Molecular techniques play a critical role in identifying breast cancer patients with overexpressed human epidermal growth factor receptor-2 (HER2). New bright field techniques such as chromogenic in-situ hybridization (CISH) and silver in-situ hybridization (SISH) have emerged to overcome some of the challenges associated with the reference standard, fluorescence in-situ hybridization (FISH). We conducted a literature review and synthesis to characterize the accuracy of HER2 tests, and inform decisions about test selection.

We searched MEDLINE and EMBASE databases using these eligibility criteria: studies evaluating invasive breast cancer samples which examined agreement between CISH or SISH, and FISH, and reported sensitivity, specificity, or concordance. We performed a bivariate meta-analysis of sensitivity and specificity using a generalized linear mixed model in Stata. We used likelihood ratio tests from meta-regression to compare accuracy between HER2 tests.

The search identified 4,475 articles, of which thirty-one were included. A total of thirteen studies (43%) evaluated dual-color SISH, twelve single-color CISH, and six dual-color CISH. The summary estimates for sensitivity and specificity were, respectively, 0.97 (95%CI 0.83–0.99) and 0.99 (95%CI 0.96–1.00) for single-color CISH, 0.98 (95%CI 0.92–0.99) and 0.98 (95%CI 0.91–0.99) for dual-color CISH; 0.92 (95%CI 0.86–0.95), and 0.96 (95%CI 0.91–0.98) for SISH. Significantly higher specificity was reported for single-color CISH than SISH (chi-square 4.12; p = 0.04), while dual-CISH had higher sensitivity than SISH (chi-square: 4.63; p = 0.03). These differences were not maintained when studies with cohorts enriched with equivocal samples were excluded.

The agreement between new bright field tests (SISH and CISH) and FISH is high (>92 percent). Indirect comparison of HER2 tests indicated that overall CISH performance exceeds that of SISH. However, low agreement between SISH and FISH in equivocal cases affects these comparative estimates. The pooled estimates from this meta-analysis can help inform future HER2 test selection decisions.