Dejan Juric, MD, leads the discussion on the role of PIK3CA mutations in HR+ metastatic breast cancer.
Andrew D. Seidman, MD: I want to come back to where we started in terms of genomic profiling. Dejan, I’ll let you take the lead on discussing how common PIK3CA mutations are, which mutations were primarily seen in SOLAR-1, and what the prognostic implications of these mutations are. What have we learned since SOLAR-1?
Dejan Juric, MD: PIK3CA is a gene that encodes PI3 kinase alpha. It’s a really important regulator of cell signaling, mediating growth signals that come from the membrane of the cell and go into the cell cycle machinery. The activating mutations in that gene and resulting activation of the protein itself are present in 40% of patients with ER [estrogen receptor]–positive metastatic breast cancer. This is the most common actionable operation we have. The only 1 that even matches their prevalence would be TP53, which is not as actionable. There are some clinical trials in that space, but at the moment, it’s not something that’s clinically actionable. As Aditya explained, we certainly should pay attention to PIK3CA from the beginning, or the first diagnostic specimen.
The mutations in PI3 kinase happen in 2 different areas. First are called hot spot mutations, which are present in these 2 domains of the proteins: helical and kinase. The vast majority of patients have 1 of those mutations, either in the helical or the kinase domain. A few patients will have both. It’s those 2 areas of the protein or gene that really matter.
In the SOLAR-1 trial, we focused on 11 mutations in total. They mostly spanned these hot spot mutations. We had another area called exon 7 that we looked at, but that was it. It was 11 mutations in exon 7, 9, and 20 of the gene. This is about 80% of all PIK3CA mutations, but it’s not all of them. We should first look at the data we have from SOLAR-1 on these hot spot mutations. The benefit of alpelisib added to fulvestrant compared with placebo plus fulvestrant is undeniable.
We should also run an additional analysis to look at what we call non–hot spot mutations. These mutations are outside these commonly mutated areas, which can also activate the protein. This is where it becomes a little challenging for a clinician. It goes back to comments made earlier by Hope and Aditya. It’s not just presence of mutation. It’s also the type of mutation that matters. For hot spot mutation, we’re on a safe ground, but with non–hot spot mutation, we have to work with our micropathologist and try to understand if this is mutation that we see. I’ll give you one example: R88Q, which was not in the initial 11 mutations of SOLAR-1. But is that mutation altering the function of the protein? Is it likely to activate the pathway? If it is, I would argue that this is the setting where we should be studying or using PI3 kinase alpha inhibitors like alpelisib. Where the difficulty continues is when you start comparing the tissue vs blood.
Andrew D. Seidman, MD: Yes. Let me ask Aditya. Do you start with looking for plasma circulating tumor DNA and resort to tissue first, or is it more of a practical decision? Maybe your patient just had a recent biopsy and you’ve got the tissue, so you won’t order the plasma. What’s your algorithm at MGH [Massachusetts General Hospital] for approaching PIK3CA testing?
Aditya Bardia, MD, MPH: If someone has a tissue biopsy, we’ll consider genotyping from the tissue biopsy, particularly in the first-line setting. After that, it’s usually plasma-based genotyping. One additional advantage of plasma-based genotyping is that if the initial tissue-based genotyping then shows a PI3 kinase mutation and the patient has progression on letrozole plus CDK4/6 inhibitor, a subset could have acquired PIK3CA mutations that you can detect in the blood. To give the patient the maximum chance of getting access to a PI3 kinase inhibitor, that’s a really good time to get plasma-based genotyping.
Andrew D. Seidman, MD: We recently saw that Fabrice André published the updated overall survival from SOLAR-1 in Annals of Oncology. There was an 8-month advantage for alpelisib vs placebo when combined with fulvestrant, which didn’t achieve statistical significance with a hazard ratio of 0.86. Yet for an interesting subset of patients with visceral disease, the curves seem to be more separate. Does anybody want to offer some thoughts on that result?
Hope S. Rugo, MD: It’s an interesting thing. I certainly don’t claim to be an expert on statistical design, but we have to distribute alpha, and the overall survival differences that are made as secondary end points are sometimes way outside what anyone would consider to be feasible in this. Although with CDK4/6 inhibitors, we saw even greater differences from what we would have guessed.
In this population, the numerical difference was interesting, but it really has to do with the calculated curves over time. The medians were not different in terms of the statistical calculations that were needed for the study. But when we looked at the visceral group of patients that we would have expected to do much more poorly, we saw a suggestion of a benefit. Even though you can’t make a statistical conclusion about that, it’s very encouraging for us in treating patients in this setting where you want to avoid going to chemotherapy. You know there’s a PIK3CA mutation, and you can treat a patient who has relatively more endocrine-resistant disease by phenotype or by clinical presentation. I thought those were helpful data clinically.
Dejan Juric, MD: Yes. I believe the number was 37 vs 23 months in the patients with visceral disease who had PIK3CA mutations. It’s a whopping difference numerically, but it’s a small study. The entire cohort for patients with PIK3CA mutations was only 341 patients, so we had to answer many questions.
What’s clear, though, Andrew, is that it looks like patients with visceral disease and PIK3CA mutation do quite poorly. We need to either deploy drugs like alpelisib there in the clinical setting or seek clinical trials for them, because they generally don’t do well. That matches some of the other analysis of CDK4/6 clinical trials, both MONARCH 2 and MONALEESA-3 trials. The abemaciclib and the ribociclib studies in ER-positive disease have shown that patients with PIK3CA mutations tend to perform less well on both the experimental arm on the CDK4/6 arm and the control arm.
Hope S. Rugo, MD: Although you benefit those patients with CDK4/6.
Dejan Juric, MD: Definitely, they do benefit.
Hope S. Rugo, MD: They have a worse outcome. Absolutely.
Dejan Juric, MD: The hazard ratio, undeniably. PIK3CA doesn’t appear to be a predictor for CDK4/6 benefit, but if you look at absolute numbers, they tend to perform less well compared with wild types in both arms.
Transcript Edited for Clarity