Post-Conference Perspectives: Hormone Receptor-Positive Advanced Breast Cancer - Episode 2
Nicholas McAndrew, MD, MSCE, reviews PIK3CA mutations in HR+/HER2- breast cancer and provides his perspective on the joint analysis of SOLAR-1 and CBYL719X2101 in which patients were given alpelisib and fulvestrant.
Nicholas McAndrew, MD, MSCE: Dejan Juric and colleagues presented a very interesting abstract regarding long-term disease control in patients with hormone receptor–positive, PIK3CA-mutated metastatic breast cancer who had been on either the SOLAR-1 trial or the CBYL719X2101 phase 1 trial with alpelisib and fulvestrant. The goal of this study was to understand who in these clinical trials was really deriving long-term disease control with the combination of alpelisib and fulvestrant. Around 40% of patients who have advanced hormone receptor–positive breast cancer have a PIK3CA mutation. The median progression-free survival on the SOLAR-1 study was improved with the addition of alpelisib and fulvestrant compared with fulvestrant alone, but it’s not clear which of these patients will have significantly improved disease control vs just moderately improved disease control. It’s really important to try to understand which patients are going to see serious long-term benefit from this.
Dr Juric and his colleagues looked at a host of secondary factors that were primarily clinical factors. They included over 40 factors in their analysis. They included patients who had long-term disease control. What is long-term disease control, and how did they come up with this number? They looked at long-term disease control as greater than 18 months. This came from looking at the average median progression-free survival in a host of studies, including SOLAR-1, trials that also had CDK4/6 inhibitors and fulvestrant, and fulvestrant monotherapy trials. They took those average median progression-free survival findings and then added 2 standard errors on top of it. So that would really be someone who’s outside the norm in terms of their degree of response. That’s how they came up with 18 months.
They took those patients, took their clinical factors, and then put them into a machine learning algorithm that would do multivariate analysis. That resulted in about 10 clinical factors that were significantly associated with this group. All in all, they had 51 patients from the SOLAR-1 study and 7 patients from the CBYL719X2101 phase 1 study who had long-term disease control. Of 51 patients in the alpelisib group, that would represent about 30% of the SOLAR-1 population who had received alpelisib. Within that, there were an additional 34 patients who had very long-term disease control, which was defined as 24 months or greater. In the multivariate analysis, they ran this by initially running a test set of 10 patients to practice the machine learning algorithm on and then applied it to the full data set.
They used receiver operating characteristics curves to help identify the combination of clinical factors that would best describe patients who had long-term disease control, which is a pretty standard method to use when doing predictive biomarker analysis. The approximately 10 clinical factors that they found associated with long-term disease control are pretty consistent with clinical features that we find represent favorable biology overall. These were patients who had ER [estrogen receptor] and, importantly, PR [progesterone receptor] positivity; a low number of metastatic sites; absence of liver metastases; bone-only disease; a decreased number of metastatic sites; and patients who were endocrine sensitive.
These are clinical factors we can easily identify within our patients. It should be very encouraging that these are the patients who—even with an aggressive PIK3CA mutation, which can represent increased or more aggressive disease and is often associated with the worst prognosis—with the combination of alpelisib, we see having significant benefit from alpelisib.
They also looked at the median dose intensity of this group. The patients in SOLAR-1 had received an 80% median dose intensity, which means that many patients were dose reduced and didn’t seem to have suffered any loss of efficacy as a result of their relative dose-intensity reductions. It’s an important study that helps us understand that patients with otherwise favorable characteristics do very well on alpelisib. If we have to dose reduce them for adverse effects, it seems to be safe to do that and not sacrifice efficacy.
Transcript edited for clarity.