® discussion, “Emerging Strategies for Renal Cell Carcinoma.” I’m Dr Michael Atkins. I’m the deputy director of the Georgetown-Lombardi Comprehensive Cancer Center in Washington, DC, and I’m a medical oncologist with a long-standing interest in melanoma, kidney cancer, and immunotherapy. I’m here today with Dr Bob Alter.

I’m Bob Alter, a medical oncologist and co-division chief of the Division of Genitourinary Malignancies at the John Theurer Cancer Center in Hackensack, New Jersey. We’re going to start today discussing the biology of renal cell carcinoma [RCC], and we’re going to have a nice conversation about the past, present, and future of renal cell carcinoma.

The most common histology of renal cell carcinoma is clear cell renal cell carcinoma, comprising about 75% of histologies of patients we see with renal cell carcinoma. There’s an increased awareness of the complexity of clear cell renal cell carcinoma, which has multiple molecular profiles, each with unique biologic properties.

Sporadic clear cell renal cell carcinoma is associated with a loss of function or alterations of the 

 gene, which is a tumor suppressor gene, and this is characterized by neoangiogenesis and upregulation of the hypoxia-inducing factor [HIF].

The wild-type VHL protein functions as a ubiquitin ligase participating in the degradation of half the patients bearing VHL alterations, resulting in high levels of HIF and resulting in the upregulation of VEGF. VEGF activates VEGFR, triggering the PI3K-AKT signaling cascade: downstream mTOR is activated, which leads to transcription of a variety of tumor-promoting factors, resulting in increased cellular migration and angiogenesis.

We have found several drivers of tumor progression, including VEGFR, other transmembrane receptors that may potentially drive metastases, including MET and AXE2, that we found more recently. Through these studies we have found therapeutic developments of anti-VEGF and anti-mTOR agents that have led to treatments for RCC. This has also improved our understanding of T-cell biology and has defined the role of immune checkpoints in curbing immune responses to prevent autoimmunity while simultaneously allowing tumor growth escape. This led to the development of immune checkpoint inhibitors designed to remove the breaks on immune inactivation against tumors. Through this science, dating to even the 1990s and the development of agents then, we’re now moving forward, which is hopefully leading to better treatments for our patients.

 That’s a good description. We’ve probably been working for 2 decades with a toolbox that includes anti-VEGF or antiangiogenic agents, with things such as sunitinib, pazopanib, axitinib, and more recently cabozantinib and lenvatinib. Potentially coming soon would be tivozanib, mTOR inhibitors such as everolimus and temsirolimus, and immunotherapies with interferon and high-dose IL-2, with kidney cancer being the disease where high-dose IL-2 was first FDA approved. What we’ve seen over the past decade is the increasing emergence of checkpoint inhibitor immune therapy with anti–PD-1 and anti–PD-L1 pathway blockers finding their way into the treatment algorithm and moving to combinations of either checkpoint inhibitors or checkpoint inhibitors with anti-VEGF agents as doublets as the preferred frontline treatments.

The Biology of RCC and Improved Treatment Outcomes

 With regard to how we use those therapies, we’re at a crossroads. We initially had some prognostic variables that came out of Memorial Sloan Kettering Cancer Center that were focused on how patients did in terms of overall survival when they were receiving interferon, which didn’t work well. It was the really the natural biology of the disease.

Through the early part of the 2000s when the VEGF receptor TKIs [tyrosine kinase inhibitors] were being used, the IMDC [International Metastatic Renal Cell Carcinoma Database Consortium] group developed new prognostic variables that predicted outcome for patients receiving VEGF receptor TKIs. It divided patients into favorable, intermediate, and poor-risk categories based on the number of poor prognostic factors that the patients had. Those factors were used as stratification factors in the phase 3 trials for checkpoint inhibitors. Those have led to results in different prognostic groups, and that has led to different approvals, but it’s unclear to what extent those clinical prognostic factors represent the true biology of the tumor.

For example, data seem to suggest that favorable-risk patients, as defined by the IMDC classification, are more likely to have tumors that have 

 loss. They are more likely to be VEGF-driven with less of an inflammatory or immunologic signature than the intermediate and poor-risk tumors. Tumors that are favorable risk are also more likely to be firm and nuclear grade 2 or 3 and lack sarcomatoid features. They more frequently have metastases to the pancreas and the thyroid gland, and on a genomic level, they are more likely to have comutations in 

, which is another gene that’s on chromosome 3 right next to 

. Those are 2 sets of mutations in chromosome 3, near 

, that are mutually exclusive. Patients with favorable risk are also less likely to have PD-L1 expression and less likely to have T cells in their tumor microenvironment. These are the patients who are rich in tumors that are likely to respond to VEGF receptor TKIs or those who could be managed by close surveillance.

However, it’s important to point out that those clinical characteristics do not always correlate directly with these biologic characteristics, meaning that you could have a patient with favorable risk based on clinical characteristics who has a tumor biology that is more likely to be responsive to immunotherapy. Therefore, I’m getting away from using those IMDC prognostic characteristics more and more in my decision-making for what therapies I should give to patients.

Tumor Biology in Favorable to Poor-Risk RCC

Let’s go back to when and how these drugs came about. I don’t want to say that I predate the Cytokine Working Group, but that was the foundation for how immunotherapies got to be. Prior to 1992, we were using agents for which we were just guessing. I know people were giving circadian rhythm 5-FU [5-fluorouracil] and materials like that, but in 1992 when the FDA approved high-dose IL-2, it was sometimes still being associated with interferon, though interferon did not get its FDA approval until it was combined with Avastin [bevacizumab]. We were working with immunotherapy for a good 13 years until December 20, 2005, when sorafenib got FDA approval. Five weeks later, sunitinib got approval. We suddenly had an embarrassment of riches. We suddenly had 3 drugs FDA approved for a disease for which we had none, and that’s where the wheels started coming off. It started becoming a population of patients to whom we were given access, and pharmaceutical companies were getting drugs by expanded access. Then mTORs came into the world. Soon enough, it was between temsirolimus and everolimus. Shortly after that, we started getting pazopanib, axitinib, and a slew of other therapies.

Historically, we’ve been watching the initial immunotherapies, followed by TKIs [tyrosine kinase inhibitors] and followed by mTORs. In 2015, we suddenly started getting second-line data on using nivolumab as a therapeutic option for patients. It comes full-circle: We now have patients getting immunotherapy all over again only to find out that immunotherapies are becoming the backbone for first-line therapy, sometimes in combination with immunotherapy or combinations with tyrosine kinase inhibitors. Through these novel therapies and combination regimens, we’ve made tremendous strides over the last 28 years, and we’re suddenly offering patients clinical trials with competitive therapies in combination compared with single-agent guessing. It’s back to the embarrassment of riches.

 Yes. I’ve frequently been asked to comment on whether there’s a need for tolerable, effective therapies or how important it is to focus on quality of life. From the beginning of my career working with high-dose IL-2, I’ve always used therapies for kidney cancer with the goal of trying to cure patients. Patients’ preferred outcome is to have their disease go away, have their treatment stopped, and have them get back to their normal life. Immunotherapy offers that possibility, whether it was a 5% to 10% chance with high-dose IL-2 or potentially as high as a 30% to 40% chance with nivolumab-ipilimumab. That is the preferred strategy. Patients returning to their normal lives ultimately results in the best quality of life, even though that may result in more toxicity for a short period of time. Curative therapies have always been my preferred strategy, and I find that patients are willing to put up with a lot of toxicity if the ultimate goal is a chance for cure.

In the absence of curative therapies, that’s when you start talking about therapeutic index being important, and one has to balance the toxicity and efficacy of treatment choices to maximize the quality of a patient’s life. That’s where one might still focus on single agents and only selectively target the tumor with minimal off-target toxicities. Fortunately, that is not a first-line consideration for most patients with kidney cancer, but it’s still something one has to take into consideration if the patients don’t respond to up-front immunotherapy regimens.

Improved Outcomes With Novel and Combination Therapies

 As we talk about how patients present and our initial assessment, we medical oncologists are getting the patient handed to us, so we have the diagnosis. Going to the primary care physicians and community oncologists when a patient presents to you as an unknown primary care patient, the typical laboratory work would be a chemistry panel, including an LDH [lactate dehydrogenase], a CBC [complete blood count], and a urinalysis. Suspicion for having a renal mass leads to a CT scan of the abdomen and pelvis, sometimes an MRI, a chest x-ray, and sometimes a CT scan of the chest. If it’s clinically appropriate based on the patient, you could do bone scanning, imaging of the head, and a core biopsy.

The challenges we have occur when patients present with a renal mass or renal tumors and have symptoms such as hematuria, IVC [inferior vena cava] thrombus, bone metastasis, painful metastasis, decreasing functional performance status, and extensive disease. Oftentimes, it’s not so simple that can just take our time and let these diagnoses evolve and have the opportunity to make therapeutic decisions. We somehow have to believe that there is a way of gauging patients by tumor volume as well as some of the parameters that are given to us.

Dating to the early use of immunotherapy, the first risk classification we had when it came to utilizing patient prognostic features would be the MSKCC [Memorial Sloan Kettering Cancer Center Risk Score], or the risk categorization on patients based on the data utilizing interferon in patients postnephrectomy. In those patients, through Dr Bob Motzer’s guidance, and through the Cytokine Working Group, they discussed the best ways of assessing these patients. They looked at parameters, including time from initial diagnosis to initial treatment, lab studies of LDH, calcium level, and hemoglobin. They were also looking at a patient’s functional performance status and using those 5 tools to categorize patients who have a favorable risk to patients who have no risk factors. Patients will be intermediate risk if they have 1 or 2 risk factors, and then patients who have more risk factors would be categorized as patients with poor risk.

Through the IMDC [International Metastatic Renal Cell Cancer Database Consortium Criteria], they substituted out the LDH, maybe through inconsistencies or through utilizing better ways of assessing the tumor burden, and they now incorporate the white blood cell count and platelet count, so neutrophilia as well as thrombocytosis would now be substituted for that. In those patient-risk categorizations, they are recognizing that these factors can sometimes predict survival as well as, through the NCCN [National Comprehensive Cancer Network] Guidelines, guide us about how to use treatments in a therapeutic fashion based on the analysis.

We’ll talk also about the CheckMate-214 trial as well as the KEYNOTE-426 trial.

 I personally reject the IMDC risk categorization for making choices between the 2 regimens that you just discussed: nivolumab-ipilimumab and axitinib-pembrolizumab. I know that the NCCN makes recommendations, but those recommendations are not based on the data; they’re based on the ways in which the trials were designed rather than the results of those trials. Once again, these models were developed for VEGF receptor TKI [tyrosine kinase inhibitor] therapies, which were almost invariably noncurative. When you’re seeing patients and you have potentially curative treatments, you should be using those treatments first rather than going to palliative therapies as your first option.

My approach when I’m seeing a patient is to ignore the NCCN Guidelines and decide first whether a patient needs therapy. If they do, then I need to determine whether they’re a candidate for either a protocol or for immunotherapy. If they need therapy and are candidates for immunotherapy, meaning they don’t have a reason why they can’t receive a checkpoint inhibitor and don’t have an alternative clinical trial for which they’re eligible, I will start them on nivolumab-ipilimumab with the hope of curing them. I will then leave other treatments that are less likely to produce durable benefit for patients who don’t have complete response or durable response to up-front immune therapy.

You mentioned active surveillance before. I find that active surveillance is probably the most difficult conversation to have with patients who are referred to a medical oncologist. They are being told by a neurologist or a primary care doctor that they already have disease with metastases, whether it’s before or after they’ve had a cytoreductive nephrectomy, and they may even be asymptomatic where the volume is nominal, and there may be small pulmonary lesions. Your first recommendation is to watch and repeat the scan in 3 months, and they are flabbergasted at how they were told they have metastatic carcinoma and you wish to observe it. You bring up a great point. Despite the painful conversation, you have to convince them that we’re taking care of the patient beforeworrying about the disease.

I agree with you. Through our high-dose IL-2 system, we were still holding on to those small groups of patients who were having complete, durable remissions with great quality of life; they had the ability not to endure toxicities of therapy while still being able to have survival. The approach should always be about reaching for the stars with the least toxic therapy. Using combination immunotherapies in an up-front regimen, even in patients who are favorable risk, may still be an excellent option as your first-line therapy.

 Yes. Going back to the days of high-dose IL-2, the patients who we would now label favorable risk had the most durable benefit with immunotherapy. In the trial I presented at ASCO [American Society of Clinical Oncology Annual Meeting] this year looking at nivolumab monotherapy in patients with metastatic kidney cancer as first-line treatment, which you participated in as well, we had a 50% response rate in the favorable-risk group of patients with over 13% complete responses with just nivolumab monotherapy. There wasn’t a single patient with favorable risk who progressed at the 12-week first scan and wouldn’t have been able to receive a subsequent therapy with a TKI after progression. If they were tolerating nivolumab well, for ipilimumab-nivolumab, we also saw another 13% of patients respond when we added ipilimumab to the nivolumab in the crossover group. A number of those patients who were who were favorable risk in that trial are likely off-treatment and doing well without evidence of disease or at least without evidence of disease progression just from nivolumab monotherapy. To me, it is still not fair that we don’t have an FDA-approved immunotherapy-alone strategy for patients with favorable risk based just on the design of the trial.

That’s right. If we want to talk about the data we’ve been seeing over the last 2 or 2½ years, in CheckMate-214, KEYNOTE-426, and JAVELIN Renal 101, and in CheckMate 9ER soon, we see that we’re not sitting here using single-agent TKIs anymore. We’ve stepped up a bit. The expectations for these patients are not just to have responses. We’re looking for durable responses, and we’re looking for overall survival. We should still think about a backbone, and we should think about how our patients are still looking for complete remission.

Risk Assessment and Diagnostic Work-up

Let’s talk about the KEYNOTE-426 trial first. It was an open-label, phase 3 trial with an untreated clear cell renal cell patient population. The study was comparing sunitinib as the standard-of-care arm, looking at patients receiving axitinib at 5-mg dosage orally twice a day and pembrolizumab 400 mg IV [intravenous] every 3 weeks. The coprimary end points are overall survival and progression-free survival [PFS] in the intent-to-treat patient population with the secondary end point being overall response rate. The breakdown was an average of 30% patients with favorable risk and 55% patients with intermediate risk with the rest of the patients being poor risk. The data were initially presented at ASCO [American Society of Clinical Oncology Annual Meeting] in 2018.

Subsequently, they recently presented the 27-month analysis, and it still shows a clear difference with axitinib and pembrolizumab being significantly better in all end points, regardless of overall survival with the coprimary end point of progression-free survival as well as the secondary end point of overall response rate. We’re seeing a median duration of response in patients getting the axitinib- pembrolizumab being close to 24 months, patients with sunitinib being close to 16 months. For the overall survival, the hazard ratio was leaning toward a 30% or 32% improvement in patients receiving axitinib-pembrolizumab compared with those receiving Sutent [sunitinib].

We’re now starting to see CR [complete remission] in patients getting axitinib-pembrolizumab at the 27-month analysis compared with patients at the 12-month analysis. The number with axitinib-pembrolizumab was up to 9% CR compared with Sutent [sunitinib], which was 3%.

Going back to CheckMate-214, Nizar [Tannir] presented the 4-year data, which were quite impressive not just because the numbers were not only beneficial but also staying straight from years 3 through 4. This was a study looking at using ipilimumab at 3 mg/kg.

I’m sorry, 1 mg/kg. It was then using nivolumab, which was 3 mg/kg, and that was a flat dose given in combination every 2 weeks for 4 cycles followed by maintenance nivolumab compared with sunitinib as the standard-of-care arm. With the data, I’m not even going to talk about the 2- and 3-year data. Let’s talk about the 4-year data: The progression-free survival is close to 60% or over 60% of patients receiving the combination of ipilimumab-nivolumab compared with 30% receiving sunitinib.

 The PFS is 35% in patients with intermediate and poor risk as well as the overall group.

That’s right, but there is still an 11% CR in patients receiving ipilimumab-nivolumab, while it’s 2% in patients receiving sunitinib. There are 2 studies that have clearly knocked off the standard-of-care arm, and we’re now utilizing combination therapy of either I/O [immuno-oncology]–I/O or I/O–TKI [tyrosine kinase inhibitor], but you have to embrace the facts. Nizar presented at ASCO GU [American Society of Clinical Oncology Genitourinary Cancers Symposium], which was the last official conference when we were still allowed to see one another in public. They presented on patients who had a CR and what happened to patients who stayed on therapy, patients who came off therapy, and patients who came off because of toxicity. For these patients, 59 are still doing significantly well in durable, complete remission 4-plus years out. That is truly a testament to how we start our conversation and what our goals are: complete remission and quality of life with patients having no toxicity long term. Unfortunately, TKIs give those challenges to the patients.

 It does raise questions: Now that we have these 2 studies, both with significant improvement in overall survival vs sunitinib, how does one choose between those 2 regimens? Both are great regimens to have available. They certainly offer advantages over sunitinib. We don’t have any head-to-head comparisons. Even though both studies used the same control arm, there are significant differences in the patient populations enrolled in the studies, in when the trials were conducted, and therefore, in what was available for salvage treatment. Where the studies were conducted also influenced what was available for salvage treatment, which made it difficult or hazardous to compare the studies.

Furthermore, the slopes of those curves for the 2 experimental arms in those studies are different, with the nivolumab-ipilimumab arm having a typical immunotherapy slope similar to sunitinib in terms of PFS for the first 12 months, and then there was a flattening of the PFS curve after 18 months while the sunitinib curve continues to fall off. The axitinib-pembrolizumab curve has an early separation between it and the sunitinib curve, and it then runs parallel to the sunitinib curve over time. It is not yet showing a tail end on that progression-free survival curve, and it remains unclear whether it’s possible to stop axitinib and have the benefit persist. We’re going to need longer follow-up and head-to-head comparisons to know whether there’s a true immune effect to the axitinib-pembrolizumab arm, whether we see that flattening of the curve and the ability to stop treatment and the ability to not just create CRs but cure patients. That’s going to take some time.

Newly Diagnosed RCC: KEYNOTE-426 and CheckMate-214

 A lot of people have asked, “What is the role of single-agent VEGF receptor TKIs [tyrosine kinase inhibitors] now that we have these 2 doublets? Which is the standard of care in the frontline treatment setting? Are there any patients who should get sunitinib or pazopanib or cabozantinib in the frontline treatment setting?” For me, the answer is very few patients. It would be the patients who can’t get immunotherapy because they have a serious autoimmune condition that, if exacerbated, could cause problems. Or it could be that this leaves them on an immunosuppressive drug that would interfere with the activity of the immunotherapy. It could also be patients who are on steroids either because of brain metastases or spinal cord compression or patients who are in the hospital where you need to get them out of the hospital. These are all patients for whom a single-agent TKI might make sense as an initial treatment, and it could potentially be the only treatment depending on whether it’s safe at any point in the future to give them immunotherapy.

We’re still going to have the winter upon us, so there may be a travel restriction as well. Even for those patients, if they’re on a TKI, you still have to monitor these patients closely with toxicity. We haven’t even talked about the avelumab-axitinib arm, and those data are very good. The overall survival [OS] is not mature, so we cannot pass judgment. When we talked about those 3 treatments, we always thought that was the third that came to the party, and there’s no OS data as of yet. That should probably come out within the next year or so, and we’ll have to see how that plays out. We’ll be following the slope to see where that takes us.

 You’re talking about the axitinib-avelumab arm?

 Unfortunately, I don’t think we’ll ever have overall survival benefit. As we’ve seen with the axitinib-pembrolizumab data, the hazard ratio for overall survival was 0.53 when the data were first reported, and it has gone up to 0.68 with longer follow-up. That’s not because axitinib-pembrolizumab doesn’t work as well over time in the next group of patients. The patients on the sunitinib arm with less aggressive disease are progressing, but they are able to receive a second-line nivolumab and get some benefit from that. Therefore, the dramatic difference that we’re seeing early on in patients alive between the 2 treatment arms tends to be attenuated over time.

The Role of VEGF TKI Monotherapy

There is also what Toni Choueiri recently presented at ESMO [European Society for Medical Oncology Congress], which was the CheckMate 9ER study. This initially had 3 arms, but it subsequently went down to 2 arms. It studied the utilization of cabozantinib and nivolumab against again the standard-of-care arm, Sutent [sunitinib]. Sutent [sunitinib] was being given at the typical 50-mg dose with 4 weeks on and 2 weeks off. The nivolumab was given every 2 weeks, and the cabozantinib was given at a starting dose of 40 mg.

The data were presented with the primary end point of progression-free survival [PFS] by blinded, independent central review. The secondary end points were overall survival, response rates, and safety. There was an 18-month median follow-up with these patients with progression-free survival as the primary end point. For patients receiving the combination, the PFS was 16.6 months compared with 8.3 months with Sutent [sunitinib] as a single agent, which is similar to what we’re seeing with how Sutent [sunitinib] has been ending.

The response rate was almost doubled: People receiving cabozantinib-nivolumab had a response rate of almost 56%, while patients taking Sutent [sunitinib] had a 27% response. There was an 8% CR [complete response] rate utilizing cabozantinib-nivolumab where we’re back to durable therapy.

The median duration of patients on the study for cabozantinib-nivolumab was more than 20 months, and for Sutent [sunitinib], it was 11 months. There were still some treatment-related adverse events similar to what one would expect. It was similar to using combination I/O [immuno-oncology]–TKI [tyrosine kinase inhibitor] therapy.

This is starting a new conversation about having a newer player where we have a little more data. It may be soon enough, but it’s still not FDA approved. It’s still going through discussions. Is this soon going to be a new practice-changing regimen that’s going to be awarded to us? This may be a conversation.

 I’m sure it will be FDA approved. Cabozantinib-nivolumab is certainly a good combination, so it’s not surprising that it’s better than sunitinib because cabozantinib is a better drug than sunitinib. The patients are getting not only cabozantinib but also nivolumab. The efficacy data appear to be at least as good as axitinib-pembrolizumab, especially when you’re considering that there are more patients with intermediate risk than there were patients with poor risk on the protocol, and the doses used were a lot less than the standard dose of cabozantinib, which is typically starting at 60 mg. Here, the patients started at 40 mg, and half the patients were down to 20 mg of cabozantinib because of toxicity. It suggests that the nivolumab was contributing in a major way to the benefit of the cabozantinib. But it was also a bit surprising how tolerable the regimen was, especially because when we were combining TKIs early on—such as sunitinib or pazopanib—with single-agent anti–PD-1 agents—such as nivolumab or pembrolizumab—those combinations were not felt to be tolerable. Therefore, we moved to things like axitinib and bevacizumab that were cleaner VEGF-blocking agents in order to avoid off-target toxicity.

Cabozantinib has a lot of extra targets that one might have thought might have contributed to toxicity without adding to efficacy, but at least at the doses used, it didn’t seem to be much of a problem, and patients were able to keep on the therapy. I’m not sure how big an impact it’s ultimately going to have on treatment use because a lot of people have gotten comfortable with axitinib-pembrolizumab if they’re going to use a TKI and an anti–PD-1 together. You can probably comment on this better than I can, but my sense is that people who are seeing maybe 5 or fewer patients with metastatic kidney cancer a year and have gotten used to giving axitinib-pembrolizumab are not going to be that quick to change to another regimen that isn’t clearly superior to it.

Impact of the Phase 3 CheckMate 9ER Trial on Practice

The sequencing issue is something we were talking about even when we first said sorafenib and sunitinib, which goes back to which agent to use first. It takes us back 14 years or so. When you deal with community oncologists, and I speak to them often, you still deal with physicians who are using ipilimumab-nivolumab for other conditions rather than just using it for renal cell carcinoma. Kidney cancer is not the most common disease that they’re dealing with; they’re commonly seeing lung cancers and melanomas as well, and they’re not using it with other cancers in combination. They may have a comfort level using combination immunotherapy, albeit with a bit of a different dosage of ipilimumab. But they’ve fallen into it, and it may be cleaner to give I/O [immuno-oncology]–I/O therapy, and then upon progression, you have a TKI [tyrosine kinase inhibitor]. This is compared with the questions we also ask: If you’re getting I/O–TKI as a first-line therapy and you progress, when do you throw the next I/O agent in? Do you throw it in as a single agent and use a TKI in the next sequence?

When you’re dealing with community oncologists, their vast knowledge about how to take care of a multitude of malignancies is significantly impressive. It sometimes gets cleaner when you lead with an I/O–I/O therapy and lead to a TKI. They may have given a little favor to axitinib-pembrolizumab, but when I speak to the community oncologist, I still think they’re using an I/O–I/O as first-line therapy.

 That’s interesting. There was a general assumption that the community may have been moving more toward TKI–I/O because of what I would consider a misconception that that therapy is better tolerated than giving ipilimumab-nivolumab because of their concern about the toxicity of ipilimumab. I always point out that ipilimumab-nivolumab is essentially for 12 weeks, and then it’s just nivolumab. Not everybody experiences the toxicity of ipilimumab-nivolumab: It’s closer to a third of patients. For most of those patients, it’s rarely treatable with immunosuppressive drugs for a short period of time. If patients end up in the hospital for colitis or elevated liver function tests, it’s usually for only a couple of days, and then they’re out on steroids and recovering. A few patients may need some extra immunosuppressive treatment like infliximab, but a majority of patients recover from those toxicities within 4 to 6 weeks. They are off immunosuppressive drugs, and many of them can go on and continue to receive nivolumab without having those toxicities recur.

This contrasts with axitinib-pembrolizumab, which probably doesn’t have the same risk of putting people in the hospital because of toxicity. However, the patient doesn’t have the opportunity to get a single-agent anti–PD-1 agent either in the future. If you’re exposed to an anti-VEGF, it may be tolerable in the beginning, but months and months of receiving a VEGF-targeted therapy is not good for your tissues. It is associated with a lot of diarrhea, a risk of cardiovascular problems, and weight loss. Patients don’t do well if they’re going to be on a TKI for a long period of time. My experience is that once you start a TKI, it’s very hard to stop it. Even if patients are not responding, you’re often going to a new TKI.

That’s right.That reminds me of the amphotericin days back in the 1980s and 1990s.

 Yes. That’s way back, but if you see a patient who is really sick, and their performance status is less than 70%, I still feel like that’s a patient for whom I want to start with immunotherapy because I’ve seen dramatic responses in some of those patients. It’s almost like the Lazarus effect. Those are patients for whom you can see major benefit with immunotherapy. For many of them, you don’t see any benefit, but you really don’t see benefit with TKIs in those patients.

That’s right. We’re coming into the time of year that is like the doughnut hole: Patients are also having to consider the cost of these oral therapies. Last week, the cost of these therapies was presented at the International Kidney Cancer Symposium. When you look at a year of I/O–I/O, it came out to about $333,000 to continue a year of initial 4 doses of ipilimumab-nivolumab followed by nivolumab, compared with almost $100,000 more having an I/O–TKI. We’re dealing with a patient population for whom, at this period of time through September and December, their out-of-pocket costs can sometimes be very financially toxic, despite the fact that they may have good tolerability. I agree with you. These patients taking pills every day offer toxicities that can be chronic. Twice today, I’ve dealt with patients who have this financial toxicity, and you hope that, through whatever foundation, you can get them money through the end of the year. They’re starting to balance it. There are sometimes medications that you can adjust the dose, but for the most part, these patients are living on the oral therapies, and it’s concerning.

The Role of Combination Treatment in RCC

 Looking at the future landscape, it’s going to get a little more complicated. There are some interesting studies going on. One is the phase 3 CLEAR study, which is looking at lenvatinib plus everolimus or lenvatinib plus pembrolizumab compared with sunitinib. We just saw a press release that the lenvatinib-plus-pembrolizumab arm is better than sunitinib. This is no surprise because lenvatinib, at least in my experience, is also a much better drug than sunitinib as a TKI [tyrosine kinase inhibitor]. That’s not a surprise, but we also saw in the press release that lenvatinib-everolimus was superior to sunitinib. I’m going to be curious to see that data to see how much better lenvatinib-pembrolizumab is than lenvatinib-everolimus and how much of the benefit that we’ve been seeing in these TKI–I/O [immuno-oncology] combinations vs sunitinib are related to the fact that axitinib, cabozantinib, and lenvatinib are all better VEGF inhibitors than sunitinib.

Another trial that is going on that will yield interesting results is the COSMIC-313 trial, which is comparing cabozantinib plus nivolumab, with nivolumab and ipilimumab, the triplet, vs nivolumab/-ipilimumab. It will be the first head-to-head comparison where we get a chance to look at the impact of a TKI on the response to an I/O agent. It will be interesting to see whether the triplet is synergistic, additive, or subadditive. There have been a lot of preclinical data that suggest that VEGF TKIs will enhance the immune response, but I always caution people that none of that data is in VHL-null immune-competent tumor models because those models don’t exist. It’s conceivable that, in kidney cancer, because of the 

 mutation, the TKI may be causing acute hypoxia, which may compromise the ability of the immune response to eliminate the last tumor cell. I’m paying particular attention to the COSMIC-313 study to see what the tail of that PFS [progression-free survival] curve looks like in the triplet arm to see whether it looks like the nivolumab-ipilimumab tail.

Finally, there is the PDIGREE study, which is going on in the Alliance group. Patients are getting 12 weeks of nivolumab-ipilimumab, and after they’ve completed 12 weeks of nivolumab-ipilimumab, if they have either a non–CR [complete response] or nonprogressive disease, which maybe will be about half the patient population, they’re being randomized to get either cabozantinib-nivolumab or nivolumab. That’s an interesting question, but it has 2 caveats in my view. No. 1, it’s a large number of patients being accrued to get to the randomized group, which is at the second-line setting. And it’s unclear how long that trial is going to take, how many people are going to be eligible for that randomization, and how many haven’t had such toxicity that they have to stop treatment or have progressed or refused the randomization at that point, and whether it’s going to be relevant if most of the patients are getting a TKI and an anti–PD-1 agent in the frontline setting.

Those 3 trials are coming, and they’re going to address some questions. Until we see those results, and even with those trials, it’s going to be a discussion about I/O–TKI versus I/O–I/O in the frontline setting.

That’s right. We then have the sequence of therapies. How do we use cabozantinib, which got it’s FDA approval in the METEOR clinical trial? CheckMate 025 brought nivolumab into our world. We mentioned the lenvatinib-everolimus combination before in the Study 111 trial, and it was the first time we saw a combination therapy that was well tolerated with significant numbers. In the second-line setting, initial presentation was a 43% response rate, but it went down to 37% as a second-line therapy with a progression-free survival of 14.6 months compared with everolimus by itself, which was the standard arm, which is 5.5 months.

We started realizing that not only do we have some excitement as a combination first-line therapy, but we may now be using them in second-line therapy. Can this lenvatinib-everolimus combination move to a first-line therapy to avoid I/O in the first line? It’s going to be very hard. We still have to believe that our patients have to have tolerable therapies, and it’s not just about giving them continuous oral therapies that will probably lead to great response rates and progression-free survivals. We’re avoiding it: The topic we’re thinking about is the CR and durable time off-therapy. When you tell a person who has metastatic cancer that you can have them on surveillance after they receive therapy, you are sometimes buying this patient years off-therapy. You have experience, and we have experience as well: For previous patients with high-dose IL-2, for as many courses as we got them through—which may have not been so many—you stop therapy. They’re on surveillance for decades. There’s nothing more fulfilling than that.

 When we’re talking about treatment sequencing, I agree: it’s always best to get the front first-line treatment that cures the patient so they never need second-line therapy, but whether I’m giving nivolumab-ipilimumab or a patient has gotten axitinib-pembrolizumab, I’m still leaning toward cabozantinib as my second-line therapy. I might occasionally give axitinib in that setting if the patient has a slow-growing disease but still wants therapy if they’re truly progressing. I’m then saving lenvatinib-everolimus for my third-line treatment because it’s 2 drugs, so it’s harder for the patients to take than cabozantinib. I find it to be very effective after cabozantinib, so I’m not surprised that lenvatinib is an active agent in the doublet with pembrolizumab in the CLEAR study, and I’m not surprised that lenvatinib-everolimus is better than sunitinib in the frontline setting. Where it may play a role in the frontline setting is in patients with non–clear cell kidney cancer, particularly chromophobe tumors or some of the aggressive papillary tumors that may not respond well to immunotherapy. That’s where the lenvatinib-everolimus combination looks like it might turn out to be the best option.

The Future Treatment Landscape of RCC

 Something I’m excited about is the HIF2-alpha inhibitor, particularly the drug that came out of Peloton Therapeutics, now being developed by Merck &Co: the MK-6482 agent. That’s going to have a potential role for some patients as a second-line treatment, particularly patients who clearly have HIF2-alpha–driven cancer because it blocks not only VEGF but also other things that are downstream of HIF2-alpha that may drive tumor progression. It’s an oral agent, it’s reasonably well tolerated, patients may be able to stay on it for a long time, and it may be a new agent that we’ll have to figure out how to use. It may have the potential to be combined with TKIs [tyrosine kinase inhibitors]. There’s a rationale for that, and there is also potential for combination with checkpoint inhibitors where it’s less likely to cause hypoxia in the tumor microenvironment. Therefore, it may be a better drug to combine with an I/O [immuno-oncology] agent than current TKIs.

That was presented at ASCO GU [American Society of Clinical Oncology Genitourinary Cancers Symposium], and there was a buzz that it was worth of having. We’re both involved in second-line therapy using MK-6482, so it’s going to be exciting to finally get to use it. It’s ultimately going to be a real player when it comes down to treating our patients not only in the second-line setting but most probably in the combination therapy setting moving forward as well.

I like what Chung-Han “Joe” Lee presented in regard to using lenvatinib-pembrolizumab as a second-line therapy in patients after failing first-line I/O therapy. It’s using it like how we see things clinically. We are definitely data driven, but after a while, we have to adapt. If we treat someone with ipilimumab-nivolumab as a first-line therapy and with cabozantinib as a second-line therapy, where do these drugs come in as third-line therapy? There are no significant data. There are small data talking about using immunotherapy as a third-line therapy, but we are secretly practicing using it. When you now have data to support it, and it’s coming in real time, we’re offering our patients 4 or 5 lines of active therapies for which we have high hopes. It’s not just palliating our patients. When you look at the IMDC [International Metastatic Renal Cell Carcinoma Database Consortium], they looked at the analysis that less than 50% of patients are receiving second-line therapy, and we’re talking about how we incorporate 5 lines of single and combination therapies. Going back to the days before 1992, it is becoming quite remarkable where the science is taking us.

 In our hands, very few patients who progress on first-line therapy don’t get a second-line therapy. That may have been the case when you had sunitinib as your first-line treatment, and there were people who had aggressive disease who didn’t respond to sunitinib who weren’t in shape to get a therapy afterward. Now that we’re giving immunotherapy as a frontline therapy, for many patients, their most aggressive disease is being eliminated. And if they do progress, it may be with slower disease. In our hands, one can always get TKIs into those patients and give them a chance if they need it.

I agree that the lenvatinib-pembrolizumab data were interesting, but I have a caveat about that. The fact that lenvatinib-pembrolizumab produced a high response rate after nivolumab-ipilimumab is not surprising because it’s the first time those patients are seeing a VEGF receptor TKI, and the patients who aren’t responding to nivolumab-ipilimumab may be enriched for patients who might respond to a VEGF receptor TKI, so a 55% response rate in that group is not surprising. You might say, “There was another group that got a VEGF receptor inhibitor and an I/O who also had a high response rate.” But those were not patients who got axitinib-pembrolizumab. Those are patients who got sunitinib often for a short period of time or at a lower dose because of toxicity. They may not truly have progressed and went off it early so they could get nivolumab, because that’s what they really wanted, and then they progressed on nivolumab and went to lenvatinib-pembrolizumab. They may not have been truly resistant to the VEGF pathway. It’s not surprising that you could see a response to a better VEGF receptor TKI in a group of patients whose disease was still responsive to VEGF inhibitors. I’m not sure what the contribution of pembrolizumab is to that combination after people have already progressed on the TKI, and that’s a question that is well worth addressing scientifically.

That’s right. As of now, we’re still using lenvatinib as your second TKI; it should not be used as a first TKI, though we do get excited about not just its science and mechanism but also about its efficacy that we’re seeing.

The last question I have to ask you is a question everyone has: Do we have a predictive biomarker that we are using yet with which we can come to the conclusion that we can predict how to treat our patients? Are we still looking for that?

 We’re getting close. As was published in 

 last week, there’s a thorough review of the biology of the tumors from the patients who went on the IMmotion151 study, comparing atezolizumab and bevacizumab to sunitinib. You could see that they divided tumors based on transcriptomics into 7 categories: 2 of which are angiogenesis driven, 3 of which were more proliferative and immunotherapy responsive, and 2 in which patients probably didn’t respond to either of those. That type of data are where we want to get to in order to help us choose treatment options.

There was then another component where they looked at next-generation sequencing, and they saw that those groups tended to divide where there’s a group that’s 

 mutated, which is more the angiogenesis sensitive, and the 

 mutated, which is more the immunotherapy sensitive. If we can boil those transcription factors and next-generation sequencing mutations down to a few factors that we can analyze on a histologic specimen, we may be able to be at a place where we can say, “Here’s a group of patients who should get TKI–I/O, and here’s a group of patients who should get I/O–I/O.” We’re probably no more than a couple of years away from being able to do that and to have the data in this manuscript validated in an independent study prospectively so we can use them as biomarkers to help us choose the right treatment.

As community oncologists in academics, we’re definitely using next-generation sequencing more. We’re putting patients on clinical trials more, which is my approach to community oncologists: If you have an academic institution that has an excellent clinical trial for your patients, it’s worth it to discuss and consider that. The pulse of what we’re trying to think about is not just immunotherapy. Do we have the potential of taking care of the patient and the cancer rather than just taking care of the cancer? Can we offer patients durable remissions and an improved functional quality of life with curability? Ultimately, the way we are looking at our patients is not about what I can do for them in the next 6 months; it’s what I can do for them for a lifetime.

 I’ve been in this business for a long time, and when I started, the median overall survival for patients with metastatic kidney cancer was 10 months. With interferon, it may have been prolonged to 13 months. With high-dose IL-2 in the patients who went on those studies, it was prolonged to 17 months. With the single-agent first-line TKIs, it may have been 25 to 28 months. We’re now at a point where we can expect a median overall survival to be as long as 5 years, and that’s all because of basic and clinical research studies where we’ve seen that tremendous improvement in overall survival over a 30-year period. That wouldn’t have happened if it wasn’t for basic science and translating that basic science into clinical trials, which led to changes in clinical practice. There are still new drugs out there that could potentially push that survival bar even higher: things like the CAR [chimeric antigen receptor] T cells that are targeting CD70 or CA9 or PSMA. Other targets include glutaminase inhibitors and CXCR4 and CDK4/6 inhibitors, which might be potential opportunities for groups of patients who don’t respond well to the standard TKI checkpoint inhibitor regimens. There is still a need and an opportunity for clinical research to raise the bar higher, and we shouldn’t stop until we’re offering curative therapies for as many patients as possible or until we have a therapy that we think is the best therapy to offer in the first-line setting for each patient we see.

We’re still very excited about going to work every day.

Emerging Strategies for Renal Cell Carcinoma | Insights