanomas. Board-certified US dermatologists (n=23), family physicians (n=7), and primary care mid-level providers (n=12)(ie, nurse practitioners, physician assistants) were asked to label the images as benign or malignant. The results from the medical providers were then compared to the performance of the AI application by looking at the sensitivity, specificity, accuracy, positive predictive value (PPV), and negative predictive value (NPV). Statistical significance was determined with a 1 sample t test run through RStudio (Posit Software, PBC), and P<.05 was considered significant.

Results

The AI application performed extremely well in differentiating between benign nevi and malignant melanomas, with a sensitivity of 80%, specificity of 95%, accuracy of 92%, PPV of 80%, and NPV of 95% (Table 1). When compared with practicing medical providers, the AI performed significantly better in almost all categories (P<.05)(Figure 1). With all medical providers combined, the AI had significantly higher accuracy, sensitivity, and specificity (P<.05). The accuracy of the individual medical providers ranged from 32% to 78%.

Compared with dermatologists, the AI was significantly more specific and accurate and demonstrated a higher PPV and NPV (P<.05). There was no significant difference between the AI and dermatologists in sensitivity or labeling the true malignant lesions as malignant. The dermatologists who participated had been practicing from 1.5 years to 44 years, with an average of 16 years of dermatologic experience. There was no correlation between years practicing and performance in determining the malignancy of lesions. Of 14 dermatologists, dermoscopy was used daily by 10 and occasionally by 3, but only 6 dermatologists had any formal training. Dermatologists who used dermoscopy averaged 11 years of use.

The AI also performed significantly better than the primary care providers, including both family physicians and mid-level providers (P<.05). With the family physicians and mid-level provider scores combined, the AI showed a statistically significantly better performance in all categories examined, including sensitivity, specificity, accuracy, PPV, and NPV (P<.05). However, when compared with family physicians alone, the AI did not demonstrate a statistically significant difference in sensitivity.

 

 

Comment

Automatic Visual Recognition Development—The AI application we studied was developed by dermatologists as a tool to assist in the screening of skin lesions suspicious for melanoma or a benign neoplasm.⁸ Developing AI applications that can reliably recognize objects in photographs has been the subject of considerable research. Notable progress in automatic visual recognition was shown in 2012 when a deep learning model won the ImageNet object recognition challenge and outperformed competing approaches by a large margin.^9,10 The ImageNet competition, which has been held annually since 2010, required participants to build a visual classification system that distinguished among 1000 object categories using 1.2 million labeled images as training data. In 2017, participants developed automated visual systems that surpassed the estimated human performance.¹¹ Given this success, the organization decided to deliver a more challenging competition involving 3D imaging—Medical ImageNet, a petabyte-scale, cloud-based, open repository project—with goals including image classification and annotation.¹²

Convolutional Neural Networks—Convolutional neural networks are computer system architectures commonly employed for making predictions from images.¹³ Convolutional neural networks are based on a set of layers of learned filters that perform convolution, a mathematical operation that reflects the relationship between the 2 functions. The main algorithm that makes the learning possible is called backpropagation, wherein an error is computed at the output and distributed backward through the neural network’s layers.¹⁴ Although CNNs and backpropagation methods have existed since 1989, recent technologic advances have allowed for deep learning–based algorithms to be widely integrated with everyday applications.¹⁵ Advances in computational power in the form of graphics processing units and parallelization, the existence of large data sets such as the ImageNet database, and the rise of software frameworks have allowed for quick prototyping and deployment of deep learning models.^16,17

Convolutional neural networks have demonstrated potential to excel at a wide range of visual tasks. In dermatology, visual recognition methods often rely on using either a pretrained CNN as a feature extractor for further classification or fine-tuning a pretrained network on dermoscopic images.^18-20 In 2017, a model was trained on 130,000 clinical images of benign and malignant skin lesions. Its performance was found to be in line with that of 21 US board-certified dermatology experts when diagnosing skin cancers from clinical images confirmed by biopsy.²¹

Triage—The AI application Triage is composed of several components contained in a web interface (Figure 2). To use the interface, the user must sign up and upload a photograph to the website. The image first passes through a gated-logic visual classifier that rejects any images that do not contain a visible skin condition. If the image contains a skin condition, the image is passed to a skin classifier that predicts the probability of the image containing 1 of 133 classes of skin conditions, 7 of which the application can diagnose with a dermoscopic image.

Image courtesy of Triage Technologies Inc and Izhaar Tejani, BA (Toronto, Ontario, Canada).

The AI application uses several techniques when training a CNN model. To address skin condition class imbalances (when more examples exist for 1 class than the others) in the training data, additional weights are applied to mistakes made on underrepresented classes, which encourages the model to better detect cases with low prevalence in the data set. Data augmentation techniques such as rotating, zooming, and flipping the training images are applied to allow the model to become more familiar with variability in the input images. Convolutional neural networks are trained using a well-known neural network optimization method called Stochastic gradient descent with momentum.²²

The final predictions are refined by a question-and-answer system that encodes dermatology knowledge and is currently under active development. Finally, the top k most probable conditions are displayed to the user, where k≤5. An initial prototype of the system was described in a published research paper in the 2019 medical imaging workshop of the Neural Information Systems conference.²³

The prototype demonstrated that combining a pretrained CNN with a reinforcement learning agent as a question-answering model increased the classification confidence and accuracy of its visual symptom checker and decreased the average number of questions asked to narrow down the differential diagnosis. The reinforcement learning approach increases the accuracy more than 20% compared with the CNN-only approach, which only uses visual information to predict the condition.²³

This application’s current visual question-answering system is trained on a diverse set of data that includes more than 20 years of clinical encounters and user-uploaded cases submitted by more than 150,000 patients and 10,000 clinicians in more than 150 countries. All crowdsourced images used for training the dermoscopy classifier are biopsy-verified images contributed by dermatologists. These data are made up of case photographs that are tagged with metadata around the patient’s age, sex, symptoms, and diagnoses. The CNN algorithm used covers 133 skin disease classes, representing 588 clinical conditions. It also can automatically detect 7 malignant, premalignant, and benign dermoscopic categories, which is the focus of this study (Table 2). Diagnoses are verified by patient response to treatment, biopsy results, and dermatologist consensus.

In addition to having improved performance, supporting more than 130 disease classes, and having a diverse data set, the application used has beat competing technologies.^20,24 The application currently is available on the internet in more than 30 countries after it received Health Canada Class I medical device approval and the CE mark in Europe.

Can AI Reliably Detect Melanoma?—In our study, of the lesions labeled benign, the higher PPV and NPV of the AI algorithm means that the lesions were more reliably true benign lesions, and the lesions labeled as malignant were more likely to be true malignant lesions. Therefore, the diagnosis given by the AI compared with the medical provider was significantly more likely to be correct. These findings demonstrate that this AI application can reliably detect malignant melanoma using dermoscopic images. However, this study was limited by the small sample size of medical providers. Further studies are necessary to assess whether the high diagnostic accuracy of the application translates to expedited referrals and a decrease in unnecessary biopsies.

Dermoscopy Training—This study looked at dermoscopic images instead of gross examination, as is often done in clinic, which draws into question the dermoscopic training dermatologists receive. The diagnostic accuracy using dermoscopic images has been shown to be higher than evaluation with the naked eye.^5,6 However, there currently is no standard for dermoscopic training in dermatology residencies, and education varies widely.²⁵ These data suggest that there may be a lack of dermoscopic training among dermatologists, which could accentuate the difference in performance between dermatologists and AI. Most primary care providers also lack formal dermoscopy training. Although dermoscopy has been shown to increase the diagnostic efficacy of primary care providers, this increase does not become apparent until the medical provider has had years of formal training in addition to clinical experience, which is not commonly provided in the medical training that primary care providers receive.^8,26

Conclusion

It is anticipated that AI will shape the future of medicine and become incorporated into daily practice.²⁷ Artificial intelligence will not replace physicians but rather assist clinicians and help to streamline medical care. Clinicians will take on the role of interpreting AI output and integrate it into patient care. With this advancement, it is important to highlight that for AI to improve the quality, efficiency, and accessibility of health care, clinicians must be equipped with the right training.^27-29

The incidence of skin cancer continues to increase, and it is by far the most common malignancy in the United States. Based on the sheer incidence and prevalence of skin cancer, early detection and treatment are critical. Looking at melanoma alone, the 5-year survival rate is greater than 99% when detected early but falls to 71% when the disease reaches the lymph nodes and 32% with metastasis to distant organs.¹ Furthermore, a 2018 study found stage I melanoma patients who were treated 4 months after biopsy had a 41% increased risk of death compared with those treated within the first month.² However, many patients are not seen by a dermatologist first for examination of suspicious skin lesions and instead are referred by a general practitioner or primary care mid-level provider. Therefore, many patients experience a longer time to diagnosis or treatment, which directly correlates with survival rate.

Dermoscopy is a noninvasive diagnostic tool for skin lesions, including melanoma. Using a handheld dermoscope (or dermatoscope), a transilluminating light source magnifies skin lesions and allows for the visualization of subsurface skin structures within the epidermis, dermoepidermal junction, and papillary dermis.³ Dermoscopy has been shown to improve a dermatologist’s accuracy in diagnosing malignant melanoma vs clinical evaluation with the unaided eye.^4,5 More recently, dermoscopy has been digitized, allowing for the collection and documentation of case photographs. Dermoscopy also has expanded past the scope of dermatologists and has become increasingly useful in primary care.⁶ Among family physicians, dermoscopy also has been shown to have a higher sensitivity for melanoma detection compared to gross examination.⁷ Therefore, both the increased diagnostic performance of malignant melanoma using a dermoscope and the expanded use of dermoscopy in medical care validate the evaluation of an artificial intelligence (AI) algorithm in diagnosing malignant melanoma using dermoscopic images.

Triage (Triage Technologies Inc) is an AI application that uses a web interface and combines a pretrained convolutional neural network (CNN) with a reinforcement learning agent as a question-answering model. The CNN algorithm can classify 133 different skin diseases, 7 of which it is able to classify using dermoscopic images. This study sought to evaluate the performance of Triage’s dermoscopic classifier in identifying lesions as benign or malignant to determine whether AI could assist in the triage of skin cancer cases to shorten time to diagnosis.

Materials and Methods

The MClass-D test set from the International Skin Imaging Collaboration was assessed by both AI and practicing medical providers. The set was composed of 80 benign nevi and 20 biopsy-verified malignant melanomas. Board-certified US dermatologists (n=23), family physicians (n=7), and primary care mid-level providers (n=12)(ie, nurse practitioners, physician assistants) were asked to label the images as benign or malignant. The results from the medical providers were then compared to the performance of the AI application by looking at the sensitivity, specificity, accuracy, positive predictive value (PPV), and negative predictive value (NPV). Statistical significance was determined with a 1 sample t test run through RStudio (Posit Software, PBC), and P<.05 was considered significant.

Results

The AI application performed extremely well in differentiating between benign nevi and malignant melanomas, with a sensitivity of 80%, specificity of 95%, accuracy of 92%, PPV of 80%, and NPV of 95% (Table 1). When compared with practicing medical providers, the AI performed significantly better in almost all categories (P<.05)(Figure 1). With all medical providers combined, the AI had significantly higher accuracy, sensitivity, and specificity (P<.05). The accuracy of the individual medical providers ranged from 32% to 78%.

Compared with dermatologists, the AI was significantly more specific and accurate and demonstrated a higher PPV and NPV (P<.05). There was no significant difference between the AI and dermatologists in sensitivity or labeling the true malignant lesions as malignant. The dermatologists who participated had been practicing from 1.5 years to 44 years, with an average of 16 years of dermatologic experience. There was no correlation between years practicing and performance in determining the malignancy of lesions. Of 14 dermatologists, dermoscopy was used daily by 10 and occasionally by 3, but only 6 dermatologists had any formal training. Dermatologists who used dermoscopy averaged 11 years of use.

The AI also performed significantly better than the primary care providers, including both family physicians and mid-level providers (P<.05). With the family physicians and mid-level provider scores combined, the AI showed a statistically significantly better performance in all categories examined, including sensitivity, specificity, accuracy, PPV, and NPV (P<.05). However, when compared with family physicians alone, the AI did not demonstrate a statistically significant difference in sensitivity.

 

 

Comment

Automatic Visual Recognition Development—The AI application we studied was developed by dermatologists as a tool to assist in the screening of skin lesions suspicious for melanoma or a benign neoplasm.⁸ Developing AI applications that can reliably recognize objects in photographs has been the subject of considerable research. Notable progress in automatic visual recognition was shown in 2012 when a deep learning model won the ImageNet object recognition challenge and outperformed competing approaches by a large margin.^9,10 The ImageNet competition, which has been held annually since 2010, required participants to build a visual classification system that distinguished among 1000 object categories using 1.2 million labeled images as training data. In 2017, participants developed automated visual systems that surpassed the estimated human performance.¹¹ Given this success, the organization decided to deliver a more challenging competition involving 3D imaging—Medical ImageNet, a petabyte-scale, cloud-based, open repository project—with goals including image classification and annotation.¹²

Convolutional Neural Networks—Convolutional neural networks are computer system architectures commonly employed for making predictions from images.¹³ Convolutional neural networks are based on a set of layers of learned filters that perform convolution, a mathematical operation that reflects the relationship between the 2 functions. The main algorithm that makes the learning possible is called backpropagation, wherein an error is computed at the output and distributed backward through the neural network’s layers.¹⁴ Although CNNs and backpropagation methods have existed since 1989, recent technologic advances have allowed for deep learning–based algorithms to be widely integrated with everyday applications.¹⁵ Advances in computational power in the form of graphics processing units and parallelization, the existence of large data sets such as the ImageNet database, and the rise of software frameworks have allowed for quick prototyping and deployment of deep learning models.^16,17

Convolutional neural networks have demonstrated potential to excel at a wide range of visual tasks. In dermatology, visual recognition methods often rely on using either a pretrained CNN as a feature extractor for further classification or fine-tuning a pretrained network on dermoscopic images.^18-20 In 2017, a model was trained on 130,000 clinical images of benign and malignant skin lesions. Its performance was found to be in line with that of 21 US board-certified dermatology experts when diagnosing skin cancers from clinical images confirmed by biopsy.²¹

Triage—The AI application Triage is composed of several components contained in a web interface (Figure 2). To use the interface, the user must sign up and upload a photograph to the website. The image first passes through a gated-logic visual classifier that rejects any images that do not contain a visible skin condition. If the image contains a skin condition, the image is passed to a skin classifier that predicts the probability of the image containing 1 of 133 classes of skin conditions, 7 of which the application can diagnose with a dermoscopic image.

Image courtesy of Triage Technologies Inc and Izhaar Tejani, BA (Toronto, Ontario, Canada).

The AI application uses several techniques when training a CNN model. To address skin condition class imbalances (when more examples exist for 1 class than the others) in the training data, additional weights are applied to mistakes made on underrepresented classes, which encourages the model to better detect cases with low prevalence in the data set. Data augmentation techniques such as rotating, zooming, and flipping the training images are applied to allow the model to become more familiar with variability in the input images. Convolutional neural networks are trained using a well-known neural network optimization method called Stochastic gradient descent with momentum.²²

The final predictions are refined by a question-and-answer system that encodes dermatology knowledge and is currently under active development. Finally, the top k most probable conditions are displayed to the user, where k≤5. An initial prototype of the system was described in a published research paper in the 2019 medical imaging workshop of the Neural Information Systems conference.²³

The prototype demonstrated that combining a pretrained CNN with a reinforcement learning agent as a question-answering model increased the classification confidence and accuracy of its visual symptom checker and decreased the average number of questions asked to narrow down the differential diagnosis. The reinforcement learning approach increases the accuracy more than 20% compared with the CNN-only approach, which only uses visual information to predict the condition.²³

This application’s current visual question-answering system is trained on a diverse set of data that includes more than 20 years of clinical encounters and user-uploaded cases submitted by more than 150,000 patients and 10,000 clinicians in more than 150 countries. All crowdsourced images used for training the dermoscopy classifier are biopsy-verified images contributed by dermatologists. These data are made up of case photographs that are tagged with metadata around the patient’s age, sex, symptoms, and diagnoses. The CNN algorithm used covers 133 skin disease classes, representing 588 clinical conditions. It also can automatically detect 7 malignant, premalignant, and benign dermoscopic categories, which is the focus of this study (Table 2). Diagnoses are verified by patient response to treatment, biopsy results, and dermatologist consensus.

In addition to having improved performance, supporting more than 130 disease classes, and having a diverse data set, the application used has beat competing technologies.^20,24 The application currently is available on the internet in more than 30 countries after it received Health Canada Class I medical device approval and the CE mark in Europe.

Can AI Reliably Detect Melanoma?—In our study, of the lesions labeled benign, the higher PPV and NPV of the AI algorithm means that the lesions were more reliably true benign lesions, and the lesions labeled as malignant were more likely to be true malignant lesions. Therefore, the diagnosis given by the AI compared with the medical provider was significantly more likely to be correct. These findings demonstrate that this AI application can reliably detect malignant melanoma using dermoscopic images. However, this study was limited by the small sample size of medical providers. Further studies are necessary to assess whether the high diagnostic accuracy of the application translates to expedited referrals and a decrease in unnecessary biopsies.

Dermoscopy Training—This study looked at dermoscopic images instead of gross examination, as is often done in clinic, which draws into question the dermoscopic training dermatologists receive. The diagnostic accuracy using dermoscopic images has been shown to be higher than evaluation with the naked eye.^5,6 However, there currently is no standard for dermoscopic training in dermatology residencies, and education varies widely.²⁵ These data suggest that there may be a lack of dermoscopic training among dermatologists, which could accentuate the difference in performance between dermatologists and AI. Most primary care providers also lack formal dermoscopy training. Although dermoscopy has been shown to increase the diagnostic efficacy of primary care providers, this increase does not become apparent until the medical provider has had years of formal training in addition to clinical experience, which is not commonly provided in the medical training that primary care providers receive.^8,26

Conclusion

It is anticipated that AI will shape the future of medicine and become incorporated into daily practice.²⁷ Artificial intelligence will not replace physicians but rather assist clinicians and help to streamline medical care. Clinicians will take on the role of interpreting AI output and integrate it into patient care. With this advancement, it is important to highlight that for AI to improve the quality, efficiency, and accessibility of health care, clinicians must be equipped with the right training.^27-29

Cutaneous angiosarcoma (cAS) is a rare malignancy arising from vascular or lymphatic tissue. It classically presents during the sixth or seventh decades of life as a raised purple papule or plaque on the head and neck areas.¹ Primary cAS frequently mimics benign conditions, leading to delays in care. Such delays coupled with the aggressive nature of angiosarcomas leads to a poor prognosis. Five-year survival rates range from 11% to 50%, and more than half of patients die within 1 year of diagnosis.^2-7

Currently, there is no consensus on the most effective treatments, as the rare nature of cAS has made the development of controlled clinical trials difficult. Wide local excision (WLE) is most frequently employed; however, the tumor’s infiltrative growth makes complete resection and negative surgical margins difficult to achieve.⁸ Recently, Mohs micrographic surgery (MMS) has been postulated as a treatment option. The tissue-sparing nature and intraoperative margin control of MMS may provide tumor eradication and cosmesis benefits reported with other cutaneous malignancies.⁹

Nearly all localized cASs are treated with surgical excision with or without adjuvant treatment modalities; however, it is unclear which of these modalities provide a survival benefit. We conducted a systematic review of the literature to compare treatment modalities for localized cAS of the head and neck regions and to compare treatments based on tumor stage.

METHODS

A literature search was performed to identify published studies indexed by MEDLINE, Cochrane Central Register of Controlled Trials (CENTRAL), Embase, and PubMed from January 1, 1977, to May 8, 2020, reporting on cAS and treatment modalities used. The search was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines.⁵ Data extracted included patient demographics, tumor characteristics (including T1 [≤5 cm] and T2 [>5 cm and ≤10 cm] based on the American Joint Committee on Cancer soft tissue sarcoma staging criteria), treatments used, follow-up time, overall survival (OS) rates, and complications.^10,11

Studies were required to (1) include participants with head and neck cAS; (2) report original patient data following cAS treatment with surgical (WLE or MMS) and/or nonsurgical modalities (chemotherapy [CT], radiotherapy [RT], immunotherapy [IT]); (3) report outcome data related to OS rates following treatment; and (4) have articles published in English. Given the rare nature of cAS, there was no limitation on the number of participants needed.

The Newcastle-Ottawa scale for observational studies was used to assess the quality of studies.¹² Higher scores indicate low risk of bias, while lower scores represent high risk of bias.

Continuous data were reported with means and SDs, while categorical variables were reported as percentages. Overall survival means and SDs were compared between treatment modalities using an independent sample t test with P<.05 considered statistically significant. Due to the heterogeneity of the data, a meta-analysis was not reported.

RESULTS

Literature Search and Risk of Bias Assessment

There were 283 manuscripts identified, 56 articles read in full, and 40 articles included in the review (Figure). Among the 16 studies not meeting inclusion criteria, 7 did not provide enough data to isolate head and neck cAS cases,^1,13-18 6 did not report outcomes related to the current review,^19-24 and 3 did not provide enough data to isolate different treatment outcomes.^25-27 Among the included studies, 32 reported use of WLE: WLE alone (n=21)^2,7,11,28-45; WLE with RT (n=24)^{2,3,11,28-31,33-36,38-41,43-51}; WLE with CT (n=7)^{2,31,35,39,41,48,52}; WLE with RT and CT (n=11)^{2,29,31,33-35,39,40,48,52,53}; WLE with RT and IT (n=3)^35,54,55; and WLE with RT, CT, and IT (n=1).⁵³ Nine studies reported MMS: MMS alone (n=5)^39,56-59; MMS with RT (n=3)^32,50,60,61; and MMS with RT and CT (n=1).⁵¹

Risk of bias assessment identified low risk in 3 articles. High risk was identified in 5 case reports,^57-61 and 1 study did not describe patient selection.⁴³ Clayton et al⁵⁶ showed intermediate risk, given the study controlled for 1 factor.

Patient Demographics

A total of 1295 patients were included. The pooled mean age of the patients was 67.5 years (range, 3–88 years), and 64.7% were male. There were 79 cases identified as T1 and 105 as T2. A total of 825 cases were treated using WLE with or without adjuvant therapy, while a total of 9 cases were treated using MMS with and without adjuvant therapies (Table). There were 461 cases treated without surgical excision: RT alone (n=261), CT alone (n=38), IT alone (n=35), RT with CT (n=81), RT with IT (n=34), and RT with CT and IT (n=12)(Table). The median follow-up period across all studies was 23.5 months (range, 1–228 months).

Comparison Between Surgical and Nonsurgical Modalities

Wide Local Excision—Wide local excision (n=825; 63.7%) alone or in combination with other therapies was the most frequently used treatment modality. The mean (SD) OS was longest for WLE with RT, CT, and IT (n=3; 39.3 [24.1]), followed by WLE with RT (n=447; 35.9 [34.3] months), WLE with CT (n=13; 32.4 [30.2] months), WLE alone (n=324; 29.6 [34.1] months), WLE with RT and IT (n=11; 23.5 [4.9] months), and WLE with RT and CT (n=27; 20.7 [13.1] months).

Nonsurgical Modalities—Nonsurgical methods were used less frequently than surgical methods (n=461; 35.6%). The mean (SD) OS time in descending order was as follows: RT with CT and IT (n=12; 34.9 [1.2] months), RT with CT (n=81; 30.4 [37.8] months), IT alone (n=35; 25.7 [no SD reported] months), RT with IT (n=34; 20.5 [8.6] months), CT alone (n=38; 20.1 [15.9] months), and RT alone (n=261; 12.8 [8.3] months).

When comparing mean (SD) OS outcomes between surgical and nonsurgical treatment modalities, only the addition of WLE to RT significantly increased OS when compared with RT alone (WLE, 35.9 [34.3] months; RT alone, 12.8 [8.3] months; P=.001). When WLE was added to CT or both RT and CT, there was no significant difference with OS when compared with CT alone (WLE with CT, 32.4 [30.2] months; CT alone, 20.1 [15.9] months; P=.065); or both RT and CT in combination (WLE with RT and CT, 20.7 [13.1] months; RT and CT, 30.4 [37.8] months; P=.204).

Comparison Between T1 and T2 cAS

T1 Angiosarcoma—There were 79 patients identified as having T1 tumors across 16 studies.^{2,31,32,34,39-41,46,48-50,53,58-60,62} The mean (SD) OS was longest for WLE with RT, CT, and IT (n=2; 56.0 [6.0] months), followed by WLE with CT (n=4; 54.5 [41.0] months); WLE with RT (n=30; 39.7 [41.2] months); WLE alone (n=22; 37.2 [37.3] months); WLE with both RT and CT (n=7; 25.5 [18.7] months); RT with IT (n=2; 20.0 [11.0] months); RT with CT (n=6; 15.7 [6.8] months); and RT alone (n=1; 13 [no SD]) months)(eTable).

T2 Angiosarcoma—There were 105 patients with T2 tumors in 15 studies.^{2,31,32,34,39-41,46,48-50,52,53,57,62} The mean (SD) OS for each treatment modality in descending order was as follows: RT with CT and IT (n=1; 36 [no SD reported] months); RT with CT (n=23; 34.3 [46.3] months); WLE with RT (n=21; 26.3 [23.8] months); WLE with CT (n=8; 21.5 [16.6] months); WLE alone (n=16; 19.8 [15.6] months); WLE with RT and CT (n=14; 19.2 [10.5] months); RT alone (n=17; 10.1 [5.5] months); CT alone (n=2; 6.7 [3.7] months); and WLE with RT, CT, and IT (n=1; 6.0 [no SD] months)(eTable).

Mohs Micrographic Surgery—The use of MMS was only identified in case reports or small observational studies for a total of 9 patients. Five cASs were treated with MMS alone for a mean (SD) OS of 37 (21.5) months, with 4 reporting cAS staging: 2 were T1^58,59 (mean [SD] OS, 37.0 [17.0] months) and 2 were T2 tumors^39,57 (mean [SD] OS, 44.5 [26.5] months). Mohs micrographic surgery with RT was used for 3 tumors (mean [SD] OS, 34.0 [26.9] months); 2 were T1^50,60 (mean [SD] OS, 42.0 [30.0] months) and 1 unreported staging (eTable).⁵⁶ Mohs micrographic surgery with both RT and CT was used in 1 patient (unreported staging; OS, 82 months).⁵¹

Complications

Complications were rare and mainly associated with CT and RT. Four studies reported radiation dermatitis with RT.^53,55,62,63 Two studies reported peripheral neuropathy and myelotoxicity with CT.^35,51 Only 1 study reported poor wound healing due to surgical complications.²⁹

COMMENT

Cutaneous angiosarcomas are rare and have limited treatment guidelines. Surgical excision does appear to be an effective adjunct to nonsurgical treatments, particularly WLE combined with RT, CT, and IT. Although MMS ultimately may be useful for cAS, the limited number and substantial heterogeneity of reported cases precludes definitive conclusions at this time.

Achieving margin control during WLE is associated with higher OS when treating angiosarcoma,^36,46 which is particularly true for T1 tumors where margin control is imperative, and many cases are treated with a combination of WLE and RT. Overall survival times are lower for T2 tumors, as these tumors are larger and most likely have spread; therefore, more aggressive combination treatments were more prevalent. In these cases, complete margin control may be difficult to achieve and may not be as critical to the outcome if another form of adjuvant therapy can be administered promptly.^24,64

When surgery is contraindicated, RT with or without CT was the most commonly reported treatment modality. However, these treatments were notably less effective than when used in combination with surgical resection. The use of RT alone has a recurrence rate reported up to 100% in certain studies, suggesting the need to utilize RT in combination with other modalities.^23,39 It is important to note that RT often is used as monotherapy in palliative treatment, which may indirectly skew survival rates.²

Limitations of the study include a lack of randomized controlled trials. Most reports were retrospective reviews or case series, and tumor staging was sparsely reported. Finally, although MMS may provide utility in the treatment of cAS, the sample size of 9 precluded definitive conclusions from being formed about its efficacy.

CONCLUSION

Cutaneous angiosarcoma is rare and has limited data comparing different treatment modalities. The paucity of data currently limits definitive recommendations; however, both surgical and nonsurgical modalities have demonstrated potential efficacy in the treatment of cAS and may benefit from additional research. Clinicians should consider a multidisciplinary approach for patients with a diagnosis of cAS to tailor treatments on a case-by-case basis.

Cutaneous angiosarcoma (cAS) is a rare malignancy arising from vascular or lymphatic tissue. It classically presents during the sixth or seventh decades of life as a raised purple papule or plaque on the head and neck areas.¹ Primary cAS frequently mimics benign conditions, leading to delays in care. Such delays coupled with the aggressive nature of angiosarcomas leads to a poor prognosis. Five-year survival rates range from 11% to 50%, and more than half of patients die within 1 year of diagnosis.^2-7

Currently, there is no consensus on the most effective treatments, as the rare nature of cAS has made the development of controlled clinical trials difficult. Wide local excision (WLE) is most frequently employed; however, the tumor’s infiltrative growth makes complete resection and negative surgical margins difficult to achieve.⁸ Recently, Mohs micrographic surgery (MMS) has been postulated as a treatment option. The tissue-sparing nature and intraoperative margin control of MMS may provide tumor eradication and cosmesis benefits reported with other cutaneous malignancies.⁹

Nearly all localized cASs are treated with surgical excision with or without adjuvant treatment modalities; however, it is unclear which of these modalities provide a survival benefit. We conducted a systematic review of the literature to compare treatment modalities for localized cAS of the head and neck regions and to compare treatments based on tumor stage.

METHODS

A literature search was performed to identify published studies indexed by MEDLINE, Cochrane Central Register of Controlled Trials (CENTRAL), Embase, and PubMed from January 1, 1977, to May 8, 2020, reporting on cAS and treatment modalities used. The search was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines.⁵ Data extracted included patient demographics, tumor characteristics (including T1 [≤5 cm] and T2 [>5 cm and ≤10 cm] based on the American Joint Committee on Cancer soft tissue sarcoma staging criteria), treatments used, follow-up time, overall survival (OS) rates, and complications.^10,11

Studies were required to (1) include participants with head and neck cAS; (2) report original patient data following cAS treatment with surgical (WLE or MMS) and/or nonsurgical modalities (chemotherapy [CT], radiotherapy [RT], immunotherapy [IT]); (3) report outcome data related to OS rates following treatment; and (4) have articles published in English. Given the rare nature of cAS, there was no limitation on the number of participants needed.

The Newcastle-Ottawa scale for observational studies was used to assess the quality of studies.¹² Higher scores indicate low risk of bias, while lower scores represent high risk of bias.

Continuous data were reported with means and SDs, while categorical variables were reported as percentages. Overall survival means and SDs were compared between treatment modalities using an independent sample t test with P<.05 considered statistically significant. Due to the heterogeneity of the data, a meta-analysis was not reported.

RESULTS

Literature Search and Risk of Bias Assessment

There were 283 manuscripts identified, 56 articles read in full, and 40 articles included in the review (Figure). Among the 16 studies not meeting inclusion criteria, 7 did not provide enough data to isolate head and neck cAS cases,^1,13-18 6 did not report outcomes related to the current review,^19-24 and 3 did not provide enough data to isolate different treatment outcomes.^25-27 Among the included studies, 32 reported use of WLE: WLE alone (n=21)^2,7,11,28-45; WLE with RT (n=24)^{2,3,11,28-31,33-36,38-41,43-51}; WLE with CT (n=7)^{2,31,35,39,41,48,52}; WLE with RT and CT (n=11)^{2,29,31,33-35,39,40,48,52,53}; WLE with RT and IT (n=3)^35,54,55; and WLE with RT, CT, and IT (n=1).⁵³ Nine studies reported MMS: MMS alone (n=5)^39,56-59; MMS with RT (n=3)^32,50,60,61; and MMS with RT and CT (n=1).⁵¹

Risk of bias assessment identified low risk in 3 articles. High risk was identified in 5 case reports,^57-61 and 1 study did not describe patient selection.⁴³ Clayton et al⁵⁶ showed intermediate risk, given the study controlled for 1 factor.

Patient Demographics

A total of 1295 patients were included. The pooled mean age of the patients was 67.5 years (range, 3–88 years), and 64.7% were male. There were 79 cases identified as T1 and 105 as T2. A total of 825 cases were treated using WLE with or without adjuvant therapy, while a total of 9 cases were treated using MMS with and without adjuvant therapies (Table). There were 461 cases treated without surgical excision: RT alone (n=261), CT alone (n=38), IT alone (n=35), RT with CT (n=81), RT with IT (n=34), and RT with CT and IT (n=12)(Table). The median follow-up period across all studies was 23.5 months (range, 1–228 months).

Comparison Between Surgical and Nonsurgical Modalities

Wide Local Excision—Wide local excision (n=825; 63.7%) alone or in combination with other therapies was the most frequently used treatment modality. The mean (SD) OS was longest for WLE with RT, CT, and IT (n=3; 39.3 [24.1]), followed by WLE with RT (n=447; 35.9 [34.3] months), WLE with CT (n=13; 32.4 [30.2] months), WLE alone (n=324; 29.6 [34.1] months), WLE with RT and IT (n=11; 23.5 [4.9] months), and WLE with RT and CT (n=27; 20.7 [13.1] months).

Nonsurgical Modalities—Nonsurgical methods were used less frequently than surgical methods (n=461; 35.6%). The mean (SD) OS time in descending order was as follows: RT with CT and IT (n=12; 34.9 [1.2] months), RT with CT (n=81; 30.4 [37.8] months), IT alone (n=35; 25.7 [no SD reported] months), RT with IT (n=34; 20.5 [8.6] months), CT alone (n=38; 20.1 [15.9] months), and RT alone (n=261; 12.8 [8.3] months).

When comparing mean (SD) OS outcomes between surgical and nonsurgical treatment modalities, only the addition of WLE to RT significantly increased OS when compared with RT alone (WLE, 35.9 [34.3] months; RT alone, 12.8 [8.3] months; P=.001). When WLE was added to CT or both RT and CT, there was no significant difference with OS when compared with CT alone (WLE with CT, 32.4 [30.2] months; CT alone, 20.1 [15.9] months; P=.065); or both RT and CT in combination (WLE with RT and CT, 20.7 [13.1] months; RT and CT, 30.4 [37.8] months; P=.204).

Comparison Between T1 and T2 cAS

T1 Angiosarcoma—There were 79 patients identified as having T1 tumors across 16 studies.^{2,31,32,34,39-41,46,48-50,53,58-60,62} The mean (SD) OS was longest for WLE with RT, CT, and IT (n=2; 56.0 [6.0] months), followed by WLE with CT (n=4; 54.5 [41.0] months); WLE with RT (n=30; 39.7 [41.2] months); WLE alone (n=22; 37.2 [37.3] months); WLE with both RT and CT (n=7; 25.5 [18.7] months); RT with IT (n=2; 20.0 [11.0] months); RT with CT (n=6; 15.7 [6.8] months); and RT alone (n=1; 13 [no SD]) months)(eTable).

T2 Angiosarcoma—There were 105 patients with T2 tumors in 15 studies.^{2,31,32,34,39-41,46,48-50,52,53,57,62} The mean (SD) OS for each treatment modality in descending order was as follows: RT with CT and IT (n=1; 36 [no SD reported] months); RT with CT (n=23; 34.3 [46.3] months); WLE with RT (n=21; 26.3 [23.8] months); WLE with CT (n=8; 21.5 [16.6] months); WLE alone (n=16; 19.8 [15.6] months); WLE with RT and CT (n=14; 19.2 [10.5] months); RT alone (n=17; 10.1 [5.5] months); CT alone (n=2; 6.7 [3.7] months); and WLE with RT, CT, and IT (n=1; 6.0 [no SD] months)(eTable).

Mohs Micrographic Surgery—The use of MMS was only identified in case reports or small observational studies for a total of 9 patients. Five cASs were treated with MMS alone for a mean (SD) OS of 37 (21.5) months, with 4 reporting cAS staging: 2 were T1^58,59 (mean [SD] OS, 37.0 [17.0] months) and 2 were T2 tumors^39,57 (mean [SD] OS, 44.5 [26.5] months). Mohs micrographic surgery with RT was used for 3 tumors (mean [SD] OS, 34.0 [26.9] months); 2 were T1^50,60 (mean [SD] OS, 42.0 [30.0] months) and 1 unreported staging (eTable).⁵⁶ Mohs micrographic surgery with both RT and CT was used in 1 patient (unreported staging; OS, 82 months).⁵¹

Complications

Complications were rare and mainly associated with CT and RT. Four studies reported radiation dermatitis with RT.^53,55,62,63 Two studies reported peripheral neuropathy and myelotoxicity with CT.^35,51 Only 1 study reported poor wound healing due to surgical complications.²⁹

COMMENT

Cutaneous angiosarcomas are rare and have limited treatment guidelines. Surgical excision does appear to be an effective adjunct to nonsurgical treatments, particularly WLE combined with RT, CT, and IT. Although MMS ultimately may be useful for cAS, the limited number and substantial heterogeneity of reported cases precludes definitive conclusions at this time.

Achieving margin control during WLE is associated with higher OS when treating angiosarcoma,^36,46 which is particularly true for T1 tumors where margin control is imperative, and many cases are treated with a combination of WLE and RT. Overall survival times are lower for T2 tumors, as these tumors are larger and most likely have spread; therefore, more aggressive combination treatments were more prevalent. In these cases, complete margin control may be difficult to achieve and may not be as critical to the outcome if another form of adjuvant therapy can be administered promptly.^24,64

When surgery is contraindicated, RT with or without CT was the most commonly reported treatment modality. However, these treatments were notably less effective than when used in combination with surgical resection. The use of RT alone has a recurrence rate reported up to 100% in certain studies, suggesting the need to utilize RT in combination with other modalities.^23,39 It is important to note that RT often is used as monotherapy in palliative treatment, which may indirectly skew survival rates.²

Limitations of the study include a lack of randomized controlled trials. Most reports were retrospective reviews or case series, and tumor staging was sparsely reported. Finally, although MMS may provide utility in the treatment of cAS, the sample size of 9 precluded definitive conclusions from being formed about its efficacy.

CONCLUSION

Cutaneous angiosarcoma is rare and has limited data comparing different treatment modalities. The paucity of data currently limits definitive recommendations; however, both surgical and nonsurgical modalities have demonstrated potential efficacy in the treatment of cAS and may benefit from additional research. Clinicians should consider a multidisciplinary approach for patients with a diagnosis of cAS to tailor treatments on a case-by-case basis.

Cutaneous angiosarcoma (cAS) is a rare malignancy arising from vascular or lymphatic tissue. It classically presents during the sixth or seventh decades of life as a raised purple papule or plaque on the head and neck areas.¹ Primary cAS frequently mimics benign conditions, leading to delays in care. Such delays coupled with the aggressive nature of angiosarcomas leads to a poor prognosis. Five-year survival rates range from 11% to 50%, and more than half of patients die within 1 year of diagnosis.^2-7

Currently, there is no consensus on the most effective treatments, as the rare nature of cAS has made the development of controlled clinical trials difficult. Wide local excision (WLE) is most frequently employed; however, the tumor’s infiltrative growth makes complete resection and negative surgical margins difficult to achieve.⁸ Recently, Mohs micrographic surgery (MMS) has been postulated as a treatment option. The tissue-sparing nature and intraoperative margin control of MMS may provide tumor eradication and cosmesis benefits reported with other cutaneous malignancies.⁹

Nearly all localized cASs are treated with surgical excision with or without adjuvant treatment modalities; however, it is unclear which of these modalities provide a survival benefit. We conducted a systematic review of the literature to compare treatment modalities for localized cAS of the head and neck regions and to compare treatments based on tumor stage.

METHODS

A literature search was performed to identify published studies indexed by MEDLINE, Cochrane Central Register of Controlled Trials (CENTRAL), Embase, and PubMed from January 1, 1977, to May 8, 2020, reporting on cAS and treatment modalities used. The search was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines.⁵ Data extracted included patient demographics, tumor characteristics (including T1 [≤5 cm] and T2 [>5 cm and ≤10 cm] based on the American Joint Committee on Cancer soft tissue sarcoma staging criteria), treatments used, follow-up time, overall survival (OS) rates, and complications.^10,11

Studies were required to (1) include participants with head and neck cAS; (2) report original patient data following cAS treatment with surgical (WLE or MMS) and/or nonsurgical modalities (chemotherapy [CT], radiotherapy [RT], immunotherapy [IT]); (3) report outcome data related to OS rates following treatment; and (4) have articles published in English. Given the rare nature of cAS, there was no limitation on the number of participants needed.

The Newcastle-Ottawa scale for observational studies was used to assess the quality of studies.¹² Higher scores indicate low risk of bias, while lower scores represent high risk of bias.

Continuous data were reported with means and SDs, while categorical variables were reported as percentages. Overall survival means and SDs were compared between treatment modalities using an independent sample t test with P<.05 considered statistically significant. Due to the heterogeneity of the data, a meta-analysis was not reported.

RESULTS

Literature Search and Risk of Bias Assessment

There were 283 manuscripts identified, 56 articles read in full, and 40 articles included in the review (Figure). Among the 16 studies not meeting inclusion criteria, 7 did not provide enough data to isolate head and neck cAS cases,^1,13-18 6 did not report outcomes related to the current review,^19-24 and 3 did not provide enough data to isolate different treatment outcomes.^25-27 Among the included studies, 32 reported use of WLE: WLE alone (n=21)^2,7,11,28-45; WLE with RT (n=24)^{2,3,11,28-31,33-36,38-41,43-51}; WLE with CT (n=7)^{2,31,35,39,41,48,52}; WLE with RT and CT (n=11)^{2,29,31,33-35,39,40,48,52,53}; WLE with RT and IT (n=3)^35,54,55; and WLE with RT, CT, and IT (n=1).⁵³ Nine studies reported MMS: MMS alone (n=5)^39,56-59; MMS with RT (n=3)^32,50,60,61; and MMS with RT and CT (n=1).⁵¹

Risk of bias assessment identified low risk in 3 articles. High risk was identified in 5 case reports,^57-61 and 1 study did not describe patient selection.⁴³ Clayton et al⁵⁶ showed intermediate risk, given the study controlled for 1 factor.

Patient Demographics

A total of 1295 patients were included. The pooled mean age of the patients was 67.5 years (range, 3–88 years), and 64.7% were male. There were 79 cases identified as T1 and 105 as T2. A total of 825 cases were treated using WLE with or without adjuvant therapy, while a total of 9 cases were treated using MMS with and without adjuvant therapies (Table). There were 461 cases treated without surgical excision: RT alone (n=261), CT alone (n=38), IT alone (n=35), RT with CT (n=81), RT with IT (n=34), and RT with CT and IT (n=12)(Table). The median follow-up period across all studies was 23.5 months (range, 1–228 months).

Comparison Between Surgical and Nonsurgical Modalities

Wide Local Excision—Wide local excision (n=825; 63.7%) alone or in combination with other therapies was the most frequently used treatment modality. The mean (SD) OS was longest for WLE with RT, CT, and IT (n=3; 39.3 [24.1]), followed by WLE with RT (n=447; 35.9 [34.3] months), WLE with CT (n=13; 32.4 [30.2] months), WLE alone (n=324; 29.6 [34.1] months), WLE with RT and IT (n=11; 23.5 [4.9] months), and WLE with RT and CT (n=27; 20.7 [13.1] months).

Nonsurgical Modalities—Nonsurgical methods were used less frequently than surgical methods (n=461; 35.6%). The mean (SD) OS time in descending order was as follows: RT with CT and IT (n=12; 34.9 [1.2] months), RT with CT (n=81; 30.4 [37.8] months), IT alone (n=35; 25.7 [no SD reported] months), RT with IT (n=34; 20.5 [8.6] months), CT alone (n=38; 20.1 [15.9] months), and RT alone (n=261; 12.8 [8.3] months).

When comparing mean (SD) OS outcomes between surgical and nonsurgical treatment modalities, only the addition of WLE to RT significantly increased OS when compared with RT alone (WLE, 35.9 [34.3] months; RT alone, 12.8 [8.3] months; P=.001). When WLE was added to CT or both RT and CT, there was no significant difference with OS when compared with CT alone (WLE with CT, 32.4 [30.2] months; CT alone, 20.1 [15.9] months; P=.065); or both RT and CT in combination (WLE with RT and CT, 20.7 [13.1] months; RT and CT, 30.4 [37.8] months; P=.204).

Comparison Between T1 and T2 cAS

T1 Angiosarcoma—There were 79 patients identified as having T1 tumors across 16 studies.^{2,31,32,34,39-41,46,48-50,53,58-60,62} The mean (SD) OS was longest for WLE with RT, CT, and IT (n=2; 56.0 [6.0] months), followed by WLE with CT (n=4; 54.5 [41.0] months); WLE with RT (n=30; 39.7 [41.2] months); WLE alone (n=22; 37.2 [37.3] months); WLE with both RT and CT (n=7; 25.5 [18.7] months); RT with IT (n=2; 20.0 [11.0] months); RT with CT (n=6; 15.7 [6.8] months); and RT alone (n=1; 13 [no SD]) months)(eTable).

T2 Angiosarcoma—There were 105 patients with T2 tumors in 15 studies.^{2,31,32,34,39-41,46,48-50,52,53,57,62} The mean (SD) OS for each treatment modality in descending order was as follows: RT with CT and IT (n=1; 36 [no SD reported] months); RT with CT (n=23; 34.3 [46.3] months); WLE with RT (n=21; 26.3 [23.8] months); WLE with CT (n=8; 21.5 [16.6] months); WLE alone (n=16; 19.8 [15.6] months); WLE with RT and CT (n=14; 19.2 [10.5] months); RT alone (n=17; 10.1 [5.5] months); CT alone (n=2; 6.7 [3.7] months); and WLE with RT, CT, and IT (n=1; 6.0 [no SD] months)(eTable).

Mohs Micrographic Surgery—The use of MMS was only identified in case reports or small observational studies for a total of 9 patients. Five cASs were treated with MMS alone for a mean (SD) OS of 37 (21.5) months, with 4 reporting cAS staging: 2 were T1^58,59 (mean [SD] OS, 37.0 [17.0] months) and 2 were T2 tumors^39,57 (mean [SD] OS, 44.5 [26.5] months). Mohs micrographic surgery with RT was used for 3 tumors (mean [SD] OS, 34.0 [26.9] months); 2 were T1^50,60 (mean [SD] OS, 42.0 [30.0] months) and 1 unreported staging (eTable).⁵⁶ Mohs micrographic surgery with both RT and CT was used in 1 patient (unreported staging; OS, 82 months).⁵¹

Complications

Complications were rare and mainly associated with CT and RT. Four studies reported radiation dermatitis with RT.^53,55,62,63 Two studies reported peripheral neuropathy and myelotoxicity with CT.^35,51 Only 1 study reported poor wound healing due to surgical complications.²⁹

COMMENT

Cutaneous angiosarcomas are rare and have limited treatment guidelines. Surgical excision does appear to be an effective adjunct to nonsurgical treatments, particularly WLE combined with RT, CT, and IT. Although MMS ultimately may be useful for cAS, the limited number and substantial heterogeneity of reported cases precludes definitive conclusions at this time.

Achieving margin control during WLE is associated with higher OS when treating angiosarcoma,^36,46 which is particularly true for T1 tumors where margin control is imperative, and many cases are treated with a combination of WLE and RT. Overall survival times are lower for T2 tumors, as these tumors are larger and most likely have spread; therefore, more aggressive combination treatments were more prevalent. In these cases, complete margin control may be difficult to achieve and may not be as critical to the outcome if another form of adjuvant therapy can be administered promptly.^24,64

When surgery is contraindicated, RT with or without CT was the most commonly reported treatment modality. However, these treatments were notably less effective than when used in combination with surgical resection. The use of RT alone has a recurrence rate reported up to 100% in certain studies, suggesting the need to utilize RT in combination with other modalities.^23,39 It is important to note that RT often is used as monotherapy in palliative treatment, which may indirectly skew survival rates.²

Limitations of the study include a lack of randomized controlled trials. Most reports were retrospective reviews or case series, and tumor staging was sparsely reported. Finally, although MMS may provide utility in the treatment of cAS, the sample size of 9 precluded definitive conclusions from being formed about its efficacy.

CONCLUSION

Cutaneous angiosarcoma is rare and has limited data comparing different treatment modalities. The paucity of data currently limits definitive recommendations; however, both surgical and nonsurgical modalities have demonstrated potential efficacy in the treatment of cAS and may benefit from additional research. Clinicians should consider a multidisciplinary approach for patients with a diagnosis of cAS to tailor treatments on a case-by-case basis.

The landscape of breast cancer care and how we're working on pushing the targeted treatment movement forward is rapidly changing, especially with antibody drug conjugates (ADCs). 

I like to think of ADCs as targeted missiles. They're essentially composed of antibodies against specific antigens, or targets of interest, and then they're combined with a linker to a chemotherapy payload. It's a way to deliver the chemotherapy in a more targeted manner than traditional chemotherapy, which is an exciting opportunity to allow us to target those patients who otherwise prefer agents that were more difficult to tolerate before this technology was invented. 

This field has grown exponentially in the last 5 to 10 years and has presented multiple new opportunities for research. What is most exciting is that we have new targets for these treatments—new antigens that we can target with novel ADCs.

The NeoSTAR trial evaluated the ADC sacituzumab govitecan (SG), which is used for patients who have earlier stage triple negative breast cancer before surgery. The idea of this is to hopefully spare patients from many of the more toxic effects of traditional chemotherapy, while still providing them precision in terms of the treatment that we're targeting in the body. 

I was involved in the ASCENT trial, which is also notable for precise treatment. It demonstrated the superiority of SG in metastatic triple negative breast cancer, and more recently the US Food and Drug Administration label has been expanded to the metastatic hormone-positive space as well. As we're developing these ADCs and broadening their use, we're able to reach larger patient populations. It's really exciting because we know there's such an appetite among our patients to use these agents, given how effective they can be and, in some situations, less toxic than the standard chemotherapy they would have otherwise gotten. 

The other big category of trials involves another ADC called trastuzumab deruxtecan (TD). Trastuzumab is conjugated against HER2, a breast cancer specific agent, and is combined with the linker, deruxtecan—a very potent chemotherapy payload. TD was initially used in patients who had HER2-positive breast cancer. In fact, trastuzumab, the first half of the drug, was used as an antibody in and of itself in a lot of earlier stage and metastatic cancer for years. We've known about that for a long time. But more recently, with the series of DESTINY trials, we have seen the major impact that TD in HER2 can have compared to other chemotherapy agents and against other ADCs as well. 

What was so exciting about the trials presented in 2022 is that they created a new category of patients called HER2-low. Before, we had always considered patients as HER2-positive or HER2-negative. We now know it's not that binary. We had already known by the way we do the pathology that people can have levels of HER2 expression. HER2-low patients are people who would have been considered HER2-negative before this—but have some HER2 expression. They have what we consider low on a scale of 0 to 3+, typically. Therefore, they're 1+ or 2+, not 0, and not 3+, because that would be considered HER2-positive. It's a little more complicated because when it's triple 2-positive, they could also do a back-up test, if a patient is positive on that, they are considered as HER2-positive. 

There is now a new category of HER2-low patients who were also shown to have a tremendous improvement benefit with TD; this new category of patients could be candidates for this treatment, although they would never have been used for HER2 targeting before. These agents can be so effective that it's even causing us to rethink our classifications of disease. 

Some of what we're working on right now is specifically looking at how patients are resistant to these agents; that's one of our major focuses as Mass General. We know these treatments are highly effective, but unfortunately, they don't last forever. Very few of these cancer treatments do, because as we know, cancer has this remarkable ability to evolve resistance to agents that we use. The amazing thing about ADCs is they've extended (in some cases) overall survival for patients, which is fantastic. But as I said, we know that they are not lasting forever. Part of what we're involved in with that clinical and research setting is to look at patients who've had success with these agents and ultimately progressed, and then figure out what changed before and after treatment (down to the single cell or genetic level) in order to continue expanding use of these treatments, extending use of the treatments, and making them more effective for more patients. 

The landscape of breast cancer care and how we're working on pushing the targeted treatment movement forward is rapidly changing, especially with antibody drug conjugates (ADCs). 

I like to think of ADCs as targeted missiles. They're essentially composed of antibodies against specific antigens, or targets of interest, and then they're combined with a linker to a chemotherapy payload. It's a way to deliver the chemotherapy in a more targeted manner than traditional chemotherapy, which is an exciting opportunity to allow us to target those patients who otherwise prefer agents that were more difficult to tolerate before this technology was invented. 

This field has grown exponentially in the last 5 to 10 years and has presented multiple new opportunities for research. What is most exciting is that we have new targets for these treatments—new antigens that we can target with novel ADCs.

The NeoSTAR trial evaluated the ADC sacituzumab govitecan (SG), which is used for patients who have earlier stage triple negative breast cancer before surgery. The idea of this is to hopefully spare patients from many of the more toxic effects of traditional chemotherapy, while still providing them precision in terms of the treatment that we're targeting in the body. 

I was involved in the ASCENT trial, which is also notable for precise treatment. It demonstrated the superiority of SG in metastatic triple negative breast cancer, and more recently the US Food and Drug Administration label has been expanded to the metastatic hormone-positive space as well. As we're developing these ADCs and broadening their use, we're able to reach larger patient populations. It's really exciting because we know there's such an appetite among our patients to use these agents, given how effective they can be and, in some situations, less toxic than the standard chemotherapy they would have otherwise gotten. 

The other big category of trials involves another ADC called trastuzumab deruxtecan (TD). Trastuzumab is conjugated against HER2, a breast cancer specific agent, and is combined with the linker, deruxtecan—a very potent chemotherapy payload. TD was initially used in patients who had HER2-positive breast cancer. In fact, trastuzumab, the first half of the drug, was used as an antibody in and of itself in a lot of earlier stage and metastatic cancer for years. We've known about that for a long time. But more recently, with the series of DESTINY trials, we have seen the major impact that TD in HER2 can have compared to other chemotherapy agents and against other ADCs as well. 

What was so exciting about the trials presented in 2022 is that they created a new category of patients called HER2-low. Before, we had always considered patients as HER2-positive or HER2-negative. We now know it's not that binary. We had already known by the way we do the pathology that people can have levels of HER2 expression. HER2-low patients are people who would have been considered HER2-negative before this—but have some HER2 expression. They have what we consider low on a scale of 0 to 3+, typically. Therefore, they're 1+ or 2+, not 0, and not 3+, because that would be considered HER2-positive. It's a little more complicated because when it's triple 2-positive, they could also do a back-up test, if a patient is positive on that, they are considered as HER2-positive. 

There is now a new category of HER2-low patients who were also shown to have a tremendous improvement benefit with TD; this new category of patients could be candidates for this treatment, although they would never have been used for HER2 targeting before. These agents can be so effective that it's even causing us to rethink our classifications of disease. 

Some of what we're working on right now is specifically looking at how patients are resistant to these agents; that's one of our major focuses as Mass General. We know these treatments are highly effective, but unfortunately, they don't last forever. Very few of these cancer treatments do, because as we know, cancer has this remarkable ability to evolve resistance to agents that we use. The amazing thing about ADCs is they've extended (in some cases) overall survival for patients, which is fantastic. But as I said, we know that they are not lasting forever. Part of what we're involved in with that clinical and research setting is to look at patients who've had success with these agents and ultimately progressed, and then figure out what changed before and after treatment (down to the single cell or genetic level) in order to continue expanding use of these treatments, extending use of the treatments, and making them more effective for more patients. 

The landscape of breast cancer care and how we're working on pushing the targeted treatment movement forward is rapidly changing, especially with antibody drug conjugates (ADCs). 

I like to think of ADCs as targeted missiles. They're essentially composed of antibodies against specific antigens, or targets of interest, and then they're combined with a linker to a chemotherapy payload. It's a way to deliver the chemotherapy in a more targeted manner than traditional chemotherapy, which is an exciting opportunity to allow us to target those patients who otherwise prefer agents that were more difficult to tolerate before this technology was invented. 

This field has grown exponentially in the last 5 to 10 years and has presented multiple new opportunities for research. What is most exciting is that we have new targets for these treatments—new antigens that we can target with novel ADCs.

The NeoSTAR trial evaluated the ADC sacituzumab govitecan (SG), which is used for patients who have earlier stage triple negative breast cancer before surgery. The idea of this is to hopefully spare patients from many of the more toxic effects of traditional chemotherapy, while still providing them precision in terms of the treatment that we're targeting in the body. 

I was involved in the ASCENT trial, which is also notable for precise treatment. It demonstrated the superiority of SG in metastatic triple negative breast cancer, and more recently the US Food and Drug Administration label has been expanded to the metastatic hormone-positive space as well. As we're developing these ADCs and broadening their use, we're able to reach larger patient populations. It's really exciting because we know there's such an appetite among our patients to use these agents, given how effective they can be and, in some situations, less toxic than the standard chemotherapy they would have otherwise gotten. 

The other big category of trials involves another ADC called trastuzumab deruxtecan (TD). Trastuzumab is conjugated against HER2, a breast cancer specific agent, and is combined with the linker, deruxtecan—a very potent chemotherapy payload. TD was initially used in patients who had HER2-positive breast cancer. In fact, trastuzumab, the first half of the drug, was used as an antibody in and of itself in a lot of earlier stage and metastatic cancer for years. We've known about that for a long time. But more recently, with the series of DESTINY trials, we have seen the major impact that TD in HER2 can have compared to other chemotherapy agents and against other ADCs as well. 

What was so exciting about the trials presented in 2022 is that they created a new category of patients called HER2-low. Before, we had always considered patients as HER2-positive or HER2-negative. We now know it's not that binary. We had already known by the way we do the pathology that people can have levels of HER2 expression. HER2-low patients are people who would have been considered HER2-negative before this—but have some HER2 expression. They have what we consider low on a scale of 0 to 3+, typically. Therefore, they're 1+ or 2+, not 0, and not 3+, because that would be considered HER2-positive. It's a little more complicated because when it's triple 2-positive, they could also do a back-up test, if a patient is positive on that, they are considered as HER2-positive. 

There is now a new category of HER2-low patients who were also shown to have a tremendous improvement benefit with TD; this new category of patients could be candidates for this treatment, although they would never have been used for HER2 targeting before. These agents can be so effective that it's even causing us to rethink our classifications of disease. 

Some of what we're working on right now is specifically looking at how patients are resistant to these agents; that's one of our major focuses as Mass General. We know these treatments are highly effective, but unfortunately, they don't last forever. Very few of these cancer treatments do, because as we know, cancer has this remarkable ability to evolve resistance to agents that we use. The amazing thing about ADCs is they've extended (in some cases) overall survival for patients, which is fantastic. But as I said, we know that they are not lasting forever. Part of what we're involved in with that clinical and research setting is to look at patients who've had success with these agents and ultimately progressed, and then figure out what changed before and after treatment (down to the single cell or genetic level) in order to continue expanding use of these treatments, extending use of the treatments, and making them more effective for more patients. 

Newly diagnosed type 2 diabetes has steadily risen by 4%-5% annually in the United States over the past 20 years; However, during the first year of the pandemic, the Journal of Pediatrics reported a possible 77% increase in youth-reported cases. Coupled with nearly 1 in 3 children being overweight, treatment for this patient population has recently evolved. 

Dr Amy S. Shah of the Cincinnati Children's Hospital Medical Center reports on the current therapeutic and lifestyle modifications guidelines for this patient population, including the consideration of prescribing glucagon-like peptide 1 receptor agonists such as liraglutide, dulaglutide, and exenatide.

--

Amy S. Shah, MD, MS, Professor of Pediatrics, Division of Endocrinology, Director of the Adolescents Type 2 Diabetes Program, Cincinnati Children's Hospital Medical Center, The University of Cincinnati, Cincinnati, Ohio

Amy S. Shah, MD, MS, has disclosed the following relevant financial relationships:

Received research grant from: National Institutes of Health

Received income in an amount equal to or greater than $250 from: Journal of the American Medical Association (Associate Editor)

Newly diagnosed type 2 diabetes has steadily risen by 4%-5% annually in the United States over the past 20 years; However, during the first year of the pandemic, the Journal of Pediatrics reported a possible 77% increase in youth-reported cases. Coupled with nearly 1 in 3 children being overweight, treatment for this patient population has recently evolved. 

Dr Amy S. Shah of the Cincinnati Children's Hospital Medical Center reports on the current therapeutic and lifestyle modifications guidelines for this patient population, including the consideration of prescribing glucagon-like peptide 1 receptor agonists such as liraglutide, dulaglutide, and exenatide.

--

Amy S. Shah, MD, MS, Professor of Pediatrics, Division of Endocrinology, Director of the Adolescents Type 2 Diabetes Program, Cincinnati Children's Hospital Medical Center, The University of Cincinnati, Cincinnati, Ohio

Amy S. Shah, MD, MS, has disclosed the following relevant financial relationships:

Received research grant from: National Institutes of Health

Received income in an amount equal to or greater than $250 from: Journal of the American Medical Association (Associate Editor)

Newly diagnosed type 2 diabetes has steadily risen by 4%-5% annually in the United States over the past 20 years; However, during the first year of the pandemic, the Journal of Pediatrics reported a possible 77% increase in youth-reported cases. Coupled with nearly 1 in 3 children being overweight, treatment for this patient population has recently evolved. 

Dr Amy S. Shah of the Cincinnati Children's Hospital Medical Center reports on the current therapeutic and lifestyle modifications guidelines for this patient population, including the consideration of prescribing glucagon-like peptide 1 receptor agonists such as liraglutide, dulaglutide, and exenatide.

--

Amy S. Shah, MD, MS, Professor of Pediatrics, Division of Endocrinology, Director of the Adolescents Type 2 Diabetes Program, Cincinnati Children's Hospital Medical Center, The University of Cincinnati, Cincinnati, Ohio

Amy S. Shah, MD, MS, has disclosed the following relevant financial relationships:

Received research grant from: National Institutes of Health

Received income in an amount equal to or greater than $250 from: Journal of the American Medical Association (Associate Editor)

The Diagnosis: Alopecic and Aseptic Nodule of the Scalp

Alopecic and aseptic nodule of the scalp (AANS) is an underdiagnosed condition presenting with one or few inflammatory nodules on the scalp with overlying nonscarring alopecia. The nodules can be soft, fluctuant, or firm and are characterized by negative fungal and bacterial stains as well as cultures.¹ Trichoscopic features such as black or yellow dots, fine vellus hairs, and broken hairs have been reported.^1-3 Dilated follicular openings may be seen and are termed the Eastern pancake sign, as they resemble the bubble cavities formed during the cooking of atayef.² The histologic features of AANS often are nonspecific but show a nodular or pseudocystic, lymphohistiocytic to acute inflammatory component centered in the dermis.¹ Granulomatous inflammation or isolated giant cells have been reported within the deep dermis.^1,4 In our patient, histopathology revealed admixed acute and granulomatous inflammation within the deep dermis (Figure). Treatment of AANS includes oral antibiotics such as doxycycline, intralesional corticosteroids, or excision.¹

Although the etiology of AANS currently is unclear, a process of follicular plugging or a deep folliculitis sparing the bulge stem cells has been theorized. Young males are disproportionately affected.¹ It is uncertain how much overlap there is, if any, between AANS and pseudocyst of the scalp, the latter of which primarily is reported in the Japanese literature and demonstrates alopecic nodules between the forehead and vertex of the scalp with pseudocystic architecture and granulomatous infiltration on histopathology.^4-7

There are several clinical and histologic differences between AANS and other diagnoses in the differential. Dermoid cysts tend to present at birth, with 70% of cases presenting before the age of 6 years, and without overlying skin changes.⁸ They represent a benign entrapment of ectoderm along embryonic closure lines during development.⁹ Histologic examination typically will show a squamous-lined cyst within the dermis with associated adnexal structures.¹⁰ Cylindromas are benign neoplasms of eccrine sweat glands named after the histologic presentation of cylinder-shaped basaloid cell populations when cross-sectioned.^11,12 When cylindromas coalesce on the scalp, they form a distinctive morphology sometimes loosely resembling a turban, giving them the previously more common name turban tumors.^11,13 Cylindromas appear as slow-growing protuberant tumors that are erythematous or flesh colored. Cylindromas are 9 times more common in females.13 Pilar cysts have a stratified squamous epithelium lining with a palisaded outer layer and are derived from the outer root sheath of hair follicles.¹⁴ Clinically, pilar cysts are smooth mobile cysts that favor skin with a dense concentration of hair follicles.^14,15 On palpation, pilar cysts are firm due to their keratinous contents and typically are nontender unless inflamed.¹⁵ Lipomas are benign mesenchymal tumors with mature adipocytes that often appear as subcutaneous nodules without overlying skin changes, though they can involve deep fascia. On palpation, lipomas generally are soft, mobile, and nontender.¹⁶

The Diagnosis: Alopecic and Aseptic Nodule of the Scalp

Alopecic and aseptic nodule of the scalp (AANS) is an underdiagnosed condition presenting with one or few inflammatory nodules on the scalp with overlying nonscarring alopecia. The nodules can be soft, fluctuant, or firm and are characterized by negative fungal and bacterial stains as well as cultures.¹ Trichoscopic features such as black or yellow dots, fine vellus hairs, and broken hairs have been reported.^1-3 Dilated follicular openings may be seen and are termed the Eastern pancake sign, as they resemble the bubble cavities formed during the cooking of atayef.² The histologic features of AANS often are nonspecific but show a nodular or pseudocystic, lymphohistiocytic to acute inflammatory component centered in the dermis.¹ Granulomatous inflammation or isolated giant cells have been reported within the deep dermis.^1,4 In our patient, histopathology revealed admixed acute and granulomatous inflammation within the deep dermis (Figure). Treatment of AANS includes oral antibiotics such as doxycycline, intralesional corticosteroids, or excision.¹

Although the etiology of AANS currently is unclear, a process of follicular plugging or a deep folliculitis sparing the bulge stem cells has been theorized. Young males are disproportionately affected.¹ It is uncertain how much overlap there is, if any, between AANS and pseudocyst of the scalp, the latter of which primarily is reported in the Japanese literature and demonstrates alopecic nodules between the forehead and vertex of the scalp with pseudocystic architecture and granulomatous infiltration on histopathology.^4-7

There are several clinical and histologic differences between AANS and other diagnoses in the differential. Dermoid cysts tend to present at birth, with 70% of cases presenting before the age of 6 years, and without overlying skin changes.⁸ They represent a benign entrapment of ectoderm along embryonic closure lines during development.⁹ Histologic examination typically will show a squamous-lined cyst within the dermis with associated adnexal structures.¹⁰ Cylindromas are benign neoplasms of eccrine sweat glands named after the histologic presentation of cylinder-shaped basaloid cell populations when cross-sectioned.^11,12 When cylindromas coalesce on the scalp, they form a distinctive morphology sometimes loosely resembling a turban, giving them the previously more common name turban tumors.^11,13 Cylindromas appear as slow-growing protuberant tumors that are erythematous or flesh colored. Cylindromas are 9 times more common in females.13 Pilar cysts have a stratified squamous epithelium lining with a palisaded outer layer and are derived from the outer root sheath of hair follicles.¹⁴ Clinically, pilar cysts are smooth mobile cysts that favor skin with a dense concentration of hair follicles.^14,15 On palpation, pilar cysts are firm due to their keratinous contents and typically are nontender unless inflamed.¹⁵ Lipomas are benign mesenchymal tumors with mature adipocytes that often appear as subcutaneous nodules without overlying skin changes, though they can involve deep fascia. On palpation, lipomas generally are soft, mobile, and nontender.¹⁶

The Diagnosis: Alopecic and Aseptic Nodule of the Scalp

Alopecic and aseptic nodule of the scalp (AANS) is an underdiagnosed condition presenting with one or few inflammatory nodules on the scalp with overlying nonscarring alopecia. The nodules can be soft, fluctuant, or firm and are characterized by negative fungal and bacterial stains as well as cultures.¹ Trichoscopic features such as black or yellow dots, fine vellus hairs, and broken hairs have been reported.^1-3 Dilated follicular openings may be seen and are termed the Eastern pancake sign, as they resemble the bubble cavities formed during the cooking of atayef.² The histologic features of AANS often are nonspecific but show a nodular or pseudocystic, lymphohistiocytic to acute inflammatory component centered in the dermis.¹ Granulomatous inflammation or isolated giant cells have been reported within the deep dermis.^1,4 In our patient, histopathology revealed admixed acute and granulomatous inflammation within the deep dermis (Figure). Treatment of AANS includes oral antibiotics such as doxycycline, intralesional corticosteroids, or excision.¹

Although the etiology of AANS currently is unclear, a process of follicular plugging or a deep folliculitis sparing the bulge stem cells has been theorized. Young males are disproportionately affected.¹ It is uncertain how much overlap there is, if any, between AANS and pseudocyst of the scalp, the latter of which primarily is reported in the Japanese literature and demonstrates alopecic nodules between the forehead and vertex of the scalp with pseudocystic architecture and granulomatous infiltration on histopathology.^4-7

There are several clinical and histologic differences between AANS and other diagnoses in the differential. Dermoid cysts tend to present at birth, with 70% of cases presenting before the age of 6 years, and without overlying skin changes.⁸ They represent a benign entrapment of ectoderm along embryonic closure lines during development.⁹ Histologic examination typically will show a squamous-lined cyst within the dermis with associated adnexal structures.¹⁰ Cylindromas are benign neoplasms of eccrine sweat glands named after the histologic presentation of cylinder-shaped basaloid cell populations when cross-sectioned.^11,12 When cylindromas coalesce on the scalp, they form a distinctive morphology sometimes loosely resembling a turban, giving them the previously more common name turban tumors.^11,13 Cylindromas appear as slow-growing protuberant tumors that are erythematous or flesh colored. Cylindromas are 9 times more common in females.13 Pilar cysts have a stratified squamous epithelium lining with a palisaded outer layer and are derived from the outer root sheath of hair follicles.¹⁴ Clinically, pilar cysts are smooth mobile cysts that favor skin with a dense concentration of hair follicles.^14,15 On palpation, pilar cysts are firm due to their keratinous contents and typically are nontender unless inflamed.¹⁵ Lipomas are benign mesenchymal tumors with mature adipocytes that often appear as subcutaneous nodules without overlying skin changes, though they can involve deep fascia. On palpation, lipomas generally are soft, mobile, and nontender.¹⁶

Up to half of patients with type 2 diabetes report poor adherence to their non-insulin medications. In this ReCAP, Dr Anne Peters, director of the University of Southern California Clinical Diabetes Programs in Los Angeles, California, offers clinicians suggestions on how to get their patients to "buy in" to these treatments, many of which do not make them feel any better. Research shows that quality of life, short- and long-term clinical outcomes, and side effects are important factors in treatment adherence. Dr Peters explains that adherence may be improved if patients understand why they need to take a medication, how it might make them feel, what to do about it, and how taking it will fit into their daily pattern. She also points out that access and affordability of the medication regimen are additional factors for which patients may need support to increase treatment adherence.

--

Anne L. Peters, MD, Professor, Department of Clinical Medicine, Clinical Scholar, Keck School of Medicine of the University of Southern California; Director, USC Clinical Diabetes Programs, University of Southern California Westside Center for Diabetes, Los Angeles, California

Anne L. Peters, MD, has disclosed the following relevant financial relationships: Serve(d) as a director, officer, partner, employee, advisor, consultant, or a trustee for: Blue Circle Health; Vertex; Abbott Diabetes Care Received research grant from: Abbott Diabetes Care; Insulet Stock options from: Teladoc; Omada Health

Up to half of patients with type 2 diabetes report poor adherence to their non-insulin medications. In this ReCAP, Dr Anne Peters, director of the University of Southern California Clinical Diabetes Programs in Los Angeles, California, offers clinicians suggestions on how to get their patients to "buy in" to these treatments, many of which do not make them feel any better. Research shows that quality of life, short- and long-term clinical outcomes, and side effects are important factors in treatment adherence. Dr Peters explains that adherence may be improved if patients understand why they need to take a medication, how it might make them feel, what to do about it, and how taking it will fit into their daily pattern. She also points out that access and affordability of the medication regimen are additional factors for which patients may need support to increase treatment adherence.

--

Anne L. Peters, MD, Professor, Department of Clinical Medicine, Clinical Scholar, Keck School of Medicine of the University of Southern California; Director, USC Clinical Diabetes Programs, University of Southern California Westside Center for Diabetes, Los Angeles, California

Anne L. Peters, MD, has disclosed the following relevant financial relationships: Serve(d) as a director, officer, partner, employee, advisor, consultant, or a trustee for: Blue Circle Health; Vertex; Abbott Diabetes Care Received research grant from: Abbott Diabetes Care; Insulet Stock options from: Teladoc; Omada Health

Up to half of patients with type 2 diabetes report poor adherence to their non-insulin medications. In this ReCAP, Dr Anne Peters, director of the University of Southern California Clinical Diabetes Programs in Los Angeles, California, offers clinicians suggestions on how to get their patients to "buy in" to these treatments, many of which do not make them feel any better. Research shows that quality of life, short- and long-term clinical outcomes, and side effects are important factors in treatment adherence. Dr Peters explains that adherence may be improved if patients understand why they need to take a medication, how it might make them feel, what to do about it, and how taking it will fit into their daily pattern. She also points out that access and affordability of the medication regimen are additional factors for which patients may need support to increase treatment adherence.

--

Anne L. Peters, MD, Professor, Department of Clinical Medicine, Clinical Scholar, Keck School of Medicine of the University of Southern California; Director, USC Clinical Diabetes Programs, University of Southern California Westside Center for Diabetes, Los Angeles, California

Anne L. Peters, MD, has disclosed the following relevant financial relationships: Serve(d) as a director, officer, partner, employee, advisor, consultant, or a trustee for: Blue Circle Health; Vertex; Abbott Diabetes Care Received research grant from: Abbott Diabetes Care; Insulet Stock options from: Teladoc; Omada Health

Endocrine therapy (ET) has long been the therapeutic mainstay for first-line treatment of HR+/HER2- breast cancer. Yet, approximately one third of patients exhibit primary or acquired ET resistance, and some patients progress within 2 years after adjuvant ET. Resistance to endocrine therapy is a major challenge for a significant number of patients who go on to develop metastatic breast cancer.

Dr Jennifer Litton from The University of Texas MD Anderson Cancer Center in Houston, Texas, reports on current treatment recommendations and supporting research on the use of endocrine therapy and CDK4/6 inhibitors in both the metastatic and, more recently, the high-risk, early HR+/HER2- breast cancer settings.

Dr Litton also reviews key drivers of endocrine resistance, including somatic mutations such as ESR1. She emphasizes the need for next-generation testing in patients with metastatic HR+/HER2- disease to look for evidence of resistance, which may have implications for the next line of therapy.

--

Professor, Department of Breast Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas

Jennifer K. Litton, MD, has disclosed the following relevant financial relationships:

Serve(d) as a director, officer, partner, employee, advisor, consultant, or trustee for: The University of Texas MD Anderson Cancer Center

Serve(d) as a speaker or a member of a speakers bureau for: Clinical Care Options; Med Learning Group; Medpage; Medscape; PRIME; Physicians Education Resource; UpToDate

Endocrine therapy (ET) has long been the therapeutic mainstay for first-line treatment of HR+/HER2- breast cancer. Yet, approximately one third of patients exhibit primary or acquired ET resistance, and some patients progress within 2 years after adjuvant ET. Resistance to endocrine therapy is a major challenge for a significant number of patients who go on to develop metastatic breast cancer.

Dr Jennifer Litton from The University of Texas MD Anderson Cancer Center in Houston, Texas, reports on current treatment recommendations and supporting research on the use of endocrine therapy and CDK4/6 inhibitors in both the metastatic and, more recently, the high-risk, early HR+/HER2- breast cancer settings.

Dr Litton also reviews key drivers of endocrine resistance, including somatic mutations such as ESR1. She emphasizes the need for next-generation testing in patients with metastatic HR+/HER2- disease to look for evidence of resistance, which may have implications for the next line of therapy.

--

Professor, Department of Breast Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas

Jennifer K. Litton, MD, has disclosed the following relevant financial relationships:

Serve(d) as a director, officer, partner, employee, advisor, consultant, or trustee for: The University of Texas MD Anderson Cancer Center

Serve(d) as a speaker or a member of a speakers bureau for: Clinical Care Options; Med Learning Group; Medpage; Medscape; PRIME; Physicians Education Resource; UpToDate

Endocrine therapy (ET) has long been the therapeutic mainstay for first-line treatment of HR+/HER2- breast cancer. Yet, approximately one third of patients exhibit primary or acquired ET resistance, and some patients progress within 2 years after adjuvant ET. Resistance to endocrine therapy is a major challenge for a significant number of patients who go on to develop metastatic breast cancer.

Dr Jennifer Litton from The University of Texas MD Anderson Cancer Center in Houston, Texas, reports on current treatment recommendations and supporting research on the use of endocrine therapy and CDK4/6 inhibitors in both the metastatic and, more recently, the high-risk, early HR+/HER2- breast cancer settings.

Dr Litton also reviews key drivers of endocrine resistance, including somatic mutations such as ESR1. She emphasizes the need for next-generation testing in patients with metastatic HR+/HER2- disease to look for evidence of resistance, which may have implications for the next line of therapy.

--

Professor, Department of Breast Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas

Jennifer K. Litton, MD, has disclosed the following relevant financial relationships:

Serve(d) as a director, officer, partner, employee, advisor, consultant, or trustee for: The University of Texas MD Anderson Cancer Center

Serve(d) as a speaker or a member of a speakers bureau for: Clinical Care Options; Med Learning Group; Medpage; Medscape; PRIME; Physicians Education Resource; UpToDate