Targeted next generation sequencing to expand HER2 status detection: Implication for newer HER2-directed agents.

Abstract

1047 Background: Establishment of ERBB2 ( HER2) amplification status in breast carcinoma (BC) and gastric carcinoma (GC) is essential for treatment selection, but no anti- HER2 therapies have been approved for tumors with low level of HER2 expression. The clinical trial ( NCT02564900 ) was initiated to evaluate the safety and efficacy of trastuzumab deruxtecan (DS-8201a) in BC patients with lower HER2 expression by current methods. This study aims to validate NGS-based detection for HER2 low expression. Methods: 275 BC and 425 GC were collected and subjected to NGS for genomic alteration detection. The testing was carried out by a College of American Pathologists (CAP) accredited and Clinical Laboratory Improvement Amendments (CLIA) certified laboratory, Shanghai, China. Protein expression was analyzed by using IHC. FISH was carried out on 108 samples, including 63 BC and 45 GC. To set up NGS cutoff, FISH was performed on additional 34 samples. Sensitivity, specificity and accuracy were evaluated based on FISH as a gold-standard reference. Results: In BC, the expression level of HER2 protein detected by IHC was overall IHC 0 in 28.7%, 1+ in 18.9%, 2+ in 27.3% and 3+ in 25.1%, respectively, while in GC, the expression level was 60.7%, 18.6%, 14.8% and 5.9%, respectively. Log2ratio was used to assess ERBB2 amplification status detected by NGS. According to the FISH results of 34 other samples with high sensitivity (98%) and specificity (100%), the threshold was determined as 0.5. 8 and 10 samples of IHC 1+/2+ met the cutoff in BC and GC, respectively. In 63 BC, there were 17 positive and 46 negative by FISH. According to the threshold, the sensitivity and specificity of NGS detection was 94.1% and 97.8%, respectively. The proportion of samples with IHC 2+ that couldn't determine NGS ERBB2 status was 49.2%. However, except for false positive and false negative, the NGS results were concordant with FISH. In 45 GC, there were 5 positive and 40 negative by FISH. The specificity and sensitivity was 97.5% and 40%, respectively. In 4 samples with IHC 2+, 2 of them were discordant with the results of NGS and FISH. All IHC 1+ didn't meet the cutoff of NGS and was FISH negative in 108 samples. The accuracy of NGS in ERBB2 detection was 96.8% and 91.1% for BC and GC, respectively. Conclusions: Our data indicated that the NGS-based detection of ERBB2 amplification had high sensitivity, specificity and accuracy. In samples with IHC 1+/2+, the results of NGS detection were high concordant with FISH detection.