Pathologic Grading of Malignant Pleural Mesothelioma: A REAL Evidence-Based Proposal

Background: A pathologic grading system (PGS) in malignant pleural mesothelioma (MPM) could be clinically warranted to better identify different risk categories of patients, plan therapy options and activate clinical trials. Design: A cohort of 328 MPM patients was raised between October 1980 and June 2015. All original slides were jointly reviewed for consistency, blindly to asbestos exposure, overall survival, staging and (neo)-adjuvant chemotherapy. Histologic scoring was constructed by attributing to each parameter, independent upon multivariate analysis, different scores based on 50% increments of the corresponding hazard ratios (HR). Accordingly, final scores ranged from 0 to 12 points for each tumor patient. Results: Histology (epithelioid, biphasic, sarcomatoid), necrosis (absent v. present), cell atypia (mild, moderate, severe), mitotic count per 1 mm2 (cut-offs: 1-2, 3-5, 6-9, 10 or more) and Ki-67 labeling index on 2000 cells (cut-off 30%) were independent factors of survival after adjusting for confounding factors (stage, age and chemotherapy). Tumor patterns in epithelioid MGM or type of material (biopsy vs. resection) did not affect survival. PGS (AUC-ROC: 0.79) outperformed mitotic count (AUC-ROC: 0.68) and Ki-67 (AUC-ROC: 0.69). Patient survival progressively deteriorated from score 0 (median: 79.2 mo.) to score 12 (median: 1.3 mo.), with median survival values being 26.6 mo. (CI: 21.6-42.6), 15.1 mo. (CI: 13.3-16.8), 8.8 mo. (CI: 6.8-10.9) and 3.9 mo. (CI: 3.3-4.9) for 0-2, 3-5, 6-7, and 8-12 score, respectively. This PGS was effective not only in MPM considered as a whole, but also within epithelioid, biphasic and sarcomatoid subgroups, with HR values being 1.34 (CI: 1.27-1.41), 1.36 (CI: 1.27- 1.46), 1.29 (1.12-1.49) and 1.44 (1.16-1.79) for each point of increase, respectively. Effectiveness of this PGS was then confirmed in an independent validation set dealing with further 60 MPM patients. Conclusions: The combination of multiple parameters outperformed each single variable to construct a simple PGS, which predicted survival in diversely featuring MPM even at the level of an individual patient’s cancer