Dosimetric comparison of the BNCT treatment planning performances when using a nnU-NET to automatically segment Glioblastoma Multiforme

Introduction: This work presents a preliminary evaluation of the use of the convolutional neural network nnU-NET to automatically contour the volume of Glioblastoma Multiforme in CT images. The goal is to assist the preparation of the Treatment Planning of patients who undergo Boron Neutron Capture Therapy (BNCT). BNCT is a binary form of radiotherapy based on the selective loading of a suitable 10-boron concentration into the tumor and on subsequent low-energy neutron irradiation. Matherials and methods: In this work a nnU-NET was employed to automatically contour the Glioblastoma tumors in a series of CT scans. The 16 medical images studied were obtained from a public database called The Cancer Imaging Archive, all of them were labeled by a radiologist. To evaluate the results different indexes commonly used in segmentation problems were employed. Results: To assess a meaningful evaluation of the nnU-NET performance for BNCT this work analyzed the difference of the clinical dosimetry in 16 patients using the manual and the automatic contoured images. Results show that the NN performs well in assisting BNCT dose calculation. Conclusions: The aim of this work was to assess the suitability of AI algorithms in contouring medical images for BNCT treatment of Glioblastoma Multiforme (GBM). The comparison between manually segmented CT images of GBM for dosimetry purposes and those segmented with the assistance of AI provides valuable insights into the effectiveness and reliability of automated segmentation techniques in radiotherapy planning.