Preliminary Report of Secondary Malignant Neoplasm Incidence in Pediatric Patients Receiving Proton versus Photon Craniospinal Irradiation for Medulloblastoma

Abstract

Modeling studies have predicted fewer secondary malignant neoplasms (SMN) following treatment with proton beam therapy (PBT) compared to photon radiotherapy (XRT). However, there are concerns regarding secondary neutron production and possible associated excess SMN with passive scatter proton therapy (PSPT) compared to pencil beam proton therapy. We reviewed our multi-institutional experience treating pediatric medulloblastoma (MB) patients with craniospinal irradiation (CSI) using either XRT or PBT. We hypothesized that excess SMN would not be observed among patients treated with PBT. From 1996 to 2014, 166 patients with MB received CSI followed by tumor bed boost using either PSPT (n = 103) or XRT (n = 63). For the proton cohort, all patients received PSPT for the CSI portion and 93.9% received PSPT for the boost. For XRT patients, all received 3-D conformal RT CSI followed by an IMRT tumor bed boost. Median age was 8.0 years at time of radiotherapy (RT), (8.0 years PBT, 8.1 years XRT). A majority of patients were male (112/166, 67.5%) and a majority had standard-risk MB (102/166, 61.4%). Median follow-up was 78.5 months for the PBT cohort and 153.2 months for the XRT cohort (p < 0.001). The 5- and 10-year overall survival rates were 82.6% and 81.2% for standard-risk and 69.8% and 64.0% for high-risk patients, respectively. No OS difference was identified by RT modality (p=0.49). The 5- and 10-year actuarial SMN rate was 2.3% and 8.1% for the entire study population (N=166). By modality, 5- and 10-year SMN rates were 2.6% and 6.0% for PBT patients, and 0.0% and 8.0% for XRT patients (p=0.71). There was no difference in incidence of SMN according to age (p=0.99), gender (p=0.50) or CSI dose (p=0.61). There were a total of 8 SMN identified in total (4 in each cohort by RT modality), with a median time from RT completion to SMN diagnosis of 71.2 months (range: 5.8 - 145.6 months). Breaking down SMNs by histology, there were 3 sarcomas, 2 malignant gliomas, 1 papillary thyroid cancer, 1 parotid mucoepidermoid carcinoma, and 1 cerebellar malignant glioneuronal tumor. All 4 SMNs in the PBT cohort occurred either within the previously-irradiated target (n=2), or in the entrance dose region (n=2). For XRT patients, 3 SMN occurred in regions receiving exit dose while 1 was within the target. Our preliminary analysis demonstrates no difference in the rate of SMN between PBT-treated and XRT-treated patients up to 10 years following RT. Given the shorter follow-up time for PBT patients, longer-term follow-up remains necessary to understand differential SMN patterns across RT modalities better. It is notable that XRT-treated patients developed a higher rate of SMNs in regions of exit dose than in the target volume.