Effect on patient care of repeat pathologic review for breast cancer.
6599 Background: Many cancer centers routinely re-review outside pathologic specimens. We hypothesized re-review rarely changes patient treatment plans. Methods: Of 1495 patients seen at the University of Virginia with a diagnosis of breast cancer from 2006-2011, the 276 cases with both internal and outside pathology reports comprised the study cohort. Interobserver agreement (kappa coefficient, K) between internal and outside diagnoses were calculated for histopathology, lymph node, margin, ER/PR, and HER2 status. We then evaluated if the change would result in a change in therapy or surveillance per the National Comprehensive Cancer Network (NCCN) guidelines. The effect of region and teaching affiliation of outside institutions was explored. Results: For the 276 cases with re-reviewed pathology at UVA there was absolute agreement for ER/PR and surgical margins, and excellent agreement for lymph node, K= 0.93, and histopathology, K=0.93. Agreement was good for HER2, K=0.83. 3 cases were changed from HER2 positive to negative (2) or intermediate (1). Of 9 changes in histopathology, 2 had a major upgrade: 1 ADH to DCIS; 1 DCIS to carcinoma. 3 had a major downgrade: 2 from DCIS to ADH; 1 from carcinoma to DCIS. 2 cases changed from ALH to LCIS. Lymph node status was changed from positive to negative in one out of 31 reviewed cases. Treatment plan would have changed for all 13, 4.7% of all patients. Changes were made almost exclusively (11/13) if referred from a hospital with no or minor teaching affiliation, including all major histopathology changes and changes in lymph node and HER2 status. Conclusions: Interobserver agreement for breast pathology between pathologists at an NCI designated cancer center and outside institutions was good. However, 4.7% of women had discordant results that would lead to a change in their care. Changes were most common for noninvasive carcinoma and benign atypia. In order to best utilize resources, referral centers may want to consider limiting re-review to the pathology from centers with high risk for discordance.