HQColon: A Hybrid Interactive Machine Learning Pipeline for High Quality Colon Labeling and Segmentation

2025-02-28

AI Summary

Existing open-source tools—particularly TotalSegmentator—exhibit low accuracy and high computational cost in colon segmentation, failing to meet the demands of digital twin applications and personalized medicine for high-resolution, clinically reliable annotations. To address this, we propose the first fully automatic, high-precision colon segmentation method for CT colonography (CTC). Our approach comprises: (i) a semi-automatic annotation pipeline integrating region-growing with interactive machine learning; (ii) the first publicly available, high-resolution CTC dataset comprising 435 annotated cases; and (iii) an end-to-end, open-source segmentation model built upon nnU-Net. Quantitative evaluation demonstrates substantial improvements: mean symmetric surface distance of 0.2 mm (a 20× reduction versus TotalSegmentator) and 95% Hausdorff distance of 1.0 mm (an 18× reduction). This work establishes a new state-of-the-art benchmark—delivering unprecedented accuracy, efficiency, and reproducibility for clinical and translational research in colorectal imaging.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Calibration & Uncertainty Quantification
• Humans and AI: Crowd Sourcing and Human Computation

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Security and Privacy: Data transparency and provenance
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

High-resolution colon segmentation is crucial for clinical and research applications, such as digital twins and personalized medicine. However, the leading open-source abdominal segmentation tool, TotalSegmentator, struggles with accuracy for the colon, which has a complex and variable shape, requiring time-intensive labeling. Here, we present the first fully automatic high-resolution colon segmentation method. To develop it, we first created a high resolution colon dataset using a pipeline that combines region growing with interactive machine learning to efficiently and accurately label the colon on CT colonography (CTC) images. Based on the generated dataset consisting of 435 labeled CTC images we trained an nnU-Net model for fully automatic colon segmentation. Our fully automatic model achieved an average symmetric surface distance of 0.2 mm (vs. 4.0 mm from TotalSegmentator) and a 95th percentile Hausdorff distance of 1.0 mm (vs. 18 mm from TotalSegmentator). Our segmentation accuracy substantially surpasses TotalSegmentator. We share our trained model and pipeline code, providing the first and only open-source tool for high-resolution colon segmentation. Additionally, we created a large-scale dataset of publicly available high-resolution colon labels.

Problem

• Develops a high-resolution colon segmentation method for clinical applications.
• Improves accuracy over existing tools like TotalSegmentator for complex colon shapes.
• Provides an open-source tool and dataset for high-resolution colon labeling.

Innovation

• Combines region growing with interactive machine learning
• Trained nnU-Net model for automatic colon segmentation
• Provides open-source tool and high-resolution dataset