ID #1120 Identification of patients with secondary malignancies after multimodal treatment for pediatric brain tumors

Background Secondary malignancies following multimodal oncological treatment are rare but represent a severe long-term complication, particularly in pediatric patients. Systematic analyses in pediatric neuro-oncology are currently lacking. Methods Based on the complete institutional pediatric neuro-oncology database, patients who developed a secondary malignancy during follow-up were identified. Extracranial secondary malignancies and meningiomas were detected using assigned ICD codes (C diagnoses excluding C71 or D32), extracted via an algorithm-based approach. Intracerebral secondary malignancies were identified through an algorithm-driven analysis of all tumor board decisions, using both keyword-based text search and a large language model. Clinical data were retrieved from the hospital patient information system. Results Among a total of 1823 patients in the pediatric neuro-oncology database, 28 patients with 34 secondary malignancies were identified. The most frequent primary tumors were medulloblastoma (n = 13), ependymoma (n = 3), and atypical teratoid/rhabdoid tumor (ATRT; n = 3). Most patients (n = 21) had been treated with craniospinal irradiation, there were only 5 patients with no documented radiotherapy. Likewise, only 7 patients had not undergone chemotherapy for the primary tumor. The most common secondary malignancies were high-grade gliomas (n = 9), meningiomas (n = 7), leukemias (n = 6), and thyroid carcinomas (n = 4). Latency periods varied substantially, ranging from 6.2 years for leukemias to 17.2 years for meningiomas. While thyroid carcinomas and meningiomas showed excellent prognosis, survival in patients with secondary high-grade gliomas was severely limited. Conclusions The applied information technology–based methods successfully identified patients with secondary malignancies within a comprehensive pediatric neuro-oncology database. The observed tumor entities correspond to expected patterns, with primary tumors amenable to curative multimodal therapy and typical secondary malignancies. Further analyses focusing on clinical risk factors and molecular characteristics of neuro-oncological secondary malignancies are planned.