Problem
Research questions and friction points this paper is trying to address.
cancer lesion segmentation
total tumour volume
user-prompted priors
whole-body CT
ground-truth generation
Effect of User-Prompted Priors on Semi-Automated Cancer Lesion Segmentation in Whole-Body Computed Tomography
🤖 AI Summary
Manual segmentation of tumor lesions in whole-body CT scans for total tumor volume (TTV) quantification is time-consuming and heavily reliant on expert expertise, while conventional RECIST criteria exhibit limited correlation with patient survival. This work proposes a semi-automatic segmentation method leveraging multiplanar orthogonal user-provided contour priors and presents the first systematic evaluation of how axial, coronal, and sagittal plane contours influence segmentation performance as spatial priors. Evaluated on an external test set comprising 3,865 lesions, the proposed approach achieves a Dice score of 0.882, significantly outperforming a no-prior baseline (0.671). These results demonstrate the efficacy and practical utility of multiplanar user prompts in generating high-quality volumetric tumor annotations.
📝 Abstract
In clinical oncology studies, metastatic cancer is commonly evaluated using "Response Evaluation Criteria in Solid Tumors" (RECIST), in which the diameter of up to five lesions is measured and followed over the course of treatment. However, RECIST shows limited correlation with overall survival. Total tumour volume (TTV) is a stronger predictor but typically relies on manual ground-truth segmentation of all lesions, which is time-consuming and requires expert domain knowledge. Semi-automated approaches leveraging user-prompted priors, such as bounding boxes and single-slice contours, as inputs to automated segmentation methods can facilitate the generation of ground-truth segmentations. This work investigates the impact of different user-prompted priors on semi-automated cancer lesion segmentation performance in whole-body computed tomography. Across 3-fold cross-validation and external testing, more complex spatial priors consistently improved performance, with contour priors from three orthogonal planes (axial, coronal and sagittal) achieving the best results. On the external test (n=3865 lesions), this approach achieved a mean Dice score of 0.882, compared to a mean Dice score of 0.671 for the baseline model with no spatial prior. These findings suggest that the use of multi-plane orthogonal user-prompted priors can improve semi-automated tumour lesion segmentation and support efficient generation of high-quality volumetric ground-truth data.
Innovation
Methods, ideas, or system contributions that make the work stand out.
user-prompted priors
semi-automated segmentation
multi-plane orthogonal contours
total tumour volume
whole-body CT
🔎 Similar Papers
No similar papers found.
💼 Related Jobs
