Discordance in pathology report after central pathology review: Implications for breast cancer adjuvant treatment

Abstract Aim Pathological predictive factors are the most important markers when selecting early breast cancer adjuvant therapy. In randomized clinical trials the variability in pathology report after central pathology review is noteworthy. We evaluated the discordance rate (DR) and inter-rater agreement between local and central histopathological report and the clinical implication on treatment decision. Methods A retrospective analysis was conducted in a series of consecutive early breast cancer tumors diagnosed by local pathologists and subsequently reviewed at the Pathology Division of European Institute of Oncology. The inter-rater agreement (k) between local and central pathology was calculated for Ki-67, grading, hormone receptors (ER/PgR) and HER2/neu. The Bland–Altman plots were derived to determine discrepancies in Ki-67, ER and PgR. DR was calculated for ER/PgR and HER2. Results From 2007 to 2013, 187 pathology specimens from 10 Cancer Centers were reviewed. Substantial agreement was observed for ER (k0.612; 95% CI, 0538–0.686), PgR (k0.659; 95% CI, 0580–0.737), Ki-67 (k0.609; 95% CI, 0.534–0.684) and grading (k0.669; 95% CI, 0.569–0.769). Moderate agreement was found for HER2 (k0.546; 95% CI, 0444–0.649). DR was 9.5% (negativity to positivity) and 31.7% (positivity to negativity) for HER2 and 26.2% (negativity to positivity) and 12.5% (positivity to negativity) for ER/PgR. According to changes in Her2 and ER/PgR status, 23 (12.2%) and 33 (17.6%) systemic prescription were respectively modified. Conclusions In our retrospective analysis, central pathological review has a significant impact in the decision-making process in early breast cancer, as shown in clinical trials. Further studies are warranted to confirm these provocative results.