Quantcast

Prognostic limitations of the Daumas-Duport grading scheme in infratentorial neuroglial tumors in children.

Research paper by Floyd H FH Gilles, C Jane CJ Tavaré, Alan A Leviton, E Tessa ET Hedley-Whyte, Anna A Sotrel, Lester L Adelman, Richard L RL Davis, Lucy B LB Rorke,

Indexed on: 03 Mar '04Published on: 03 Mar '04Published in: Pediatric and Developmental Pathology



Abstract

The Daumas-Duport grading scheme (DDGS) utilizes four histologic features in an additive method (grade 1 if none present, grade 2 if only one is present, etc.). Its efficacy in achieving prognostically homogeneous groups of childhood infratentorial neuroglial tumors and its concordance with World Health Organization (WHO) diagnoses has not been evaluated. We investigated these questions using the Childhood Brain Tumor Consortium (CBTC) database of 1241 neuroglial tumors limited to the infratentorial compartment. We calculated survival function estimates for various DDGS grades as well as the histologic features within each grade. The feature of endothelial prominence improved survival expectation, whereas the remaining three features of nuclear atypia, mitoses, and necrosis were associated with worsened survival. Survival estimates for tumors with DDGS grades 2 and 3 did not differ. Some grades contained feature subsets with significantly different survival distributions. The survival distributions of DDGS grade 1, DDGS grade 2 with only endothelial prominence, and DDGS grade 3 with nuclear atypia and endothelial prominence were not significantly different. DDGS grade within WHO diagnoses had no significant effect on survival expectation. We conclude that grading by summation of only four histologic features, as in the DDGS, is inappropriate for assessment of childhood neuroglial tumors. A classification scheme considering the complete histologic content is more likely to provide clinically useful diagnoses. Such a scheme, based on the CBTC database is available. This scheme uses 26 histologic features identified as reliable in read-reread studies.