• Sofie Lunden Nielsen
5. Term (Master thesis), Medicine, Master (Master Programme)
Background: Manual assessment of HER2 protein expression in gastric and gastroesophageal junction (GEJ) adenocarcinoma is prone to inter-observer variability and hampered by tumor heterogeneity and different scoring criteria. This study aimed to evaluate the accuracy of digital image analysis (DIA) as a more objective and precise method for the assessment of HER2 protein expression.
Methods: Hundred and ten gastric and GEJ adenocarcinomas were included applying a tissue micro array (TMA) approach with three cores per case. Two immunohistochemistry (IHC) assays, PATHWAY® and HercepTest™, and fluorescence in situ hybridization (FISH) were performed for all cases. Manual interpretation of IHC slides followed guidelines for both resection and biopsy specimens. The HER2 CONNECT™ DIA software as designed for breast carcinoma was applied. Connectivity, calculated by the software, was converted to the standard IHC scores (negative, equivocal, positive) applying predetermined cut-off values for breast carcinoma as well as novel cut-off values. Cases with excessive cytoplasmic and nuclear staining as well as HER2 amplified IHC negative cases were excluded from HER2 CONNECT™ analysis.
Results: Manual scoring of IHC slides, using criteria for biopsies, achieved the most optimal combination of specificity and sensitivity using FISH as the reference. Applying HER2 CONNECT™ with established connectivity cut-off values designed for breast carcinoma resulted in 72.7% sensitivity and 100% specificity for the identification of HER2 amplified cases. The sensitivity was increased to 100% when the new cut-off values were applied, while the specificity remained 100%. With the new cut-off values, a statistically significant reduction of IHC equivocal cases (50.0% for PATHWAY® and 36.4% for HercepTest™) was observed. Three cores with HER2 protein overexpression and amplification were classified as negative by HER2 CONNECT™. However, the other cores from the specific cases ensured an accurate classification.
Conclusion: HER2 CONNECT™ seems to be an effective tool for assessment of HER2 protein expression and gene amplification in gastric and GEJ adenocarcinoma.
Publication date2015
ID: 224275226