Abstract BACKGROUND Detection of a PIK3CA mutation (PIK3CAmut) is a prerequisite for treatment with currently approved PI3K inhibitors in breast cancer (BC). However, this was only recently introduced in testing guidelines, and cross-assay concordance studies are limited. Using available blood and tissue specimens from participants (pts) screened for the Phase 3 INAVO120 trial (NCT04191499) of the recently approved PI3Kɑ inhibitor inavolisib, we compared the ability of various molecular tests to identify PIK3CAmut. METHODS Data from FoundationOne®Liquid CDx (F1LCDx®; Foundation Medicine, Inc.) profiling on circulating tumor (ct)DNA was available for 1422 pts, either from central screening for study eligibility (n = 1383) or post-enrollment testing of ctDNA from local blood / tumor tissue test eligible pts (n = 39). Due to data / sample sharing restrictions, 43 pts from China were excluded from an otherwise global population. Retrospective to enrollment, DNA extracted from available archival / fresh tumor tissue was assessed by FoundationOne®CDx (F1CDx®; n = 323) sequencing and / or the Cobas® PIK3CA Mutation Test (a polymerase chain reaction PCR test; n = 84). Depending on the sample data available, assay concordance was evaluated for samples from both enrolled and unenrolled individuals. For certain assay comparisons, the fraction of samples in which a PIK3CAmut was detected by both assays is reported; where otherwise appropriate, traditional positive (+), negative (-), and overall percent agreement (PPA, NPA, OPA) between assays are reported. RESULTS Of the 1383 individuals screened centrally with F1LCDx, 500 (36.2%) had an enrollment-eligible PIK3CAmut; 242 of these plus the 39 local test-identified pts (n = 281) were enrolled in the study and comprised the efficacy-evaluable population for this analysis. F1LCDx was performed on available ctDNA from 37 / 39 local test-identified pts; 27 / 37 samples (73.0%) had a PIK3CAmut identified by both assays. Of the ten discordances, most could be explained by low ctDNA tumor fraction (median 0.0% in discordant, 6.0% in remainder). F1CDx was performed on available tumor tissue from 29 local test-identified pts; a PIK3CAmut was identified in all 29 (100%). F1LCDx and F1CDx data were available for 315 pts (214 enrolled and treated, 101 screened but not enrolled / treated). For this F1LCDx vs F1CDx analysis, PPA (F1LCDx+ and F1CDx+ / F1CDx+), NPA (F1LCDx- and F1CDx- / F1CDx-), and OPA (F1LCDx+, F1CDx+, F1LCDx-, F1CDx- / all data) were 93.0% (226 / 243), 73.6% (53 / 72), and 88.6% (279 / 315), respectively. Cobas PCR and F1CDx data were available for 84 pts (42 enrolled and treated, 42 screened but not enrolled / treated). PPA (Cobas+ and F1CDx+ / F1CDx+), NPA (Cobas- and F1CDx- / F1CDx-), and OPA were 90.2% (46 / 51), 100% (33 / 33), and 94.0% (79 / 84), respectively. With regard to efficacy, pts enrolled by a local test (hazard ratio for progression-free survival PFS 0.7; 95% CI 0.3-1.5; n = 39) or the F1LCDx central test (hazard ratio for PFS 0.5; 95% CI 0.3-0.7; n = 242) showed a similar response to inavolisib plus palbociclib and fulvestrant over placebo plus palbociclib and fulvestrant. CONCLUSIONS This comprehensive concordance analysis utilizing INAVO120 specimens demonstrated the ability of both tissue- and blood-based PCR and next-generation sequencing assays to robustly identify a pt population with PIK3CA-mutated, hormone receptor-positive, HER2-negative advanced BC who may consistently benefit from this INAVO-based regimen. Citation Format: S. Hilz, K. Kalinsky, D. Juric, N. Turner, K. Jhaveri, S. Im, S. Loibl, S. Loi, C. Saura, P. Schmid, E. Shaddox, N. Pantoja Galicia, C. Guo, Y. Li, W. Darbonne, R. Desai, M. Pagan, P. Karavagul, J. Aimi, S. Royer-Joo, J. Schutzmann, T. Stout, C. Song, P. Bedard, K. Hutchinson. Comparative analysis of blood- and tissue-based PIK3CA mutation detection methods in the pivotal Phase 3 INAVO120 trial of palbociclib + fulvestrant ± inavolisib in hormone receptor-positive, HER2-negative advanced breast cancer abstract. In: Proceedings of the San Antonio Breast Cancer Symposium 2025; 2025 Dec 9-12; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2026;32(4 Suppl):Abstract nr PS1-10-26.
Hilz et al. (Tue,) studied this question.