Improving Deep Learning-Based Target Volume Auto-Delineation for Adaptive MR-Guided Radiotherapy in Head and Neck Cancer: Impact of a Volume-Aware Dice Loss

arXiv cs.CV / 4/14/2026

💬 OpinionIdeas & Deep AnalysisModels & Research

共有:

Key Points

The study proposes integrating a Volume-Aware (VA) Dice loss into a self-configuring deep learning (nnU-Net ResEnc M) framework to improve automatic target auto-delineation for adaptive MR-guided radiotherapy in head and neck cancer.
Using the HNTS-MRG 2024 dataset, the authors compare three loss setups: a standard Dice baseline, a VA “Dual Mask” approach for both primary tumor (PT) and metastatic lymph nodes (LN), and a VA “Selective LN Mask” approach that applies VA loss to LN only.
The Selective LN Mask configuration delivers the best nodal performance, improving LN volumetric Dice (0.758 vs. 0.734) and LN lesion-wise sensitivity (84.93% vs. 81.80%), especially for small, complex nodal metastases.
A key trade-off appears in the selective configuration: PT detection precision drops significantly (63.65% vs. 81.27%), indicating that focusing volume awareness only on LN can harm primary tumor labeling.
Overall, the Dual Mask setup provides the most balanced multi-label segmentation, improving LN sensitivity while maintaining PT precision (82.04% PT precision with improved LN detection to 83.46%).

Abstract

Background: Manual delineation of target volumes in head and neck cancer (HNC) remains a significant bottleneck in radiotherapy planning, characterized by high inter-observer variability and time consumption. This study evaluates the integration of a Volume-Aware (VA) Dice loss function into a self-configuring deep learning framework to enhance the auto-segmentation of primary tumors (PT) and metastatic lymph nodes (LN) for adaptive MR-guided radiotherapy. We investigate how volume-sensitive weighting affects the detection of small, anatomically complex nodal metastases compared to conventional loss functions. Methods: Utilizing the HNTS-MRG 2024 dataset, we implemented an nnU-Net ResEnc M architecture. We conducted a multi-label segmentation task, comparing a standard Dice loss baseline against two Volume-Aware configurations: a "Dual Mask" setup (VA loss on both PT and LN) and a "Selective LN Mask" setup (VA loss on LN only). Evaluation metrics included volumetric Dice scores, surface-based metrics (SDS, MSD, HD95), and lesion-wise binary detection sensitivity and precision. Results: The Selective LN Mask configuration achieved the highest LN Volumetric Dice Score (0.758 vs. 0.734 baseline) and significantly improved LN Lesion-Wise Detection Sensitivity (84.93% vs. 81.80%). However, a critical trade-off was observed; PT detection precision declined significantly in the selective setup (63.65% vs. 81.27%). The Dual Mask configuration provided the most balanced performance across both targets, maintaining primary tumor precision at 82.04% while improving LN sensitivity to 83.46%. Conclusions: A volume-sensitive loss function mitigated the under-representation of small metastatic lesions in HNC. While selective weighting yielded the best nodal detection, a dual-mask approach is required in multi-label tasks to maintain segmentation accuracy for larger primary tumor volumes.