Bishal Gyawali, MD, PhD

This transcript has been edited for clarity.

Hello, everyone. This is Dr Bishal Gyawali, from Queens University, Kingston, Canada. Today, I’m back with you to talk about some of the fallacies that I have seen in medicine, oncology, and the drug regulatory space. I wanted to clarify some of these fallacies.

In my last video, I talked about the FDA denying the approval of a new cancer drug. Let me start with one of the fallacies that is pertinent to that, which is that some people make an argument that patients are dying from a certain condition, such as cancer, or even any other disease besides cancer. That is an absolutely true statement, but that does not necessarily mean there should be a lower bar for drug approvals or we should be approving any drug that has a hint of benefit.

In fact, if we have increased mortality rates and our patients are dying from a certain condition, that means we actually need to have good drugs. We need to have drugs that prevent mortality. We need to have drugs that improve outcomes. Just having any drug out there, if we lower our threshold and are letting any drug be used in these patients because the argument is that people are dying, then in fact, it can have negative consequences.

First, there will be opportunity costs. If you can get any lousy drug into the market and make billions of dollars out of it, then there is no strong motivation to produce drugs that actually remarkably improve outcomes.

Second, patients will also be misled. It’s the patient’s opportunity cost in that they will use whatever time they have remaining to pursue these treatments that were not going to improve their outcomes anyway. This is time they could have better spent either in pursuing better treatments, if those treatments are out there, or to prioritize their time accordingly. This rather gives them a false hope, which can be harmful in the long term.

The first fallacy is that just because people are dying does not necessarily mean we should have more new drugs with a lower bar for approval.

The second fallacy I want to talk about, which is also related to this, is that if a certain cancer is rare, the bar for new drug approval should be quite low.

Of course, rare cancers are a special category, and rare cancers should be treated differently from a regulatory perspective. Absolutely. If the cancer is rare, we cannot have trials with large sample sizes to generate evidence. That problem is there, but that does not necessarily translate to the decision that we should approve anything, even something with a small hint of benefit.

There are other methods to make sure that, even in rare cancers, we can generate good-quality evidence. In fact, from an equity perspective, why should patients with rare cancer not deserve drugs that have good-quality evidence?

We can’t tell someone that, “Your cancer is rare, so you should get drugs that only have a benefit in terms of response rate whereas other cancers that are not rare will have drugs based on survival.”

Going back to the point about the difficulty in doing big trials in patients with rare cancers, that is absolutely true and there should be regulatory flexibility in this. I think accelerated approval is a pathway that allows for this regulatory flexibility, which allows access to these drugs early on based on earlier signals of benefit. You can continue to generate evidence in the future and confirm the clinical benefit.

There are also other nuances to this. One is that we should also make sure that this regulatory flexibility with rare cancers should not be misused. What do I mean by that? First, all rare cancers are not the same. There are some cancers that are ultra rare, and then there are some cancers that technically might fit the definition of rare, but trials are possible. Case in point: adrenocortical cancer. It is considered to be a very rare cancer, but there have been randomized trials in adrenocortical cancer.

Our efforts should be to maximize our collaboration globally so that a cancer that is rare locally will still not be so rare globally when we collect all these patients.

In certain situations, like let’s say, based on the molecular subtypes, any common cancer can be sliced and diced into a rare subtype: MSI-high, BRAF-negative, HER2-positive, right-sided colon cancer. If you start to slice cancers into these smaller and smaller molecular subtypes, you can consider anything as a rare cancer. That should not be misused as an excuse to get away from doing proper trials and generating adequate evidence for our patients.

The third fallacy I want to talk about is that increasing cancer incidence in a certain subgroup of population does not automatically translate into, “We should start screening this subgroup of population.”

A certain cancer — let’s say cancer X or cancer Y — is increasing in a young population, so therefore, we should lower the age of the screening of young populations. This cancer is increasing in this ethnic population, so therefore we should start screening this ethnic population more frequently. This cancer is increasing in this type of minority, so therefore, we should start screening this minority more.

No, it does not work like that. Increasing incidence will make us concerned, of course, but that does not necessarily translate into, “We should start screening them.” In order for a screening test to be useful, it has to fulfill a number of criteria.

The goal is not to detect cancers. The goal is to detect cancers that are not indolent enough that they would have never caused problems, nor speed up the diagnosis of aggressive cancers that are going to be lethal pretty soon anyway. The goal is to detect those cancers in the middle, so that by detecting early, we can intervene and improve the outcomes and improve the mortality from that cancer.

This type of intervention requires a thoughtful consideration of the increasing incidence of the cancer, of course, but also the utility of the screening test in that subgroup of population; the life expectancy of this subgroup of patients with and without cancer; the interventions available to address that increasing burden of cancer; and whether by intervening we are going to reduce the mortality rates.

Just because we can detect cancers does not mean we should detect cancers. That’s the third fallacy I wanted to talk about.

The fourth fallacy is related to when someone is asking for more evidence for anything. There is a new drug for this cancer,so what is the evidence? Or there is this new intervention that will detect ctDNA or whatever before the cancer relapses, or before the cancer even shows up as a screening test.

Whenever there is any treatment that is being promoted and someone asks for evidence, people sometimes try to make personal attacks by saying, “Oh, so, you’re okay with patients dying. You don’t want to save lives.”

Absolutely we want to save lives. That’s why we’re in this field, and that’s why we’re asking for more evidence. You should not consider someone who is asking for evidence as evil or that this person does not want this new drug, or this person does not want this innovation. No, that person actually wants to make sure that that innovation actually helps people. That’s why that person is asking for more evidence.

If we stop asking for evidence, then our whole practice becomes based on emotions, faith, and trust rather than science. You could extrapolate it to the other extreme, like if you are not asking for evidence. If it is interpreted as someone who is asking for evidence is evil, or someone who does not want patients to get new drugs, then you could extrapolate these to people also making claims about alternative medicines or ivermectin nowadays, and claiming that this cures cancer.

Science is science. You need to be the same no matter the circumstances. If you are asking for data for ivermectin, you should also be asking data for your cancer drug that you think is going to work. We should always ask for evidence.

Asking for evidence is not a sign that whoever is asking for evidence does not want the patient to have access to the drug. It is showing that the person who is asking for evidence actually wants to make sure that the patients who get this drug are actually being helped by the drug rather than being harmed.

I’m talking about 5 fallacies today. The final, fifth fallacy is that clinical expertise does not equal expertise in making public health decisions or even expertise in critical appraisal. Someone can be a fantastic breast cancer doctor, the best oncologist for breast cancer. That does not automatically make that person the best person to evaluate clinical trials of breast cancer drugs.

Someone can be a fantastic colon cancer doctor. That does not make that person automatically the best person to evaluate whether or not colonoscopy or colon cancer screening is indicated in a certain patient population.

These population-level decisions — including should this drug be approved, should this drug be funded, and should this screening test be made a public health measure, all of these public health decisions that are done at a population level — require different expertise in critical appraisal, clinical epidemiology, and public health.

Just because someone is a fantastic clinician does not make that person a fantastic public health expert. I see on social media often that a famous doctor with expertise in their domain, let’s say a famous neurosurgeon, might say, “I think brain tumors are increasing in incidence in young persons, so we should be targeting an MRI screening for everyone over the age of 30.”

I’m just making this up, but we see examples of things not dissimilar to this. Just because someone is a neurosurgeon does not make them an expert on brain tumor epidemiology, surveillance, or screening. We should separate clinical expertise from public health expertise.

Thank you.

A version of this article first appeared on Medscape.com.

This transcript has been edited for clarity.

Hello, everyone. This is Dr Bishal Gyawali, from Queens University, Kingston, Canada. Today, I’m back with you to talk about some of the fallacies that I have seen in medicine, oncology, and the drug regulatory space. I wanted to clarify some of these fallacies.

In my last video, I talked about the FDA denying the approval of a new cancer drug. Let me start with one of the fallacies that is pertinent to that, which is that some people make an argument that patients are dying from a certain condition, such as cancer, or even any other disease besides cancer. That is an absolutely true statement, but that does not necessarily mean there should be a lower bar for drug approvals or we should be approving any drug that has a hint of benefit.

In fact, if we have increased mortality rates and our patients are dying from a certain condition, that means we actually need to have good drugs. We need to have drugs that prevent mortality. We need to have drugs that improve outcomes. Just having any drug out there, if we lower our threshold and are letting any drug be used in these patients because the argument is that people are dying, then in fact, it can have negative consequences.

First, there will be opportunity costs. If you can get any lousy drug into the market and make billions of dollars out of it, then there is no strong motivation to produce drugs that actually remarkably improve outcomes.

Second, patients will also be misled. It’s the patient’s opportunity cost in that they will use whatever time they have remaining to pursue these treatments that were not going to improve their outcomes anyway. This is time they could have better spent either in pursuing better treatments, if those treatments are out there, or to prioritize their time accordingly. This rather gives them a false hope, which can be harmful in the long term.

The first fallacy is that just because people are dying does not necessarily mean we should have more new drugs with a lower bar for approval.

The second fallacy I want to talk about, which is also related to this, is that if a certain cancer is rare, the bar for new drug approval should be quite low.

Of course, rare cancers are a special category, and rare cancers should be treated differently from a regulatory perspective. Absolutely. If the cancer is rare, we cannot have trials with large sample sizes to generate evidence. That problem is there, but that does not necessarily translate to the decision that we should approve anything, even something with a small hint of benefit.

There are other methods to make sure that, even in rare cancers, we can generate good-quality evidence. In fact, from an equity perspective, why should patients with rare cancer not deserve drugs that have good-quality evidence?

We can’t tell someone that, “Your cancer is rare, so you should get drugs that only have a benefit in terms of response rate whereas other cancers that are not rare will have drugs based on survival.”

Going back to the point about the difficulty in doing big trials in patients with rare cancers, that is absolutely true and there should be regulatory flexibility in this. I think accelerated approval is a pathway that allows for this regulatory flexibility, which allows access to these drugs early on based on earlier signals of benefit. You can continue to generate evidence in the future and confirm the clinical benefit.

There are also other nuances to this. One is that we should also make sure that this regulatory flexibility with rare cancers should not be misused. What do I mean by that? First, all rare cancers are not the same. There are some cancers that are ultra rare, and then there are some cancers that technically might fit the definition of rare, but trials are possible. Case in point: adrenocortical cancer. It is considered to be a very rare cancer, but there have been randomized trials in adrenocortical cancer.

Our efforts should be to maximize our collaboration globally so that a cancer that is rare locally will still not be so rare globally when we collect all these patients.

In certain situations, like let’s say, based on the molecular subtypes, any common cancer can be sliced and diced into a rare subtype: MSI-high, BRAF-negative, HER2-positive, right-sided colon cancer. If you start to slice cancers into these smaller and smaller molecular subtypes, you can consider anything as a rare cancer. That should not be misused as an excuse to get away from doing proper trials and generating adequate evidence for our patients.

The third fallacy I want to talk about is that increasing cancer incidence in a certain subgroup of population does not automatically translate into, “We should start screening this subgroup of population.”

A certain cancer — let’s say cancer X or cancer Y — is increasing in a young population, so therefore, we should lower the age of the screening of young populations. This cancer is increasing in this ethnic population, so therefore we should start screening this ethnic population more frequently. This cancer is increasing in this type of minority, so therefore, we should start screening this minority more.

No, it does not work like that. Increasing incidence will make us concerned, of course, but that does not necessarily translate into, “We should start screening them.” In order for a screening test to be useful, it has to fulfill a number of criteria.

The goal is not to detect cancers. The goal is to detect cancers that are not indolent enough that they would have never caused problems, nor speed up the diagnosis of aggressive cancers that are going to be lethal pretty soon anyway. The goal is to detect those cancers in the middle, so that by detecting early, we can intervene and improve the outcomes and improve the mortality from that cancer.

This type of intervention requires a thoughtful consideration of the increasing incidence of the cancer, of course, but also the utility of the screening test in that subgroup of population; the life expectancy of this subgroup of patients with and without cancer; the interventions available to address that increasing burden of cancer; and whether by intervening we are going to reduce the mortality rates.

Just because we can detect cancers does not mean we should detect cancers. That’s the third fallacy I wanted to talk about.

The fourth fallacy is related to when someone is asking for more evidence for anything. There is a new drug for this cancer,so what is the evidence? Or there is this new intervention that will detect ctDNA or whatever before the cancer relapses, or before the cancer even shows up as a screening test.

Whenever there is any treatment that is being promoted and someone asks for evidence, people sometimes try to make personal attacks by saying, “Oh, so, you’re okay with patients dying. You don’t want to save lives.”

Absolutely we want to save lives. That’s why we’re in this field, and that’s why we’re asking for more evidence. You should not consider someone who is asking for evidence as evil or that this person does not want this new drug, or this person does not want this innovation. No, that person actually wants to make sure that that innovation actually helps people. That’s why that person is asking for more evidence.

If we stop asking for evidence, then our whole practice becomes based on emotions, faith, and trust rather than science. You could extrapolate it to the other extreme, like if you are not asking for evidence. If it is interpreted as someone who is asking for evidence is evil, or someone who does not want patients to get new drugs, then you could extrapolate these to people also making claims about alternative medicines or ivermectin nowadays, and claiming that this cures cancer.

Science is science. You need to be the same no matter the circumstances. If you are asking for data for ivermectin, you should also be asking data for your cancer drug that you think is going to work. We should always ask for evidence.

Asking for evidence is not a sign that whoever is asking for evidence does not want the patient to have access to the drug. It is showing that the person who is asking for evidence actually wants to make sure that the patients who get this drug are actually being helped by the drug rather than being harmed.

I’m talking about 5 fallacies today. The final, fifth fallacy is that clinical expertise does not equal expertise in making public health decisions or even expertise in critical appraisal. Someone can be a fantastic breast cancer doctor, the best oncologist for breast cancer. That does not automatically make that person the best person to evaluate clinical trials of breast cancer drugs.

Someone can be a fantastic colon cancer doctor. That does not make that person automatically the best person to evaluate whether or not colonoscopy or colon cancer screening is indicated in a certain patient population.

These population-level decisions — including should this drug be approved, should this drug be funded, and should this screening test be made a public health measure, all of these public health decisions that are done at a population level — require different expertise in critical appraisal, clinical epidemiology, and public health.

Just because someone is a fantastic clinician does not make that person a fantastic public health expert. I see on social media often that a famous doctor with expertise in their domain, let’s say a famous neurosurgeon, might say, “I think brain tumors are increasing in incidence in young persons, so we should be targeting an MRI screening for everyone over the age of 30.”

I’m just making this up, but we see examples of things not dissimilar to this. Just because someone is a neurosurgeon does not make them an expert on brain tumor epidemiology, surveillance, or screening. We should separate clinical expertise from public health expertise.

Thank you.

A version of this article first appeared on Medscape.com.

This transcript has been edited for clarity.

Hello, everyone. This is Dr Bishal Gyawali, from Queens University, Kingston, Canada. Today, I’m back with you to talk about some of the fallacies that I have seen in medicine, oncology, and the drug regulatory space. I wanted to clarify some of these fallacies.

In my last video, I talked about the FDA denying the approval of a new cancer drug. Let me start with one of the fallacies that is pertinent to that, which is that some people make an argument that patients are dying from a certain condition, such as cancer, or even any other disease besides cancer. That is an absolutely true statement, but that does not necessarily mean there should be a lower bar for drug approvals or we should be approving any drug that has a hint of benefit.

In fact, if we have increased mortality rates and our patients are dying from a certain condition, that means we actually need to have good drugs. We need to have drugs that prevent mortality. We need to have drugs that improve outcomes. Just having any drug out there, if we lower our threshold and are letting any drug be used in these patients because the argument is that people are dying, then in fact, it can have negative consequences.

First, there will be opportunity costs. If you can get any lousy drug into the market and make billions of dollars out of it, then there is no strong motivation to produce drugs that actually remarkably improve outcomes.

Second, patients will also be misled. It’s the patient’s opportunity cost in that they will use whatever time they have remaining to pursue these treatments that were not going to improve their outcomes anyway. This is time they could have better spent either in pursuing better treatments, if those treatments are out there, or to prioritize their time accordingly. This rather gives them a false hope, which can be harmful in the long term.

The first fallacy is that just because people are dying does not necessarily mean we should have more new drugs with a lower bar for approval.

The second fallacy I want to talk about, which is also related to this, is that if a certain cancer is rare, the bar for new drug approval should be quite low.

Of course, rare cancers are a special category, and rare cancers should be treated differently from a regulatory perspective. Absolutely. If the cancer is rare, we cannot have trials with large sample sizes to generate evidence. That problem is there, but that does not necessarily translate to the decision that we should approve anything, even something with a small hint of benefit.

There are other methods to make sure that, even in rare cancers, we can generate good-quality evidence. In fact, from an equity perspective, why should patients with rare cancer not deserve drugs that have good-quality evidence?

We can’t tell someone that, “Your cancer is rare, so you should get drugs that only have a benefit in terms of response rate whereas other cancers that are not rare will have drugs based on survival.”

Going back to the point about the difficulty in doing big trials in patients with rare cancers, that is absolutely true and there should be regulatory flexibility in this. I think accelerated approval is a pathway that allows for this regulatory flexibility, which allows access to these drugs early on based on earlier signals of benefit. You can continue to generate evidence in the future and confirm the clinical benefit.

There are also other nuances to this. One is that we should also make sure that this regulatory flexibility with rare cancers should not be misused. What do I mean by that? First, all rare cancers are not the same. There are some cancers that are ultra rare, and then there are some cancers that technically might fit the definition of rare, but trials are possible. Case in point: adrenocortical cancer. It is considered to be a very rare cancer, but there have been randomized trials in adrenocortical cancer.

Our efforts should be to maximize our collaboration globally so that a cancer that is rare locally will still not be so rare globally when we collect all these patients.

In certain situations, like let’s say, based on the molecular subtypes, any common cancer can be sliced and diced into a rare subtype: MSI-high, BRAF-negative, HER2-positive, right-sided colon cancer. If you start to slice cancers into these smaller and smaller molecular subtypes, you can consider anything as a rare cancer. That should not be misused as an excuse to get away from doing proper trials and generating adequate evidence for our patients.

The third fallacy I want to talk about is that increasing cancer incidence in a certain subgroup of population does not automatically translate into, “We should start screening this subgroup of population.”

A certain cancer — let’s say cancer X or cancer Y — is increasing in a young population, so therefore, we should lower the age of the screening of young populations. This cancer is increasing in this ethnic population, so therefore we should start screening this ethnic population more frequently. This cancer is increasing in this type of minority, so therefore, we should start screening this minority more.

No, it does not work like that. Increasing incidence will make us concerned, of course, but that does not necessarily translate into, “We should start screening them.” In order for a screening test to be useful, it has to fulfill a number of criteria.

The goal is not to detect cancers. The goal is to detect cancers that are not indolent enough that they would have never caused problems, nor speed up the diagnosis of aggressive cancers that are going to be lethal pretty soon anyway. The goal is to detect those cancers in the middle, so that by detecting early, we can intervene and improve the outcomes and improve the mortality from that cancer.

This type of intervention requires a thoughtful consideration of the increasing incidence of the cancer, of course, but also the utility of the screening test in that subgroup of population; the life expectancy of this subgroup of patients with and without cancer; the interventions available to address that increasing burden of cancer; and whether by intervening we are going to reduce the mortality rates.

Just because we can detect cancers does not mean we should detect cancers. That’s the third fallacy I wanted to talk about.

The fourth fallacy is related to when someone is asking for more evidence for anything. There is a new drug for this cancer,so what is the evidence? Or there is this new intervention that will detect ctDNA or whatever before the cancer relapses, or before the cancer even shows up as a screening test.

Whenever there is any treatment that is being promoted and someone asks for evidence, people sometimes try to make personal attacks by saying, “Oh, so, you’re okay with patients dying. You don’t want to save lives.”

Absolutely we want to save lives. That’s why we’re in this field, and that’s why we’re asking for more evidence. You should not consider someone who is asking for evidence as evil or that this person does not want this new drug, or this person does not want this innovation. No, that person actually wants to make sure that that innovation actually helps people. That’s why that person is asking for more evidence.

If we stop asking for evidence, then our whole practice becomes based on emotions, faith, and trust rather than science. You could extrapolate it to the other extreme, like if you are not asking for evidence. If it is interpreted as someone who is asking for evidence is evil, or someone who does not want patients to get new drugs, then you could extrapolate these to people also making claims about alternative medicines or ivermectin nowadays, and claiming that this cures cancer.

Science is science. You need to be the same no matter the circumstances. If you are asking for data for ivermectin, you should also be asking data for your cancer drug that you think is going to work. We should always ask for evidence.

Asking for evidence is not a sign that whoever is asking for evidence does not want the patient to have access to the drug. It is showing that the person who is asking for evidence actually wants to make sure that the patients who get this drug are actually being helped by the drug rather than being harmed.

I’m talking about 5 fallacies today. The final, fifth fallacy is that clinical expertise does not equal expertise in making public health decisions or even expertise in critical appraisal. Someone can be a fantastic breast cancer doctor, the best oncologist for breast cancer. That does not automatically make that person the best person to evaluate clinical trials of breast cancer drugs.

Someone can be a fantastic colon cancer doctor. That does not make that person automatically the best person to evaluate whether or not colonoscopy or colon cancer screening is indicated in a certain patient population.

These population-level decisions — including should this drug be approved, should this drug be funded, and should this screening test be made a public health measure, all of these public health decisions that are done at a population level — require different expertise in critical appraisal, clinical epidemiology, and public health.

Just because someone is a fantastic clinician does not make that person a fantastic public health expert. I see on social media often that a famous doctor with expertise in their domain, let’s say a famous neurosurgeon, might say, “I think brain tumors are increasing in incidence in young persons, so we should be targeting an MRI screening for everyone over the age of 30.”

I’m just making this up, but we see examples of things not dissimilar to this. Just because someone is a neurosurgeon does not make them an expert on brain tumor epidemiology, surveillance, or screening. We should separate clinical expertise from public health expertise.

Thank you.

A version of this article first appeared on Medscape.com.

The use of surrogate endpoints to predict overall survival in oncology trials has been vigorously debated in the oncology community. 

In multiple studies, we have shown that the correlation of surrogate endpoints with overall survival has been weak and, in most cases, inadequate. However, others have made arguments against using overall survival as the gold standard endpoint for cancer drug trials. These arguments include claims that overall survival is not a feasible endpoint for pivotal cancer drug trials and that negative overall survival results may be unreliable.

In this column, we counter some of these arguments from the perspective of an oncologist who treats cancer and is engaged in oncology research and a patient advocate with lived experience of cancer.

Argument: Overall survival is not the only meaningful clinical endpoint. Response rates and progression-free survival also are important clinical endpoints because patients feel good when their tumor shrinks or they have more time without disease progression. Both situations can improve patients’ quality of life. 

Response: On the face of it, this argument sounds reasonable, but it treats response rates and progression-free survival as clinical endpoints rather than surrogate endpoints that need to be validated. A patient would surely welcome news that treatment has reduced the size of their tumor and would feel dismayed if the tumor is growing. However, these reactions rest on the belief that these changes signal recovery in the case of a shrinking tumor, or worsening symptoms in the case of a growing tumor, neither of which may be true. 

“Disease progression” and “progression-free survival” were developed as technical terms to signify whether a therapy in a phase 2 clinical trial was sufficiently promising to advance to phase 3. The terms were not intended to inform clinical practice or establish whether a new therapy provides clinically meaningful benefits for patients. In clinical trials, a progression event is defined as an increase in the sum of tumor diameters by more than 20%, or the appearance of a new lesion. These are both arbitrary benchmarks that in no way imply that a patient’s subjective experience of symptoms or their survival time would differ widely between a tumor growth of 19% vs 21%, or if a new, tiny lesion develops.

Of course, disease progression sometimes does affect a patient’s quality of life, but trials treat all progression events equally. An event in the brain that affects the patient’s ability to walk and an asymptomatic tumor growth in the lung from 1 cm to 1.3 cm are both considered disease progression in trials.

Patients care about progression-free survival because they mistakenly believe that an improvement will probably lead to improved overall survival or quality of life. But overall survival and quality of life can actually be affected detrimentally, despite progression-free survival gains, because the patient may experience treatment-related physical and financial toxicity without clinical benefit. Conversely, it is possible to have better overall survival and quality of life without any improvement in progression-free survival.

With response rates, the relationships to overall survival and quality of life are even weaker. Because progression-free survival and response rates do not represent clinical benefit, they cannot be considered clinical endpoints. They are surrogate endpoints that need to be validated.

A patient will reasonably welcome a treatment-related change that improves the prospect of living a longer life or enduring less suffering but should not be misled when these outcomes have not been demonstrated as being clinically beneficial.

Argument: It is not ethical to wait for overall survival when there is an unmet need.

Response: The phrase “unmet need” is used to represent the lack of effective treatment options for a particular condition. Regulators may expedite approval of a new drug that is intended to fulfil the need because waiting for overall survival may not seem ethical in these settings. 

In reality, however, the term is used inconsistently, both when the need is either questionable (the indication is indolent) or the need is real but one or more treatment options already exist. 

When the need for a drug with clinical benefit is genuinely unmet, presenting patients with just any drug is unethical; their need will continue to exist despite drug approvals if survival is not improved. In fact, approving ineffective drugs can impede the development of drugs that work.

Argument: Overall survival is not a practical endpoint for drug trials for rare cancers. 

Response: This argument has some practical validity because the low prevalence of a rare cancer may limit study recruitment opportunities. From an equity perspective, however, patients with rare cancers are no less deserving of treatments based on meaningful evidence of safety and efficacy than are patients with common cancers. Like any other patient with cancer, they should know those treatments have shown the potential for an overall survival or quality of life benefit before undergoing potentially toxic treatments. 

Several trials, such as FIRM-ACT, VIT-0910, AEWS0031, and rEECur, have shown that international collaboration makes randomized clinical trials with an endpoint of overall survival possible for most rare cancers. We support offering early access via accelerated approval when scientific evidence suggests a drug has potential benefit for patients, but the trial should be followed to confirm an overall survival benefit and the drug should be withdrawn if an overall survival benefit is lacking. 

Argument: The point of measuring overall survival is to ensure that the treatment does not worsen overall survival. 

Response: This argument is frequently made, even by regulators. However, unless a drug is tested in a justified noninferiority design trial where the aim is to determine whether it provides compensatory benefits — easier administration, lower cost, or fewer side effects than the alternative — the goal of new cancer drugs is to improve survival. Merely “not worsening survival” is incongruent with medicine’s fundamental goal of benevolence. All drugs have toxicities, some of which can be fatal. Recommending a treatment with toxicities because it did not worsen survival would harm patients and go against the Hippocratic Oath.

Argument: Overall survival in this trial was negative, but this was probably a false-negative finding. We expect future trials will be positive. 

Response: When a drug fails to improve overall survival, some experts will argue that it was a false-negative result and a future trial will be positive. For instance, to explain the negative overall survival findings in a 2020 confirmatory trial, the researchers cited a subgroup analysis in which the drug received accelerated approval based on early exploratory results. However, the subgroup analysis violated the statistical plan and raised the likelihood that the early trial’s favorable outcome was false-positive.

A second trend that increases the likelihood of false-positives is when the same anticancer medications are studied across multiple trials. If we run dozens of trials of the same drug, some of them will eventually be positive by chance alone. 

Dismissing a negative overall survival finding as a statistical artifact puts patients at risk because oncologists may continue to use the drug, despite it providing no discernable clinical benefit to patients. 

Argument: The overall survival was negative because of crossover so that patients in the control arm received the therapy at progression. 

Response: Highly effective drugs will continue to show overall survival gains despite crossover — that is, when patients in the control arm receive the experimental drug at the time of progression.

For example, a trial evaluating the addition of pertuzumab in first-line HER2-positive metastatic breast cancer allowed crossover and still showed a significant overall survival benefit. Another trial evaluating sunitinib vs placebo in advanced gastrointestinal stromal tumors showed a significant overall survival improvement despite 80% of patients in the placebo arm crossing over to sunitinib at progression.

However, in the SONIA trial, receiving cyclin-dependent kinase 4/6 inhibitors as first-line treatment of metastatic hormone receptor–positive breast cancer did not improve survival and added toxicities compared with receipt of the same agents as second-line treatment.

If crossover negates an overall survival benefit, this implies that patients in the experimental arm who received the drug upfront and patients in control arm who get the drug only at progression have no difference in survival times. If the survival outcomes in the groups are similar, that is evidence that the new drug should be given only to those who need treatment at the time of progression and not to everyone upfront. 

Giving patients the drug at the time of progression would spare patients toxicities, preserve efficacy, and save health systems costs.

Argument: Although overall survival is negative in the confirmatory trial, the clinical practice guidelines recommend using the drug, and the guidelines should be followed.

Response: We cannot ignore results of a confirmatory trial just because the drug is already in the guidelines. Guidelines should be updated on the basis of the latest evidence, rather than clinical practice being dictated by outdated guidelines. When a drug given accelerated approval appears in the guidelines, the confidence level of the evidence should be rated low until confirmatory results are available. If a confirmatory trial shows that overall survival does not improve, the guidelines should be quickly updated. However, even if they are not, physicians should not just blindly follow the guidelines.

Furthermore, practitioners should be aware that members of oncology guidelines committees often have financial conflicts of interest with pharmaceutical companies, which the literature shows can bias guideline recommendations.

 

The Bottom Line

We believe that despite arguments put forward against relying on or measuring overall survival, it continues to be the most patient-centric marker of clinical benefit. Patients deserve to know that the cancer drugs their oncologists recommend will contribute to a longer or better life.

Bishal Gyawali, Associate Professor, Department of Oncology and Department of Public Health Sciences; Scientist, Division of Cancer Care and Epidemiology, Queen’s University, Kingston, Ontario, Canada; Affiliated Faculty at the Program on Regulation, Therapeutics, and Law, Department of Medicine, Brigham and Women’s Hospital, Boston, Massachusetts, has disclosed the financial relationship by receiving consulting fees from Vivio Health. Sharon Batt, Cancer survivor and patient advocate; Adjunct professor, Department of Bioethics and Department of Political Science, Dalhousie University, Halifax, Nova Scotia, Canada, has disclosed no relevant financial relationships.

A version of this article first appeared on Medscape.com.

The use of surrogate endpoints to predict overall survival in oncology trials has been vigorously debated in the oncology community. 

In multiple studies, we have shown that the correlation of surrogate endpoints with overall survival has been weak and, in most cases, inadequate. However, others have made arguments against using overall survival as the gold standard endpoint for cancer drug trials. These arguments include claims that overall survival is not a feasible endpoint for pivotal cancer drug trials and that negative overall survival results may be unreliable.

In this column, we counter some of these arguments from the perspective of an oncologist who treats cancer and is engaged in oncology research and a patient advocate with lived experience of cancer.

Argument: Overall survival is not the only meaningful clinical endpoint. Response rates and progression-free survival also are important clinical endpoints because patients feel good when their tumor shrinks or they have more time without disease progression. Both situations can improve patients’ quality of life. 

Response: On the face of it, this argument sounds reasonable, but it treats response rates and progression-free survival as clinical endpoints rather than surrogate endpoints that need to be validated. A patient would surely welcome news that treatment has reduced the size of their tumor and would feel dismayed if the tumor is growing. However, these reactions rest on the belief that these changes signal recovery in the case of a shrinking tumor, or worsening symptoms in the case of a growing tumor, neither of which may be true. 

“Disease progression” and “progression-free survival” were developed as technical terms to signify whether a therapy in a phase 2 clinical trial was sufficiently promising to advance to phase 3. The terms were not intended to inform clinical practice or establish whether a new therapy provides clinically meaningful benefits for patients. In clinical trials, a progression event is defined as an increase in the sum of tumor diameters by more than 20%, or the appearance of a new lesion. These are both arbitrary benchmarks that in no way imply that a patient’s subjective experience of symptoms or their survival time would differ widely between a tumor growth of 19% vs 21%, or if a new, tiny lesion develops.

Of course, disease progression sometimes does affect a patient’s quality of life, but trials treat all progression events equally. An event in the brain that affects the patient’s ability to walk and an asymptomatic tumor growth in the lung from 1 cm to 1.3 cm are both considered disease progression in trials.

Patients care about progression-free survival because they mistakenly believe that an improvement will probably lead to improved overall survival or quality of life. But overall survival and quality of life can actually be affected detrimentally, despite progression-free survival gains, because the patient may experience treatment-related physical and financial toxicity without clinical benefit. Conversely, it is possible to have better overall survival and quality of life without any improvement in progression-free survival.

With response rates, the relationships to overall survival and quality of life are even weaker. Because progression-free survival and response rates do not represent clinical benefit, they cannot be considered clinical endpoints. They are surrogate endpoints that need to be validated.

A patient will reasonably welcome a treatment-related change that improves the prospect of living a longer life or enduring less suffering but should not be misled when these outcomes have not been demonstrated as being clinically beneficial.

Argument: It is not ethical to wait for overall survival when there is an unmet need.

Response: The phrase “unmet need” is used to represent the lack of effective treatment options for a particular condition. Regulators may expedite approval of a new drug that is intended to fulfil the need because waiting for overall survival may not seem ethical in these settings. 

In reality, however, the term is used inconsistently, both when the need is either questionable (the indication is indolent) or the need is real but one or more treatment options already exist. 

When the need for a drug with clinical benefit is genuinely unmet, presenting patients with just any drug is unethical; their need will continue to exist despite drug approvals if survival is not improved. In fact, approving ineffective drugs can impede the development of drugs that work.

Argument: Overall survival is not a practical endpoint for drug trials for rare cancers. 

Response: This argument has some practical validity because the low prevalence of a rare cancer may limit study recruitment opportunities. From an equity perspective, however, patients with rare cancers are no less deserving of treatments based on meaningful evidence of safety and efficacy than are patients with common cancers. Like any other patient with cancer, they should know those treatments have shown the potential for an overall survival or quality of life benefit before undergoing potentially toxic treatments. 

Several trials, such as FIRM-ACT, VIT-0910, AEWS0031, and rEECur, have shown that international collaboration makes randomized clinical trials with an endpoint of overall survival possible for most rare cancers. We support offering early access via accelerated approval when scientific evidence suggests a drug has potential benefit for patients, but the trial should be followed to confirm an overall survival benefit and the drug should be withdrawn if an overall survival benefit is lacking. 

Argument: The point of measuring overall survival is to ensure that the treatment does not worsen overall survival. 

Response: This argument is frequently made, even by regulators. However, unless a drug is tested in a justified noninferiority design trial where the aim is to determine whether it provides compensatory benefits — easier administration, lower cost, or fewer side effects than the alternative — the goal of new cancer drugs is to improve survival. Merely “not worsening survival” is incongruent with medicine’s fundamental goal of benevolence. All drugs have toxicities, some of which can be fatal. Recommending a treatment with toxicities because it did not worsen survival would harm patients and go against the Hippocratic Oath.

Argument: Overall survival in this trial was negative, but this was probably a false-negative finding. We expect future trials will be positive. 

Response: When a drug fails to improve overall survival, some experts will argue that it was a false-negative result and a future trial will be positive. For instance, to explain the negative overall survival findings in a 2020 confirmatory trial, the researchers cited a subgroup analysis in which the drug received accelerated approval based on early exploratory results. However, the subgroup analysis violated the statistical plan and raised the likelihood that the early trial’s favorable outcome was false-positive.

A second trend that increases the likelihood of false-positives is when the same anticancer medications are studied across multiple trials. If we run dozens of trials of the same drug, some of them will eventually be positive by chance alone. 

Dismissing a negative overall survival finding as a statistical artifact puts patients at risk because oncologists may continue to use the drug, despite it providing no discernable clinical benefit to patients. 

Argument: The overall survival was negative because of crossover so that patients in the control arm received the therapy at progression. 

Response: Highly effective drugs will continue to show overall survival gains despite crossover — that is, when patients in the control arm receive the experimental drug at the time of progression.

For example, a trial evaluating the addition of pertuzumab in first-line HER2-positive metastatic breast cancer allowed crossover and still showed a significant overall survival benefit. Another trial evaluating sunitinib vs placebo in advanced gastrointestinal stromal tumors showed a significant overall survival improvement despite 80% of patients in the placebo arm crossing over to sunitinib at progression.

However, in the SONIA trial, receiving cyclin-dependent kinase 4/6 inhibitors as first-line treatment of metastatic hormone receptor–positive breast cancer did not improve survival and added toxicities compared with receipt of the same agents as second-line treatment.

If crossover negates an overall survival benefit, this implies that patients in the experimental arm who received the drug upfront and patients in control arm who get the drug only at progression have no difference in survival times. If the survival outcomes in the groups are similar, that is evidence that the new drug should be given only to those who need treatment at the time of progression and not to everyone upfront. 

Giving patients the drug at the time of progression would spare patients toxicities, preserve efficacy, and save health systems costs.

Argument: Although overall survival is negative in the confirmatory trial, the clinical practice guidelines recommend using the drug, and the guidelines should be followed.

Response: We cannot ignore results of a confirmatory trial just because the drug is already in the guidelines. Guidelines should be updated on the basis of the latest evidence, rather than clinical practice being dictated by outdated guidelines. When a drug given accelerated approval appears in the guidelines, the confidence level of the evidence should be rated low until confirmatory results are available. If a confirmatory trial shows that overall survival does not improve, the guidelines should be quickly updated. However, even if they are not, physicians should not just blindly follow the guidelines.

Furthermore, practitioners should be aware that members of oncology guidelines committees often have financial conflicts of interest with pharmaceutical companies, which the literature shows can bias guideline recommendations.

 

The Bottom Line

We believe that despite arguments put forward against relying on or measuring overall survival, it continues to be the most patient-centric marker of clinical benefit. Patients deserve to know that the cancer drugs their oncologists recommend will contribute to a longer or better life.

Bishal Gyawali, Associate Professor, Department of Oncology and Department of Public Health Sciences; Scientist, Division of Cancer Care and Epidemiology, Queen’s University, Kingston, Ontario, Canada; Affiliated Faculty at the Program on Regulation, Therapeutics, and Law, Department of Medicine, Brigham and Women’s Hospital, Boston, Massachusetts, has disclosed the financial relationship by receiving consulting fees from Vivio Health. Sharon Batt, Cancer survivor and patient advocate; Adjunct professor, Department of Bioethics and Department of Political Science, Dalhousie University, Halifax, Nova Scotia, Canada, has disclosed no relevant financial relationships.

A version of this article first appeared on Medscape.com.

The use of surrogate endpoints to predict overall survival in oncology trials has been vigorously debated in the oncology community. 

In multiple studies, we have shown that the correlation of surrogate endpoints with overall survival has been weak and, in most cases, inadequate. However, others have made arguments against using overall survival as the gold standard endpoint for cancer drug trials. These arguments include claims that overall survival is not a feasible endpoint for pivotal cancer drug trials and that negative overall survival results may be unreliable.

In this column, we counter some of these arguments from the perspective of an oncologist who treats cancer and is engaged in oncology research and a patient advocate with lived experience of cancer.

Argument: Overall survival is not the only meaningful clinical endpoint. Response rates and progression-free survival also are important clinical endpoints because patients feel good when their tumor shrinks or they have more time without disease progression. Both situations can improve patients’ quality of life. 

Response: On the face of it, this argument sounds reasonable, but it treats response rates and progression-free survival as clinical endpoints rather than surrogate endpoints that need to be validated. A patient would surely welcome news that treatment has reduced the size of their tumor and would feel dismayed if the tumor is growing. However, these reactions rest on the belief that these changes signal recovery in the case of a shrinking tumor, or worsening symptoms in the case of a growing tumor, neither of which may be true. 

“Disease progression” and “progression-free survival” were developed as technical terms to signify whether a therapy in a phase 2 clinical trial was sufficiently promising to advance to phase 3. The terms were not intended to inform clinical practice or establish whether a new therapy provides clinically meaningful benefits for patients. In clinical trials, a progression event is defined as an increase in the sum of tumor diameters by more than 20%, or the appearance of a new lesion. These are both arbitrary benchmarks that in no way imply that a patient’s subjective experience of symptoms or their survival time would differ widely between a tumor growth of 19% vs 21%, or if a new, tiny lesion develops.

Of course, disease progression sometimes does affect a patient’s quality of life, but trials treat all progression events equally. An event in the brain that affects the patient’s ability to walk and an asymptomatic tumor growth in the lung from 1 cm to 1.3 cm are both considered disease progression in trials.

Patients care about progression-free survival because they mistakenly believe that an improvement will probably lead to improved overall survival or quality of life. But overall survival and quality of life can actually be affected detrimentally, despite progression-free survival gains, because the patient may experience treatment-related physical and financial toxicity without clinical benefit. Conversely, it is possible to have better overall survival and quality of life without any improvement in progression-free survival.

With response rates, the relationships to overall survival and quality of life are even weaker. Because progression-free survival and response rates do not represent clinical benefit, they cannot be considered clinical endpoints. They are surrogate endpoints that need to be validated.

A patient will reasonably welcome a treatment-related change that improves the prospect of living a longer life or enduring less suffering but should not be misled when these outcomes have not been demonstrated as being clinically beneficial.

Argument: It is not ethical to wait for overall survival when there is an unmet need.

Response: The phrase “unmet need” is used to represent the lack of effective treatment options for a particular condition. Regulators may expedite approval of a new drug that is intended to fulfil the need because waiting for overall survival may not seem ethical in these settings. 

In reality, however, the term is used inconsistently, both when the need is either questionable (the indication is indolent) or the need is real but one or more treatment options already exist. 

When the need for a drug with clinical benefit is genuinely unmet, presenting patients with just any drug is unethical; their need will continue to exist despite drug approvals if survival is not improved. In fact, approving ineffective drugs can impede the development of drugs that work.

Argument: Overall survival is not a practical endpoint for drug trials for rare cancers. 

Response: This argument has some practical validity because the low prevalence of a rare cancer may limit study recruitment opportunities. From an equity perspective, however, patients with rare cancers are no less deserving of treatments based on meaningful evidence of safety and efficacy than are patients with common cancers. Like any other patient with cancer, they should know those treatments have shown the potential for an overall survival or quality of life benefit before undergoing potentially toxic treatments. 

Several trials, such as FIRM-ACT, VIT-0910, AEWS0031, and rEECur, have shown that international collaboration makes randomized clinical trials with an endpoint of overall survival possible for most rare cancers. We support offering early access via accelerated approval when scientific evidence suggests a drug has potential benefit for patients, but the trial should be followed to confirm an overall survival benefit and the drug should be withdrawn if an overall survival benefit is lacking. 

Argument: The point of measuring overall survival is to ensure that the treatment does not worsen overall survival. 

Response: This argument is frequently made, even by regulators. However, unless a drug is tested in a justified noninferiority design trial where the aim is to determine whether it provides compensatory benefits — easier administration, lower cost, or fewer side effects than the alternative — the goal of new cancer drugs is to improve survival. Merely “not worsening survival” is incongruent with medicine’s fundamental goal of benevolence. All drugs have toxicities, some of which can be fatal. Recommending a treatment with toxicities because it did not worsen survival would harm patients and go against the Hippocratic Oath.

Argument: Overall survival in this trial was negative, but this was probably a false-negative finding. We expect future trials will be positive. 

Response: When a drug fails to improve overall survival, some experts will argue that it was a false-negative result and a future trial will be positive. For instance, to explain the negative overall survival findings in a 2020 confirmatory trial, the researchers cited a subgroup analysis in which the drug received accelerated approval based on early exploratory results. However, the subgroup analysis violated the statistical plan and raised the likelihood that the early trial’s favorable outcome was false-positive.

A second trend that increases the likelihood of false-positives is when the same anticancer medications are studied across multiple trials. If we run dozens of trials of the same drug, some of them will eventually be positive by chance alone. 

Dismissing a negative overall survival finding as a statistical artifact puts patients at risk because oncologists may continue to use the drug, despite it providing no discernable clinical benefit to patients. 

Argument: The overall survival was negative because of crossover so that patients in the control arm received the therapy at progression. 

Response: Highly effective drugs will continue to show overall survival gains despite crossover — that is, when patients in the control arm receive the experimental drug at the time of progression.

For example, a trial evaluating the addition of pertuzumab in first-line HER2-positive metastatic breast cancer allowed crossover and still showed a significant overall survival benefit. Another trial evaluating sunitinib vs placebo in advanced gastrointestinal stromal tumors showed a significant overall survival improvement despite 80% of patients in the placebo arm crossing over to sunitinib at progression.

However, in the SONIA trial, receiving cyclin-dependent kinase 4/6 inhibitors as first-line treatment of metastatic hormone receptor–positive breast cancer did not improve survival and added toxicities compared with receipt of the same agents as second-line treatment.

If crossover negates an overall survival benefit, this implies that patients in the experimental arm who received the drug upfront and patients in control arm who get the drug only at progression have no difference in survival times. If the survival outcomes in the groups are similar, that is evidence that the new drug should be given only to those who need treatment at the time of progression and not to everyone upfront. 

Giving patients the drug at the time of progression would spare patients toxicities, preserve efficacy, and save health systems costs.

Argument: Although overall survival is negative in the confirmatory trial, the clinical practice guidelines recommend using the drug, and the guidelines should be followed.

Response: We cannot ignore results of a confirmatory trial just because the drug is already in the guidelines. Guidelines should be updated on the basis of the latest evidence, rather than clinical practice being dictated by outdated guidelines. When a drug given accelerated approval appears in the guidelines, the confidence level of the evidence should be rated low until confirmatory results are available. If a confirmatory trial shows that overall survival does not improve, the guidelines should be quickly updated. However, even if they are not, physicians should not just blindly follow the guidelines.

Furthermore, practitioners should be aware that members of oncology guidelines committees often have financial conflicts of interest with pharmaceutical companies, which the literature shows can bias guideline recommendations.

 

The Bottom Line

We believe that despite arguments put forward against relying on or measuring overall survival, it continues to be the most patient-centric marker of clinical benefit. Patients deserve to know that the cancer drugs their oncologists recommend will contribute to a longer or better life.

Bishal Gyawali, Associate Professor, Department of Oncology and Department of Public Health Sciences; Scientist, Division of Cancer Care and Epidemiology, Queen’s University, Kingston, Ontario, Canada; Affiliated Faculty at the Program on Regulation, Therapeutics, and Law, Department of Medicine, Brigham and Women’s Hospital, Boston, Massachusetts, has disclosed the financial relationship by receiving consulting fees from Vivio Health. Sharon Batt, Cancer survivor and patient advocate; Adjunct professor, Department of Bioethics and Department of Political Science, Dalhousie University, Halifax, Nova Scotia, Canada, has disclosed no relevant financial relationships.

A version of this article first appeared on Medscape.com.