Background: Neoadjuvant therapy is currently standard of care for patients with HER2-positive (HER2+) breast cancer. Patients with residual disease after neoadjuvant therapy have a higher risk of metastatic recurrence compared to patients achieving pathologic complete response (pCR). While both primary breast cancer and residual disease may harbor cancer cell subpopulations with high metastatic potential, residual disease after neoadjuvant targeted therapy may include larger subpopulations of therapy-resistant cells. A better understanding of the differences in genomic aberrations in residual disease following neoadjuvant therapy in comparison to primary breast cancer is essential to choice of novel post-neoadjuvant treatment strategy for HER2+ patients with residual disease.

Methods: We are currently building a large ethnically diverse, breast cancer cohort with integrated genomic data. We analyzed clinical data of stage 1-3, HER2+ breast cancer. pCR after neoadjuvant therapy was defined as ypT0N0 or ypTisNo. Using the xT 595-gene panel (Tempus Labs, Inc.) with matched tumor-normal samples in a subset of the cohort, we evaluated genomic heterogeneity between HER2+ primary breast cancers and residual disease following neoadjuvant therapy.

Results: There were 469 HER2+ early stage invasive breast cancer patients in the study cohort with a median follow-up of 6.0 years. The mean age at diagnosis was 53.6 (SD=13.4) years. 52.7% are European Americans and 39.2% are African Americans. 61.8% were hormone receptor-positive (HR+)/HER2+ and 37.7% were HR-/HER2+. Although patients undergoing neoadjuvant chemotherapy (n=191) had lager tumor and more nodal positive disease compared to patients undergoing adjuvant therapy (n=213), the two groups have similar recurrence-free survival rate (adjusted hazard ratio 0.88, 95% CI 0.50-1.55). In patients undergoing neoadjuvant therapy, 72 (37.7%) achieved pCR. Women with HR-/HER2+ tumors had higher pCR rate (58.9%), compared with 24.6% in women with HR+/HER2+ tumors (p<0.001). pCR was a very strong predictor for recurrence: the 5-year recurrence-free survival was 0.97 for patients with pCR in contrast to 0.75 for patients without pCR. Of the 38 patients whose tumor DNAs were sequenced to date, there were 23 primary tumors and 15 residual tumors. Compared to primary HER2+ tumors, residual tumors had lower frequency of somatic alterations in ERBB2 (33% vs. 74%, p=0.02), CDK12 (13% vs. 48%, p=0.039), and ATM (0% vs. 26%, p=0.063) compared to the primary tumors. The overall tumor mutational burden was 2.1 m/MB (range: 0.4-42.5) in primary tumors and 2.9 m/MB (0-12.5) in residual tumors (p=0.94). Median ERBB2 copy number is 10 (range: 2-20) in primary tumors and 3 (2-20) in residual tumors. Alterations in the CDK4 pathway (CDK4, CCND1, CDKN2A, and RB1) were more frequently detected among tumors with loss of ERBB2 amplification (50%) than tumors with ERBB2 amplification (10%, p=0.011), suggesting they may be mutually exclusive. In particular, 4 of the residual tumors had mutations in the CDK4 pathway and none co-occurred with ERBB2 amplification.

Conclusions: These preliminary data support the importance of integrating genomic sequencing of residual tumors to identify potentially non-HER2 druggable targets as a post-neoadjuvant treatment strategy. Alternative approaches including addition of novel therapies such as CDK4 inhibitors may have the potential to reduce costs associated with extended use of HER2-targeted therapies when unlikely to improve overall survival for non-HER2 amplified tumors.

VIEW THE PUBLICATION