Expert perspectives on the current paradigm of PI3K inhibition in the setting of relapsed/refractory HR+/HER2- breast cancer.
Hope S. Rugo, MD, FASCO: After we move on and talk about PI3 kinase mutations, PIK3CA mutations and PI3 kinase inhibitors, there were a couple of interesting analyses. There was a subgroup analysis at ESMO [European Society for Medical Oncology] Breast from Fatima Cardoso [, MD, director of the Breast Unit of the Champalimaud Clinical Center in Lisbon, Portugal,] looking at the BYLieve trial [r: NCT03056755]. Then there was an additional analysis at ASCO [American Society of Clinical Oncology] this year of the BYLieve trial looking at subsets in a poster discussion. It was interesting because it included duration of CDK4/6 inhibitor and other areas, the biomarkers, et cetera. But then there was an oral presentation by Dejan Juric [, MD,] also from SOLAR-1 [clinical trial; NCT02437318] looking at a dizzying array of different mutations and analyzing at what the outcome is. What’s your combined take on that?
Mark Pegram, MD: Well, BYLieve was the follow-on to SOLAR-1, because SOLAR-1 looked at a patient population that had not had prior exposure to CDK4/6 inhibitors. BYLieve was the perfect parallel to look at the modern era with patients where they’re treated now and see what their activity of alpelisib is in the PIK3CA-mutant subset. That established a proof of concept that the drug is quite active in a post–CDK4/6-exposed patient population. The first abstract presentation that you mentioned was a subgroup analysis to expand on the BYLieve data. The subgroups of interest that they looked at included menopausal status, bone-only disease patients; those patients with high burdened visceral metastasis. The bottom line is the data looked good for all those subsets. Although some of the subsets were small, there were clear trends showing in general better outcomes for the PFS [progression-free survival] endpoint when adding alpelisib into the treatment armamentarium. I think that’s gratifying to see that in most of these clinical subsets that you can readily identify, the drug seems to have activity without question. Another abstract that you mentioned looked more at the genomic level, at some of these patients with PIK3CA mutations and compared the genomic complexity of patients depending on how long they’ve been exposed to prior CDK4/6 inhibitors. Greater than or less than 6 months was their cutoff. Perhaps not surprisingly, patients with longer exposure to CDK4/6 inhibitors seemed to have a more complex genomic landscape compared to patients with shorter duration. The shorter-duration patients also had lower levels of circulating tumor DNA for example, and had a better prognosis, not surprisingly. I thought that was some interesting findings. Nothing that is ready to hang your hat on yet in clinical practice on Monday morning, mind you, but interesting science and more work needs to be done in that area. Then as you mentioned, the oral presentation today that did next-generation sequencing on patients from the BYLieve study. I thought for sure it was going to show some new genes when altered that would be associated with resistance clearly and that would be hypothesis-generating. Then you can do a prospective trial to critically evaluate that hypothesis. Unfortunately, when looking at a very wide range of genes including TP53, ESR1, CCND1, MAP3K1, etc, the benefits seem to be maintained with alpelisib irrespective of all these complex mutations that were found with the next-gene sequencing campaign. Gratifying that the drug is so active; it seems to be agnostic to many of these alterations that we find on next-gen. But at the same time, it is a bit disappointing because we didn’t find something that is really a card-carrying resistance factor that we can measure and be confident with in clinical decision-making. We are not there yet.
Hope S. Rugo, MD, FASCO: There is an identified series of altered genes that seem to correlate with shorter time on CDK4/6 inhibitors. I don’t know if that really qualifies as resistance per se, but that’s been looked at in several different settings. Then it seemed like MYC amplification correlated with worse outcome as it has in the neoadjuvant setting with triple-negative disease. It’s not good to MYC amplification, but as you have pointed out to me before, Mark, it doesn’t have a druggable target yet. I think that is really interesting, and I thought another thing that was interesting in the alpelisib data was that even if you had shorter exposure to CDK4/6 it didn’t really impact the duration of response to or PFS to alpelisib in combination with endocrine therapy. That is intriguing because it’s not what we thought about before which is ever decreasing PFS. There was a discussion at the post-HER discussion about hyperglycemia with alpelisib, which can be a big issue, and this lovely young woman who is graduating from fellowship at Memorial [Memorial Sloan Kettering Cancer Center] presented some institutional data. One of the things that was discussed in the discussion was that you should get your NGS [next-generation sequencing] relatively early because somebody’s going to be on their CDK4/6 inhibitor if they have a high hemoglobin. If they have a PIK3CA mutation, they have a high hemoglobin A1C or high sugar, you could refer them early and maybe get them into shape in advance. Is that anything that you do Aditya? How do you manage that?
Aditya Bardia, MD, MPH: It’s an interesting idea. To be honest it’s not something done in clinical practice and usually we check for PIK3CA mutations generally when patients have progression in first-line therapy because we are big believers in liquid biopsies. At that time, we tend to do liquid biopsy, look at mutations, PIK3CA, select alpelisib, some other alteration, then choose other therapy. We’ve discovered HER2 mutations and their ongoing studies looking at both TKIs [tyrosine kinase inhibitors] as well as antibody-drug conjugates for HER2-mutant cancers. It’s an interesting concept if the idea is to select PI3 kinase inhibitors in the future; maybe you should plan before?
Mark Pegram, MD: Though it’s a nice idea at least in our clinics the practicality is it’s hard to get an endocrine referral in a timely fashion and get those patients. You have to plan way in advance.
Hope S. Rugo, MD, FASCO: Eighteen months then you could do it. I thought it was a great idea.
Mark Pegram, MD: We might need a fraction of that.
Anne O’Dea, MD: At the time of diagnosis of metastatic disease.
Hope S. Rugo, MD, FASCO: Get your NGS. You’re doing your liquid biopsy and then refer to endocrinology.
Komal Jhaveri, MD, FACP: I completely agree Aditya. I think the practical and the standard of care used for biomarker directed therapy post CDK4/6 makes the most sense, just because we have therapies to offer based on that data. Having said that, we were just talking about the triplets strategies with PI3K inhibitors and AKT inhibitors. We all do biopsies from a metastatic site when a patient has newly diagnosed metastatic suspicion; we are getting the ER/PR [estrogen receptor/progesterone receptor] HER2 status. We at Memorial have been doing NGS from that biopsy because we’re already planning a biopsy. Then upon progression and CDK4/6 it would be nice to get a biopsy, but at the very minimum you could do a liquid biopsy. With that original biopsy, because PIK3CA mutations at majority truncal mutations and clonal mutations that information isn’t going to change, you will have enough time and you can offer them first line therapy. We use that for dual purposes, it’s not necessarily a standard-of-care decision that we make, but that comes in handy if you have a trial to offer in that setting. Then the planning can come along. I agree with you, getting an endocrinologist is harder than getting an oncologist. You want to plan ahead of time and get that for your patient.
Hope S. Rugo, MD, FASCO: The other suggestion from the audience that we should send them back to their primary care doctor. I was like, “What primary care doctor?” It’s very hard to find primary care doctors too; they’re all very busy. There are new PI3 kinase inhibitors that are being studied. I was interested to see that Inavolisib has hyperglycemia too. I think they reported something like 64% to 68% with of inavolisib and then there was really interesting data from faction about the AKT inhibitor capivasertib, a randomized phase 2 trial where originally, they said it worked and it was really nice. It led to the CAPItello-291 phase 3 trial [NCT04305496] that has completed accrual. We’re waiting for data but AKT inhibitor capivasertib with fulvestrant vs fulvestrant alone, and they said it worked in everybody; you didn’t need an alteration in the PIK3CA pathway. Then they went back and did NGS, and it only worked in the people with mutated pathways. Mark, they went back, and they looked at the people who were said to be wild type and about 25% of them had mutated pathways. The question is, do we trust our results? They looked only at the mutations and then they did a broader evaluation. So, some people wonder should they just be looking for that mutation, or should they do this broader look?
Mark Pegram, MD: Tumors are very heterogenous and very dynamic under selection pressure from various therapeutics you get outgrowth of underlying clones that may have been a very low fraction, perhaps even undetectable at a prior point in time. Therefore, I think you’re right. We need the most recent up to date information possible and cast the widest net to make sure that we get some things that though they may be low frequency now they might emerge as a resistant dominant clone later.
Hope S. Rugo, MD, FASCO: They used the same sample. It just was the test didn’t pick it up.
Mark Pegram, MD: Other differences in NGS technologies for sure as well.
Hope S. Rugo, MD, FASCO: When, if you’re going to do NGS and you’re looking broadly presumably you’ll pick them up, but to me it was more of a lesson not to just look for that narrow group of PIK3CA mutations that the test that was approved with alpelisib.
Mark Pegram, MD: Those are too narrow, I agree. And even though the labeling, which I think calls those out, in all practicality all of us are using the reports from next-generation commercial houses rather than just the FDA [Food and Drug Administration] approved.
Hope S. Rugo, MD, FASCO: And that, I think we can be confident and know.
Transcript edited for clarity.