Andrew J. Armstrong, MSc: In the first-line setting, abiraterone and enzalutamide are highly effective. More than 80% or 90% of patients will respond. And so, there isn’t a major unmet need for a biomarker that will distinguish responders and nonresponders, because most patients do respond. However, one of the biggest disappointments of each of these drugs is that they share resistance mechanisms. So, when you use one drug followed by the other, the second drug tends to be less effective, and vice versa. If you start with enzalutamide and then switch to abiraterone later, abiraterone is much less effective than the same drug if you’d never had enzalutamide.

And so, the cross-resistance became a very important topic of discussion and management. There is a need for a predictive biomarker that indicates that cross-resistance, and the measurement of these androgen receptor variants becomes increasingly prevalent as the disease progresses and resistance develops.

For example, if you do a circulating tumor cell test and measure AR-V7 in the blood of men who have not yet had abiraterone or enzalutamide, the prevalence of a positive test is pretty low. It’s about 10% to 12%, meaning that most patients who have the test are negative, and most patients likewise will benefit from these therapies. So, the clinical utility of such a test is fairly limited in that frontline setting.

However, if you look after abiraterone or enzalutamide, the prevalence of a positive test goes up to about 20% to 25%. And those patients tend to really have no benefit from the crossover therapy to the oral drug, abiraterone or enzalutamide, after the frontline oral therapy. These patients who are AR-V7—positive tend to benefit from other therapies though. Docetaxel, cabazitaxel, and radium are all effective therapies in patients who have AR-V7-driven tumors. The AR-V7 patients do have a poor prognosis. Having AR-V7 is associated with other poor prognostic factors such as having widespread metastatic disease and high-grade disease; many other genomic alterations are likely present in these cells. So, we haven’t yet figured out a way to kill the AR-V7–positive cells other than with chemotherapy and radiation.

When a physician uses enzalutamide and abiraterone in the clinic, and in the laboratory, we see a wide range of resistance mechanisms, not just AR-V7. For example, with enzalutamide, there are agonistic mutations in the ligand-binding domain that can turn enzalutamide from a blocker into a stimulant of the cancer. That is apparently very uncommon. Under 10% of patients studied to date have these mutations, and it seems fairly rare. Other pathways that may be a variant include loss of AR function. So, this neuroendocrine phenotype of prostate cancer where the tumor essentially no longer cares about the androgen receptor. And drugs that block the androgen receptor essentially have no effect on these tumors. For these antiplastic or aggressive variant tumors, really the indication for these treatment options include docetaxel, chemotherapy, platinum-based chemotherapy, and novel approaches.

Other more transcriptional pathways that are upregulating include the glucocorticoid receptor, which has been implicated in enzalutamide resistance, as well as other pathways. There are about 12 to 15 different mutations in each prostate cancer patient that confer a selective growth advantage or the resistance to cell death. Every patient with prostate cancer has his own unique mutational spectrum of disease modifying mutations. So, this is where the personalized medicine approach does come in because each patient carries with him a set of unique mutations that make the tumor aggressive.

Androgen receptor variants may be associated with different mutational signatures; we just haven’t discovered what those are yet. AR-V7 is associated with a poor prognosis whether it’s with abiraterone on enzalutamide. And trying to block AR-V7 is really an ongoing active-area investigation. With abiraterone, as opposed to enzalutamide, you have agonistic mutations where the prednisone that you use as part of the abiraterone can turn the androgen receptor into a stimulating receptor. And so, there are some unique mechanisms that are unique to abiraterone versus enzalutamide, and vice versa.

The cross resistance is one of the biggest issues in clinical practice, and community oncologists recognize this by using these oral—oral sequences. When you use abiraterone after enzalutamide, the response rate is down to about 20% or 30%, lasting 3 to 5 months. Same for enzalutamide after abiraterone—the response rates are just very low and very short-lived. And so, the decision to switch a therapy becomes very important. We generally recommend, according to the Working Group guidelines, not to switch a therapy based on PSA changes alone, but when other manifestations of the disease start to worsen, such as radiographic changes, new metastases, growth of existing metastases, symptomatic progression. That becomes a real trigger and a need to switch therapy.

But often, that is dependent on what that switch is. If the switch is to docetaxel, chemotherapy, there may be a reluctance to switch right away to that therapy because of the adverse effects of chemotherapy. Often, we’re now looking at combination approaches, and so the switching question is a little less important as we’re looking to add on therapies such as radium-223.

I think the important point of identifying resistance is that every patient is likely to have his own unique set of mutations that created the tumor in the first place. For example, about 20% of men have DNA repair alterations—BRCA1, BRCA2, ATM. These gatekeeper mutations unleash a wide number of additional mutations that make these tumors highly aggressive, such a p53 mutation, MIC gains, neuroendocrine transformations. These are very difficult to target but are an active area of research.

It’s important, when you’re seeing a patient with prostate cancer, to consider a mutational profile such as through FoundationOne or a number of other commercial platforms that can really look at a host of mutations that are highly individualized to that patient. AR-V7 is not presently part of those biopsy-based mutational profiles. The AR-V7 assays are really moving into the clinic more as a liquid biopsy rather than a solid biopsy.

Transcript Edited for Clarity