Conservation of pyramidal tract in radiosurgery for brain metastases of lung adenocarcinoma: Three-dimensional analysis of biologically effective dose.

Abstract

Gamma knife surgery (GKS) for brain metastases (BMs) adjacent to the pyramidal tract (PT) is still a challenge to conduct. PT visualization and biologically effective dose (BED) calculation on a voxel-by-voxel basis may provide data to establish clinically safe values. We aimed to assess the relationship of parameters extracted from the BED-volume histogram with outcomes of PT after GKS-treating target (adjacent BM of lung adenocarcinoma). We formed BED-volume histograms for 672 BMs in a retrospective cohort, using 3-dimensional (3D) coordinate values of PT, target, and each iso-centre to calculate the 3D BED distribution in a 200 × 200 × 200 matrix. PT conservation failure (PTCF) was judged clinically and radiologically and classified as lesion progression and radionecrosis. Cox proportional hazards models were used to analyse 3D BED parameters. Internal validation of models was performed by bootstrapping. There were 116 (17.3 %) subjects with PTCF in the cohort, of which 74 (11.0 %) and 42 (6.3 %) were caused by lesion progression and radionecrosis, respectively. Multivariate analysis showed that D_{Lesion_min} BED and D_{Lesion_90%} BED significantly predicted lesion progression (P <.001). D_{PT_Max} BED and V_{PT_ BED40} significantly predicted radionecrosis (P <.001). The model predicting PTCF showed fair discrimination and calibration of D_{Lesion_min} BED + D_{Lesion_90%} BED and D_{PT_Max} BED + V_{PT_ BED40}. The conservation of PT in GKS for BMs of lung adenocarcinoma depends on the combination of PT-tolerated BED and target effective control BED. Therefore, a BED-volume histogram with a 3D BED algorithm is proposed to assess plan quality.