Recurrence is the most common cause of late mortality in pediatric brain tumor survivors. However, it is unclear how long such patients should be monitored with periodic neuroimaging. Therefore, we investigated the utility of neuroimaging surveillance for recurrence ≥5 years post diagnosis in survivors of pediatric medulloblastoma and ependymoma. We conducted a retrospective study of survivors of medulloblastoma or ependymoma treated between 2000 and 2017. Eligible survivors were disease-free 5 years after diagnosis and underwent magnetic resonance imaging surveillance ≥5 years after diagnosis. Medulloblastoma survivors with a history of recurrence <5 years after diagnosis were excluded. Of 302 children diagnosed in the study period, 129 met inclusion criteria (89 medulloblastoma/40 ependymoma; 77 (59.7%) male; median age at diagnosis 6 years (range < 1-13); median time from diagnosis to last scan 134 months (61-283)). Four medulloblastoma patients had late recurrent disease, one of which was detected on routine neuroimaging (asymptomatic). All medulloblastoma patients with late recurrence died, except for one previously unirradiated patient who was disease-free 29 months after recurrence. Nine ependymoma patients had late recurrence of which 7 were detected on routine neuroimaging. Six out of seven asymptomatic late recurrent ependymoma patients remain alive with a median time after recurrence of 45.5 months (range: 3-121). Both symptomatic patients died. Among ependymoma survivors, asymptomatic detection of late recurrence by surveillance neuroimaging was associated with better survival than symptomatic detection, supporting the continuation of surveillance for at least 10 years after diagnosis. The benefit of prolonged surveillance in medulloblastoma survivors remains uncertain.
Keywords: ependymoma; medulloblastoma; neuroimaging; recurrence; surveillance.
© 2025 UICC.