Abstract
Cardiac substructures are essential in thoracic radiation therapy planning to minimize risk of radiation-induced heart disease. Deep learning (DL) offers efficient methods to reduce contouring burden but lacks generalizability across different modalities and overlapping structures. This work introduces and validates a Modality-AGnostic Image Cascade (MAGIC) for comprehensive and multi-modal cardiac substructure segmentation. MAGIC is implemented through replicated encoding and decoding branches of an nnU-Net-based, U-shaped backbone conserving the function of a single model. Twenty cardiac substructures (heart, chambers, great vessels (GVs), valves, coronary arteries (CAs), and conduction nodes) from simulation CT (Sim-CT), low-field MR-Linac, and cardiac CT angiography (CCTA) modalities were manually delineated and used to train (n=76), validate (n=15), and test (n=30) MAGIC. Twelve comparison models (four segmentation subgroups across three modalities) were equivalently trained. All methods were compared for training efficiency and against reference contours using the Dice Similarity Coefficient (DSC) and two-tailed Wilcoxon Signed-Rank test (threshold, p<0.05). Average DSC scores were 0.75(0.16) for Sim-CT, 0.68(0.21) for MR-Linac, and 0.80(0.16) for CCTA. MAGIC outperforms the comparison in 57% of cases, with limited statistical differences. MAGIC offers an effective and accurate segmentation solution that is lightweight and capable of segmenting multiple modalities and overlapping structures in a single model. MAGIC further enables clinical implementation by simplifying the computational requirements and offering unparalleled flexibility for clinical settings.
Abstract (translated)
心脏亚结构在胸部放射治疗规划中至关重要，可降低由辐射引起的疾病风险。深度学习（DL）提供有效的方法来减轻轮廓绘制负担，但缺乏跨不同模态和重叠结构的泛化能力。本工作引入并验证了一种无模式感知图像级联（MAGIC），用于全面且多模态的心脏亚结构分割。MAGIC通过复制基于nnU-Net、呈U形架构模型的编码和解码分支来实现，该架构保留了单一模型的功能。 从模拟CT（Sim-CT）、低场MR-Linac和心脏CT血管造影（CCTA）模态中手动勾画出20个心脏亚结构（包括心脏、心室、大血管(GVs)、瓣膜、冠状动脉(CAs)以及传导节点），用于训练(n=76)，验证(n=15)和测试MAGIC(n=30)。同样，十二种对比模型（四种分割子组跨三种模态）被同等培训。所有方法均根据训练效率进行比较，并通过Dice相似系数(DSC)与双尾Wilcoxon符号秩检验（阈值p<0.05）参考轮廓进行比对。 Sim-CT、MR-Linac和CCTA的平均DSC分数分别为0.75(0.16)、0.68(0.21)和0.80(0.16)。在57%的情况下，MAGIC优于对比模型，但统计差异有限。 MAGIC提供了一种有效的且准确的分割解决方案，它是轻量级的，并能够在单一模型中对多种模态及重叠结构进行分割。此外，MAGIC通过简化计算需求和为临床环境提供无与伦比的灵活性而促进了其在临床上的应用。
URL
https://arxiv.org/abs/2506.10797
