Twenty-four patients met eligibility criteria for inclusion. The list of all patients with medulloblastoma treated with two or more courses of radiotherapy is summarized in table 1. All patients received upfront radiotherapy as part of the curative-intent first radiotherapy, either craniospinal radiotherapy followed by posterior fossa boost in 20 patients (12 SR and 8 HR) or focal posterior fossa radiation in 4 infants. Re-irradiation timing per risk is described in table 2. Two patients received a third course of radiation at 3.4 and 1.8 years after RT2 (SR=1, HR=1). Except for 2, all patients received additional treatment after RT2 as low dose chemotherapy, anti VEGF or multiagent chemotherapy combination with antiangiogenic effect.
The re-irradiation followed surgery and/or chemotherapy in 20 patients. The four patients who were treated by radiotherapy immediately after recurrence, 3 had an isolated spinal cord relapse and 1 an isolated ST lesion (infant=3 and HR=1). Among the 9 patients (SR=5, HR=3, Infant=1) first treated with surgery (biopsy=1, STR=4, GTR=4), 3 (HR) patients with isolated recurrence (ST=2, SC=1) did not received chemotherapy prior to RT2. Among the 15 patients who did not undergo surgery at time of relapse, 11 received chemotherapy prior to RT2. The second course of radiation consisted of CSI in 15 (infant=4, SR=7, HR=4) and focal in 9 (SR=5, HR=4). One patient did not complete the focal re-irradiation course due to a disseminated systemic infection by herpes zoster.
Twelve patients received 2 courses of CSI as part of RT1 and RT2 or RT3. The dose of CSI at RT2/RT3 for SR and HR was 21.6Gy, except for one patient (SR) who received 19.8Gy. The median time between two courses of CSI (RT1-RT2 or RT1-RT3) was 3.3 years (range 0.7-5 years), while the median time between CSI and focal radiotherapy was 2.2 years (range 1-4.8 years). The median cumulative maximum dose applied to the spinal cord for the subgroup who received two courses of CSI was 52.2Gy (range 43.2-59.4), being 45Gy for SR and 57.6 for the HR. Radiotherapy details are described in table 3 and supplementary table 1.
The overall PFS rates for the SR and HR group were 28% at 3 years and 0% at 5 years compared with the OS rates at 3 years of 80% and 35% at 5 years (Figure 1, a). The 3-year post-initial failure OS was 58% for the subgroup who received 2 courses of CSI while 0% for the subgroup with focal re-irradiation (p=0.001 HR 0.1082 95% CI [0.02867-0.4085]) (Figure 1, b). Median OS from RT2 for patients who received 2 courses of CSI was 19.3 months, while patients who received a focal RT2 was 5 months. The 3-year post-re-RT PFS (31% vs. 0%; p=0.0005 HR 0.0938 95% CI [0.02476-0.3561]) and OS (25% vs. 0%; p=0.0003 HR 0.0792 95% CI [0.02023-0.3104]) were significantly improved for patients who received CSI as RT2 compared to patients who received focal re-irradiation (Figure 1, c and d). Cumulative radiotherapy dose and overall survival correlation are described in table 3. Survival outcome is summarized in supplementary table 2. Apart from re-CSI, no other statistically significant predictor of survival improvement was identified including risk at diagnosis, IV topotecan, chemotherapy prior to RT2 or anti VEGF. Moreover, surgery at relapse did not offer an overall survival benefit in the present study (p=0,52) (Supplementary table 3).
All 4 infants medulloblastoma were excluded from above survival analysis considering the difference between upfront radiation field (focal vs. CSI) and the distinctive biology. Median time of survival from diagnosis was 6.15 years. Median OS from recurrence for the infant subgroup was 4.35 years while the median OS from RT2 was 3.45 years. The 3-year post-initial failure OS was 100% while 3-year post-re-RT PFS and OS were 66%.
At the time of the analysis, 6 patients were alive; 5 (infant=2, HR=3) with no evidence of disease and 1 with tumor progression (SR). All of them had received CSI as part of RT2/RT3. The median follow-up for the alive patients was 7 years (range 5.5-11.2). Details of the survivors are included in supplemental table 5
Toxicity
Toxicities were identified through a retrospective chart review. No patients discontinued therapy due to adverse effects suggesting acceptable tolerability. Haematology toxicities without transfusion requirement were observed in 3 patients who received CSI as RT2 (infant=1, SR=1, HR=1); CTCAE version 4.03, grade 2 and grade 3. Other toxicities included asymptomatic hypothyroidism in 5 patients diagnosed after the second course of radiation. No other adverse events were reported, specifically no symptomatic intratumoral haemorrhagic events or symptomatic radionecrosis were observed in our cohort. None of them required hospitalization during the re-irradiation period. Three asymptomatic radionecrosis were found at 3 and 6 months after RT2. Only 1 patient received a high dose of dexamethasone at the time of initial RT2 and was able to taper it upon 2 weeks of initiation of therapy.
Assessment of intellectual functioning
Assessment of neurocognitive outcome was performed in the 6 alive patients with a median time from RT1 of 6.26 years (range 4.45-10.13) and from RT2 of 4.22 (range 0.49-6.7). All the survivors fell within mild to moderate intellectual disability range in the Wechsler Intelligence Scale, with a median IQ at last assessment of 58 (range 43-69) (Figure 2 and supplementary table 4).