JCO Precision Oncology Conversations
Health & Fitness:Medicine
Tumor Mutational Burden as a Predictor of First-Line Immune Checkpoint Inhibitor Versus Carboplatin Benefit in Cisplatin-Unfit Patients With Urothelial Carcinoma, with Dr. Shilpa Gupta
JCO PO author Dr. Shilpa Gupta, MD, Associate Professor of Medicine at the Cleveland Clinic and GU Medical Oncology Director, shares analysis on outcomes in real-world settings for metastatic urothelial carcinoma (mUC) patients. Host Dr. Rafeh Naqash and Dr. Gupta discuss the utility of tumor mutational burden (TMB) to determine treatment, and mUC patient response from immune checkpoint inhibitors (ICPI) as compared with carboplatin. Click here to read the article!
TRANSCRIPT
Dr. Rafeh Naqash: Hello and welcome to JCO Precision Oncology Conversations. I am Dr. Rafeh Naqash, assistant professor of medicine at OU Stephenson Cancer Center. You're listening to the JCO Precision Oncology Conversations podcast.
Today I'll be talking with Dr. Shilpa Gupta, who is an associate professor of medicine at the Cleveland Clinic and also the GU Medical Oncology Director. And we'll be talking about their group's recent paper, ‘Tumor Mutational Burden as a Predictor of First-Line Immune Checkpoint Inhibitor Versus Carboplatin Benefit in Cisplatin-Unfit Patients With Urothelial Carcinoma’.
Full disclosures for our guest can be found on the article’s publication page.
Hello and welcome to the podcast, Dr. Gupta. It's nice to have you here. For the sake of this podcast, we'll be referring to each other using our first names. So welcome and thanks for joining us today.
Dr. Shilpa Gupta: It's my pleasure to be here, Rafeh, I’m really excited about chatting about this paper with you. Thank you for the opportunity.
Dr. Rafeh Naqash: Thank you so much. So today we'll be discussing this interesting publication of yours, talking about biomarkers. And I often refer to biomarkers as the Pandora's Box for immune checkpoint inhibitors because definitely one size does not fit all. And reading through your paper, I saw a lot of interesting findings that you have defined in this publication. But for starters, what was the premise and background of why you wanted to study this question of tumor mutational burden as a biomarker in this patient population?
Dr. Shilpa Gupta: Yeah, that's a great question, Rafeh. The treatment paradigm for urothelial cancer patients has really evolved over the last many years. For example, patients who are eligible to receive cisplatin-based chemotherapy, that's the treatment of choice. And for patients who are not eligible to receive cisplatin due to a variety of reasons like chronic kidney disease, heart failure, peripheral neuropathy, poor performance status, or hearing loss, in the past, we used to treat them with gemcitabine and carboplatin, but outcomes were quite dismal with median overall survival less than six months or so. And then in 2017, the approval of pembrolizumab and atezolizumab as single agents was welcome news because these patients had more durable responses and survival was longer than historically with gemcitabine-carboplatin. And this is what became the standard of care based on the FDA expedited approval.
However, in 2018, the FDA restricted the use of immunotherapy only to those patients whose tumors had high PD-L1 or who were not eligible to receive carboplatin, based on the interim analysis from the phase three trials IMvigor130, which compared atezolizumab to gemcitabine-carboplatin, one of the cohorts for cis-ineligible patients, and KEYNOTE-361, which compared pembrolizumab to gemcitabine-carboplatin in the cis-ineligible cohort. And furthermore, recently, the FDA actually further restricted the label for pembrolizumab, because in the phase 3 study, even in high PD-L1 subgroups, pembrolizumab did worse than gemcitabine-carboplatin, regardless of their PD-L1 status. There were early deaths, lower response rates, and in the IMvigor130 study, we recently saw that atezolizumab was actually withdrawn for this indication altogether.
So there has been this attraction for PD-L1 for a long time, but now multiple studies in urothelial cancer have shown that PD-L1 is not a durable biomarker. And we wanted to see if there's other biomarkers which can be accessible at the point of care. And we wanted to study how tumor mutational burden can or cannot pan out as a treatment selection or complementary to clinical criteria. Right now, there's no biomarkers to guide treatment for patients in urothelial cancer for carboplatin or immunotherapy use. And that was the premise for the study.
Dr. Rafeh Naqash: Excellent. Thank you so much for that detailed understanding of why you decided to pursue this.
Now, from the listener standpoint when you define cisplatin-ineligible patients, in your practice, what is the percentage of patients that you see who are technically cisplatin-ineligible? Does comorbidity play an important role in determining which patients, or does it depend on your discussion with the patient? What are those factors that you would describe to define what cisplatin-ineligibility would constitute?
Dr. Shilpa Gupta: So historically, Matt Galsky and colleagues described cisplatin-ineligibility as patients with a creatinine clearance less than 60 mLs per minute, hearing loss greater than grade two, poor ECOG performance status two or higher, peripheral neuropathy, which is significant or significant heart failure. Now, those all make patients ineligible for cisplatin.
Now, more recently, we know that we can safely give cisplatin as long as creatinine clearance is above 50. So for the real world, 50 is a threshold where we can use split dose cisplatin. And I'll say, given that bladder cancer or urothelial cancer is a disease of the elderly, median age being 71 years, a lot of our patients have these comorbidities, chronic kidney disease, diabetes, and whatnot, which precludes us from using cisplatin. So in the real world, I would say that around 50% of patients are ineligible to receive cisplatin.
Dr. Rafeh Naqash: Interesting. And that goes back to the point where not everything that resulted from clinical trials, or the data that we get, may not be exactly applicable to the real world patient population, as you have pointed out in this interesting paper. So going back to the manuscript now from a methodology perspective, what kind of data did you include to get to the results that we'll talk about next? What was the inclusion and what was the patient population in this analysis?
Dr. Shilpa Gupta: So the patient population basically were patients who had a confirmed diagnosis of metastatic urothelial cancer. And the databases we used were the US-wide Flatiron Foundation Medicine Clinical Genomic Database, which has patients who were listed as metastatic urothelial cancer. But in addition, they also had genomic testing performed from their tumors, and results were available. And we accessed the database between 2011 until April 2021. And all these patients had had genomic testing using Foundation Medicine assay. And this de-identified data was basically US-wide across 280 cancer clinics and that's around 800 sites of care. And there's a whole range of retrospective longitudinal clinical data that was available, derived from the electronic health records comprising patient-level structured and unstructured data and also their genomic information from the tumors. And there was clinical data including demographics, lab values, performance status, timing of treatment, exposure, as well as time of progression and survival.
We decided to include patients if they received a frontline single agent immunotherapy, no matter what it was, whether pembrolizumab, atezolizumab, Nivolumab, durvalumab or avelumab, or a carboplatin-based chemotherapy. And just for the readers, this is a retrospective review. So we just used these selected patients who got in these therapies. We also required that these patients had tumor mutational burden information available through the tissue biopsy and patients who received chemotherapy and immunotherapy together were excluded and details are present in the manuscript, but this was pretty much the broad selection criteria.
Dr. Rafeh Naqash: Thank you so much. And definitely a very representative patient population from a real world setting with different therapy and different other clinical variables that are relevant in the real world setting.
So from an analysis standpoint, you, from what I read, define both a predictive and a prognostic aspect to tumor mutational burden. Could you tell us more about those results and highlight some of the interesting findings from that perspective?
Dr. Shilpa Gupta: Yes, absolutely. So as you know, tumor mutation burden cut off of ten mutations per megabase is currently utilized by the FDA, whereby approval of pembrolizumab for tumor agnostic condition was made. So that's what we considered high versus low. And we found that in this, after propensity weighing in, the tumor mutational burden less than ten group, basically those patients did not benefit from checkpoint inhibitor single agent as compared to tumor mutational burden of ten or greater. And so basically, we found that patients who had tumor mutational burden ten or higher overall had more favorable progression-free survival time to next treatment, as well as overall survival when they got a single agent immune checkpoint inhibitor, as opposed to those who got carboplatin, and also when compared to those who had tumor mutational burden less than ten. So we also looked at PD-L1 information available from the genomic database, but it was only available for around 35% of patients and still we were able to see that PD-L1 did not correlate with any of these outcomes as we show in the paper.
Dr. Rafeh Naqash: I see. And as you mentioned, you show both time to treatment failure PFS being better in TMB high patients defined as ten mutations per megabase. I didn't specifically see results related to TMB high versus low in a carboplatin specific cohort. Is that analysis something that was looked at and trying to understand whether neoantigens in a platinum-based setting specifically make a difference whether high TMB is predictive there in the carboplatin setting. Was that looked at?
Dr. Shilpa Gupta: So yes, we looked at, in the Figure 4, for the comparison of the TMB and which we were looking at the checkpoint inhibitor versus chemo. So for TMB low the chemotherapy cohort had more favorable results. Is this what you were getting at?
Dr. Rafeh Naqash: Yeah, I think what I was specifically trying to look at, like you have shown in the paper, is TMB is predictive of benefit with checkpoint inhibitors and is also prognostic in the checkpoint inhibitor setting. So my question was more whether it had a prognostic implication in a carboplatin specific cohort. So meaning high TMB, whether it correlated with better outcomes with carboplatin therapy versus low TMB. So if that was looked at.
Dr. Shilpa Gupta: We didn't look at that specifically, we only compared whether high TMB did better with the immunotherapy or chemotherapy.
Dr. Rafeh Naqash: And some of the correlation of this in my mind comes from some data that people have looked at in the lung cancer setting, whether high TMB makes a difference and for example, resected lung cancer patients, which usually gets platinum-based adjuvant therapies. So that's why I was wondering if there's any correlation there. But this is definitely interesting.
Now, my next question was going to be in your manuscript you mentioned around 30% of patients had tumor mutational burden more than or equal to ten. Did you identify any other unique characteristics from any other mutational standpoint or a PD-L1 standpoint in the high TMB cohort?
Dr. Shilpa Gupta: Yes. So PD-L1 didn't really stand out to be a very steady biomarker in our experience. And this is also what was reflected in the phase three trials like DANUBE where they looked at the durvalumab and tremelimumab, IMvigor130 or KEYNOTE-361. So that was pretty consistent that these studies also showed TMB to be more useful in exploratory analysis. Of course, these patients were not stratified based on that. And we also looked at other emerging biomarkers, for example, F-TBRS and angiogenesis gene expression signatures as well as tGE3. And we need to evaluate them in a separate study to see what pans out.
But for now, I think as far as in the real world, we are looking at a lot of genomic testing being done and right now we really don't know how to use that for making treatment decisions, right? PD-L1 has really phased out as of any utility whatsoever. And using TMB; I think in addition to the clinical characteristics, like when possible, we should be offering patients carboplatin. There's no doubt about that for cisplatin ineligible patients. But there's those patients who, if they're refusing chemotherapy and we really can't make a case for giving them single agent immunotherapy, I think TMB can come in handy to justify and make sure that we're not doing them a disservice by not giving carboplatin. And I think future trials need to use this biomarker in a prospective setting to further establish its utility.
Dr. Rafeh Naqash: Definitely, I agree it's a case-by-case situation from a patient standpoint to determine what therapy is appropriate for the patient and what is most realistic, what is the expectation that the patient has, from that treatment.
Now, from a TMB standpoint, one of the ongoing debates is if it is a binary cut off or whether it could be tertiles for a certain tumor type or quartiles. Was there any subanalysis or any subsequent study that your team would be looking at from a TMB cut off standpoint? Maybe a higher cut off would mean a better outcome and maybe lesser duration of therapy in those patients. Is that somewhat of a consideration?
Dr. Shilpa Gupta: Yeah, that's a great question, Rafeh. And I think the reason we stuck to it as a binary end point is because that's the FDA definition, so people don't try to extrapolate based on anything higher or lower. But yeah, that's a great question. And I know in lung cancer they're looking at different ranges. As far as urothelial cancer, we just stuck to the ten mutations per megabase for now.
Dr. Rafeh Naqash: Of course. And one of the other interesting things I really like to see in the paper is your figure specifically on the ECOG performance status and how clinical trials sometimes do not include patients on the higher ECOG performance status spectrum. And your study obviously had a good representation on that standpoint. What were some of the findings from the ECOG standpoint that were somewhat different in your cohort than what you would see in clinical trials in general?
Dr. Shilpa Gupta: Yes, as we've shown in Figure 5, the ECOG in real world, it was quite an eye opener to see that there was a considerable number of patients who were documented as ECOG performance status three. And if you see the ECOG performance status two bar was around 50% and ECOG performance status one was also lower than what has traditionally been included in the phase three trials. And in the phase three trials, there's hardly any patients with ECOG performance status two compared to what we saw in the real world. And very few patients, in fact, hardly any had ECOG performance status zero in our real world analysis. So clearly the trials need to be more inclusive, as has been the ASCO message all along. And it's always very surprising to see the big gap between the real world and the clinical trial patient population.
Dr. Rafeh Naqash: Definitely, I think more and more, especially cooperative group trials that you and many others are leading, are trying to be as inclusive as possible, which is important to get a better understanding of how these therapies do in different patient populations. And one of the questions I wanted to ask you, and I've seen this a few times in different checkpoint therapy treated tumors, is this initial rapid progression in some patients where the chemotherapy arm does better, but the immunotherapy arm kind of falls rapidly and then starts plateauing. In your clinical experience, have you seen that? And if yes, what are the features of some of those patients that have this rapid progression from a clinical and both from a biomarker standpoint?
Dr. Shilpa Gupta: That's a great question, Rafeh, and we do see that every now and then, and especially in my experience, we've seen that in women in particular who have bone metastases are really challenging to treat with immunotherapy. And sometimes we find that the disease just rapidly blows through immunotherapy and we really need to do more biomarker work to understand what determines these biomarkers of hyper-progression, so to speak. I know there's a lot of work going on in the field and we are also trying to understand these by serially collecting blood and circulating tumor DNA from our patients during their treatment journey.
Dr. Rafeh Naqash: Exactly. Definitely work in progress and another unique patient population where more needs to be done to understand what are the events that lead to these hyper-progression aspects, whether it's in the bone or brain or any other compartment in the body.
Well, this has been exciting and interesting, but before we end, we try to know a little bit more about the investigator, the author. So, Shilpa, can you tell us a little bit about your journey in oncology and your journey as a trainee, your journey as faculty, as a clinical trialist, as a successful clinical trialist? And any advice for junior investigators listening to this conversation?
Dr. Shilpa Gupta: Yeah, thank you for asking. I think oncology always struck me as a very exciting field back in my residency days, 2005, 2006. And at the time, so much was going on, like just drugs like bevasizumab were just coming around for colorectal cancer and in lung cancer drugs like EGFR inhibitors were coming around. And that kind of really excited me. And talking with my mentor at the time, who was a really well-renowned transplanter, he said to me that if he had to do it all over again, he would love to get into solid tumor oncology with all the excitement that's going on. I was drawn to oncology also because of, not only it's a learning experience every day, but it can be very gratifying to see amazing responses and patients living longer despite having advanced disease, and also provides a lot of challenges every day when every patient is not the same. So I think that was the reason why I was drawn to oncology and provides us an opportunity to really develop new therapies as opposed to some of the other specialties because of how challenging the patient population is.
And as far as my journey, you know, I've now been in the US for almost 18 years and have been in a variety of places, and I think it's been a very rewarding journey despite multiple bumps along the way. And I'm really glad to be doing what I'm doing and trying to advance the field, clinical trials, and learning from people around me.
Dr. Rafeh Naqash: Thank you so much for giving us a little glimpse into your journey and your experiences. And it's always inspiring to listen to successful investigators and also try to emulate in some ways what you have done and what you've achieved. And thank you again for coming on this podcast. And thank you for choosing JCO Precision Oncology as a destination for your manuscript, and hopefully we'll see more of the same from you and your group in the subsequent years to come and more in this field of biomarkers.
Thank you for listening to JCO Precision Oncology Conversations. You can find all our shows, including this one, at ASCO.org/podcasts or wherever you get your podcasts. To stay up to date, be sure to follow and share JCO PO content on Twitter @JCOPO_ASCO. All JCO PO articles and series can be found at ascopubs.org/journal/PO.
The purpose of this podcast is to educate and to inform. This is not a substitute for professional medical care and is not intended for use in the diagnosis or treatment of individual conditions.
Guests on this podcast express their own opinions, experience, and conclusions. Guest statements on the podcast do not express the opinions of ASCO. The mention of any product, service, organization, activity, or therapy should not be construed as an ASCO endorsement.
Guest Bio
Shilpa Gupta, MD, is Associate Professor of Medicine at the Cleveland Clinic and GU Medical Oncology Director.
Guest disclosures
Stock and Other Ownership Interests: Nektar, Moderna Therapeutics
Honoraria: Bristol Myers Squibb
Consulting or Advisory Role: Gilead Sciences, Guardant Health, AVEO, EMD Serono, Pfizer, Merck, Loxo/Lilly
Speakers' Bureau: Bristol Myers Squib
Create your
podcast in
minutes
It is Free