Joyce O’Shaughnessy, MD: I want to transition to something we can definitely use after progression on CDK4/6, and that is alpelisib. First, we’ve got to find the right patient population. I’ll say a few words about looking for and finding the patients with the PIK3CA mutation in their breast cancer. Then we’re going to turn it over to you, Hope, to update us on your really nice data, the BYLieve data, from ASCO [American Society of Clinical Oncology Annual Meeting].

This is put against the need for NGS [next-generation sequencing], as Denise said, on the map. We’ve got to find these patients with ER [estrogen receptor]–positive, HER2 [human epidermal growth factor receptor 2]–negative metastatic breast cancer, whose breast cancers have a PIK3CA mutation. You know it’s a truncal mutation that happens very early in the development of primary breast cancer.

By and large, if you look for it in the primary breast cancer, you’ll find it in about 40% or so of patients. In the metastatic setting you should find it, and about 40% of patients. Many of us biopsy the first recurrence. Sometimes we don’t get that much tissue. It’s not that easy to assay for NGS all our metastatic biopsies. Here we have our patients; they’re on a CDK4/6 inhibitor, and we need to be thinking about what the next step is for those patients.

And we need to see if alpelisib is going to be an option, so many of us will get ctDNA [circulating tumor DNA] on those patients either upon their initial presentation with metastatic disease or upon progression on a CDK4/6 inhibitor. In data from the SOLAR-1 trial, which was fulvestrant plus or minus the alpelisib, only 6% had a prior CDK4/6 inhibitor, hence the larger experience that Hope found from the BYLieve trial. We found in the SOLAR-1 that patients in the metastatic setting, in whom circulating tumor DNA identified the PIK3CA mutation, did marvelously well with very large separation in the progression-free survival curve. That was very impressive. If we do find PIK3CA mutations in ctDNA in our metastatic patients, that’s a great patient for alpelisib.

If we don’t find it in the circulating tumor DNA, I certainly go back to the patients’ primary breast cancer, and I test it there as well. The NCCN [National Comprehensive Cancer Network] really does recommend that particular algorithm. If I don’t find it in either place, I will check for it again further down the road in the metastatic setting because the PALOMA-3 trial—which gave fulvestrant plus or minus palbociclib—got ctDNA before treatment and upon progression found about 10% more patients had to PIK3CA mutation upon progression on a CDK 4/6 inhibitor.

I’ll keep looking because over time, as Hope said, they’ll be more clonal heterogeneity that evolves, and you might still be able to help the patient. It’s been pretty easy to find the patients with the PIK3CA mutations. Hope, what do we do with those patients who are progressing on CDK4/6 inhibitors? Tell us about the BYLieve data that will help us understand who will benefit from alpelisib.

Hope S. Rugo, MD, FASCO: These are really fascinating data. Alpelisib was first approved based on the data from SOLAR-1 where patients who had progressed on an aromatase inhibitor had to receive fulvestrant and were either placebo or alpelisib, and then the PIK3CA-mutant cohort. The median progression-free survival was almost doubled with a hazard ratio of 0.65, leading to regulatory approval of alpelisib.

The big issue is that SOLAR-1 was enrolled completely before CDK4/6 inhibitors were really approved in most places. We really didn’t have enough data to know how alpelisib would work in this situation; 20 patients in SOLAR-1 had prior CDK4/6 inhibitor treatment. In that group with PIK3CA mutations, the median progression-free survival went from 1.8 to 5.5 months, in such a tiny group of patients.

We designed the BYLieve trial with the idea to really treat those patients. At ASCO we presented cohort A; there were 3 cohorts. The third cohort after chemotherapy is still accruing. The second cohort hasn’t reached the 6-month landmark. In the cohort we presented, patients had to have progressed on a CDK4/6 inhibitor and an aromatase inhibitor as their immediate prior treatment.

Then they were all treated with alpelisib and fulvestrant. They all had to have a PIK3CA mutation and HER-positive disease. Our final analysis of these data included patients who also had central confirmation of PIK3CA mutations. Of 127 patients, 6 of them didn’t have central confirmation, which is kind of interesting. We’re always going to see small differences. Our end point was the number of patients who are alive without progression at 6 months, and that was 50.4%.

We had to have a lower limit of the 95% confidence interval greater than 30%, and it was 41%. We definitely met the end point. If you looked at that tiny group of patients who had prior CDK4/6 inhibitors— in SOLAR-1, 44% of them were alive with disease progression at 6 months—it seems pretty good to me. Progression-free survival in this group of patients was 7.3 months, which is also pretty good.

Response rates were comparable, and most patients had a decrease in tumor size. Importantly, we looked at toxicity, and we saw similar issues as we had seen in SOLAR-1 with hyperglycemia and rash. We did again see that patients who took prophylactic antihistamines before they started or around the time they started the therapy had a marked decrease in severity rash and incidence of rash. Many fewer patients discontinued from because of hyperglycemia, clearly showing that people—when they get used to the treatment—are better able to manage the toxicity.

The dose intensity was higher actually in BYLieve than it had been in SOLAR-1. It wasn’t blinded, so everybody knew they were on it. The single-arm trials are important because we need to know how drugs that have gotten approved fare in the real-world situation, where everybody is getting CDK4/6 inhibitors.

We don’t want to randomize patients again to receive endocrine therapy alone. What we did was we used a real-world evidence cohort of patients who all had PIK3CA-mutated advanced breast cancer, and had all received prior CDK4/6 inhibitor therapy from the US Flatiron Health–Foundation Medicine clinico-genomic database to better understand how our data from BYLieve compared with the real-world setting. A few patients received capecitabine, and most received endocrine therapy with or without another targeted agent, like a CDK4/6 inhibitor on progression.

What we saw was that regardless of the way we weighted these different analyses and matched to them, we used a different way of matching. The progression-free survival appeared to be twice as long in almost all. Essentially, all cases were at least twice as long if not longer in the BYLieve cohort compared with the Flatiron cohort of patients, which included almost 100 patients. These are really encouraging data and really inform us in clinical practice about how we can use this PIK3CA inhibitor, alpha-specific inhibitor, and also that experience and education really help us manage the toxicity.

Transcript edited for clarity.