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
Authors: Sogand Beirami, Zahra Esmaeilzadeh, Ahmed Gomaa, Pluvio Stephan, Ishita Sheth, Thomas Weissmann, Juliane Szkitsak, Philipp Schubert, Yixing Huang, Annette Schwarz, Stefanie Corradini, Florian Putz
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.
Figures & tables
Figure 1: Examples of better performance of the VA dice loss to the baseline
Figure 2: Segmentation masks produced using different configurations on the multi-label dataset. The visualization illustrates the influence of the Volume-Aware Dice loss under various settings. (a) Ground Truth: PT is shown in green, and LN in yellow. (b) Baseline Model: accurate segmentation for LN, but underperformance in PT segmentation. (c) Volume-Aware Dice Loss - Dual Mask Configuration: segmentation quality for both PT (green) and LN (yellow) is well maintained. (d) Volume-Aware Dice Loss - Selective LN Mask Configuration: achieves improved LN segmentation compared to the dual mask setup, while PT performance decreases slightly, yet remains better than in the base model.
Figure 3: Plots showing volumetric and surface distance similarities for primary tumor and lymph node metastases across evaluated configurations.
Metric
Baseline Model
Volume-Aware Dice Loss Dual Mask
Dual Mask p value
Volume-Aware Dice Loss Selective LN Mask
Selective LN Mask p value
PT Dice Score
0.622
0.624
0.949
0.602
0.184
LN Dice Score
0.734
0.747
0.064
0.758
0.019
LN Adaptive Dice Score
0.707
0.719
0.078
0.730
0.020
LN Lesion-Wise Dice Score
0.710
0.722
0.235
0.726
0.224
PT SDS (mm)
0.646
0.645
0.752
0.608
0.018
LN SDS (mm)
0.790
0.806
0.040
0.816
0.014
Table 1: Comparison of segmentation performance across models.
Baseline Model
Volume-Aware Dice Loss Dual Mask
Dual Mask p value
Volume-Aware Dice Loss Selective LN Mask
Selective LN Mask p value
PT Lesion-Wise Binary Precision
0.813
0.820
0.510
0.636
0.000
LN Lesion-Wise Binary Precision
0.736
0.736
0.983
0.760
0.104
PT Lesion-Wise Binary Sensitivity
0.837
0.848
0.298
0.870
0.038
LN Lesion-Wise Binary Sensitivity
0.818
0.835
0.078
0.849
0.010
Table 2: Lesion-wise binary detection precision and sensitivity compared across models.
Figure 4: Plots showing binary detection scores (Sensitivity and Precision) for primary tumor and lymph node metastasis across evaluated configurations.
Figure 5: A comprehensive visualization showcasing all evaluated configurations on a single patient. (a) Ground Truth: PT is displayed in green, and LN in yellow. (b, c, d): (c) achieving the best overall performance, exceeding both the base model (b) and the Volume-Aware Dice loss in selective LN mask approach (d). Arrow in ground truth mask: False positive lymph node metastasis corresponding to an anatomical structure with suspicious imaging appearance.
Medical image segmentation delineates organs, lesions, and other structures in scans to support measurement, diagnosis, treatment planning, and research.
Purpose: Deep-learning-based three-dimensional (3D) dose prediction is widely used in automated radiotherapy workflows. However, most existing models are trained with voxel-wise regression losses, which are poorly aligned with clinical plan evaluation criteria based on dose-volume histogram (DVH) metrics. This study aims to develop a clinically guided loss formulation that directly optimizes clinically used DVH metrics while remaining computationally efficient for head and neck (H&N) dose prediction. Methods: We propose a clinical DVH metric loss (CDM loss) that incorporates differentiable D-metrics and surrogate V-metrics, together with a lossless bit-mask region-of-interest (ROI) encoding to improve training efficiency. The method was evaluated on 174 H&N patients using a temporal split (137 training, 37 testing). Results: Compared with MAE- and DVH-curve based losses, CDM loss substantially improved target coverage and satisfied all clinical constraints. Using a standard 3D U-Net, the PTV Score was reduced from 1.544 (MAE) to 0.491 (MAE + CDM), while OAR sparing remained comparable. Bit-mask encoding reduced training time by 83% and lowered GPU memory usage. Conclusion: Directly optimizing clinically used DVH metrics enables 3D dose predictions that are better aligned with clinical treatment planning criteria than conventional voxel-wise or DVH-curve-based supervision. The proposed CDM loss, combined with efficient ROI bit-mask encoding, provides a practical and scalable framework for H&N dose prediction.
Ruochen Gao, Marius Staring, Frank Dankers
Division of Image Processing, Department of Radiology, Leiden University Medical Center, Leiden, the Netherlands. · Department of Radiation Oncology, Leiden University Medical Center, Leiden, the Netherlands.
Head and neck cancers (HNC) represent a significant global health burden, with accurate tumor delineation being essential for effective radiotherapy planning. The complexity of the oropharyngeal anatomy, combined with the heterogeneous appearance of tumors on imaging, makes manual segmentation time-intensive and subject to inter-observer variability. Beyond segmentation, predicting long-term clinical outcomes, such as recurrence-free survival (RFS), and determining human papillomavirus (HPV) status from noninvasive imaging, remain challenging yet clinically valuable goals. The HECKTOR 2025 challenge addresses these needs by establishing a comprehensive benchmark for automated HNC analysis using multimodal PET/CT imaging and electronic health records. Building on previous editions (2020-2022), this challenge features an expanded multi-institutional dataset comprising over 1,100 patients from 10 centers worldwide. Participants were tasked with three complementary objectives: (1) segmenting primary gross tumor volumes (GTVp) and metastatic lymph nodes (GTVn), (2) predicting recurrence-free survival, and (3) classifying HPV status. The challenge attracted 35 registered teams, with 15 final submissions evaluated on a held-out test set. Top-performing algorithms achieved a mean Dice similarity coefficient of 0.75 for segmentation, a concordance index of 0.66 for survival prediction, and a balanced accuracy of 0.56 for HPV classification. This paper presents a comprehensive analysis of the submitted methodologies, evaluates their performance across different lesion characteristics, and discusses their implications for clinical translation in automated oncology workflows and decision support systems.
Numan Saeed, Salma Hassan, Shahad Hardan +27
Mohamed bin Zayed University of Artificial Intelligence (MBZUAI), Abu Dhabi, UAE. · Amsterdam UMC, Amsterdam, The Netherlands. · The Netherlands Cancer Institute, Amsterdam, The Netherlands. +22
Multimodal MRI provides complementary information for brain tumor segmentation, but high accuracy alone does not reveal whether a model relies on clinically appropriate imaging evidence. Radiologists use different sequences for different findings, such as T1ce for enhancing tumor and FLAIR for peritumoral edema. Yet most 3D segmentation networks fuse MRI sequences through fixed connections, regardless of patient, region, or depth. We frame multimodal segmentation as grounded multimodal reasoning and introduce (i) a model agnostic probing framework combining per region modality ablation, Dominant Sensitivity Enhancement (DSE), and gate trajectory analysis to reveal sequence reliance, and (ii) evidence that adaptive Hyper-Connections (HC) induce clinically aligned modality reliance without modality supervision. Across nnU Net, SwinUNETR, and VT-UNet on BraTS 2021, most HC configurations increase reliance on clinically appropriate sequences. This interpretable pattern recurs across all three backbones, suggesting it arises from adaptive connections rather than a single architecture. HC also improves segmentation accuracy at negligible parameter cost. On validation, the best HC configuration increases mean Dice by 0.43 to 0.81 percentage points across the three backbones. On the held-out test set, it reduces mean HD95 by 1.17 to 1.78 mm (approximately 20 to 27 percent) for nnU-Net and VT UNet, adding at most 0.53 percent parameters. Our framework offers an inspectable measure of clinical grounding and extends naturally to multimodal medical foundation models.