HER2 Immunohistochemistry Significantly Overestimates HER2 Amplification in Uterine Papillary Serous Carcinomas

Abstract

Recently, there have been numerous reports showing that HER2 overexpression or amplification occurs in a variable number of uterine papillary serous carcinoma (UPSC) cases, leading to a current clinical trial targeting this pathway. Although approved algorithms exist for scoring HER2 overexpression/amplification in breast and gastroesophageal carcinomas, scoring criteria and the optimal methodology for assessing HER2 in UPSC are currently unknown. Most frequently, the American Society of Clinical Oncology/College of American Pathologists (ASCO/CAP) breast carcinoma algorithms have been utilized for UPSC, wherein cases are screened with immunohistochemistry (IHC), followed by fluorescence in situ hybridization for equivocal cases. However, interpreting HER2 IHC can be prone to significant subjectivity, often leading to false-positive results. To better correlate HER2 IHC results with underlying amplification in UPSC, we compared HER2 overexpression by IHC with HER2 amplification with chromogenic in situ hybridization (CISH). A total of 69 cases of UPSC-57 pure and 12 mixed-were identified over a 10-year period. All were included in a tissue microarray, and HER2 IHC and CISH were performed. Each case was scored according to the most recent 2013, as well as the 2007, ASCO/CAP scoring guidelines for breast carcinoma. Whole-tissue sections were also examined in cases with amplification by CISH on initial screening, as well as an equal number of negative cases, to account for intratumoral heterogeneity. Nine (13%) cases showed HER2 amplification by CISH, whereas 14 (20%) and 28 (40%) cases showed overexpression with IHC when the 2007 or 2013 ASCO/CAP criteria were utilized, respectively. The overall concordance rate between CISH and IHC was 64% (9/14) with the 2007 ASCO/CAP criteria and 32% (9/28) with the 2013 ASCO/CAP criteria. Intratumoral heterogeneity was seen in 3 (33%) amplified cases. No additional amplified cases were identified on subsequent whole-section examination after the initial negative tissue microarray screening. While confirming that HER2 amplification is found in a subset of UPSC, our results also show that screening with IHC will overestimate the number of cases showing underlying HER2 gene amplification. The significant discordance between methods in our study suggests that optimal HER2 testing parameters in UPSC are yet to be defined. Future clinical trials should incorporate both IHC and ISH results for each patient in their study design, so that conclusions can finally be made about which method is ultimately a better predictor of treatment response.