LMAP-20 EVIDENCE-BASED MULTIVARIATE NORMAL TISSUE COMPLICATION PROBABILITY (NTCP) STUDY OF DOMAIN-SPECIFIC COGNITIVE DECLINE AFTER PARTIAL BRAIN RT

Abstract Beyond the hippocampus, there are no evidence-based dose constraints for eloquent structures which subserve memory and attention/processing speed. We performed a multivariate normal tissue complication probability analysis of post-RT neurocognitive decline, examining dosimetric predictors brain regions in prospective longitudinal clinical trial. 78 patients with primary tumors receiving fractionated RT completed comprehensive evaluation high-resolution volumetric diffusion MRI at baseline 6 months post-RT. Image processing using validated automated segmentation parcellated individual white matter (WM) tracts, cortical regions, hippocampi. Neurocognitive test performance was scored reliable change indices adjusted practice effects: negative scores indicate decline. Univariate logistic regression model building were performed. On univariate mean max to multiple corpus callosum (CC) correlated speed decline including Dmax anterior CC (p=0.011) central (p=0.010), Dmax/Dmean mid-anterior (p=0.006 0.010). Dmean left fornix associated (p=0.023), as increasing age chemotherapy. attention, bootstrapped showed most frequently selected variable CC. Performance nested cross-validation by AUC 0.80 (0.75-0.84); LASSO 0.76 (0.72-0.81) being frequent variable. The top five important variables Random Forest CC, all WM, combined posterior Model 0.66 (0.60-0.71). present first, our knowledge, NTCP speed, beyond hippocampus. Dose several ROIs significantly impairment. These data can guide future cognitive-sparing strategies RT.