Purpose: HER2 may be a relevant biomarker in Barrett's cancer. We compared three HER2 laboratory methods, standard fluorescence in situ hybridization (FISH), image-based three-dimensional FISH in thick (16 microm) sections, and immunohistochemistry, to predict patient outcome.
Experimental design: Tissue microarray sections from 124 Barrett's cancer patients were analyzed by standard FISH on thin (4 microm) sections and by image-based three-dimensional FISH on thick (16 microm) sections for HER2 and chromosome-17, as well for p185(HER2) by immunohistochemistry. Correlations with clinical and follow-up data were examined.
Results: Only three-dimensional FISH on thick (16 microm) sections revealed HER2 gene copy gain to be associated with increased disease-specific mortality (relative risk, 2.1; 95% confidence interval, 1.06-4.26; P = 0.033). In contrast, standard FISH on thin (4 microm) sections and immunohistochemistry failed to predict clinical outcome. Low-level gain of HER2 occurred frequently in Barrett's cancer (>or=2.5-4.0 HER2 copies, 59.7%; HER2-to-chromosome-17 ratio, >or=1.1-2.0; 61.2%) and defined a subpopulation for patient outcome as unfavorable as HER2 gene amplification [disease-free survival, P = 0.017 (HER2 copies)]. This low-level group was neither definable by standard FISH nor immunohistochemistry. No prognostic significance was found for chromosome-17 aneusomy.
Conclusions: Low-level copy gains of HER2 define a biologically distinct subpopulation of Barrett's cancer patients. Importantly, these subtle copy number changes are not reliably detected by standard FISH in thin (4 microm) tissue sections, highlighting a thus far unrecognized weakness in HER2 FISH testing. These results should be taken into account for accurate evaluation of biomarkers by FISH and for HER2 FISH testing in tissue sections.