In screening patients with esophageal cancers for HER2 status, the relative efficiency of immunohistochemistry (IHC) vs fluorescence in situ hybridization (FISH) has been debated. Researchers from the Mayo Clinic compared the testing strategies and have proposed an algorithm that puts IHC up front, with FISH restricted to cases with an indeterminate (2+) IHC score.1
Harry H. Yoon, MD, of the Department of Oncology, who presented the results at the 2013 Gastrointestinal Cancers Symposium, said, “Our findings support a testing algorithm for resected esophageal adenocarcinomas where IHC is used for initial screening, and FISH testing is restricted to cases with equivocal IHC results.”
HER2 overexpression, a known driver of tumor aggressiveness in esophageal adenocarcinoma, is found in 7% to 22% of cases. For HER2-positive patients, trastuzumab (Herceptin) improves survival; therefore, testing for HER2 is important.
Concordance between HER2 Testing Methods
While exploratory research in upper digestive tumors has suggested that IHC is more predictive of trastuzumab benefit than FISH, the concordance between these two tests has not been established, and each testing method has its advantages. IHC is faster, less labor-intensive, and less expensive, while FISH is less vulnerable to tissue artifacts and offers a more objective scoring system, according to Dr. Yoon.
Dr. Yoon and colleagues evaluated 673 surgically resected esophageal adenocarcinoma specimens for the concordance of HER2 testing between IHC and FISH, in a blinded manner using FDA-approved assays. A consensus IHC score was determined by two pathologists using tumor-specific criteria (negative, 0 or 1+; indeterminate, 2+; positive, 3+). Gene amplification by FISH was defined as a HER2/CEP17 (chromosome 17) ratio ≥ 2.
By FISH, 17% of tumors tested positive. By IHC, 13% tested positive, 25% were indeterminate (ie, 2+), 23% were negative by 1+, and 39% were negative by a 0 score.
Among the 89 patients with IHC 3+ score, 79 were FISH-positive and 10 were FISH-negative. Of the 167 indeterminate (2+) patients, 21 were FISH-positive and 146 were FISH-negative. Of the 417 IHC-negative patients, 16 were FISH-positive and 401 were FISH-negative.
This yielded a concordance rate of 95% between FISH and patients who were IHC 3+ or IHC-negative, and a 74% concordance rate when IHC 2+ patients were included.
HER2 amplification was detected in 89% of IHC 3+ cases, 13% of IHC 2+ cases, and 4% of IHC 0 to 1+ cases. Accordingly, using FISH as the reference standard, the positive predictive value of a positive IHC test (3+) was 89%, and the negative predictive value of a negative test was 96%. Importantly, the positive predictive value of an indeterminate IHC score (2+) for detecting HER2 amplification was 13%, he said.
“In the largest study to date comparing HER2 testing methods in esophageal adenocarcinoma, a negative IHC result nearly excludes the presence of gene amplification by FISH,” Dr. Yoon said. “A positive IHC result (3+) strongly predicts for the presence of amplification (89%), and an indeterminate IHC result (2+) is a weak predictor for amplification (13%).”
“Ideally, the clinical impact of HER2 expression or amplification is best studied as a predictor of benefit from trastuzumab therapy,” he said. “However, the clinical data that can address this question more robustly than the ToGA trial [which established the benefit of adding trastuzumab to chemotherapy for gastric cancer] are, at best, many years in the future. Our proposed algorithm is intended to serve as a helpful guide for clinicians in the meantime. Integrating our findings with ToGA data, we believe that IHC is the preferred screening test in surgical esophageal adenocarcinoma specimens to determine HER2 status, with referral to FISH restricted to cases with an indeterminate IHC score.” ■
Disclosure: Dr. Yoon has received honoraria and research funding from Genentech/Roche.
1. Yoon HH, Shi Q, Sukov WR, et al: HER2 testing in esophageal adenocarcinoma using parallel tissue-based methods. 2013 Gastrointestinal Cancers Symposium. Abstract 2. Presented January 24, 2013.
J. Randolph Hecht, MD, Professor of Clinical Medicine and Director of the Gastrointestinal Oncology Program at the University of California, Los Angeles, commented to The ASCO Post that it is premature to accept this algorithm in the absence of its correlation with clinical outcomes. The one...