Comparison of visual and automated assessment of HER2 status and their impact on outcome in primary operable invasive ductal breast cancer

Aims: To compare visual and computerized image analysis of HER2 immunohistochemistry (IHC) with fluorescence in-situ hybridization (FISH) for HER2 status, and to examine the relationships with outcome in patients with primary operable invasive ductal breast cancer. Methods and results: Tissue microarrays for 431 breast cancer patients were used to compare different approaches to the assessment of HER2 status. The cores were scored visually and with the Slidepath Tissue IA system, using the NICE-approved scoring system for the HercepTest, as well as by FISH. The agreement between visual and image analysis of HER2 IHC was excellent [interclass correlation coefficient (ICCC) = 0.95, r(s) = 0.90, r = 0.91, k = 0.81, and P < 0.001]. The agreement of HER2 FISH with visual and image analysis of HER2 IHC was also excellent (ICCC = 0.95 and ICCC = 0.92, respectively). Univariate survival analysis showed equivalent associations of visual and image analysis of HER2 and HER2 FISH with both recurrence-free survival (all P < 0.01) and cancer-specific survival (all P < 0.05) in patients with invasive ductal breast cancer. Conclusions: Computerized image analysis of HER2 IHC gives results comparable to those obtained with visual assessment, with possible advantages in diagnostic pathology.