In terms of molecular analysis, this was an unselected patient population. Whole exome sequencing was done in 17 patients who had soft tissue for analysis. We found 1 patient with a BRIP1
alteration. That being said, we did not definitively find HRD in any of the patients. Additional molecular analyses with sequencing on more patients with soft tissue disease is needed. We will also need to look at other factors like microsatellite instability and tumor mutational burden (TMB). This cohort is being expanded to 100 patients; these data [that were presented] involved 42 patients. A phase III trial is being launched in patients with mCRPC who received prior docetaxel and either abiraterone acetate or enzalutamide. They will be randomized to receive pembrolizumab plus olaparib versus the alternate hormone agent. For example, if you had abiraterone, you would get enzalutamide, and vice versa.
What was the rationale for combining a checkpoint inhibitor with a PARP inhibitor?
Since this was an unselected trial, we did think there was a good chance we would find patients with HRD. There are data in breast and ovarian cancers showing that when you add a PARP inhibitor to patients with HRD, you increase PD-L1 expression and TMB. There is quite a bit of theory there and data to support it. In the HRD-proficient population, there even is growing literature in breast cancer for patients who do not have BRCA
mutations that if you give them a PARP inhibitor you can increase PD-L1 expression. More recent publications show that because you are giving a PARP inhibitor and you have these unresolved DNA fragments that accumulate in the cytosol, you can induce type-1 interferon and lead to anti-tumor immunity. Following up with the PD-1/PD-L1 antibody makes a lot of sense there.
Is there anything you would like to add about KEYNOTE-365?
It is an interesting study. Because it is a nonrandomized study, it is interestingly hypothesis-generating. Acquiring more patients, and understanding the demographics and molecular data, will help us move this treatment forward. Of course, the only way to prove this is beneficial in patients is to do the randomized control trial, which is being launched.
You also presented the preliminary results of the GALAHAD study of niraparib in mCRPC. What were the key takeaways from this research?
GALAHAD is using niraparib, which is another PARP inhibitor, in this setting. So far, there are no data to suggest that one PARP is better than the other. Some have more enzymatic activity, others have better PARP-tracking capability, but we don't know if one is generally better than the other. At this point, there are multiple PARP inhibitors being studied, with niraparib being one of them. This trial was a very interesting and educational one in the fact that when it first started, they used a circulating tumor DNA assay and accrued patients very rapidly. As time goes on, however, I think it makes sense to focus on patients with biallelic alterations.
That being said, the results on patients with biallelic alterations—about 50 patients—looks pretty promising. We are seeing good composite response rates, and the only reason I am focusing on this is because it focuses on RECIST evaluation, PSA decline, and circulation tumor cell enumeration. Again, these are open-label phase II nonrandomized trials, but they certainly look promising.
How rare are these biallelic gene defects?
If you look at the Stand Up to Cancer studies, it was found that about 23% of patients harbored biallelic alterations in DNA repair genes. That being said, as time moves forward, we may find that those rates are not quite that high. It might be lower than that because we are at tertiary cancer centers and we may be seeing a different patient population. Do I think it is a significant rate? Yes. Do I think it is ultimately going to be as high as 23%? Maybe, on the lower side.