User login
NEWPORT BEACH, CALIF. — A model that mimics the recommender system used by Netflix and Amazon can help predict outcomes of lenalidomide treatment in patients with non–deletion 5q (non-del[5q]) myelodysplastic syndromes (MDS), according to new research.
The model was used to identify genomic biomarkers that were associated with resistance or response to lenalidomide. Researchers found these associations in 39% of patients with non-del(5q) MDS, and the model predicted response or resistance with 82% accuracy.
Yazan Madanat, MD, of the Cleveland Clinic, and his colleagues presented these findings at the Acute Leukemia Forum of Hemedicus.
Dr. Madanat explained that his group’s model is similar to the recommender system used by Netflix and Amazon, which makes suggestions for new products based on customers’ past behavior. Dr. Madanat and his colleagues used their model to show that patients with certain molecular or cytogenetic abnormalities are likely to respond or not respond to lenalidomide.
The researchers began by looking at 139 patients who had received at least two cycles of lenalidomide treatment. There were 118 patients with MDS, and 108 who had received lenalidomide monotherapy. However, the team focused on the 100 patients who had non-del(5q) MDS, 58 of whom had normal karyotype (NK) and 19 of whom had complex karyotype (CK).
The model revealed several combinations of genomic/cytogenetic abnormalities that could predict resistance to lenalidomide, including the following:
- DNMT3A and SF3B1
- EZH2 and NK
- ASXL1, TET2, and NK
- STAG2, IDH1/2, and NK
- TP53, del(5q), and CK
- BCOR/BCORL1 and NK
- JAK2, TET2, and NK
- U2AF1, +/– ETV6, and NK
However, only the following two combinations could predict response to lenalidomide:
- DDX41 and NK
- MECOM and KDM6A/B
These combinations could be applied to 39% of the patients with non-del(5q) MDS, and the model predicted response or resistance to lenalidomide with 82% accuracy.
Although the biomarkers were found in only a subset of patients, Dr. Madanat said these findings may help physicians tailor therapy for MDS patients, given the high level of accuracy the researchers observed.
“It’s really important to validate the results in a prospective manner and to ensure that we’re able to apply them clinically and potentially change the way we’re treating our patients,” he added.
Dr. Madanat and his colleagues reported having no relevant conflicts of interest.
The Acute Leukemia Forum is held by Hemedicus, which is owned by the same company as this news organization.
NEWPORT BEACH, CALIF. — A model that mimics the recommender system used by Netflix and Amazon can help predict outcomes of lenalidomide treatment in patients with non–deletion 5q (non-del[5q]) myelodysplastic syndromes (MDS), according to new research.
The model was used to identify genomic biomarkers that were associated with resistance or response to lenalidomide. Researchers found these associations in 39% of patients with non-del(5q) MDS, and the model predicted response or resistance with 82% accuracy.
Yazan Madanat, MD, of the Cleveland Clinic, and his colleagues presented these findings at the Acute Leukemia Forum of Hemedicus.
Dr. Madanat explained that his group’s model is similar to the recommender system used by Netflix and Amazon, which makes suggestions for new products based on customers’ past behavior. Dr. Madanat and his colleagues used their model to show that patients with certain molecular or cytogenetic abnormalities are likely to respond or not respond to lenalidomide.
The researchers began by looking at 139 patients who had received at least two cycles of lenalidomide treatment. There were 118 patients with MDS, and 108 who had received lenalidomide monotherapy. However, the team focused on the 100 patients who had non-del(5q) MDS, 58 of whom had normal karyotype (NK) and 19 of whom had complex karyotype (CK).
The model revealed several combinations of genomic/cytogenetic abnormalities that could predict resistance to lenalidomide, including the following:
- DNMT3A and SF3B1
- EZH2 and NK
- ASXL1, TET2, and NK
- STAG2, IDH1/2, and NK
- TP53, del(5q), and CK
- BCOR/BCORL1 and NK
- JAK2, TET2, and NK
- U2AF1, +/– ETV6, and NK
However, only the following two combinations could predict response to lenalidomide:
- DDX41 and NK
- MECOM and KDM6A/B
These combinations could be applied to 39% of the patients with non-del(5q) MDS, and the model predicted response or resistance to lenalidomide with 82% accuracy.
Although the biomarkers were found in only a subset of patients, Dr. Madanat said these findings may help physicians tailor therapy for MDS patients, given the high level of accuracy the researchers observed.
“It’s really important to validate the results in a prospective manner and to ensure that we’re able to apply them clinically and potentially change the way we’re treating our patients,” he added.
Dr. Madanat and his colleagues reported having no relevant conflicts of interest.
The Acute Leukemia Forum is held by Hemedicus, which is owned by the same company as this news organization.
NEWPORT BEACH, CALIF. — A model that mimics the recommender system used by Netflix and Amazon can help predict outcomes of lenalidomide treatment in patients with non–deletion 5q (non-del[5q]) myelodysplastic syndromes (MDS), according to new research.
The model was used to identify genomic biomarkers that were associated with resistance or response to lenalidomide. Researchers found these associations in 39% of patients with non-del(5q) MDS, and the model predicted response or resistance with 82% accuracy.
Yazan Madanat, MD, of the Cleveland Clinic, and his colleagues presented these findings at the Acute Leukemia Forum of Hemedicus.
Dr. Madanat explained that his group’s model is similar to the recommender system used by Netflix and Amazon, which makes suggestions for new products based on customers’ past behavior. Dr. Madanat and his colleagues used their model to show that patients with certain molecular or cytogenetic abnormalities are likely to respond or not respond to lenalidomide.
The researchers began by looking at 139 patients who had received at least two cycles of lenalidomide treatment. There were 118 patients with MDS, and 108 who had received lenalidomide monotherapy. However, the team focused on the 100 patients who had non-del(5q) MDS, 58 of whom had normal karyotype (NK) and 19 of whom had complex karyotype (CK).
The model revealed several combinations of genomic/cytogenetic abnormalities that could predict resistance to lenalidomide, including the following:
- DNMT3A and SF3B1
- EZH2 and NK
- ASXL1, TET2, and NK
- STAG2, IDH1/2, and NK
- TP53, del(5q), and CK
- BCOR/BCORL1 and NK
- JAK2, TET2, and NK
- U2AF1, +/– ETV6, and NK
However, only the following two combinations could predict response to lenalidomide:
- DDX41 and NK
- MECOM and KDM6A/B
These combinations could be applied to 39% of the patients with non-del(5q) MDS, and the model predicted response or resistance to lenalidomide with 82% accuracy.
Although the biomarkers were found in only a subset of patients, Dr. Madanat said these findings may help physicians tailor therapy for MDS patients, given the high level of accuracy the researchers observed.
“It’s really important to validate the results in a prospective manner and to ensure that we’re able to apply them clinically and potentially change the way we’re treating our patients,” he added.
Dr. Madanat and his colleagues reported having no relevant conflicts of interest.
The Acute Leukemia Forum is held by Hemedicus, which is owned by the same company as this news organization.
REPORTING FROM ALF 2019
