Recommendations for modification of terminology of neuroblastic tumors and prognostic significance of Shimada classification. A clinicopathologic study of 213 cases from the pediatric oncology group

To develop consistency in terminology and pathologic criteria, the authors reviewed the literature and 213 cases of neuroblastic tumors (NT) registered with Pediatric Oncology Group (POG) protocols 8104 and 8441. The patients were given standardized therapy stratified according to POG stage and patient age, and four or more histologic sections of primary tumor resected before therapy were available in each of these 213 cases. All stages were represented. The recommended nomenclature combines conventional terms and criteria with those used by Bove and McAdams and Shimada et al. The main features of the recommended nomenclature are as follows: (1) the terms neuroblastoma (NB) and ganglioneuroblastoma (GNB) are retained instead of stroma‐poor NB and stroma‐rich NB, recommended by Shimada et al.; (2) undifferentiated NB is considered a subtype separate from poorly differentiated NB; and (3) the term GNB is used only when there is a predominant ganglioneuromatous component admixed with the minor neuroblastomatous component. With the use of these criteria and terms, the Shimada classification was determined in the 213 cases. The results showed that, even after stratification for age, POG stage, and primary site, there is a statistically significant difference in survival rate between favorable histologic and unfavorable histologic prognostic subgroups. The authors recommend that definitive prognostic categorization of an NT according to Shimada classification should be done only when adequate histologic material is available from a primary tumor resected before any other therapy. Categorization done on histologic material from small biopsy specimens, previously treated primary tumors, or meta‐static sites should be considered tentative.