Arjun V. Balar, MD: The next several years will be a hotly contested and anticipated area in terms of bladder cancer research, because as these agents bring their way into the first-line setting—we now have pembrolizumab, as well as atezolizumab in the first-line setting—we will see that an increasing proportion of patients with advanced bladder cancer will be treated with single-agent immunotherapy. The challenge that remains is, how do we incorporate chemotherapy into this treatment paradigm? What we do know, is that chemotherapy does extend survival in the metastatic setting in the first-line setting, specifically, with cisplatin-based chemotherapy.
We also know that up to 5% to 10% of patients can actually be cured with cisplatin-based chemotherapy. So, I think it’s too early to say that chemotherapy is going to be replaced. I think chemotherapy still has a strong role in advanced bladder cancer and will still be used. But the key questions that we have to test over the next several years are: what proportion of patients, and which patients in particular, are those who need to be treated with chemotherapy? There are several trials that will address this question. KEYNOTE-361 and IMvigor 130 are both frontline, randomized trials testing pembrolizumab and atezolizumab, IMvigor 130 is testing atezolizumab, whereas KEYNOTE-361 is testing pembrolizumab. And in both trials, they are testing them in combination with chemotherapy and as single agents versus chemotherapy on its own. These 2 trials will help definitively establish the role for single-agent versus combination of therapy with chemotherapy.
Currently, there are several strategies in place for testing immunotherapy in the first-line setting. What stands out, is the DANUBE study, which is testing single-agent durvalumab versus durvalumab in combination with tremelimumab versus standard-of-care chemotherapy. And this particular trial is well poised to test whether combination immunotherapy with CTLA4 compares against single-agent durvalumab and also chemotherapy.
What we are observing from initial phase I trials of CTLA4 plus PD-1 antibodies, is that the response rate may actually be higher. However, this does come at the price of excessive toxicity. In fact, up to 30% to 40% of patients will have a grade 3 or 4 toxicity with combination CTLA4 and PD-1, which is in stark contrast to the roughly 15% to 16% of grade 3 and 4 toxicity that we have thus far observed with single-agent PD-1 or PD-L1 antibodies.
So, CheckMate-032 is a very large phase I/II trial that is testing ipilimumab in combination with nivolumab at various dose schedules in advanced bladder cancer, amongst other solid tumor types. And in this particular study, 2 groups of patients were enrolled. One group was treated with ipilimumab at 3 mg/kg and with nivolumab at 1 mg/kg; whereas the opposite group, or the other group, was treated with ipilimumab at 1 mg/kg and nivolumab at 3 mg/kg. And what we saw in this particular study, is that the patients who were treated at the higher dose of ipilimumab had a response rate of 38%. However, only 26 patients were treated in that arm, suggesting that we may need to enroll more patients in that particular study. However, the initial feeling is that the combination strategy of ipilimumab and nivolumab, which is a CTLA4 and PD-1 antibody combination, may lead to higher responses. But of course, the challenge here is the toxicity with this combination therapy.
The combination of chemotherapy and immunotherapy is a challenging question. We are certainly learning from studies in advanced lung cancer, now that pembrolizumab has recently been approved in combination with chemotherapy in non–small cell lung cancer. We need to test that question in the bladder cancer patient population as well, and 2 studies are poised to test that question—KEYNOTE-361 and IMvigor 130 are both positioned to test that. Additionally, there are a number of investigator-initiated trials that are testing the combination of cisplatin-based chemotherapy in combination with immunotherapy, to test whether chemotherapy in certain agents can actually induce an immune response, that is then further propagated by the addition of checkpoint blockade. However, more time is needed and more data will need to be looked at to see really what the role is.
What I’m particularly excited about in the era of immunotherapy and combinations—not only for bladder cancer but a variety of solid tumors—is, what existing drugs can we use to make immunotherapy work better? One of my personal areas of interest is combining immunotherapy in earlier lines of disease. In fact, in looking at immunotherapy in patients who have localized muscle-invasive bladder cancer, some of our patients are not candidates for surgery.
We commonly treat them with radiation and low-dose chemotherapy. We know that radiation can be immunogenic, and one of my trials is testing whether we can add immune checkpoint blockade to chemotherapy and radiation to improve the percentage of patients who are cured with that combination approach, in those who have muscle-invasive bladder cancer. Additionally, we’re looking at the combination of VEGF-targeted drugs with checkpoint blockade because we know that VEGFs, which stands for vascular endothelial growth factor, can be highly immunosuppressive in the immune microenvironment of the tumor. If we lift that immunosuppression, that might help in making immune checkpoint blockade work better. And lastly, another exciting combination that I’m interested in seeing, is the combination of immune checkpoint blockade with IDO inhibitors, which appear to be well tolerated and efficacious in advanced melanoma, and we’re testing those combinations now in advanced bladder cancer as well.
The choice of chemotherapy versus immunotherapy remains a challenging one. And today, we do not have a good way of deciding which patient should be treated with chemotherapy first versus immunotherapy first.
I’ll summarize it by saying that so far PD-L1 expression has been loosely associated with response probability, and certainly we see slightly higher responses in patients who are PD-L1-positive. However, PD-L1–negative patients still have responses, and certainly it is not sufficient to exclude patients from being treated with immunotherapy. In my personal experience, what I feel is that patients who have a high volume of disease burden, who may be symptomatic from their disease, and may have low PD-L1 expression, these are the patients who may be best treated with chemotherapy first—who would get their treatment or their disease under control—and follow that with immunotherapy.
The other flip side is the patients who have a relatively low volume of disease burden and perhaps, their disease is restricted to lymph nodes only. These are the patients who do the best with immunotherapy, and these patients may best be treated with immunotherapy upfront.
The development of predictive biomarkers in advanced bladder cancer has been a hotly contested area. We have first looked at PD-L1 expression, and a variety of assays, antibodies, and scoring algorithms have been tested. Essentially, the summary of the data show that higher PD-L1 levels do correlate with response probability. But, however, patients who have low levels, and even absent levels of PD-L1 expression still do respond. And in fact, if we look at all the second-line trials, even the patients who had low PD-L1 levels had response rates in the range of 10%, which is what we actually see with chemotherapy. And so, we know that PD-L1 expression is something that does not correlate strongly enough to help us make treatment decisions. Other areas of interest include: mutational load, gamma-interferon gene signature, and TCJ subtyping. And these are some of the exploratory biomarkers that we’re looking at that might be helpful in the treatment decision making.