Abstract
Signal counting in the HER2-FISH test is carried out manually in most institutions. Although it is critical in breast cancer treatment, interobserver reproducibility is inevitable. In addition, manual counting requires hard labor from examiners, which might produce false data. To avoid them, we adopt an automated counting device for the HER2-FISH test, which has been approved by FDA. This system is expected to be superior to manual counting in terms of objectivity and reproducibility. However, its applicability is not well documented. From April 2014 to March 2015, we collected 198 breast cancer specimens, including both needle biopsy and mastectomy specimens, and carried out the HER2-FISH test. We examined the applicability of our automatic counting system according to specimen sample type (needle biopsy vs mastectomy), formalin fixation time (16–27 h vs 64–75 h vs 88–99 h) and histological type of breast cancer (papillotubular carcinoma vs solid-tubular carcinoma vs scirrhous carcinoma). In this study, 142 cases (71.7%) were evaluated automatically and 198 cases (100.0%) were evaluated manually. Biopsy specimens (86/110 cases, 78.2%) demonstrated significantly higher performance than mastectomy specimens (56/88 cases, 63.6%). On the other hand, both formalin fixation time and histological type did not show statistically significant differences between biopsy and mastectomy specimens in the automatic counting system. Our data suggest that our automated counting system could be a useful tool for the HER2-FISH test in both biopsy and mastectomy specimens. Formalin fixation time and histological type of breast cancer had no effect on automatic counting.
