2571 - Clinical Target Volumes for Glioma with Automatic Integration of Tumor Infiltration Pathways

Purpose/Objective(s): Defining radiation target volumes with accurate integration of the neuroanatomy is one of the major difficulties in designing treatments for glioma patients. We have developed a deep learning network for normal brain anatomy that learns neural connections between brain substructures that may serve as tumor infiltration pathways. The brain substructure predictions are applied to an automatic workflow that defines the gross tumor volume (GTV) to clinical target volume (CTV) expansion.

Materials/Methods: Two radiation oncologists delineated six brain substructures (hemispheres, brainstem, cerebellum, optic chiasm, optic nerves, ventricles, and midline) for 65 glioma patients. Expert knowledge of white matter tracts was incorporated into the delineations by overlapping structures at the location of the neuronal connections (e.g. cerebellum and brainstem fiber crossing at the cerebellar peduncles). NnU-Net was used for model training (52 patients for training, 13 for testing) with CT image as input and segmentation map as output. Automatic CTVs were generated for all patients by expanding manually-contoured GTVs by 15 mm using a shortest-path solver that constrains the expansion so that it cannot leak outside the brain tissue, using predicted brain substructures as no-flux boundary conditions. Ground truth CTVs were defined as a 15 mm expansion of the GTV adjusted for manually-contoured brain substructures.

Results: Table 1 reports the similarity scores for 13 patients in the test set for the hemispheres, brainstem, cerebellum, optic chiasm, optic nerves, and ventricles. The automatic CTVs generated showed excellent similarity to the ground truth CTVs with mean (± std) DSC, Surface DSC (with 2 mm tolerance) and HD95 scores of 96.90% ± 1.37%, 92.23% ± 6.21% and 2.38 mm ± 1.04 mm, respectively.

Conclusion: We have successfully developed an automated workflow for radiation target volume definition that incorporates the prior knowledge of neuronal connections that can serve as pathways for tumor infiltration. The method is designed to be generalizable to the latest European and American consensus/trial guidelines, and flexible to accommodate for differences in delineation practice between low and high-grade glioma.