OBJECTIVE Ovarian germ cell tumors (GCT) show greater histologic and biologic heterogeneity than their testicular counterparts and remain poorly understood. Ploidy analysis was performed on ovarian GCT registered on Pediatric Oncology Group germ cell tumor protocols 9048 and 9049 to distinguish biologically distinct subsets of immature teratomas and malignant ovarian germ cell tumors. METHODS Tumors from 22 patients (mean age 12 years) were analyzed and classified according to the submitting diagnosis; when pure samples of different histologic subtypes within a single tumor were possible, these were analyzed separately. Archival tissue was disaggregated and Feulgen stained; DNA index (DI) was determined by static image analysis utilizing internal normal cells as diploid controls. RESULTS 26 histologic subtypes from 22 patients were analyzed. The tumors of 18 patients were composed of a single histologic subtype according to the submitting institution, including 6 dysgerminomas, 8 immature teratomas (IT), and 4 endodermal sinus tumors (EST). Two tumors contained both IT and EST components that were separately analyzed. Two tumors were classified as mixed germ cell tumors; 1 showed multiple intermingling subtypes unable to be separately analyzed and the second showed three histologic subtypes separately analyzed (IT, EST, embryonal carcinoma). From a total of 15 malignant histologic GCT subtypes in 14 patients, all but 2 demonstrated a DI of 1.4-2.4 (mean 1.85). Two diploid malignant GCT (1 EST, 1 dysgerminoma) were both associated with gonadoblastoma. Overall, 11 IT subtypes were analyzed and 9 were diploid (2 grade 1, 5 grade 2, and 2 grade 3). Two tumors originally submitted and classified as pure IT (grades 2 and 3) were aneuploid with a dominant diploid and a secondary aneuploid peak (both DI 1.7). On central review, both of these tumors demonstrated the presence of subtle patterns of EST that were unrecognized by the submitting institution and were much too small for separate analysis. Analysis of the 3 patients containing sufficient IT and EST to be separately analyzed all showed a diploid IT component and an aneuploid EST component. CONCLUSIONS Analysis of ploidy data suggests that polyploidization is a consistent finding in malignant ovarian GCT arising in normal patients, similar to the data for adult testicular GCT. Immature teratomas in this pediatric population, however, are most commonly diploid, regardless of grade. The development of EST within an IT is associated with the development of an aneuploid clone. Therefore, the finding of such a clone in an IT may be of diagnostic utility, as EST may be difficult to recognize. Last, the development of a malignant GCT in patients with gonadal dysgenesis may be pathogenetically different from those arising in normal patients, in that polyploidization is not required.