Michael J. Hall, MD, Fox Chase Cancer Center, discusses the emergence of PARP inhibition in the armamentarium for patients with pancreatic cancer.
Michael J. Hall, MD
PARP inhibitors, which have approved indications in breast and ovarian cancer, are expanding to additional treatment paradigms, including pancreatic cancer.
Data from the phase III POLO trial presented at the 2019 ASCO Annual Meeting showed that the PARP inhibitor olaparib (Lynparza) significantly improved progression-free survival (PFS) compared with placebo among patients with germline BRCA-mutated metastatic pancreatic cancer. The median PFS with olaparib was 7.4 months compared with 3.8 months with placebo (HR, 0.53; 95% CI, 0.35-0.82; P = .0038).1,2
In an interview with OncLive, POLO study investigator Michael J. Hall, MD, Fox Chase Cancer Center, discussed the emergence of PARP inhibition in the armamentarium for patients with pancreatic cancer.
OncLive: What was the rationale for looking at a PARP inhibitor as maintenance in this setting?
Hall: Pancreas cancer is 1 of the most dismal tumors in all of oncology, particularly in gastrointestinal oncology, and what we have discovered over the years is that the only option to control disease once it spreads outside of the pancreas, which is 80% of patients when they present, is to use big time multiagent chemotherapies that are toxic. What’s challenging about that is you’re already taking someone who is often very sick from the disease, and to be able to extend their life sometimes by 6 to 8 months, you’re giving them very toxic chemotherapies. That’s not necessarily something that’s attractive to a patient. They want extended disease-free survival and QoL; they want to have both.
The idea behind this was that there were some preliminary data several years ago that showed that in [patients with] pancreas tumors that have a germline BRCA mutation, and at that point we knew the range was maybe 5% to 15%, we knew that those tumors responded differently than the other pancreas tumors. They seemed to be especially sensitive to platinum-based agents. Talia Golan, MD, who is the senior author on the study, had a paper in the Journal of Clinical Oncology that showed that platinum agents were great in this subgroup of patients with pancreas cancer.
Over time, we understood it was the deficiency in homologous recombination and BRCA that was driving this sensitivity, so when the PARP inhibitor class of agents was developed, naturally this became something [of interest]. I think a lot of people first looked at pancreas cancer and said there is no way an oral agent will ever be able to control this disease if FOLFIRINOX can only hold off PFS by a few months, how can this oral agent ever do this? That was a real gamble. I think there were a lot of naysayers that felt it probably wouldn’t work. The trial was born of that. How was this trial designed?
To get on this trial, there were a few different steps. First, you had to undergo germline BRCA testing. I’ll talk a little later on in regard to germline versus somatic, but over 3000 people worldwide were screened for this study to find the roughly 90 patients that went on the olaparib arm and the roughly 60 patients who went on the placebo arm. That was the first part.
Among that group, you had to be a responder to platinum therapy. We know that’s a big part of understanding who is going to respond to PARP inhibitors, so on this trial you actually had a choice of whether it was platinum-based like oxaliplatin, some were on gemcitabine, but frankly most people were on a modified FOLFIRINOX regimen since that is the standard US regimen. You had to be a responder on that for at least a particular amount of time. Once you were a responder for at least 16 weeks or more showing a response, your doctor could actually leave you on that regimen as long as they wanted and as long as you continued to respond, but there was a minimal requirement to demonstrate response.
Once that response was demonstrated, then people were randomized 3:2 to olaparib versus placebo. Everyone asks why 3:2, but that’s because the developers of the trial knew it was not going to be attractive to someone who had an aggressive disease to consider being randomized to placebo, so we tried to make the numbers a little better to go on the study drug. We ultimately randomized around 90 patients to the olaparib arm and around 60 patients to the placebo arm.
What were the findings that were presented at ASCO from the POLO trial?
Hall: PFS was our major endpoint. What we saw and what we presented was that we saw a nearly doubling of PFS with a hazard ratio of progression of 0.53, which I think blew everyone away. The P value is very nice, and it demonstrates this is something that works.
There are other endpoints that we are going to be waiting for, so we will be waiting for OS. It’s going to take a while, though, because some of the responders are still responding for a very long period of time. The median duration of response right now is about 24 months. That shows you that for the people where it is working, we are seeing [good responses]. To get the 106 deaths for OS, that may take some time.
We’ve done some subgroup analyses on the PFS, I think the 1 of the high points of that is BRCA1 and BRCA2 responded roughly the same way. We didn’t see any differences in the subgroups, and I thought the other exciting thing was roughly 50% of people had either a partial response or complete response before they went on the study and roughly 50% were at stable disease. Those 2 groups actually responded the same way, which shows you it’s not just about picking out super responders and giving an icing-on-the-cake drug. It’s about demonstrating some response, and then the olaparib takes over on that. That’s super exciting.
Toxicity data was exactly as we predicted. Olaparib had somewhat more toxicity than being on placebo, naturally, but we didn’t see any extremely high rates of grade 3/4 toxicities, so that fit what we were estimating.
We also did some health-related QoL analyses. The punchline of that was that olaparib had no significant decline of health-related QoL compared to being on a placebo or nothing. I think that’s a strong [finding] because the last thing you would want to do is give someone a maintenance therapy, but you just make them sick again. That’s not what we saw, and we are very happy about that.
Where do we go from here?
We need to wait for the rest of the data to come out. We need the rest of the data to mature, so I think that will be exciting. The other question is what about somatic BRCA? We know germline BRCA[mutations are in about] 6% to 7% [of patients], but somatic is probably another 3% to 4% of patients. The FDA approval will first be for the germline as the study showed, but the questions will be asked, and the research will be done looking at somatic BRCA. There are also other genes and mutations in the BRCA family, like PALB2 gene, the CHEK2 gene, ATM—they are all part of the Fanconi anemia pathway. Those are mutations based on some research from other groups, and we know those mutations are common. The question is if we can also get the same response.
There is this bigger group of people that are called the homologous recombination deficient pancreas cancers. The rough number is about 30% of pancreas cancers that are homologous recombination deficient. Can we use olaparib in that group as well? Can we combine olaparib with other agents? I think that’s the future. That’s what we will see at ASCO 2020, 2021.