Dr. Ali Raza Khaki – Loose Regulatory Standards Portend a New Era of Imprecision Oncology – Part 1/2

Jul 1, 2021 | Rising Tide Podcast


On The Rising Tide Podcast, scientists from the independent, nonprofit Center for Genomic Interpretation discuss with leading experts the need to further raise the bar on accuracy and quality in clinical genetics, genomics, and precision medicine. Only through improving accuracy and quality can the promise of precision medicine be realized. CGI’s ELEVATEGENETICS services are made available to health insurers and stakeholders, to help them identify the most accurate and clinically useful genetic and genomic tests, and to help them steer clear of low quality tests that have the potential to lead to patient harm and wasted expenditure.


GARLAPOW: What happens when a drug company requests FDA approval for use of a precision medicine cancer drug that goes beyond the scope of its clinical trials or has no demonstrated benefit? And then what happens when the FDA approves these requests, and why should anybody care? Today, on The Rising Tide Podcast we talk with Dr. Ali Khaki of Stanford University, who in 2021 authored the peer-reviewed article with the ominous title of “Loose Regulatory Standards Portend a New Era of Imprecision Oncology.” Dr. Khaki is a hematologist/oncologist with board certification in oncology, hematology, and internal medicine. He is also a Clinical Assistant Professor of Oncology at Stanford University School of Medicine. And what’s most important? Dr. Khaki is paying attention and publishing about the advances in his field, and some of what he sees worries him. In preview, Dr. Khaki has some concerns about the current state of precision medicine in treating cancer patients. 

KHAKI: My concern with precision medicine is that it’s perceived to be more precise than it truly is, and when this happens, this can lead people to argue against performing randomized clinical trials or producing high quality evidence. Or in some cases even disregarding the proven therapies that have been shown to work in the past, in favor of unproven therapies because of precision medicine. You’re taking something that has not been proven to have a benefit compared to things that have been shown in rigorous studies to actually help patients live longer, live better, and I thought that these broad drug approvals contradicted the principles of precision medicine where we’re trying to be more precise and not less precise.

GARLAPOW: So if some precision medicine is over-promising, what was the promise to begin with?

KHAKI: But more specifically for oncology I think we think of precision medicine when we reference the idea of using molecular or genomic markers or data of a patient’s cancer or their germline to guide therapy decisions. I think this definition is a pretty accurate representation of how we think about it today in 2021. And this would be in contrast to I guess what you’d say is non-precision medicine or regular oncology, which is the world before the modern definition of precision medicine where we would use histology or pathology to guide therapy. I think the promise of precision medicine oncology is the idea that we could test the patient’s tumor for molecular change, and then use that information to pick the best treatment for that patient. Ideally that treatment would be highly effective with minimal toxicity. 

GARLAPOW: Okay so precision medicine is meant to be precise and accurate, but what’s the harm for patients if drugs are FDA approved beyond the scope of what they’ve been demonstrated to be able to do? 

KHAKI: I think the harm is three things: one, they’ll experience the side effects of treatment; two, they’ll experience the expensive cost of therapy; and three, all of that is with a low likelihood of benefit.  

GARLAPOW: And beyond the patient level, what is the impact of making precision medicine not so precise? 

KHAKI: The big consequence of all this is uncertainty. We are welcoming a lot more uncertainty into the practice of medicine. Uncertainty has consequences for patients, consequences for our society. 

GARLAPOW: I’m Dr. Megan Garlapow with The Center for Genomic Interpretation and you’re listening to The Rising Tide Podcast, where we learn from experts about improving the accuracy and quality of precision medicine and clinical genetics and genomics. Please note that this podcast does not provide medical advice, and is not intended to be a substitute for professional medical advice, diagnosis, or treatment. Always seek the advice of your physician or other qualified healthcare provider with any questions you may have regarding a medical condition. Additionally, comments of The Rising Tide’s guests are their own and do not necessarily reflect the position of The Center for Genomic Interpretation. 

Today I’m joined by Dr. Ali Raza Khaki, genitourinary oncologist at Stanford Cancer Institute and Clinical Assistant Professor at Stanford University. This is the first of a two-part series with Dr. Khaki discussing his recent publication in cancer investigation “Loose Regulatory Standards Portend a New Era of Imprecision Oncology” about the promises and pitfalls of precision medicine; particularly the potential consequences for patients when FDA approvals may reach beyond the mark as they approve drugs for treating cancer. Please note that Dr. Khaki’s views are his own and do not necessarily represent the views of his employer. Additionally some of what we discuss is speculative in nature and/or describes use of treatment that has not been approved by the FDA. Welcome Dr. Khaki thanks for joining The Rising Tide. 

KHAKI: Thanks Megan for having me on this podcast and for taking interest in my recent publication. I look forward to this conversation. 

GARLAPOW: Absolutely. So Ali, what is precision medicine? 

KHAKI: It’s interesting that the definition of precision medicine has evolved over time. So I thought it would be helpful to just look up a definition to see how others are defining it today. So I looked at MedlinePlus, which is a website from the National Library of Medicine under the NIH. And the definition they have is, “Precision medicine is an emerging approach for disease treatment and prevention that takes into account individual variability in genes, environment, and lifestyle for each person.” I think this definition is a pretty accurate representation of how we think about it today in 2021. But more specifically for oncology, I think we think of precision medicine when we reference the idea of using molecular or genomic markers or data of a patient’s cancer or their germline to guide therapy decisions. And this would be in contrast to I guess what you’d say is non-precision medicine or regular oncology, which is the world before the modern definition of precision medicine where we would use histology or pathology to guide therapy. And so we were still using molecular markers, it’s just that we’ve gotten more precise or I guess to smaller measurements, right into the genome. And that’s what we call precision medicine today. 

GARLAPOW: Okay so what are the biggest promises of precision medicine? 

KHAKI: So I think the promise of precision medicine oncology is the idea that we could test the patient’s tumor for molecular change and then use that information to pick the best treatment for that patient. And then ideally that treatment would be highly effective with minimal toxicity. So there have been some examples of this that have been true success stories. The best example is probably one of the first, which is imatinib or chronic myelogenous leukemia or CML.

Prior to the discovery of imatinib  the average life expectancy for a CML patient was measured in years after diagnosis. And treatment included things like toxic therapies like a bone marrow transplant. And now with the identification of BCR-ABL fusion gene and the approval of imatinib and other similar medications, patients can live almost a near-normal lifespan which is remarkable on a single oral medication. There have been other success stories with targeted therapy things like in lung cancer we have specific mutations EGFR ALK and others, HER2 positivity and breast cancer, but even with these breakthroughs only a minority of patients benefit from precision medicine or precision oncology today. In a study from a few years ago published by Vinay Prasad and colleagues at the Oregon Health Sciences University at the time, it was published in JAMA Oncology, they estimated about eight percent of patients with cancer were eligible for a genome targeted therapy. And this was up from five percent in 2006. So making progress between 2006 and 2018 when this was published, but it’s not growing at an exponential rate. It’s sort of more of an iterative process of growth. 

GARLAPOW: Okay so taking all that into account, what are some of the pitfalls of precision medicine? 

KHAKI: Yeah we’ll be talking about the pitfalls today when we get into my article and some of the limitations of recent drug approvals. But I think in general, my concern with precision medicine is that it’s perceived to be more precise than it truly is. And when this happens, this can lead people to argue against performing randomized clinical trials, or producing high quality evidence, or in some cases even disregarding the proven therapies that have been shown to work in the past in favor of unproven therapies because of precision medicine. And ultimately I think biological plausibility can be quite compelling. And with that and the hype around some of this precision medicine, I worry that we are putting sometimes the cart before the horse and are leading some in the biomedical and scientific community to be less rigorous in the empirical evaluation of therapies that have been considered under this umbrella.

GARLAPOW: I totally agree, thank you. So can you please provide a general overview of your January 2021 commentary “Loose Regulatory Standards Portend a New Era of Imprecision Oncology” published in Cancer Investigation. What motivated you to write it? What happened? And why does it matter? 

KHAKI: Yeah so in this commentary, I reviewed two recent drug approvals from the FDA. They were both approved in 2020, that I thought the drug approval indications were broader than the studied population. Or in the population that was shown to have benefit in the clinical trial. And I thought that these broad drug approvals contradicted the principles of precision medicine where we’re trying to be more precise and not less precise, and the idea of identifying a specific group that would benefit from a therapy and only treating those patients. So when you are approving it for more than those patients, then that’s a complete contradiction. So I wanted to highlight that and discuss this. 

GARLAPOW: Okay. So as a reminder for everyone, clinical trials are research studies performed on people to assess a medical, surgical, or behavioral intervention, and are used in cancer to determine the safety and efficacy of things such as new drugs or even drugs that have been around but are being tested in diseases or types of tumors or cancer in which they have not previously been tested. FDA approvals beyond the trial populations have occurred in more than one therapy in precision medicine. How does this happen, Ali? And how can the FDA approve beyond the trial population? 

KHAKI: Yeah so I think in the two examples that I cite that we’ll get into here, two different things are happening which is interesting. So in the prostate cancer example with olaparib, the label indication reflects the studied population, but the population noted to have benefit was only a subset of the entire study population. So under the auspices of precision medicine, we should focus on the population that benefited is what I argue. Conversely, in the pembrolizumab approval, which we’ll also talk about, the FDA label is broader than the studied population because the belief is that the marker of interest, in this case the tumor mutational burden (TMB), may be predictive regardless of cancer type even though all cancer types were not included. So in one setting the population was narrower than the indication, the study population was narrower than the eventual indication. And the other one, the study population was the eventual indication, but the population benefiting was a narrower population. 

GARLAPOW: Okay. So it’s only been a few weeks at the time of the recording of this podcast since the publication of your commentary. Have you gotten any feedback from your colleagues or from the FDA?

KHAKI: Well, to be honest with you, I wasn’t sure if anyone had even read it, so before you reached out to me I was very excited to hear from you. But I will say it did take me some time to get published. I was rejected by a few journals before Cancer Investigation accepted it, and this could be because there are others who are speaking about similar issues. There was a good debate in the Annals of Oncology between Vivek Subbiah and Vinay Prasad regarding the pembrolizumab approval published sometime in the fall of 2020. And then also Mike Schweitzer and some of my old colleagues at the University of Washington had written a nice piece in the JCO about the olaparib approval. There are others who have been speaking about these issues. I think one thing that I tried to do in my commentary was try to lace these two issues together and make a broader statement about the problematic theme that this sort of trajectory can set up for precision oncology. 

GARLAPOW: Okay. Very interesting. So let’s jump into the trial on the drug called pembrolizumab. You use two different trials and approvals in your commentary. The first on pembrolizumab or pembro, discusses this immunotherapeutic agent. Immunotherapies are drugs that have transformed some areas of cancer care. There are several approved immunotherapeutic agents, and sometimes patients are eligible to receive an immunotherapy such as pembro based not on the type of cancer that they have, but rather on certain markers present on the tumor. In the case of pembro the FDA approved it for tumor agnostic treatment of metastatic or unresectable solid cancers with tumor mutation burden levels above a certain level. Tumor agnostic means that the treatment is not based on type of cancer. Tumor mutation burden is the number of mutations present in a tumor per stretch of the genome. The FDA approval is for solid tumors with at least 10 mutations per megabase or chunk of genome. You describe the trial KEYNOTE-158 that the FDA used to make this approval of pembro. What happened in the trial? 

KHAKI: Yeah so the KEYNOTE-158 trial was an 11-cohort, single-arm, Phase 2 trial that assessed pembrolizumab in patients with advanced, less frequently occurring solid tumors. The cancer types included biliary adenosine carcinoma, neuroendocrine tumors, endometrial carcinoma, cervical and vulvar squamous cell carcinoma, small cell lung cancer, malignant pleural mesothelioma, capillary and follicular thyroid carcinoma, and salivary gland cancers. 10 of the 11 cohorts that were used in the trial were part of a retrospective analysis that led to this approval. So patients in the trial were treated with pembrolizumab for up to two years or until disease progression or unacceptable toxicity. And as part of the trial, tissue samples have been collected. And then for this exploratory analysis these samples underwent FoundationOne testing, which is a genomic testing platform which includes as part of the evaluation tumor mutational burden. So the exploratory analysis that led to the approval compared the tumor response rate or what percent of patients had their tumor shrink by at least 30 percent between those with a TMB of 10 mutations per megabase or higher, compared to those with a TMB of less than 10. And the analysis identified 102 patients who were considered TMB high and the response rate in this population was 29.60% or about a third of patients had had their tumors shrink by at least 30 percent. Compared to only 6 percent 6.7 percent in the 688 patients who were considered non-TMB high. So while the response rate was higher in the TMB high population, the durability response, or how long a patient who had a response maintained that response, as well as the overall survival was the same between the two groups, the TMB high and TMB low group. So it’s really only a greater proportion of patients were having their tumor shrink. What’s interesting is that in patients that were enrolled in this trial between 2016 and 2019 but since this time or in this time, some of the cancers included in this trial have also had approvals based on other clinical trials. So among the 102 patients who were TMB high, 34 had small cell lung cancer which has subsequently had immune checkpoint inhibitor approval. 16 had cervical cancer and 15 had endometrial cancer, so 65 those 65 patients actually all had a pembrolizumab approval for their histology and not necessarily for the TMB. And the next two cancers actually, anal and vulvar, also had other immune checkpoint inhibitor approval. So drugs like pembrolizumab that work the same way but are not pembrolizumab. So of the 102 patients who were considered TMB high, 89% of them had tumors which had already shown benefits to checkpoint inhibitors. So, why I bring this up is that it may be that the histology, what we already knew about the cancer before looking at TMB, is telling us who is going to benefit from the treatment, and the TMB may not be adding that much to what we already know from the histology. 

GARLAPOW: Okay that’s very interesting and I think that’s very important to note. So we have this trial population that you just described for us. How does it compare to the FDA drug approval population, what we see now on the pembro label from the FDA? 

KHAKI: So I mentioned the trial population includes these 10 broad tumor types of some of the more rare tumors but the FDA drug approval did not restrict the approval to these cancer types. So the FDA drug approval is what as you said tumor agnostic was the word you used, and so it didn’t specify any cancer type. It said that anyone who’s got a solid tumor that’s metastatic or advanced, could get this treatment as long as their TMB was greater than 10 mutations per megabase. So what’s notably absent from the trial population which I think is important is that patients with prostate and breast cancer, the two most prevalent cancer types in the country, were not represented in this trial population.

GARLAPOW: Okay so patients with breast cancer and with prostate cancer, they weren’t represented in KEYNOTE- 158. So what? Why does this matter? 

KHAKI: Yeah I focus on these populations because in addition to them being the two most prevalent cancers, they’re the two most common cancers in America, neither of these cancers has shown much efficacy. The immune checkpoint inhibitor class of drugs, like pembrolizumab, has not been shown to really work well in these two cancers. There are some exceptions, Triple-negative breast cancer has shown some benefit to this, but in general these two cancer types have not responded well to these therapies. In a prior Phase 1 study of immune checkpoint inhibitor, a different immune checkpoint inhibitor called nivolumab in breast cancer, the response rate, the percent that had that 30 percent shrinkage, was only three percent. Similarly in a prostate cancer trial, the KEYNOTE-199 trial investigating pembrolizumab in prostate cancer, the response rate again was three to five percent. So this is similar to the non-TMB-high response rates that we’re seeing in the re-analysis of the KEYNOTE-158 trial. So what this means is that this broad TMB-high pembrolizumab of approval includes these common cancers even though patients with these cancers weren’t included in this analysis. So if the medication doesn’t work for these cancers, which we don’t know, since it wasn’t included, there is potential for significant harm to a lot of people. But then the next question becomes, so what do we know about TMB for these cancers? And the data I mentioned before were just general clinical trials for those populations of breast and prostate. There have been some analyses looking at TMB in these populations and there the data is a little bit more promising so maybe we’re not completely lost here. There was a presentation from something called the TAPUR Registry at ASCO in 2019 that reported outcomes for breast cancer patients based on TMB, and they used a cut point of nine mutations per megabase. And they reported that breast cancer patients had a 20 percent response rate with pembrolizumab in that registry, so it’s not even a clinical trial. But they’re capturing the data after the fact and following those patients forward. But what we don’t know about this population, based on that presentation, is what subtype of breast cancer they had. Were these Triple-negative breast cancer patients that would have already had an FDA indication for a checkpoint inhibitor? And were there other types of breast cancer and did all the different subtypes of breast cancer benefit equally with the TMB-high? Or was histology or other markers more predictive than TMB? Similarly in prostate cancer, there was a follow-up biomarker analysis of the KEYNOTE-199 trial that I mentioned that had the low response rate. The three to five percent response rate. I did note that people who had a higher TMB was associated with a better PSA response, meaning a higher proportion of those patients had their PSA decline while on pembrolizumab. However, the trial did not show an association between TMB-high and radiographic progression pre-survival. Meaning even though the PSA was declining, it didn’t necessarily mean that the tumors were shrinking or patients were necessarily living longer. That’s not been shown. And so there may be a benefit, but there’s a lot of uncertainty around whether patients with prostate breast cancer benefit from these therapies.

GARLAPOW: So what could this mean for a patient with non-TNBC breast cancer or with prostate cancer? They go into their oncologist and they’re prescribed pembro. What could happen to that patient? 

KHAKI: Yeah I mean it’s a concern. I mean it’s possible that they could benefit to pembrolizumab, but we don’t know that. There’s a lot of uncertainty about what’s going to happen to them. And for example if you’re a patient with castration-resistant prostate cancer who has, you know, there’s five FDA approved medications for castration-resistant prostate cancer. If you go on pembrolizumab before receiving these therapies, you’re taking something that has not been proven to have a benefit compared to things that have been shown in rigorous studies to actually help patients live longer, live better. And so that’s a decision a patient and their oncologists have to make together, but it comes with a lot of risk and a lot of uncertainty. And then similarly, instead of us doing this off the cuff sort of in the wild west, wouldn’t it be better for us to do this on a clinical trial? There’s plenty of breast and prostate cancer patients out there. They’re the two most common cancers in America. So why not measure TMB in these patients and do a rigorous study to actually show that there’s a benefit? 

GARLAPOW: I agree. So let’s talk about tumor agnostic approvals, where it’s not based on tumor type, it’s just based on solid tumors, those approvals just based on the TMB, tumor mutation burden. There are other drugs with other indications that are sort of a tumor agnostic approach. Does this work to have one threshold of a biomarker to cover, say, all of the solid tumors? Or if it does work in some circumstances like what do we need to know about that marker to apply it across different tumor types. Is it appropriate at all? 

KHAKI: Yeah so as you mentioned there, this was the fourth tumor agnostic approval that I’m aware of. In addition to pembrolizumab for TMB-high tumors, there’s pembrolizumab that was approved for MSI-high or mismatch repair deficient tumors. So there’s other molecular markers that have also been used to approve pembrolizumab. And then there are two tumor agnostic approvals for these tropomyosin receptor kinase inhibitors, the TRK kinase inhibitors, the larotrectinib and entrectinib. In all four of these approvals, a couple of things to note: one, these approvals are all based on single arm studies, so studies without randomization. So there already exists uncertainty of how well they work because without randomization there’s a lot more uncertainty about the true benefit in a population. In addition, I think anytime you have a tumor agnostic approval, if you are from a tumor that was not well represented in that study population there’s additional uncertainty regarding whether the medication will have the same effect on you that it showed in the broader studied population. Cancer biology is very complicated and in general things that work in one setting don’t necessarily work in another setting. For example, HER2 targeted agents are much more effective for HER2-positive breast cancer than for HER2-positive gastric cancer. Or in the example of TMB, a wide variety of different cut points for TMB have been utilized for different cancers and in different settings. It isn’t clear whether there’s anything magical about the 10 threshold in pembrolizumab. Maybe 10 is an important number for lung cancer but the number could be three or five for prostate cancer or for something else. So if we are developing these things with a heterogeneous population, then there’s a lot more uncertainty about the benefit in a specific patient or population. So I think that tumor agnostic approvals, I’m not sure if they’re ever appropriate, but I think that if they are appropriate, there has to be this understanding that there has to be a decision made between patients, payers, regulatory bodies, doctors, that they’re accepting of more uncertainty of the benefit. And it depends on how you weigh that uncertainty because we’re never going to be able to do a study on  every patient in a single trial. So we can take two paths forward. One is we can identify the patients who will benefit and we can study them and then show efficacy and approve medications. Or if we think that this is the way to do things, then we have to accept the uncertainty that comes with this decision.

GARLAPOW: Okay. That’s a really good perspective for considering this and navigating that uncertainty. Let’s go on to the second part of your commentary on the PARP inhibitor olaparib and its approval. You discuss its approval in metastatic castration-resistant prostate cancer, or mCRPC, based on results from the Phase 3, PROfound trial. In general terms, what happened in the PROfound trial? 

KHAKI: So the PROfound trial was a randomized Phase 3 trial that investigated this PARP inhibitor olaparib in patients with metastatic castration-resistant prostate cancer who had alterations in a pre-specified gene involved in homologous recombination repair and who had progression on a novel hormonal therapy, which is enzalutamide or abiraterone. That’s a lot of different terms and words thereDR. KHAKIlet me break it down. So put simply, in this trial, men who had advanced prostate cancer, they progressed on at least one oral anti-testosterone agent before and had a mutation in one of a number of pre-specified genes. Patients were randomized in a two-to-one fashion. The two patients got olaparib to one patient getting the physician’s choice of a novel hormonal therapy, enzalutamide or abiraterone. And the primary endpoint of the trial was progression-free survival, the time until the cancer starts growing or someone dies. And the trial was designed with two cohorts; the first cohort had patients with one of three different types of mutations which is BRCA1, BRCA2, or ATM. And these are probably the best established genes in this homologous recombination repair pathway and have been shown in other cancer types as well to have a benefit with PARP inhibitors. In cohort 2 the patients had alterations in one of 12 other genes and the trial overall was a positive trial which showed that progression-free survival was longer in patients who got olaparib compared to the control arm which is the physician’s choice. And it was about just under six months in the olaparib arm and three and a half months in the control arm, so just a couple months longer. And more recently there was a follow-up publication at the end of 2020 that showed that there’s also improvement in overall survival for patients who were treated with olaparib. However, the pre-specified subgroup analysis that was presented in the initial paper showed that the patients who benefited most from these therapies were mostly just the BRCA1 and BRCA2 patients. And most of the other subgroups didn’t seem to benefit. And, in fact, one of those subgroups, the PPP2R2A mutation group, seem to do worse than they did with hormonal therapy.

GARLAPOW: Okay, so as a reminder for our audience, olaparib is one of a class of drugs called PARP inhibitors. They function by blocking the PARP enzyme which can cause death of tumor cells in some types of cancer by blocking the tumor’s ability to repair its DNA. Cancers with mutations in certain genes in these repair pathways can be particularly susceptible to treatment with PARP inhibitors such as olaparib. The most well-known mutations are in BRCA1 BRCA2 or BRCA1/BRCA2 genes. But there’s a whole suite of genes involved in DNA repair, some of which might have mutations that could make them sensitive to treatment with PARP inhibition. Olaparib has been approved to treat breast, ovarian, and pancreatic cancer with mutations in BRCA1 and/or BRCA2 with some other details in those approvals. Can you please describe olaparib’s approval in metastatic castration-resistant prostate cancer? 

KHAKI: Yeah, so the FDA approval for olaparib states, this is quoting right from the label, “For the treatment of adult patients with deleterious or suspected deleterious germline or somatic homologous recombination repair gene-mutated metastatic castration-resistant prostate cancer who have progressed following prior treatment with enzalutamide or abiraterone.” It’s a mouthful. But the big thing here to note is that the FDA didn’t limit the approval to just the patients that benefit in the trial, which is the BRCA1 and BRCA2 patients. And that’s the population that has been shown in other cancers to benefit is the BRCA1 and BRCA2 patients. And it allowed the approval for the full patient population study. So this is unique to prostate cancer. All the other cancer types that this approval has taken place has been restricted to the BRCA population.

GARLAPOW: So the approval in mCRPC, that prostate cancer approval, it includes more genes than in olaparib’s other approvals. But as you described previously, an analysis of patients based on the gene with the mutation showed that patients with mutations in BRCA1 or BRCA2 BRCA1 or 2, responded well to olaparib compared to the physician’s choice. But other groups either did not respond any differently, or that one group, they responded worse than physician’s choice. What does this mean? Why does this matter? 

KHAKI: What this means to me is that when the FDA approves it for all these other genes, that means that many more men are eligible for treatment with olaparib than were shown to benefit in the trial. And so a patient who is, for example, averse to chemotherapy, he may be inclined to turn to Dr. Google and find out that olaparib is approved. And he’s got one of those mutations. And he might be inclined to choose that over chemotherapy or other proven therapies that have been shown to save lives or extend life in prostate cancer. Or there might be a community oncologist who’s seeing 20, 30, 40 patients a day and can’t keep up. The first patient he sees is a breast, and then he sees a prostate, then you’re seeing a sarcoma, and you’re seeing a lymphoma, and leukemia. There’s no way an oncologist can keep up with all the different cancer types and all different therapies. He or she may not know the details of which genes were associated with a benefit. They may not have read this PROfound trial in such great detail to see that one figure that shows that stratified response rate. So this could mean that there’s patients like these patients who experience harm because they’re being given a therapy that’s not been proven to work for their cancer. That harm can happen for a period of months until they have evidence of progression and what is the harm? I think the harm is three things. One, they’ll experience the side effects of treatment. Two, they’ll experience the expensive cost of therapy. And three, all that is with a low likelihood of benefit. So instead of that, we could restrict the approval to the population that showed where they had a benefit, and then it makes it easier for patients, for clinicians, and for the entire oncology community to use these medications responsibly. 

GARLAPOW: Okay. I think it is notable too that in Europe the approval was limited to BRCA1 and BRCA2. I was listening recently to a tumor board, a virtual tumor board, and they described an mCRPC patient with a mutation in ATM, one of those DNA repair genes that is now on the prostate label for olaparib. But one of the clinicians, he commented on how since olaparib’s expanded approval beyond BRCA1 and BRCA2, he’s had a few patients with pancreatic cancer, which is on the BRCA1/2 olaparib label. But he’s had a few patients with pancreatic cancer with mutations, potentially pathogenic mutations, in their ATM genes, and he’s given them olaparib off-label. So I’m going to be a bit of a devil’s advocate here. Pancreatic cancer is terrible. I mean the best cancer is the cancer that never occurs, all cancer is terrible. But isn’t it compassionate to just try something for this patient? What is compassion in this sort of a situation?  

KHAKI: I think compassion is giving patients options of treatments that have been shown to work. Or things that can help them. And I’m not convinced that this is helpful. I think that if you are preventing patients from getting treatments that have been proven to be helpful, or if you are promising them, or even they perceive a promise that this medication is going to be artificial and it just takes them away from other meaningful things in their life like spending time with their family or just dealing with life without dealing with the cost and the toxicity of a therapy that’s not been proven to work, that’s what I think compassion is. I don’t think compassion is giving people therapies just because we have them and because we can. Because that’s why we do studies in the first place. So we can have confidence that what we are prescribing can work in patients that we give them. I think in general, when we are lenient in how we approve medications across the board, we are welcoming a lot more uncertainty into the practice of medicine, and that uncertainty has consequences for patients, it has consequences for our society because of the cost of these medications to the country. And in general I think that the consequences are, one is toxicity, the other thing is that time away from life is also a big consequence for patients. So even if there are no other proven therapies and one may be inclined to be like, “Oh 

I can give them something more,” but maybe the more compassionate thing to do is to be honest with your patient. Go spend the rest of your life with your family. That’s probably the more compassionate thing to do in my eyes.

GARLAPOW: Thank you Ali. So in your commentary you write, and I thought that this was really poignant, “Therefore, a blanket tumor agnostic approval without evidence of efficacy in the two most prevalent solid tumors, and at a single cut point, strays from the precision aspired with precision medicine.” Both of these trials taken together, your commentary, the points you are driving home here, what is this going to do to patients? And what is the long-term outlook of this if we don’t get this under control? Do doctors understand this?  

KHAKI: I think that I’m being too redundant here, but I think the big consequence of all this is uncertainty. That’s what I always say is that we’re just introducing more and more uncertainty into the practice of medicine. So in the blanket approval scenario, we don’t know if the drugs are going to work in the populations that are not studied. So if we approve these medications without knowing how well they work, then we have to accept that we are exposing patients to toxicity and costs with a low likelihood of benefit. The history of medicine will tell you the majority of things that are studied, don’t end up working. So we should assume that things are not going to work until they’ve been proven to work. And so once approved, it limits our ability, once the drug is approved, it limits our ability to capture the data to better understand how well the drugs are working. In the long-term outlook, if this continues, is what I call the wild west, where we have lots of drugs, lots of tests, but not really knowing how much anything works really well. So do doctors understand this? I think that it’s hard to say. I think that there are many doctors who do understand this. I think the people who understand this best are probably your community oncologist who are seeing all those patients and they’re on the front lines and they’re the ones who are understanding the limitations of our therapeutics as well as the burden that this can put on patients both in terms of costs as well as sort of additional toxicity. I think some of us in academia can sometimes be blinded by the lure of precision oncology. I think one of the limitations in medical education broadly is that doctors aren’t well trained in the critical appraisal of clinical trials, so if we don’t improve this training in medical education, we also are not equipping future physicians to have the skills they need to make the decisions in the world. Especially the regulatory bodies not doing so. So if the FDA is there to protect you and approve only things that have a good efficacy, then you can get away with not having critical appraisal in your physicians because there’s other skills that physicians have. But once you take away that regulatory body actually regulating, then you’re putting that onus on doctors, patients, and other members in the healthcare community.

GARLAPOW: I agree. Absolutely. So, this is The Rising Tide Podcast. A rising tide lifts all boats. What’s the way forward from here? What improvements are needed to support meaningful advances in precision medicine? How can we raise the tide so that precision medicine can improve? 

KHAKI: Yeah, it’s a great question and I think we can raise the tide by doing the following. First, I think we have to conduct precise, randomized clinical trials of novel therapies in the population of interest. So precision medicine should start with precision in study design. I think using clinical trial tools is important. So using randomization. If you want to, you can use randomized Phase 2 trials so you have smaller trials that can help give you a no-go decision about whether to do a larger Phase 3 study. We can use stratification. So we sort of stratify a randomized trial by the populations of interest, so we can be more precise in generating the evidence that we need to be confident about the production of improving benefit. I think number two, we can train future physicians with the skills of critical appraisal, so they are equipped to counsel their patients appropriately. And then ideally I think that the third thing I’d say is that we should ideally remove the heavy hand of private enterprise that’s in biomedicine today. The more we let pharmaceutical companies conduct their own clinical trials, and we allow clinical trialists to have conflicts of interest, or in academia to have conflicts of interest with these companies, or we take away the FDA’s power to regulate, the more we are building a future where we are allowing all this to happen, what we’re seeing happen. We’re allowing poorly designed clinical trials without  physicians able to speak up about the flaws, and without a regulatory body stopping them, this all leads to this lenient approval without necessarily clear benefit to patients.

GARLAPOW: Wow. Ali thank you so much for sharing your insights, knowledge, and experience. This is incredible. This whole conversation has been excellent. This concludes the first of a two-part series we’re doing with Ali. So stay tuned for the next episode when we get to dive into more details of these trials and approvals and look for opportunities to improve, so that the right patients can receive the right treatment at the right time. 


If you loved this episode of The Rising Tide podcast, please support us on Patreon and select us on Amazon Smile. The Center for Genomic Interpretation can help your organization identify high quality clinical genetics and genomics partners, or can help a laboratory identify areas for quality improvement. Through Elevate Genetics, we engage with laboratories, encourage genetic and genomic test validity, and assess test efficacy. Find us at clarifygenetics.org and genomicinterpretation.org.

For our listeners following along on YouTube, remember to hit subscribe and to activate alerts for when we post new episodes. You don’t want to miss any of these important conversations with leading experts. Together we can raise the tide.


Narrated by Dr. Megan Garlapow

Produced and edited by Kathryn Mraz, Hunter Giles, and Brynlee Buhler