Novel Immunotherapy Approaches for Urothelial Cancer - Episode 2
Jonathan Rosenberg, MD: While immune checkpoint blockade therapy represents a normal advance in the treatment of advanced bladder cancer, unfortunately only a minority of patients respond to the treatment. Approximately 30% of patients will respond to pembrolizumab in the first-line setting. About 24% of patients will respond to atezolizumab in the first-line setting, in cisplatin-ineligible patients. In the second-line setting, the response rates are approximately 15% to 21%.
Clearly, we need to do better. Seventy-five percent to 80% of patients are not deriving substantial benefit from these treatments. But the people who do derive benefit have derived a dramatic long-term benefit—often with durable responses. Many of the trials that are ongoing have not yet reached the median duration of responses for these patients, suggesting that these are in the order of 2 to 3 years, if not longer. Certainly, in my own practice, I have patients who were on the original clinical trials with these agents. They are still responding to therapy 3 and 4 years later. So, it’s very exciting for the people in whom it works. But we’re left with a very large group of patients in whom the immune drugs are not providing dramatic benefit.
There are certain strategies that we can think about, to try to combine agents, that might improve the likelihood of response and clinical benefit with immunotherapy in bladder cancer. One obvious one is with chemotherapy. Multiple trials are testing standard chemotherapies in combination with immunotherapy. There are phase II pilot studies asking particular questions, and larger randomized phase III trials are going on, that are trying to definitively answer whether or not it improves survival in patients who receive the immune checkpoint blockade. Hopefully we’ll know the answer to this in the next several years.
There are also many efforts looking at other approaches to combine novel targeted agents. These focus on either immunotherapy or other approaches that may lead to improved outcomes. There are agents targeting angiogenesis, such as bevacizumab, targeted therapies, tyrosine kinase inhibitors, that are going to be tested in bladder cancer, in combination with immunotherapy. There are immune—immune combinations of multiples types, including drugs that are focused on the indoleamine dioxygenase pathway—IDO1 inhibitors. Several of them have actually shown proof-of-concept activity in combination with an immune checkpoint blockade in bladder cancer, showing response rates that were substantially higher than what you might expect with the single drug by itself. Fortunately, this has been seen without tremendous toxicity. Another very interesting area are DNA vaccines, where DNA is injected into the patient, along with immune-stimulatory molecules, to lead to a new immune response against cancer-related antigens.
Biomarkers to select patients for treatment of bladder cancer remain to be a very controversial area. PD-L1 testing on tumor cells or immune cells by immunohistochemistry has not proven useful, at this point, in bladder cancer. It’s been proven more useful in lung cancer and other diseases. In 2 randomized phase III trials, PD-L1 testing failed to really demonstrate a useful effect. In the pembrolizumab randomized phase III trial, which supported the drug approval in the second-line setting, PD-L1 testing did, in fact, show that the drug worked better in PD-L1—positive than PD-L1–negative patients. But, in fact, when you look at the trial, the PD-L1–negative patients who got pembrolizumab did the best. So, it’s really not a helpful biomarker to pick patients. If you use that biomarker, you’re not going to pick patients properly.
In the atezolizumab phase III trial, which was a negative phase III trial, the high levels of PD-L1 expression predicted response to both chemotherapy and immunotherapy. It used a different test, the SP142, from the pembrolizumab trial. That trial also showed that it’s not a useful biomarker. It probably highlights a better prognosis group of patients, or is predictive for both chemotherapy and immunotherapy response. And so, PD-L1 testing, at the moment, is not useful in bladder cancer. I don’t recommend it or use it in my practice. It may highlight groups of patients that might be useful in the future with other targeted combinations—in particular, subpopulations. But right now, it’s not really a useful test.
Tumor mutation burden, on the other hand, is a very interesting concept that has been shown, in many disease types, to be associated with improved outcome. This has been shown in bladder cancer. In our study of atezolizumab, the phase II study that led to FDA approval, we showed that the total number of mutations was highest in the responders. The top quartile of patients had improved overall survival. So, it not only led to improved tumor shrinkage, but it also improved long-term outcomes.
Other studies have shown this. Another interesting potential biomarker of response to immune checkpoint blockade is the presence of DNA damage repair mutations beyond things we know about—like MLH1, MSH2, and the mismatch repair pathway. The presence of DNA damage repair mutations in bladder cancer, in tumors, appears to be associated with substantially higher response rates as well. In part, it’s probably mediated through tumor mutation burden. But it provides a very discreet way of measuring whether or not a patient’s tumor is likely to respond to immunotherapy. Tumor mutation burden is a continuous variable. It’s not clear if someone who has 7 mutations is going to do worse than someone who has 9 mutations, if your threshold is 8. The presence of a DNA damage repair mutation might actually demonstrate the fact that the tumor is poised to respond because there are more neoantigens available.
Transcript Edited for Clarity