Poster lunch (ID 46) Poster display session

20P - ERBB2 mRNA as predictor of response to anti-HER2 antibody-drug conjugates (ADC) in breast cancer (BC) (ID 333)

Presentation Number
20P
Lecture Time
12:15 - 12:15
Speakers
  • Gaia Griguolo (Bellinzona, Switzerland)
Session Name
Poster lunch (ID 46)
Location
Exhibition area, MARITIM Hotel Berlin, Berlin, Germany
Date
03.05.2019
Time
12:15 - 13:00

Abstract

Background

Quantitative measurements of HER2 might identify the probability to respond to anti-HER2 ADCs better than semi-quantitative methods such as immunohistochemistry (IHC).

Methods

Expression of 55 BC-related genes was determined by nCounter in tumor samples from a retrospective series of 58 T-DM1-treated patients (pts) with advanced BC from two independent institutions. Genes associated with T-DM1 overall response (OR; partial or complete) were identified using two-class unpaired SAM. Cutoff Finder (Plos One 2012) was used to optimize cutoff to predict OR. Uni- and multi-variate analysis for OR were performed using logistic regression. Nine BC cell lines (4 HER2-) were treated with T-DM1 (1.25 µg/ml) for 72 hours; treatment response (reduction in cell viability) was correlated to ERBB2 levels. Finally, % of ERBB2-high tumors across HER2 IHC groups was explored in 392 and 368 primary BCs from an in-house (nCounter) and the TCGA (RNAseq) datasets, respectively.

Results

HER2 IHC status in T-DM1-treated pts was: IHC0 (n = 5), IHC+1 (n = 3), IHC+2 (n = 12), IHC+3 (n = 35). 60% of pts were HR+, 72% received 1 prior line of trastuzumab and 93% had visceral metastases. OR rate was 43%, median PFS was 5.3 months (95% CI 3.9-7.6). GRB7 and ERBB2 were the top 2 genes associated with OR (FDR<1%). Although HER2 IHC (3+ vs ≤ 2+) was associated with OR in univariate analysis (p = 0.027), only ERBB2 expression (ORR 1.93, p = 0.021) and number of prior therapy lines (0-1 vs ≥ 2) were independently associated with OR. An ERBB2 cutoff was optimized to predict OR (100% sensitivity, 36.4% specificity, AUC 0.69). In BC cell lines, the ERBB2 cutoff predicted T-DM1 activity (median cell viability 43% vs. 100%, p = 0.015). Importantly, T-DM1 response across BC cell lines strongly correlated with ERBB2 continuous expression (r = 0.88). Finally, ERBB2-high distribution within HER2 IHC groups in both datasets (in-house and TCGA) was: 0% and 0% (IHC0), 1.10% and 0.59% (IHC+1), 0% and 0% (IHC+2/non-amplified), 9.38% and 25% (IHC+2/amplified) and 76.68% and 79.63 (IHC+3).

Conclusions

ERBB2 expression is associated with T-DM1 efficacy in BC regardless of HER2 IHC status. Lower ERBB2 cutoffs might be needed for highly potent anti-HER2 ADCs such as DS-8201.

Legal entity responsible for the study

Translational Genomics and Targeted Therapeutics in Solid Tumors, August Pi i Sunyer Biomedical Research Institute (IDIBAPS), Barcelona, Spain.

Funding

Asociación Española Contra El Cancer (AECC, Postdoctoral 2017) University of Padua (DOR 1721185/17 - DOR 1830512/18).

Disclosure

G. Griguolo, T. Giarratano: Travel grants: Pfizer. T. Pascual: Advisory board: Roche. M.V. Dieci: Consultancy role: Eli Lilly, Genomic Health; Advisory boards: Eli Lilly, Celgene. V. Guarneri: Institutional research grant: Roche; Advisory boards: Eli Lilly, Roche, Novartis; Speaker’s bureau: Eli Lilly, Novartis. A. Prat: Advisory role, speaker fees: Daiichi Sankyo, NanoString, Roche, Novartis, Pfizer, Lilly; Funding: NanoString Technologies, Roche, Novartis. All other authors have declared no conflicts of interest.

Collapse