Real-World OS Outcomes Are Similar for Frontline CDK4/6 Inhibitor Combos in HR+, HER2– Breast Cancer

Findings from the real-world, retrospective P-VERIFY study (NCT06495164) showed that differences in overall survival (OS) were not observed for 3 combination therapies featuring CDK4/6 inhibitors in the first-line treatment of patients with hormone receptor–positive, HER2-negative metastatic breast cancer. Findings were presented at the 2024 San Antonio Breast Cancer Symposium.

“To date, we have no head-to-head phase 3 studies comparing the CDK4/6 inhibitors,” noted Kari B. Wisinski, MD, professor of medicine at the University of Wisconsin Carbone Cancer Center in Madison, Wisconsin, in a presentation of the data.

P-VERIFY used the US nationwide Flatiron Health electronic health record-derived deidentified panoramic database to compare effectiveness among 3 CDK4/6 inhibitors plus an aromatase inhibitor (AI). Patients included in the trial had hormone receptor–positive, HER2-negative metastatic breast cancer, were 18 years or older, and started index treatment with either palbociclib (Ibrance) plus AI (n=6831; PAL+AI), ribociclib (Kisqali) plus AI (n=1279; RIB+AI), or abemaciclib (Verzenio) plus AI (n=1036; ABE+AI).

OS was determined using Kaplan-Meier estimates (unadjusted analysis). The investigators then applied standardized inverse probability of treatment weighting (sIPTW) to compare OS rates between treatment groups during the index period, from February 2015 to November 2023 (adjusted analysis).

“There was no difference in OS both in the unadjusted and adjusted methodologies,” Wisinski stated. The median follow-up was 33 months for the PAL+AI group, 16 months for the RIB+AI group, and 21 months for the ABE+AI group, she said.

The median OS prior to applying sIPTW (unadjusted analysis) in PAL+AI was 54.4 months (95% CI, 52.4-56.1), 60.3 months in RIB+AI (95% CI, 54.7-68.5), and not reached (NR) in ABE+AI (95% CI, 55.4- not evaluated [NE]). The OS rates in PAL+AI at 12, 24, and 30 months were 89.6%, 77.4%, and 71.4%, respectively. For RIB+AI these OS rates were 90.0%, 78.0%, and 73.3%, respectively. In ABE+AI, the OS rates were 88.4% at 12 months, 76.1% at 24 months, and 71.5% at 30 months.

For the unadjusted analysis, when comparing ABE+AI with PAL+AI the HR was 0.94 (95% CI, 0.83-1.06; P=.3205), comparing RIB+AI vs PAL+AI the HR was 0.93 (95% CI, 0.83-1.04; P=.2012), comparing ABE+AI vs RIB+AI the HR was 1.01 (95% CI, 0.87-1.19); P=.8698).

After applying sIPTW (adjusted analysis), the median OS for PAL+AI was 54.6 months (95% CI, 52.6-56.4), 59.0 months for RIB+AI (95% CI, 50.9-66.1), and 64.5 months for ABE+AI (95% CI, 55.4-NE). The OS rates at 12, 24, and 30 months for PAL+AI were 89.7%, 77.5%, and 71.4%, respectively. These were 89.2%, 77.3%, and 72.2% for RIB+AI, respectively. For ABE+AI the OS rates were 88.2%, 76.1%, and 71.5%, respectively.

In the adjusted analysis, for ABE+AI vs PAL+AI the HR was 0.95 (95% CI, 0.84-1.08; P=.4292), for RIB+AI vs PAL+AI the HR was 0.98 (95% CI, 0.87-1.10; P=.7531), and for ABE+AI vs RIB+AI the HR was 0.97 (95% CI, 0.82-1.14; P=.6956).

Also in the adjusted analysis, 3714 deaths were reported across all 3 groups: 3096 in PAL+AI (45.3%), 328 in RIB+AI (25.6%), and 290 in ABE+AI (28.0%). The median follow-up duration in group 1 was 33.0 months, 15.7 months in RIB+AI, and 21.5 months in ABE+AI. In addition, there were 5432 patient records who were censored across groups: 3735 (54.7%), 951 (74.4%), and 746 (72.0%), respectively.

“Now there are some limitations for this analysis. The RIB+AI and ABE+AI cohorts had smaller numbers and shorter follow-up, and the point estimates were relatively (or possibly) unstable after 30 months. In addition, as always with real-world analysis there is potential for unmeasured confounders,” Wisinski noted.

Wisinski applied these real-world data to the following phase 3 trials: PALOMA-2 (NCT01740427), MONALEESA-2 (NCT01958021), MONALEESA-3 (NCT02422615), MONALEESA-7 (NCT02278120), and MONARCH-3 (NCT02246621). The primary end point for each study was progression-free survival (PFS), and all 3 demonstrated benefit, Wisinski explained. “The hazard ratios were also quite consistent among these studies,” she said. When comparing the median OS, Wisinski stated that “median OS has been improved with the addition of ribociclib, abemaciclib demonstrated a clinical meaningful benefit but it was not statistically significant, and palbociclib did not provide benefit.” In the real-world analysis the median OS rates were quite consistent with what was observed in these studies, Wisinski added.

According to investigators, after applying sIPTW, demographics and clinical characteristics were balanced across the 3 groups (with standardized differences less than 0.1). This included exact 1:1 matching on the following: age, gender, race and ethnicity, practice type (community or academic), disease stage at initial diagnosis, ECOG, time from initial to metastatic breast cancer diagnosis, visceral disease, bone-only disease, number of metastatic sites, and menopausal status at initial diagnosis.

For example, in PAL+AI, 62.5% of patients were White, 9.3% were Black, and 28.1% were of other ethnicities; For RIB+AI, 62.6% were White, 9.2% Black, and 28.3% were of other ethnicities. For ABE+AI, 63.0% were White, 9.1 were Black, and 27.9% were of other ethnicities. In addition, 99.0% were female in PAL+AI, 99.1% in RIB+AI, and 98.9% in ABE+AI.

Regarding adverse effects (AEs), there were higher rates of neutropenia in the PAL+AI group (67%) than RIB+AI (60%) and ABE+AI (21.9%). However, anemia was greater in the PAL+AI group (5.4%) and ABE+AI groups (6%) groups vs the RIB+AI (0.9%). The occurrence of thrombocytopenia was consistent across the 3 groups: 1.6% for PAL+AI, 0.9% for RIB+AI, and 1.9% for ABE+AI.

“The P-VERIFY study suggests that there is no difference in OS among first-line CDK4/6 inhibitors and this would align with the PFS data that we have from those first-line studies; However, the statistically OS benefit that we have seen has only been with ribociclib. I think that longer follow-up would be very valuable and although real-world analyses have limitations it is unlikely that we will have a head-to-head study, and thus, these data add to the information that we have,” Wisinski concluded.

Reference

Rugo HS, Layman RM, Lynce F, et al. Comparative overall survival of CDK4/6 inhibitors plus an aromatase inhibitor (AI) in HR+/HER2– MBC in the US real-world setting. Presented at: 2024 San Antonio Breast Cancer Symposium; December 10-14, 2024; San Antonio, TX. Abstract PS2-03.