Biomarker Data from the Phase III KATHERINE Study of Adjuvant T-DM1 versus Trastuzumab for Residual Invasive Disease after Neoadjuvant Therapy for HER2-Positive Breast Cancer

Abstract

Purpose: In KATHERINE, adjuvant T-DM1 reduced risk of disease recurrence or death by 50% compared with trastuzumab in patients with residual invasive breast cancer after neoadjuvant therapy (NAT) comprised of HER2-targeted therapy and chemotherapy. This analysis aimed to identify biomarkers of response and differences in biomarker expression before and after NAT.

Experimental design: Exploratory analyses investigated the relationship between invasive disease-free survival (IDFS) and HER2 protein expression/gene amplification, PIK3CA hotspot mutations, and gene expression of HER2, PD-L1, CD8, predefined immune signatures, and Prediction Analysis of Microarray 50 intrinsic molecular subtypes, classified by Absolute Intrinsic Molecular Subtyping. HER2 expression on paired pre- and post-NAT samples was examined.

Results: T-DM1 appeared to improve IDFS versus trastuzumab across most biomarker subgroups, except the HER2 focal expression subgroup. High versus low HER2 gene expression in residual disease was associated with worse outcomes with trastuzumab [HR, 2.02; 95% confidence interval (CI), 1.32-3.11], but IDFS with T-DM1 was independent of HER2 expression level (HR, 1.01; 95% CI, 0.56-1.83). Low PD-L1 gene expression in residual disease was associated with worse outcomes with trastuzumab (HR, 0.66; 95% CI, 0.44-1.00), but not T-DM1 (HR, 1.05; 95% CI, 0.59-1.87). PIK3CA mutations were not prognostic. Increased variability in HER2 expression was observed in post-NAT versus paired pre-NAT samples.

Conclusions: T-DM1 appears to overcome HER2 resistance. T-DM1 benefit does not appear dependent on immune activation, but these results do not rule out an influence of the tumor immune microenvironment on the degree of response.

Figure 1.

Study populations for biomarker analyses. ^aTwo patients (both in the trastuzumab arm) are not included: one had HER2-positive status that was determined locally but that was not centrally confirmed; the other was randomized twice in error, with a missing HER2 status at the time of the first randomization and a positive HER2 status during rerandomization. HR, hormone receptor.

Figure 2.

Forest plot of IDFS by HER2 expression subgroups assessed at eligibility. Results for HER2 expression by IHC and ISH are from pre-NAT biopsies in 80.4% (1,195/1,486) of patients and from surgical tissue in the remaining 19.4% (289/1,486) of patients. Data for groups with fewer than 15 patients are not shown. Two patients did not have confirmed HER2-positive disease. NE, not estimable.

Figure 3.

T-DM1 improved IDFS regardless of PIK3CA mutation status, and PIK3CA mutations were not prognostic overall. Of the 1,363 samples available, 1,027 (75.3%) were post-NAT surgical samples and 336 (24.7%) were pre-NAT samples. A,PIK3CA mutation analysis by treatment arm. B,PIK3CA mutation analysis in pooled-treatment arms.

Figure 4.

Predictive value of mRNA expression levels on IDFS. Forest plot of treatment effect on IDFS by mRNA expression level subgroups (>median vs. ≤median) in RNA-evaluable surgical tissue samples. Dashed line indicates the overall treatment effect in the population with biomarker-evaluable surgical samples (n = 814). The analysis is adjusted for tumor content (percentage of tumor in the marked area). One sample could not be included because there was insufficient information on tumor content. Teff, T-effector signature (CD8/granzymeA/granzymeB/perforin/IFNγ); three-gene signature (PD-L1/IFNγ/CXCL9); five-gene signature (PD-L1/granzymeB/CD8/IFNγ/CXCL9); Th1 cytokine signature (CXCL9/CXCL10/CXCL11); checkpoint inhibitor signature (PD-L1/PD-L2/IDO); B-cell signature (CD19/CD79A/FCRL5/MS4A1/POU2AF1/STAP1).

Figure 5.

Biomarker gene expression analysis in surgical tissue samples. Kaplan–Meier plots of IDFS by (A) HER2 mRNA expression level (>median vs. ≤median) and treatment group, (B) AIMS PAM50 intrinsic subtypes in trastuzumab-treated patients, (C) AIMS PAM50 intrinsic subtypes in T-DM1–treated patients, and (D) PD-L1 mRNA expression level (>median vs. ≤median) and treatment group. One sample could not be included because there was insufficient information on tumor content. AIMS-HER2-E, AIMS HER2-enriched.