Modeling the Risk of Neurocognitive Effects from Radiation Therapy in Childhood Cancer Survivors: Initial Results From the Pediatric Normal Tissue Effects in the Clinic (PENTEC) CNS Task Force

Abstract

Brain injury from radiation therapy (RT) includes necrosis and neurocognitive (NC) impairment that can affect childhood cancer survivors’ quality-of-life. PENTEC seeks to quantitate RT dose-dependence of adverse effects and the risk modification by age and other factors. The PENTEC CNS task force reports preliminary results of RT related NC outcomes. A literature search for studies reporting incidence of RT associated NC sequelae (impaired IQ, NC function, or need for special education) was performed. Prescribed doses were converted to equivalent 2 Gy/fraction (EQD2). The Lyman-Kutcher-Burman formalism with generalized equivalent uniform dose (gEUD) was used for modeling normal tissue complication probability (NTCP) of NC effects. For whole brain (WB)/craniospinal or partial brain (PB) RT it was assumed that the prescribed dose uniformly irradiated the entire brain or 1/8 of the brain, respectively. In this analysis, a simple composite variable reflecting various NC endpoints was used. The model was fitted using a weighted least squares method and corrected for the patient number in each study. Best-fit parameter confidence intervals were determined using a bootstrap resampling technique. The search identified 1173 abstracts; each was evaluated by 2 reviewers. Two hundred nineteen studies with useful information were short-listed. The 62 highest scored articles were abstracted and the 19 studies (21 data sets, 1498 patients) with objective NC testing and RT data were analyzed (Table). Preliminary results show a sigmoidal dose-response for NC impairments: D50 corresponding to 50% risk of complication of 26.6 Gy (68% CI 25.5-27.9 Gy); slope of dose response 0.35 (0.31-0.40) and volume parameter 0.46 (0.40-0.59). The model indicates a 5% NC risk associated with 11.3 Gy WB and 29.8 Gy PB RT.Abstract 1032; TableStudyYearPtsWB/CS Dose (Gy)Focal Dose (Gy)Neurocognitive Function DeclineIncidence of Endpoint %Fouladi200331-52.2IQ0Fouladi Focal200538-54IQ < 7023.6Fouladi CSI20051418-3051.9IQ71.4Waber19952718-IQ0Krull201310224-IQ52Vern-Gross201434-59.4IQ0Agbahiwe2017483656IQ54.1Merchant2000823.4-3250.4IQ12.5Merchant200488-23.4-32IQ0Merchant200978-54IQ100Merchant201477-52.13IQ0Merchant20145823.4-39.655.8NC100Edelstein20112418-General50Waber200412518-General0Christopherson20143928.854General49Krull 18 Gy201316718-General12.3Krull 24 Gy201318624-General27Jakacki200461854Status50Cheng20162223.440Process Speed33.3Van Dijk201228521-24.849.3-51.3Disorder14.4Yeom20134118-39.654Special Ed70 Open table in a new tab The initial results of this meta-analysis confirm a RT dose and volume relationship to NC injury risk. Further model refinement will include the effect of age, chemotherapy and include a formal evaluation of heterogeneity across studies.