Colorectal Cancer Tumor Grade Segmentation in Digital Histopathology Images: From Giga to Mini Challenge

2025-07-07

AI Summary

Rectal cancer histopathological grading suffers from high subjectivity, substantial inter-observer variability, and a critical shortage of expert pathologists—necessitating automated, standardized solutions. To address this, we introduce METU CCTGS, the first large-scale, publicly available whole-slide image dataset specifically designed for colorectal cancer grading, developed in conjunction with the ICIP Challenge. Leveraging a multi-team competitive framework, we benchmark state-of-the-art deep learning models—including Swin Transformer—for tumor region semantic segmentation. Evaluation was conducted uniformly on the Codalab platform using macro-averaged F-score and mean Intersection-over-Union (mIoU). All six participating teams surpassed the baseline model (top F-score ≥ 62.92), demonstrating the feasibility and clinical promise of deep learning for automated rectal cancer grading. This work significantly improves segmentation consistency and reproducibility, paving the way for robust, scalable digital pathology tools.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Multi-instance/Multi-view Learning
• Natural Language Processing: Interpretability, Analysis, and Evaluation of NLP Models

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

Colorectal cancer (CRC) is the third most diagnosed cancer and the second leading cause of cancer-related death worldwide. Accurate histopathological grading of CRC is essential for prognosis and treatment planning but remains a subjective process prone to observer variability and limited by global shortages of trained pathologists. To promote automated and standardized solutions, we organized the ICIP Grand Challenge on Colorectal Cancer Tumor Grading and Segmentation using the publicly available METU CCTGS dataset. The dataset comprises 103 whole-slide images with expert pixel-level annotations for five tissue classes. Participants submitted segmentation masks via Codalab, evaluated using metrics such as macro F-score and mIoU. Among 39 participating teams, six outperformed the Swin Transformer baseline (62.92 F-score). This paper presents an overview of the challenge, dataset, and the top-performing methods

Problem

• Automate colorectal cancer tumor grading in histopathology images
• Reduce subjectivity in CRC diagnosis due to pathologist variability
• Standardize segmentation using public dataset and evaluation metrics

Innovation

• Used METU CCTGS dataset for CRC grading
• Evaluated segmentation masks with macro F-score
• Swin Transformer baseline outperformed by six teams