Abstract

590 Background: pCR successfully identified candidates for (chemotherapy) de-escalation in the ADAPT HER2+ trial program. Here, we report biomarker analysis for pCR in the single arm, hypothesis-generating Keyriched-1 trial (NCT03988036) in HER2-E eBC treated with pembrolizumab and dual HER2 blockade. The trial rationale is based on high efficacy of anti-HER2 agents and potentially high immunogenicity and immune checkpoint protein expression in HER2-E eBC. Methods: 48 patients with HER2 2+ (ISH+) or 3+ eBC (stage I-III) and HER2-E subtype received pembrolizumab (200 mg), trastuzumab biosimilar (loading dose, LD 8 mg/kg, maintenance dose, MD 6 mg/kg), and pertuzumab (LD 840 mg/kg, MD 420 mg/kg) q3w for 12 weeks. Primary outcome was pCR (ypT0/is ypN0). Multiomic factor analysis was performed using different combinations of biomarker datasets: baseline BC360 single genes and gene signatures, multiplexed immunohistochemistry (mIHC), stromal tumor infiltrating lymphocytes (sTILs), somatic mutations and tumor mutational burden, and changes in circulating tumor DNA status under therapy. Factors representing combined markers developed from these pooled datasets were identified based on eigenvalue >1 criteria. Logistic regression with factors dichotomized by median score was conducted to analyze their association with pCR. Furthermore, logistic regression with elastic net regularization was performed using all single markers, factors from multiomic factor analysis, and clinical characteristics. Results: Factor involving gene expression (mainly related to immune response and tumorigenesis), mIHC markers (immune and HER2) in stroma, and sTILs datasets was significantly associated with pCR. Using the median as cutoff, higher scores increased the odds of pCR (unadj. OR 5.00, 95%CI 1.35, 18.56; p=.016). pCR rates were 66.7% and 28.6% in the high and low score groups, respectively. Effect of the factor (high vs low; OR 2.19) was also important in a model that additionally included clinical markers: grade (3 vs 2; OR 2.57), ER status (ER+ vs ER-; OR 0.89), cT (cT 2/3 vs 1; OR 0.69), cN (cN 1/2 vs 0; OR 0.51), and PR status (PR+ vs PR-; OR 0.34). When standardized values for all individual markers from gene expression, mIHC, and sTIL datasets were analyzed, markers with greatest impact on pCR vs. non-pCR were genes FNBP1, CD36, MYCN, and SIX1 (OR 1.43-1.62) and PR status (OR 0.65), with stronger impact than grade (OR 1.22) or nodal status (OR 0.90). Conclusions: Our multiomic analysis uncovered potential biomarkers for pCR in HER2-E eBC treated with immune therapy and dual-HER2 blockade alone. These results together with the results from the ADAPT translational research program support validation in larger trials investigating chemotherapy-free regimens in selected patients with HER2+ eBC. Clinical trial information: NCT03988036 .