Optimized Atlas-Based Auto-Segmentation of Bony Structures from Whole-Body Computed Tomography

Purpose: To develop and test a method for fully automated segmentation of bony structures from whole-body computed tomography (CT) and evaluate its performance compared with manual segmentation.Methods and Materials: We developed a workflow for automatic whole-body bone segmentation using atlas-based segmentation (ABS) method with a postprocessing module (ABSPP) in MIM MAESTRO software. Fifty-two CT scans comprised the training set to build the atlas library, and 29 CT scans comprised the test set. To validate the workflow, we compared Dice similarity coefficient (DSC), mean distance to agreement, and relative volume errors between ABS(PP) and ABS with no postprocessing (ABS(NPP)) with manual segmentation as the reference (gold standard).Results: The ABS(PP) method resulted in significantly improved segmentation accuracy (DSC range, 0.85-0.98) compared with the ABS(NPP) method (DSC range, 0.55-0.87; P < .001). Mean distance to agreement results also indicated high agreement between ABS(PP) and manual reference delineations (range, 0.11-1.56 mm), which was significantly improved compared with ABS(NPP) (range, 1.00-2.34 mm) for the majority of tested bony structures. Relative volume errors were also significantly lower for ABS(PP) compared with ABSNPP for most bony structures.Conclusions: We developed a fully automated MIM workflow for bony structure segmentation from whole-body CT, which exhibited high accuracy compared with manual delineation. The integrated postprocessing module significantly improved workflow performance. (c) 2023 American Society for Radiation Oncology. Published by Elsevier Inc. All rights reserved.

Automated summary

This study developed an atlas-based method for fully automated segmentation of bony structures from whole-body CT and evaluated its performance compared to manual segmentation. The results showed that the segmentation method with a postprocessing module had higher accuracy and agreement compared to the method without postprocessing, with lower relative volume errors.