Although precise anatomic staging is prognostically important in neuroblastoma, most widely employed staging systems remain incompatible. The International Neuroblastoma Staging System (INSS) was formulated to incorporate the basic elements of several systems to and define the significance of tumor resectability, anatomic "midline," and lymph node involvement. The authors sought to determine the applicability and value of the INSS compared with the classic Evans system. Between 1980 and 1992, 424 children with the diagnosis of local or regional neuroblastoma were entered in Childrens Cancer Group (CCG) clinical trials. The patients were assigned to Evans stage I, II, or III, by clinical and surgicopathologic assessment, and were treated uniformly by Group-wide therapy protocols. INSS stage 1, 2A, 2B, or 3, was applied, by retrospective analysis, to the children in the earlier studies, and by prospective evaluation of recent patients in the current studies. Survival and relapse-free survival (RFS) rates were determined and compared, based on age at diagnosis, extent of resection, and staging reassignment. All 87 Evans stage I patients were classified as INSS stage 1 and had a 92% 3-year RFS rate. Of the 144 Evans stage II patients, 65 also qualified as INSS stage 1 patients, with an 82% RFS rate. The other 79 stage II children remained in INSS stage 2A or 2B and had a 70% RFS rate (P = .10). Of the 193 Evans stage III patients, 24 were reassigned to INSS stage 1 (85% RFS rate) and 33 to stage 2A or 2B (65% survival rate; 61% RFS rate).(ABSTRACT TRUNCATED AT 250 WORDS)