A Large Convolutional Neural Network for Clinical Target and Multi-organ Segmentation in Gynecologic Brachytherapy with Multi-stage Learning

2025-06-01

AI Summary

To address the low segmentation accuracy of clinical target volumes (CTVs) and critical organs (bladder, rectum, uterus) in gynecological brachytherapy (GYN-BT) CT imaging—attributed to poor soft-tissue contrast, large anatomical variability, and scarce annotated data—this study proposes a novel three-stage hierarchical learning framework: (1) self-supervised pretraining using sparse submanifold convolution, (2) multi-organ supervised fine-tuning, and (3) GYN-BT–specific fine-tuning. The resulting 3D network, GynBTNet, significantly enhances robustness for complex boundary segmentation. On a public GYN-BT dataset, it achieves Dice scores of 0.837±0.068 (CTV), 0.940±0.052 (bladder), 0.842±0.070 (rectum), and 0.871±0.047 (uterus). Moreover, GynBTNet outperforms nnU-Net and Swin-UNETR across all Hausdorff distance (HD95) and average surface distance (ASD) metrics.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Calibration & Uncertainty Quantification
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Purpose: Accurate segmentation of clinical target volumes (CTV) and organs-at-risk is crucial for optimizing gynecologic brachytherapy (GYN-BT) treatment planning. However, anatomical variability, low soft-tissue contrast in CT imaging, and limited annotated datasets pose significant challenges. This study presents GynBTNet, a novel multi-stage learning framework designed to enhance segmentation performance through self-supervised pretraining and hierarchical fine-tuning strategies. Methods: GynBTNet employs a three-stage training strategy: (1) self-supervised pretraining on large-scale CT datasets using sparse submanifold convolution to capture robust anatomical representations, (2) supervised fine-tuning on a comprehensive multi-organ segmentation dataset to refine feature extraction, and (3) task-specific fine-tuning on a dedicated GYN-BT dataset to optimize segmentation performance for clinical applications. The model was evaluated against state-of-the-art methods using the Dice Similarity Coefficient (DSC), 95th percentile Hausdorff Distance (HD95), and Average Surface Distance (ASD). Results: Our GynBTNet achieved superior segmentation performance, significantly outperforming nnU-Net and Swin-UNETR. Notably, it yielded a DSC of 0.837 +/- 0.068 for CTV, 0.940 +/- 0.052 for the bladder, 0.842 +/- 0.070 for the rectum, and 0.871 +/- 0.047 for the uterus, with reduced HD95 and ASD compared to baseline models. Self-supervised pretraining led to consistent performance improvements, particularly for structures with complex boundaries. However, segmentation of the sigmoid colon remained challenging, likely due to anatomical ambiguities and inter-patient variability. Statistical significance analysis confirmed that GynBTNet's improvements were significant compared to baseline models.

Problem

• Accurate segmentation of CTV and organs in gynecologic brachytherapy
• Challenges from anatomical variability and low CT soft-tissue contrast
• Limited annotated datasets for multi-organ segmentation in GYN-BT

Innovation

• Multi-stage learning framework for segmentation
• Self-supervised pretraining with sparse convolution
• Hierarchical fine-tuning on specialized datasets