Risk of Postsurgical Disease Recurrence in RCC
Transcript:Robert A. Figlin, MD: Sandy, there are some subtle differences in these trials. Let me see if I can summarize a couple of those. Imaging evaluations were not standard across all the trials. Some of them were investigator evaluated, and some of them were independently reviewed. Doses and schedules were a bit different across the drugs, which may or may not have an effect.
Presented at the 2017 ASCO Annual Meeting was an interesting observation in the PROTECT trial, where patients that got full doses had a significant benefit with respect to disease-free survival, as compared with the trial, as Michael points out, where, oftentimes, doses needed to be reduced because of tolerance to therapy and also the populations. When we generated the University of California, Los Angeles Integrated Staging System [UISS] in 2002 and published it, we still didn’t have a good way of knowing how to pick out the poorest prognosis patients with high-risk resected disease.
At the 2017 ASCO Annual Meeting, there was a presentation by Dr. Bernard Escudier that talked about a gene signature that has been validated based upon some work from Dr. Brian Rini. How are you putting this together? On the heels of not only evaluation of targeted therapy in this space but also the knowledge that, right around the corner, we’re going to be asking hundreds, if not thousands, of patients to consider the role of immuno-oncology therapies in this space, what are your thoughts?
Sandy Srinivas, MD: We really have to go back to, what do we include for risk assessment? You spoke about the UCLA system [UISS], but there is the Kattan nomogram. There is the Mayo Clinic’s assessment. A lot of things are taken into consideration—size of tumor, vascular invasion, presence or absence of necrosis, the TNM stage. If you actually take the same patient and try to apply that to these 3 groups, you get a different risk assessment. So, I think it really comes down to some level of harmonization as to what we even consider as high risk. That’s the first thing.
ASSURE had a different assessment. PROTECT has a different one, and I think SORCE is different. Each of the trials that we have today has a different inclusion of whom the patient is. So, I think the first step as we start thinking about how to interpret these trials and, maybe, as we get ready for immuno-oncology, would be to get some clarity about which assessment we are going to use to get these patients enrolled.
Robert A. Figlin, MD: Thai, I don’t want to put you on the spot, but I will. A patient comes into your office in Arizona. They’ve just had a T3 lesion resected. You’ve done imaging, and they have no evidence of advanced disease. What’s the conversation going to be like?
Thai H. Ho, MD, PhD: I agree with Sandy. One of the things is, how do we better define what’s considered high risk? I like the Mayo Clinic’s SSIGN score. I like the UCLA’s scoring because these are things you can pull off the pathology report. You don’t need fancy DNA sequencing or RNA sequencing. You don’t have to have the patient pay extra money, because this is information that our pathologists have provided to us for free. So, I’d look at that, and I also break it down. For instance, I’d use the SSIGN score, and I’d break it down into low risk, intermediate risk, and high risk. I define “low risk” as a SSIGN score of 0 to 3. Historically, when you follow patients out over 20 years, at 15 years, greater than 90% of those patients with a SSIGN score of 0 to 3 will still be alive. If you look at the high-risk population, these are patients with a SSIGN score of greater than 8. At 15 years, only 6% to 7% of them will be alive. I use that SSIGN score first to figure out where that patient falls into.
The other thing that I’m also looking for is, is there a patient who has a small tumor but could conversely die of their disease? This is where a lot of our research is actually focused on, particularly with the BAP1 mutations. By these different SSIGN score stratifications, we can actually identify patients who have a small tumor but are at a higher risk of dying of their disease.
In some cases, with BAP1, they’re 3 times as likely to die of their disease as they are of having a BAP1 wild-type mutation. I discuss that with the patient. We do offer them molecular profiling, and to give them a sense of the frequency of surveillance, it’s tailored to the molecular profiling in combination with the clinical pathologic characteristics.
Robert A. Figlin, MD: Have you found in your patients that some accept adjuvant targeted therapy? Or is your recommendation obviously to put them on a trial first—but, absent of a trial availability, are you offering that to your patients?
Thai H. Ho, MD, PhD: I’m very upfront with them. I always tell them that we didn’t find that patients live longer as a result of being on these therapies. You have to be on these therapies for a year, and that’s difficult for a lot of them because, if you think about the other solid tumors, adjuvant therapies usually last about 6 months. This is a whole year. It’s a significant cost to take some of these medications. I present them the data, and most of them, knowing that they don’t think they’re going to live longer as a result, would like to have the molecular profiling. But they choose not to pursue targeted therapy unless it’s part of a clinical trial.
Robert A. Figlin, MD: Well, that’s very helpful. Martin, what are your thoughts on this?
Martin H. Voss, MD: The question that has come up now a couple of times in everyone’s thoughts is whether, in the future, we will be adding any type of molecular testing to help stratify these patients. It’s been done successfully in other cancers. Many community oncologists use the Oncotype DX assay regularly for patients with breast cancer to make decisions on adjuvant therapy. And there is some data, that you alluded to before, that tries to take that same approach in kidney cancer now. The Cleveland Clinic group had a publication several years back where they looked at nonmetastatic disease and the likelihood of disease recurrence based on a 16-gene score that is RNA based. And what they found—and it was subsequently validated by an independent European data set—was that, independent of the T stage, molecular information can help predict the likelihood of recurrence to some extent. Just like BAP1 can identify some patients with small tumors that have a higher likelihood of recurrence, the 16-gene score can do the same.
Those investigators have now taken things a step further. Dr. Bernard Escudier had an abstract at the 2017 ASCO Annual Meeting where he tried to apply that knowledge, now, on the S-TRAC data set—meaning that if we look at patients treated with sunitinib in the adjuvant setting versus those that receive placebo, is there any useful information in the 16-gene score to help us determine who the patients are that may or may not benefit?
What we can gather from the abstract on both sunitinib- and placebo-treated patients is that the gene score helps us to understand who is at higher and lower risk of recurring. What we don’t know is whether that helps with treatment decision making. Are there patients who we clearly think should have sunitinib therapy based on this or not?
This may all go back to what Michael mentioned before: Clear cell kidney cancer, universally, is driven by loss of VHL, up-regulation of HIF [hypoxia-inducible factor], and active VEGF signaling. So, maybe for this class of agents, molecular testing is not going to be so helpful to stratify things. But we are now thinking about targeted immunotherapy in the perioperative space, and there might be a bigger difference there.
We don’t have that signal yet in the metastatic setting, but we are going to investigate that further, and we should apply what we learned in the metastatic setting—quickly—to the perioperative trials, because that was a shortcoming of how the targeted therapies were tested in the perioperative space. Technology wasn’t available, but now, really, there’s no excuse to not do it. We have next-generation sequencing available to us, and there are many strategies to look into the tumor microenvironment. We should do that as we test immuno-oncology agents in the perioperative setting.
Transcript Edited for Clarity
Brought to you in part by Eisai
