We have 2 recent trials that have changed the treatment paradigm in the first-line setting, which is a big change for the field. Both of these studies have successfully challenged sunitinib (Sutent) as the standard of care in the treatment of metastatic kidney cancer. And Sutent has been one of the gold standards over 10 years now since its original approval back in 2006. Now, we&rsquo;re going to be discussing both of these studies today. First, a TKI study, CABOSUN, which is a randomized phase II study that tested a novel tyrosine kinase inhibitor, cabozantinib, in comparison with sunitinib. And second, an immunotherapy/TKI comparison in the phase III CheckMate-214 study that led to the approval of ipilimumab plus nivolumab in this disease.

With cabozantinib, it&rsquo;s important to know it had previously been FDA approved in TKI-pretreated patients, which predates the analysis of this study. This trial, the CABOSUN trial, is a smaller randomized phase II study that was conducted in the Cooperative Group mechanism. It was an Alliance study that was published by Toni Choueiri, et al, in the

last year. And this study was a randomized phase II trial that enrolled about 150 patients and randomized them 1:1 to receive either sunitinib or cabozantinib for the treatment of metastatic disease. This trial limited enrollment to patients with IMDC-intermediate or poor-risk disease. It was unique in that way, and it pursued a primary endpoint of progression-free survival.

So, it was a positive study that demonstrated superior PFS in patients treated with a novel tyrosine kinase inhibitor, cabozantinib, which is distinctly different from sunitinib in that it not only blocks VEGF receptor signaling but also AXL and MET, both of which are thought to mediate angiogenesis alternative to VEGF-mediated neovasculature. Both AXL and MET signaling have been shown to be upregulated in patient samples treated with sunitinib with acquired resistance to that drug. So, in this trial, it was now shown that, indeed, cabozantinib was superior. The FDA approved the drug for the first-line indication regardless of IMDC risk status. Although in the trial, it&rsquo;s important to know it was conducted exclusively in patients with intermediate- or poor-risk IMDC status.

The second study we&rsquo;re going to be discussing is the CheckMate-214 study that led to the FDA approval of ipilimumab/nivolumab combination in treatment-na&iuml;ve patients with intermediate- or poor-risk disease. This was a very large phase III global trial that randomized patients 1:1 to receive either standard sunitinib or the combination of ipilimumab/nivolumab for the treatment of metastatic disease. This brought on about 1100 patients, and patients were stratified by IMDC risk and by geographic region.

This was an interesting trial in its statistical design. It had 3 coprimary endpoints and it was powered to demonstrate a difference in all 3, the first being overall survival, the second being objective response rate, and the third being progression-free survival. And in order to accommodate all 3, the alpha was split between the 3. So, there was a higher threshold or rather a lower

value necessary than what we&rsquo;re used to seeing in order to demonstrate significance for each of these 3. Patients were treated on a standard sunitinib schedule with 4-week, 2-week, 50 mg daily. And ipilimumab/nivolumab treatment schedule, it&rsquo;s important to note it&rsquo;s actually different from what is approved for melanoma and that&rsquo;s based on the phase I work that had been done in kidney cancer specifically.

This study was successful in demonstrating superiority in 2 of the 3 primary endpoints. So, overall survival was superior for the immunotherapy combination with a hazard ratio of 0.63. And landmark analyses had also suggested that more patients were alive at 12 and 18 months in the immunotherapy arm than the TKI arm. There were also superior objective responses for the immunotherapy combination. That, importantly, included a 9% complete remission rate in patients with intermediate- or poor-risk disease. It&rsquo;s important to note that these primary analyses were conducted in the intermediate- and poor-risk population only. So, while the trial brought on patients with favorable intermediate- and poor-risk disease, the primary endpoint analyses were conceptualized so they were limited to patients with intermediate- and poor-risk disease. And that was done based on a prior experience with a combination that suggested higher efficacy in this population. The third primary endpoint, progression-free survival, was numerically in favor of the immunotherapy combination but did not meet the prespecified threshold for significance.

mRCC: Impact of CABOSUN and CheckMate-214

Following up on these new approvals, the question that obviously arises is, how do we make decisions in newly diagnosed patients who are being considered for frontline therapy in the metastatic setting? And there are certain criteria that can help with decision making. We have now touched upon risk stratification criteria several times in discussing the trials, and it&rsquo;s interesting to note that the 2 new approvals in the first-line space, cabozantinib and ipilimumab/nivolumab, were granted based on trials that had a primary endpoint specifically looking at intermediate- and poor-risk patients. And, indeed, using the risk stratification of the International Metastatic Database Consortium, or the IMDC, is very helpful in patients.

The first question that I usually ask myself when I see a patient with newly diagnosed metastatic disease is, do they in fact need treatment at that point in time? And the IMDC risk stratification tool is helpful as we sometimes consider patients for active surveillance in the first-line setting. We can delay the onset of targeted therapy for some of our patients, those with indolent disease and a smaller disease burden. And the IMDC group in the past has presented data and published on this, and it&rsquo;s helpful to know that patients with a low IMDC score and limited organ involvement can be considered safely. Some remain on active surveillance for extended periods of time.

Once one decides whether or not to start therapy, we can consider different standards now that are available to us. We have single-agent TKI therapy, and there are 3 TKIs that are commonly considered, now based on the approvals we havecabozantinib, which we&rsquo;ve heard about, sunitinib of course, and then lastly pazopanib. And then we now have the immunotherapy combination. So, CheckMate-214 now has introduced I-O therapy into the first-line space, which constitutes a huge change to the paradigms that we go by when we treat patients in the clinics.

To me, what is most relevant in considering all the trial data that are out there right now in the first-line space is the high complete remission rate that has been seen on the CheckMate-214 study. So, what distinctly sets this apart from the various TKI trials is that patients treated with ipilimumab/nivolumab, specifically those with intermediate- or poor-risk disease, can achieve complete remissions. And the CR rate in that patient population on CheckMate-214 was 9%, which is something that we&rsquo;ve never seen in TKI trials.

The other thing that we haven't discussed yet that is very relevant is that on CheckMate-214, about 20% of the patient population had favorable-risk disease. These were not included in the primary endpoint analysis per how the trial had been set up, but, of course, in the secondary analysis, these patients were looked at separately. And much to everyone&rsquo;s surprise, these patients actually did better with a kinase inhibitor. So, within CheckMate-214, about 200 patients on each arm had favorable-risk disease. And when we only look at those patients then, the objective response rate was significantly higher, north of 50%, in the sunitinib-treated patients, lower with ipilimumab/nivolumab, in the 20% range, and the same for progression-free trial. So, there were significantly better progression-free survivals for the TKI-treated patients with favorable-risk disease.

And then lastly, even if you look at overall survival landmark analysis, 12 and 18 months overall survival rates numerically favored the TKI arm, so sunitinib did better. That now is very relevant as we look at first-line therapy patients before we decide on which agent to put them on. Indeed, the first instance where now risk stratification with IMDC goes into decision making. So, a new patient should be assessed by IMDC risk stratification, patients with intermediate- and poor-risk disease, in my mind, should certainly be considered for combination immunotherapy based on the strong overall survival endpoint and the relatively high CR rates that have been seen. But it&rsquo;s important to take a thorough history in terms of the patients&rsquo; prior medical problems. Patients with autoimmune disease, we know, are at higher risk for autoimmune toxicities with immune-activating agents like this and were excluded from participation in trials like CheckMate-214. So, that&rsquo;s the second thing that is very relevant. Comorbidities, prior medical history should be assessed for all patients.

And then thirdly, of course, the overall burden, pace of disease, and maybe also organ distribution. Cabozantinib, which now is FDA approved in the first-line setting, has been studied extensively by its manufacturers, not only in kidney cancer but also in prior efforts in prostate cancer for its effect on bone metastases. And there is a lot of preclinical work but also clinical work that has gone into correlates of bone turnover, including osteoclast biology and so forth, that suggest that this agent, more so than others, has an effect on kidney cancer metastases in the bone. So, certainly for patients with a high burden of disease, osseous metastatic disease, this would be a very attractive agent.

Factors for Choosing Frontline mRCC Treatment

IMDC risk status, as we have now said, becomes a very relevant piece of information, especially in the first-line setting for our patients now. And it&rsquo;s interesting to note that for the 2 most recent approvals in that space, cabozantinib but also ipilimumab/nivolumab studies were analyzed specifically in that setting. Now, it is relevant to understand that the FDA approvals don&rsquo;t integrate IMDC risk status for both of these. So, rightfully for ipilimumab/nivolumab, the FDA has limited its approval to patients with IMDC intermediate- or poor-risk disease, and they do so because the benefit over sunitinib was not seen in favorable-risk patients, indeed the opposite was the case as we have discussed.

Cabozantinib is FDA approved now for untreated and pretreated metastatic kidney cancer, and there is no limitation set or no specification given for risk status. Meaning, the CABOSUN trial in the first-line setting was conducted only in intermediate- and poor-risk disease. However, the FDA approval is not limited to that group but specifies that patients only have to have metastatic RCC. Similarly, both of these regimens were tested in the trials that we have discussed in patients with clear-cell histology. They have not looked at patients with less common variants that we often summarize as nonclear cell renal cell cancer. The FDA approvals for both of these drugs, just like for all other agents we have, are not limited to clear-cell kidney cancer, meaning insurance approval can be requested for patients with non&ndash;clear cell histology. So, also with the limitation that we do not have dedicated data from prospective clinical trials yet, all of that is ongoing presently.

We, of course, have all the data to guide choice of tyrosine kinase inhibitors in the first-line setting. For many years, we have discussed sunitinib and pazopanib as the 2 standards that have been used most widely really across the globe. They, too, of course have been compared in a large randomized trial. The COMPARZ study published in the

in 2012, at its time, was the largest trial ever conducted in kidney cancer. About 1200 patients were randomized in that trial, which was a noninferiority study, to prove that pazopanib was not inferior to sunitinib in terms of progression-free survival in treatment-na&iuml;ve patients. It was a positive study in that there was no inferiority for pazopanib in that setting.

There were a lot of secondary analyses built in to that trial, many of which aim toward the tolerability of drugs, toxicity profiles for the 2 agents in comparison, and the quality of life scores. And on this trial, most of these favor pazopanib, as was concluded to be the more tolerable agent between the 2. Proponents of sunitinib since then have argued that the treatment schedule for sunitinib can be adjusted easily in order to make this a more tolerable, more manageable agent. So, sunitinib, when it was originally approved based on its pivotal trial back in 2006, has always been given on a 50 mg 4-week on, 2-week off schedule. And ever since it hit the market, there are many efforts, initially retrospectively and then also some prospective efforts, that have looked into alternative schedules to make this agent more tolerable. The most commonly used one is 150 mg on a 2-week on, 1-week off schedule, which, in the span of 6 weeks, delivers the same amount of medication but gives more frequent and shorter breaks for toxicity not to build up for so long and for the treatment intermission not to be as long.

There&rsquo;s a small randomized phase II study from Korea that has actually looked at this with a primary efficacy endpoint of progression-free survival in treatment-na&iuml;ve patients. And that study met its primary endpoint, suggesting that the 2-week on, 1-week off schedule was in fact more efficacious than 4-week on, 2-week off. And such as all the prior retrospective efforts, it had also demonstrated better tolerability for the novel schedule. So, certainly, many investigators now have more and more been using the 2-week on, 1-week off schedule for sunitinib-treated patients.

The Canadian investigators have been very invested in sunitinib, and there have been data to suggest individualizing the treatment schedulenot only in terms of adjustment for toxicity but also dose escalation in patient who do not show any signs of toxicity&mdash;can be helpful, both in terms of tolerance and efficacy. So, sunitinib is not often used entirely, and if it is being used, there is certainly more and more evidence to suggest that individualizing schedules are helpful.

mRCC: Risk Status and Other Factors for Therapy

PD-L1 testing has been investigated intensely in kidney cancer and has been in other diseases, and this actually goes beyond the application of checkpoint inhibitor therapy. There are data from the COMPARZ trial, which was a tyrosine kinase inhibitor study that compared sunitinib with pazopanib using PD-L1 immunohistochemistry and demonstrated that patients who are treated with tyrosine kinase inhibitors and have higher tumor expression of PD-L1 levels have adverse outcomesmeaning patients do worse with high PD-L1 levels&mdash;suggesting that this aspect of tumor biology may in fact add to the aggressiveness of cancer. Now several years later, we have data from the I-O trials that have helped us to understand that it not only is an adverse prognostic marker in terms of the biology of disease, but it is a predictive marker that shows favorable outcome in the immunotherapy-treated patients, which makes sense because it&rsquo;s the target.

The early studies for PD-1 and PD-L1 inhibitors in this disease suggested that response rates were higher in patients who were PD-L1expressing on tumor cells and on immune cells. It&rsquo;s important to know, and has always been stressed, that lack of PD-L1 expression does not preclude patients from having benefit from these agents, so it should not lead to these patients not being considered for these trials.

The randomized phase III trial that led to the approval of nivolumab in the second-line setting, the CheckMate-025 study, compared nivolumab with everolimus in TKI-pretreated patients. And, importantly, PD-L1 expression status did not seem to affect the superiority over everolimus in these patients, meaning whether or not a patient was PD-L1positive, they did better with immunotherapy than with the mTOR inhibitor. So, at the time, PD-L1 testing was deemed not a relevant marker to make treatment decisions. Since then, we now have the CheckMate-214 studywhich, again, is the ipilimumab/nivolumab versus sunitinib trial that compared tyrosine kinase inhibitor therapy with combination I-O therapy. And, of course, PD-L1 expression status was investigated on that trial&mdash;again, we tested for expression on tumor cells, not on immune cells. And the best way to summarize the data is that we again saw an adverse signal for the TKI-treated patients, meaning the cancer-specific survival with sunitinib therapy was worse in PD-L1&ndash;expressing tumors. And on the flip side, patients who were treated with an immunotherapy combination were more likely to benefit from immunotherapy if they were high expressers.

So, for example, the CR rate in intermediate- and poor-risk patients was 9% across the board, but if we looked at patients who were PD-L1positive and had intermediate- or poor-risk disease, the CR rate went up to 16%. If we looked at overall survival benefit of ipilimumab/nivolumab over sunitinib in PD-L1&ndash;negative versus PD-L1&ndash;positive patients, then we noteand Bob Motzer presented this at SITC in 2017&mdash;that superiority was seen in both instances, but it was much more marked in patients who were PD-L1&ndash;positive. So, the hazard ratio went down to below 0.5 in patients who were PD-L1&ndash;high expressors, comparing the I-O combination over sunitinib.

So, in my mind, it is not a marker that will help us make decisions between 2 regimens, but it&rsquo;s certainly prognostic in terms of the tumor biology if we do not consider therapy. I think it is predictive for combination immunotherapy with ipilimumab/nivolumab. And knowing that a patient is PD-L1positive up front should be a strong motivator to pursue ipilimumab/nivolumab combination therapy. That being said, as stressed before, a PD-L1&ndash;negative patient can still benefit and is still, according to the CheckMate-214 data, more likely to benefit from that combination than from sunitinib alone.

Role of PD-L1 Status in mRCC

If we look beyond the first-line setting, we have a number of agents to consider in patients who have progressed on first-line therapy, be it kinase inhibitor therapy or I-O therapy. And it has become somewhat difficult for physicians to decide between the many options that we have on the shelf. The NCCN committee gives category 1 support to 4 regimens in the second-line setting presently, and that is based on randomized data that have previously been published. The 4 choices that they endorse are single-agent cabozantinib in patients who have not previously received cabozantinib; single-agent axitinib, based on the AXIS trial that was published many years ago; the combination of lenvatinib/everolimus, based on a phase II/III arm randomized trial, the Eisai 205 study, that put that combination on the market; and then lastly, nivolumab monotherapy based on CheckMate-025.

What&rsquo;s confusing here is that these endorsementsand more relevantly, these trials&mdash;were conducted prior to the new approvals in the first-line setting. So, we have discussed now that we have cabozantinib and ipilimumab/nivolumab as considerations in the first-line therapy. We have no data and no expert recommendations to guide for any of these second-line regimens in patients who have received one of these two in the first-line setting. Now, nonetheless, we can use the second-line data for decision making. We can think about how these kinase inhibitors are different. So, the different TKIs that are out there all inhibit VEGFR-2 as their principal target. Additionally, they have other targets, at least some of them do. Axitinib, which has been on the market the longest, is a very pure VEGF receptor-2 kinase inhibitor. It&rsquo;s kind of like a sniper gun that goes very potently and very selectively after VEGFR-2.

Cabozantinib, which we&rsquo;ve talked about quite a bit now, in addition to VEGFR-2, inhibits 2 relevant kinases, AXL and MET, both of which have been shown to mediate angiogenesis and VEGF-independent ways. So, the rationale here is that cancer cells, once they become independent of VEGF signaling, need to find other ways to build neovasculature. And it has been shown in preclinical models but also based on patient samples that kinase inhibitor pretreated tumors upregulate AXL and MET as mechanisms of resistance. So, cabozantinib has exploited these as targets, and based on the randomized METEOR study, it was much more efficacious than the mTOR inhibitor, everolimus, in TKI-pretreated patients.

Lenvatinib, which is FDA approved in combination with everolimus, in addition to VEGF receptor-2, blocks fibroblast growth factor receptor, so FGFR-1 through FGFR-4. Again, FGFR signaling is thought to mediate angiogenesis, independent of VEGF. As I mentioned, it has been approved in the combination with the mTOR inhibitor, everolimus, and it&rsquo;s the only targeted combination that is currently on the market. When we think about these 4 options that are supported by the NCCN and that all have randomized data that led to their approval, it&rsquo;s important to note that none of these 4 have been compared with one another. So, they have all been compared to older standards. Axitinib was superior to sorafenib, and the 3 other regimens that we&rsquo;ve talked about were all superior to everolimus. But none of the 4 have been compared with one another.

mRCC: Decision Making for VEGF TKIs

It&rsquo;s difficult to decide between the many options that are there, and I&rsquo;ve mentioned that out of the efficacious regimens that are there and that are being commonly used, there isn&rsquo;t one clear winner that the data suggest should be uniformly used in the second-line setting. But the 1 message that has come across quite clearly is that single-agent everolimus therapyeverolimus being an oral mTOR inhibitor that has been on the market for TKI-pretreated kidney cancer for many years&mdash;should no longer be considered a standard of care in the second-line setting and likely also not in the third-line setting. Everolimus was inferior on randomized trials when compared with cabozantinib. It was inferior to single-agent nivolumab. It was inferior to the combination of lenvatinib/everolimus, and indeed also to lenvatinib monotherapy.

So, presently, it does not have a defined role in TKI-pretreated patients. It has been tested also in the first-line setting. There was the phase II RECORD-3 trial, that was published in the

a couple of years ago, that tested its utility in the first-line setting and that was a noninferiority trial that tried to prove its noninferiority to sunitinib in untreated patients regardless of risk status. That was a negative study. It was indeed found to be inferior.

Everolimus has been tested intensely in nonclear cell kidney cancer in that it&rsquo;s back to 2 randomized studies previously published. The ASPEN and the ESPN trials both tested sunitinib versus everolimus in untreated metastatic non&ndash;clear cell kidney cancer variants. Both of these studies statistically were conceptualized to prove superiority to everolimus, and neither of them were able to do so. So, everolimus was not deemed superior to sunitinib in the primary endpoint analyses for both of these randomized studies. Again, while it has shown some efficacy across various non&ndash;clear cell phenotypesspecifically chromophobe kidney cancers and also, to some extent, papillary kidney cancers&mdash;there was not the sense that it was clearly superior to kinase inhibitor therapy. And, again, it would not be considered a standard in that setting.

Role for Everolimus in mRCC Treatment?

How do we then decide which second-line regimen to use? I think there are several factors that can be helpful, and it&rsquo;s not that there&rsquo;s the 1 decision factor that should drive a decision. Certainly, the experience that patients have had with first-line therapy would go into that decision making. So, for example, patients who have been on kinase inhibitor therapy for an extended period of time with significant toxicity may be better suited to then sequence over to I-O therapy with single-agent nivolumab, which tends to be tolerated quite well. Vice versa, patients who have done very well with TKI therapy with limited toxicity may benefit from continuing that strategy with another VEGF-directed kinase inhibitor, as it has been shown that these can be applied serially.

Again, we can return to cabozantinib as a drug that may have unique activity in patients with bone metastases. And I would not base that so much on the subgroup analyses that have been done off of METEOR because there are subgroup analyses off of other trials like CheckMate-025 that also suggest superiority over everolimus for immunotherapy. But I do think that all the data that are there from preclinical modeling and from serum samples that have been collected as part of cabozantinib development suggest an effect on osteoclast biology, and the remodeling in the bone should be considered. So, I do think that patients with a big burden of disease, including osseous metastatic disease, might be good candidates for cabozantinib therapy. There are no biomarkers that really drive decision making in that space. It has not been proven that immunohistochemistry for PD-L1 must drive decision making towards single-agent nivolumab over these other agents that we have discussed.

c-MET signaling, MET being a target for cabozantinib, has been investigated in terms of the outcomes seen on the various cabozantinib trials. But, again, that has not been compared to other kinase inhibitor trials or immunotherapy. So, there is no biomarker that helps with decision making. I do think that comorbidities are quite relevant. Patients with poorly controlled hypertension are not great candidates for kinase inhibitor therapy. Patients with poorly controlled diabetes may not be the best candidates for lenvatinib/everolimus combination therapy knowing that Afinitor (everolimus) can cause worsening of the glucose intolerance. Again, as mentioned before, patients with an autoimmune disease, specifically those with systemic autoimmune disease that have acquired immunosuppression in the past, should not be considered up front as candidates for I-O therapy if other alternatives can be considered.

mRCC: Factors for Choosing Second-Line Therapy

There remains several unmet needs in the management of metastatic kidney cancer, and we have revisited rare kidney cancer variants several times now. And there&rsquo;s little that we know with how effective indeed these newer drugs are in the nonclear cell setting. I do think it&rsquo;s important to emphasize clinical trial enrollment for patients with rare histologies, and I do think it&rsquo;s important for our pharmaceutical partners to show interest in conducting trials in that space rather than us only pursuing retrospective research in this setting.

I do think that all the optimism we have developed seeing what immunotherapy can do in untreated and pretreated patients points out a new unmet need, and that is to actually increase the number of complete remissions and the hope to achieve long-term disease control, possibly cure, for our patients. And I do think that that emphasizes the need for biomarker development, not only for the agents that we have on the market now, but also for drugs to come. It&rsquo;s exciting to see that there are so many new therapies underway right now, and the emphasis certainly remains on combination therapy.

We have not talked yet about I-O/TKI combinations that now pair the old and the new and show that VEGF-directed kinase inhibitors can be paired safely with immuno-oncology drugs. That is certainly where the field is going in terms of the large registration trials that are ongoing right now. There are several combinations that are worth watching out for, and some of these, per their earlier development efforts, have now been presented in print. The combination of axitinib plus pembrolizumab seems to be highly active, and the phase Ib study was published in the

not too long ago and demonstrated very high objective response rate, north of 70%, which is higher than what we have seen in untreated patients with ipilimumab/nivolumab. This is a small trial, it was only about 50 patients, but there was a notable CR rate included with that, and certainly it makes us interested to see if it might be even more effective. It is interesting to note also that on that trial, which granted brought on the more favorable risk type population, the benefit of the combination seems to extend to not only intermediate- and poor-risk but also favorable-risk patients who also had a response rate of about 75%. And that now has, of course, led to a large randomized study that has completed accrual, which compares the combination of axitinib/pembrolizumab with sunitinib in untreated patients, regardless of IMDC risk status.

There are various other TKI/I-O combinations that are showing similarly favorable early signals and are worth watching, another one being axitinib/avelumab, which has been very promising in the phase I setting and that&rsquo;s also led to a randomized phase III study. And then thirdly, the combination of lenvatinib plus pembrolizumab. Again, this is the combination that in treatment-na&iuml;ve patients has reported the highest objective response rate to date of over 80% in a very small sample size. Again, it is now being tested in a large randomized trial in comparison to sunitinib and on a third arm to lenvatinib plus everolimus. So, certainly, a lot to look forward to as we see these combinations mature and hopefully push the boundaries for our patients even further.

Upfront Therapy for mRCC