Critical combinations of radiation dose and volume predict intelligence quotient and academic achievement scores after craniospinal irradiation in children with medulloblastoma

Abstract

Purpose: To prospectively follow children treated with craniospinal irradiation to determine critical combinations of radiation dose and volume that would predict for cognitive effects.

Methods and materials: Between 1996 and 2003, 58 patients (median age 8.14 years, range 3.99-20.11 years) with medulloblastoma received risk-adapted craniospinal irradiation followed by dose-intense chemotherapy and were followed longitudinally with multiple cognitive evaluations (through 5 years after treatment) that included intelligence quotient (estimated intelligence quotient, full-scale, verbal, and performance) and academic achievement (math, reading, spelling) tests. Craniospinal irradiation consisted of 23.4 Gy for average-risk patients (nonmetastatic) and 36-39.6 Gy for high-risk patients (metastatic or residual disease >1.5 cm(2)). The primary site was treated using conformal or intensity modulated radiation therapy using a 2-cm clinical target volume margin. The effect of clinical variables and radiation dose to different brain volumes were modeled to estimate cognitive scores after treatment.

Results: A decline with time for all test scores was observed for the entire cohort. Sex, race, and cerebrospinal fluid shunt status had a significant impact on baseline scores. Age and mean radiation dose to specific brain volumes, including the temporal lobes and hippocampi, had a significant impact on longitudinal scores. Dichotomized dose distributions at 25 Gy, 35 Gy, 45 Gy, and 55 Gy were modeled to show the impact of the high-dose volume on longitudinal test scores. The 50% risk of a below-normal cognitive test score was calculated according to mean dose and dose intervals between 25 Gy and 55 Gy at 10-Gy increments according to brain volume and age.

Conclusions: The ability to predict cognitive outcomes in children with medulloblastoma using dose-effects models for different brain subvolumes will improve treatment planning, guide intervention, and help estimate the value of newer methods of irradiation.