Validity and reproducibility of histologic diagnosis and grading for adult soft-tissue sarcomas

Abstract

Soft-tissue sarcomas in adults show great variations in histologic type and grade. A valid and reproducible prognostication system is needed to select patients with soft-tissue sarcomas who could benefit from adjuvant chemotherapy. This study was conducted to assess the validity and reproducibility of diagnosis of the histologic type, MIB-1 grade, and mitosis grade, as well as of the 3 components of these grading systems. MIB-1 grade is a recently proposed grading system for predicting the prognosis of patients with adult soft-tissue sarcomas on the basis of 3 criteria (tumor differentiation, necrosis, and MIB-1 score) and replaces the mitotic count in the French system with MIB-1 immunohistochemical staining. Four surgical pathologists from 4 institutions who had experience in diagnostic soft-tissue tumor pathology reviewed 130 cases of soft-tissue sarcoma and independently determined histologic type and grade. The validity of histologic diagnosis was measured by sensitivity and specificity, and that of grading was measured by κ statistics and percentage agreement with the diagnosis of the expert panel at the National Cancer Center, which was defined as a gold standard. Interobserver reproducibility was measured by κ and by percentage agreement between the diagnoses of the 4 pathologists. The validity of the diagnosis of histologic type was high for synovial sarcoma, small round-cell sarcoma, and liposarcoma (sensitivity 89% to 100%; specificity, 98% to 100%) but low for malignant fibrous histiocytoma (MFH) and spindle-cell sarcoma (73% to 75%; 93% to 95%). For the grading, the validity of the MIB-1 grade was substantial (κ = 0.68; agreement, 79%) and higher than that of the mitosis grade (0.54; 69%). The most valid component was tumor differentiation (κ = 0.79), followed by tumor necrosis (0.66), MIB-1 score (0.59), and mitotic score (0.37). Interobserver reproducibility of histologic diagnosis was high for small round-cell sarcoma, synovial sarcoma, and liposarcoma (κ = 0.92, 0.90, 0.87; percentage agreement = 99%, 97%, 96%, respectively); for grading, reproducibility was highest for tumor differentiation (0.78; 87%) and second highest for MIB-1 grade (0.68; 79%). We conclude that diagnosis of the type of soft-tissue sarcoma for synovial sarcoma, small round-cell sarcoma, and liposarcoma and the MIB-1 grading system based on tumor differentiation are highly valid and reproducible among Japanese pathologists who are familiar with the grading system, whereas re-evaluation of histologic criteria is essential for other histologic types such as MFH and spindle-cell sarcoma. HUM PATHOL 33:111-115. Copyright © 2002 by W.B. Saunders Company