Relation of tumor-cell ploidy to survival in children with medulloblastoma

To assess the value of tumor-cell ploidy as a predictor of survival in medulloblastoma. Ploidy determinations were based on the flow-cytometric analysis of cellular DNA content in fresh tumor specimens taken from 34 consecutively treated children with newly diagnosed medulloblastoma. Patients were assigned a high or low risk of failure depending on tumor size and invasiveness, and the presence or absence of metastatic disease. Treatment consisted of radiotherapy, with or without chemotherapy, according to institutional or cooperative group protocols. Univariate analysis of candidate prognostic factors showed that only tumor-cell ploidy and clinical risk group had a statistically significant influence on survival. Patients with hyperdiploid stem lines (n = 9) had significantly longer survival times (P = .04) than did those with diploid lines (n = 20). The estimated 5-year survival probabilities (+/- SE) for these two subgroups were 89% +/- 11% and 48% +/- 13%, respectively. Although clinical risk status (high v low) showed essentially the same predictive strength as ploidy, the two features identified largely nonoverlapping subgroups. Thus, within the clinical high-risk group, it was possible to distinguish hyperdiploid patients whose 5-year survival rate (83% +/- 15%) was comparable to that of patients with localized, low-risk tumors. This prospective study indicates that both ploidy and clinical risk group are important prognostic factors in medulloblastoma. Their combined use at diagnosis would distinguish patients who require more aggressive therapeutic intervention (diploid, clinical high-risk group) from those who could be expected to benefit most from standard treatment.