Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Tumor Detection, Segmentation and Classification Challenge on Automated 3D Breast Ultrasound: The TDSC-ABUS Challenge

Gongning Luo, Mingwang Xu, Hongyu Chen, Xinjie Liang, Xing Tao, Dong Ni, Hyunsu Jeong, Chulhong Kim, Raphael Stock, Michael Baumgartner, Yannick Kirchhoff, Maximilian Rokuss, Klaus Maier-Hein, Zhikai Yang, Tianyu Fan, Nicolas Boutry, Dmitry Tereshchenko, Arthur Moine, Maximilien Charmetant, Jan Sauer, Hao Du, Xiang-Hui Bai, Vipul Pai Raikar, Ricardo Montoya-del-Angel, Robert Marti, Miguel Luna, Dongmin Lee, Abdul Qayyum, Moona Mazher, Qihui Guo, Changyan Wang, Navchetan Awasthi, Qiaochu Zhao, Wei Wang, Kuanquan Wang, Qiucheng Wang, Suyu Dong

TL;DR

The paper tackles the problem of reliable tumor analysis in 3D ABUS by introducing the first public benchmark (TDSC-ABUS2023) that jointly evaluates detection, segmentation, and classification. It details a 200-volume ABUS dataset with expert annotations and a rigorous, Docker-based evaluation framework across three tasks, including specific metric formulas and inf-value handling. The results section summarizes participant submissions and provides in-depth analyses of 18 participating approaches, highlighting top performers and key design strategies. Overall, the work demonstrates the feasibility and value of a multi-task ABUS benchmark and offers practical insights for developing robust, clinically relevant CAD systems. The benchmark is openly accessible to spur future advances in ABUS tumor analysis and facilitate cross-method comparisons.

Abstract

Breast cancer is one of the most common causes of death among women worldwide. Early detection helps in reducing the number of deaths. Automated 3D Breast Ultrasound (ABUS) is a newer approach for breast screening, which has many advantages over handheld mammography such as safety, speed, and higher detection rate of breast cancer. Tumor detection, segmentation, and classification are key components in the analysis of medical images, especially challenging in the context of 3D ABUS due to the significant variability in tumor size and shape, unclear tumor boundaries, and a low signal-to-noise ratio. The lack of publicly accessible, well-labeled ABUS datasets further hinders the advancement of systems for breast tumor analysis. Addressing this gap, we have organized the inaugural Tumor Detection, Segmentation, and Classification Challenge on Automated 3D Breast Ultrasound 2023 (TDSC-ABUS2023). This initiative aims to spearhead research in this field and create a definitive benchmark for tasks associated with 3D ABUS image analysis. In this paper, we summarize the top-performing algorithms from the challenge and provide critical analysis for ABUS image examination. We offer the TDSC-ABUS challenge as an open-access platform at https://tdsc-abus2023.grand-challenge.org/ to benchmark and inspire future developments in algorithmic research.

Tumor Detection, Segmentation and Classification Challenge on Automated 3D Breast Ultrasound: The TDSC-ABUS Challenge

TL;DR

The paper tackles the problem of reliable tumor analysis in 3D ABUS by introducing the first public benchmark (TDSC-ABUS2023) that jointly evaluates detection, segmentation, and classification. It details a 200-volume ABUS dataset with expert annotations and a rigorous, Docker-based evaluation framework across three tasks, including specific metric formulas and inf-value handling. The results section summarizes participant submissions and provides in-depth analyses of 18 participating approaches, highlighting top performers and key design strategies. Overall, the work demonstrates the feasibility and value of a multi-task ABUS benchmark and offers practical insights for developing robust, clinically relevant CAD systems. The benchmark is openly accessible to spur future advances in ABUS tumor analysis and facilitate cross-method comparisons.

Abstract

Breast cancer is one of the most common causes of death among women worldwide. Early detection helps in reducing the number of deaths. Automated 3D Breast Ultrasound (ABUS) is a newer approach for breast screening, which has many advantages over handheld mammography such as safety, speed, and higher detection rate of breast cancer. Tumor detection, segmentation, and classification are key components in the analysis of medical images, especially challenging in the context of 3D ABUS due to the significant variability in tumor size and shape, unclear tumor boundaries, and a low signal-to-noise ratio. The lack of publicly accessible, well-labeled ABUS datasets further hinders the advancement of systems for breast tumor analysis. Addressing this gap, we have organized the inaugural Tumor Detection, Segmentation, and Classification Challenge on Automated 3D Breast Ultrasound 2023 (TDSC-ABUS2023). This initiative aims to spearhead research in this field and create a definitive benchmark for tasks associated with 3D ABUS image analysis. In this paper, we summarize the top-performing algorithms from the challenge and provide critical analysis for ABUS image examination. We offer the TDSC-ABUS challenge as an open-access platform at https://tdsc-abus2023.grand-challenge.org/ to benchmark and inspire future developments in algorithmic research.
Paper Structure (37 sections, 5 equations, 11 figures, 7 tables)

This paper contains 37 sections, 5 equations, 11 figures, 7 tables.

Table of Contents

  1. Introduction
  2. Methods
  3. Challenge dataset
  4. Challenge organization
  5. Evaluation metrics and ranking scheme
  6. Results
  7. Challenge participants and submissions
  8. Algorithms summarization of teams
  9. T1: Shenzhen University (SZU)
  10. T2: Pohang University of Science and Technology (POSTECH)
  11. T3: Heidelberg University(HU)
  12. T4: KTH Royal Institute of Technology (KTH)
  13. T5: École pour l'informatique et les techniques avancées (EPITA)
  14. T6: FathomX (FX)
  15. T7: Philips Research (PR)
  16. ...and 22 more sections

Figures (11)

  • Figure 1: Representative ABUS Image Cases. (a) A small-sized tumor exhibiting a jagged boundary. (b) A comparatively larger tumor characterized by a smooth periphery.
  • Figure 2: Summary of TDSC-ABUS 2023 Challenge participants and submissions. There were 563 teams registering on the official grand-challenge webpage and 107 of them signed the agreement. Finally, 14 teams submitted segmentation validation results, 9 teams submitted classification validation results, and 10 teams submitted detection validation results.
  • Figure 3: The model structure of T1
  • Figure 4: The model structure of T2
  • Figure 5: Network structure diagram of T11
  • ...and 6 more figures