ID #1210 Associations Between Radiation Dose to Critical Brain Structures and Neurocognitive Outcomes in Patients Treated for Medulloblastoma on COG ACNS0331: A Report from the Children’s Oncology Group.
Matthew Ladra, Kristina Hardy, Lan Lin, Yimei Li, Jeff MichalskiAbstract
Background
The COG study ACNS0331 randomized medulloblastoma patients to receive an involved field (IF) vs a whole posterior fossa (PF) boost following craniospinal irradiation (CSI) and children under 8 years of age were randomized to either standard dose CSI to 23.4Gy or a reduced dose CSI to 18Gy.Children aged 3+ participated in neuropsychological testing.
Methods
We identified 128 consecutive patients from ACNS0331 with both neurocognitive testing and complete radiation plan data.CNS regions including whole brain, supratentorial brain (ST), temporal lobes (TL), cerebellum, and brainstem were re-contoured in all patients for consistency.
Results
At T2 (2-3yrs post-dx), estimated IQ decreased by 2.4 points for every 1 Gy increase to the mean ST dose in pts ages 3-7 (p = 0.01), decreased by 1.3 points (ages 3-7) and 1 point (all ages) for every 1 Gy increase to the mean LTL dose (p = 0.01; p = 0.02), and decreased by 1 points for every 4% increase to LTL V30 and 3% increase to LTL V40 in pts of all ages (p = 0.04; p < 0.01).RT dose to CNS structures was examined by modality (3DCRT vs IMRT/VMAT vs Protons). Patients receiving PF boost with 3DCRT had the highest mean ST/LTL doses (32Gy/44Gy) and those receiving IF boost and protons had the lowest mean ST/LTL doses (27Gy/30Gy).Mean LTL V50 was: PF + 3D (40%), PF+IMRT (25%), PF+protons (10%), IF + 3D (18%), IF+IMRT (15%), IF+protons (8%).
Late Breaking Results
T3 (5 year) timepoints have been updated and will be analyzed.Correlations between dose to Hippocampus, Corpus Callosum, Cochlea and neurocognitive outcomes will be presented, which were not included in the 2018 T2 analysis.Supratentorial and Temporal Lobe correlations will be updated.Critical structure dosimetry from proton cohort compared to photon cohort will be presented as well.
Conclusions
Relatively small increases in dose across sensitive CNS regions significantly increased the risk of neurocognitive dysfunction and radiation modality influenced structural doses significantly, independent of randomization.