Abstract Background I-SPY2.2 is a multicenter phase 2 platform sequential multiple assignment randomized trial (SMART) in the neoadjuvant breast cancer (BC) setting, evaluating novel regimens in Block A, followed by standard therapies in Blocks B/C if indicated. Therapy in Block B is guided by tumor Response Predictive Subtype (RPS; Wolf, Cancer Cell 2022), which integrates immune, DNA repair, and luminal gene expression signatures with HR and HER2 status, classifying BC into 6 subtypes. ARX788 is an anti-HER2 Antibody Drug Conjugate (ADC) with site-specific conjugation of the noncleavable tubulin polymerization inhibitor Amberstatin (AS269). Patients (pts) with RPS S5 (HER2+/non-Luminal) and S6 (HER2+/Luminal) were eligible for assignment to ARX788 Q3W x 4 cycles in Block A. The primary endpoint was pCR following investigational Block A or after the entire therapy sequence. Methods Pts underwent serial MRI scans during therapy. Projected responders after Block A or B could proceed to surgery early, while others continued to Blocks B/C, which included TCHPx6 or THP±AC. Efficacy of ARX788 followed by surgery was evaluated using a Bayesian covariate-adjusted model estimating pCR and compared to fixed subtype-specific thresholds. To estimate pCR rate in the context of a multi-decision treatment regimen (Blocks A/B/C), we estimated pCR using a Bayesian model that considers timing of pCR and compares rates to a subtype-specific Dynamic Control (DC) modeled from I-SPY data (N=1,818). Results From Sep 2022 to Dec 2024, 100 pts were randomized to ARX788. In sequence with chemotherapy, 63/100 pts achieved pCR, and 82/100 had RCB 0/1, all without AC. Notably, 40 pts proceeded to surgery after Block A alone; of these, 33 (82%) had RCB 0/1, avoiding prolonged conventional chemotherapy. The RPS S6 (HER2+/Luminal) subgroup had a pCR rate of 39%, compared to a 17% DC rate, exceeding the predefined probability threshold of superior pCR rate to DC. Receptor status alone did not correlate with efficacy, emphasizing the importance of molecular subtyping in treatment selection. There were no RCB-3 cases. Ocular toxicity was reported in 95% of patients (9% grade 3), and pneumonitis in 6% (2% grade 3), with no treatment related deaths. 8 pts did not complete all planned ARX788 due to adverse events. The results for ARX788 as a stand-alone therapy or sequential strategy are summarized in the Table. Conclusions The sequential combination of ARX788 with chemo/anti-HER2 therapy demonstrated high efficacy, particularly in the HER2+/Luminal subtype compared to DC, suggesting that a significant proportion of pts could avoid prolonged chemotherapy. Ongoing studies are underway to define and implement strategies to prevent treatment associated ocular toxicity. I-SPY2.2 validates the use of personalized neoadjuvant therapy guided by molecular diagnostics and response-adaptive strategies tailored to each individual patient. NCT01042379 Citation Format: P. R. Pohlmann, H. S. Rugo, C. Yau, D. Yee, A. J. Chien, N. O. Williams, A. M. Wallace, J. C. Boughey, C. Vaklavas, M. Arora, V. Borges, A. S. Clark, C. Omene, C. Isaacs, E. Stringer-Reasor, R. Nanda, C. Falkson, K. S. Albain, N. Chan, E. T. Roussos Torres, M. Rozenblit, J. Tseng, S. Bommakanti, C. Nangia, L. Brown-Swigart, G. L. Hirst, N. Pasricha, C. H. Kretzer, J. Perlmutter, A. Borowsky, W. F. Symmans, L. van ‘t Veer, N. Hylton, L. J. Esserman. Pathologic complete response rates (pCR) after the novel HER2 ADC ARX788: Results from the I-SPY2.2 trial 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 RF5-06.
Pohlmann et al. (2026) studied this question.