Abstract Purpose During follow-up after radiotherapy for pediatric brain tumors, multiple cerebral microbleeds and/or cavernous malformations (CCMs) are sometimes detected on MRI. Although most cases remain asymptomatic, some lesions increase in number or size, and rarely cause hemorrhage with clinical symptoms. We investigated the characteristics of radiation-induced microbleeds/CCMs in pediatric patients treated at our institution. Patients and Methods We retrospectively reviewed pediatric patients who underwent whole-brain irradiation or whole-ventricular irradiation. Patients were included if they had no microbleeds/CCMs prior to radiotherapy and had a follow-up period of at least two years. A total of 44 patients treated since 2001 were eligible, including 21 patients with germ cell tumors and 23 with embryonal tumors, for whom follow-up imaging was available. Results The age at irradiation ranged from 2 to 17 years (mean, 10 years). Microbleeds/CCMs were identified during follow-up in 26 patients (59%). Multiple microbleeds were observed in 21 patients, and solitary lesions in 5 patients. The mean interval from irradiation to the detection of microbleeds/CCMs was 9.1 years (range, 1–21 years). The mean interval was 8.4 years for whole-brain irradiation (17 patients) and 10.2 years for whole-ventricular irradiation (6 patients). No patients developed clinical symptoms attributable to microbleeds/CCMs. During follow-up, overt intracerebral hemorrhage arising from microbleeds/CCMs occurred in 3 patients (12%), with a mean latency of 5.7 years after irradiation. None of these patients required surgical intervention. Conclusion Radiation-induced microbleeds/CCMs were observed after both whole-brain and whole-ventricular irradiation, with many cases detected more than 10 years after treatment. Although these lesions may increase in number or size and rarely result in hemorrhage, most remain asymptomatic. Therefore, careful long-term follow-up is warranted; however, the optimal duration of surveillance remains a subject of debate.
Miwa et al. (Tue,) studied this question.