Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

KPIs 2024 Challenge: Advancing Glomerular Segmentation from Patch- to Slide-Level

Ruining Deng, Tianyuan Yao, Yucheng Tang, Junlin Guo, Siqi Lu, Juming Xiong, Lining Yu, Quan Huu Cap, Pengzhou Cai, Libin Lan, Ze Zhao, Adrian Galdran, Amit Kumar, Gunjan Deotale, Dev Kumar Das, Inyoung Paik, Joonho Lee, Geongyu Lee, Yujia Chen, Wangkai Li, Zhaoyang Li, Xuege Hou, Zeyuan Wu, Shengjin Wang, Maximilian Fischer, Lars Kramer, Anghong Du, Le Zhang, Maria Sanchez Sanchez, Helena Sanchez Ulloa, David Ribalta Heredia, Carlos Perez de Arenaza Garcia, Shuoyu Xu, Bingdou He, Xinping Cheng, Tao Wang, Noemie Moreau, Katarzyna Bozek, Shubham Innani, Ujjwal Baid, Kaura Solomon Kefas, Bennett A. Landman, Yu Wang, Shilin Zhao, Mengmeng Yin, Haichun Yang, Yuankai Huo

TL;DR

The KPIs 2024 Challenge broadens kidney pathology benchmarking by providing a PAS-stained, rodent CKD-focused dataset with over 10,000 annotated glomeruli across 60+ WSIs, targeting both patch-level and WSI-level segmentation and detection. The study demonstrates that patch-level methods achieve high Dice scores through ensembling and multi-scale processing, while WSI-level results require advanced stitching, post-processing, and hybrid architectures to handle gigapixel images and diverse disease models. Key contributions include a comprehensive dataset, dual-task evaluation, and comparative leaderboards that highlight the effectiveness of transformer-based hybrids and SAM-based detection frameworks, offering a scalable path toward large-scale quantification in CKD research. The work establishes new benchmarks, promotes robust methodologies for kidney pathology analysis, and motivates future data diversification and reduction of preprocessing burdens to improve real-world clinical utility.

Abstract

Chronic kidney disease (CKD) is a major global health issue, affecting over 10% of the population and causing significant mortality. While kidney biopsy remains the gold standard for CKD diagnosis and treatment, the lack of comprehensive benchmarks for kidney pathology segmentation hinders progress in the field. To address this, we organized the Kidney Pathology Image Segmentation (KPIs) Challenge, introducing a dataset that incorporates preclinical rodent models of CKD with over 10,000 annotated glomeruli from 60+ Periodic Acid Schiff (PAS)-stained whole slide images. The challenge includes two tasks, patch-level segmentation and whole slide image segmentation and detection, evaluated using the Dice Similarity Coefficient (DSC) and F1-score. By encouraging innovative segmentation methods that adapt to diverse CKD models and tissue conditions, the KPIs Challenge aims to advance kidney pathology analysis, establish new benchmarks, and enable precise, large-scale quantification for disease research and diagnosis.

KPIs 2024 Challenge: Advancing Glomerular Segmentation from Patch- to Slide-Level

TL;DR

The KPIs 2024 Challenge broadens kidney pathology benchmarking by providing a PAS-stained, rodent CKD-focused dataset with over 10,000 annotated glomeruli across 60+ WSIs, targeting both patch-level and WSI-level segmentation and detection. The study demonstrates that patch-level methods achieve high Dice scores through ensembling and multi-scale processing, while WSI-level results require advanced stitching, post-processing, and hybrid architectures to handle gigapixel images and diverse disease models. Key contributions include a comprehensive dataset, dual-task evaluation, and comparative leaderboards that highlight the effectiveness of transformer-based hybrids and SAM-based detection frameworks, offering a scalable path toward large-scale quantification in CKD research. The work establishes new benchmarks, promotes robust methodologies for kidney pathology analysis, and motivates future data diversification and reduction of preprocessing burdens to improve real-world clinical utility.

Abstract

Chronic kidney disease (CKD) is a major global health issue, affecting over 10% of the population and causing significant mortality. While kidney biopsy remains the gold standard for CKD diagnosis and treatment, the lack of comprehensive benchmarks for kidney pathology segmentation hinders progress in the field. To address this, we organized the Kidney Pathology Image Segmentation (KPIs) Challenge, introducing a dataset that incorporates preclinical rodent models of CKD with over 10,000 annotated glomeruli from 60+ Periodic Acid Schiff (PAS)-stained whole slide images. The challenge includes two tasks, patch-level segmentation and whole slide image segmentation and detection, evaluated using the Dice Similarity Coefficient (DSC) and F1-score. By encouraging innovative segmentation methods that adapt to diverse CKD models and tissue conditions, the KPIs Challenge aims to advance kidney pathology analysis, establish new benchmarks, and enable precise, large-scale quantification for disease research and diagnosis.

Paper Structure

This paper contains 23 sections, 4 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Related Works
  3. Patch-Level Glomerular Segmentation
  4. Whole Slides Image-Level Glomerular Detection and Segmentation
  5. Challege description
  6. Overview
  7. Data description
  8. Data overview
  9. Challenge Setup
  10. Metrics and evaluation
  11. Choice of metrics
  12. Participating methods
  13. Task 1: Patch-level Glomeruli Segmentation
  14. Task 2a: WSI-level Glomeruli Segmentation
  15. Task 2b: WSI-level Glomeruli Detection
  16. ...and 8 more sections

Figures (4)

  • Figure 1: Overview of the challenge task setup. The KPIs challenge involves two tasks: (1) Patch-level Diseased Glomeruli Segmentation, focusing on the precise delineation of glomeruli in high-resolution image patches, and (2) Whole Slide Image (WSI)-level Diseased Glomeruli Segmentation, requiring robust segmentation across entire tissue sections. Example images illustrate the input data, ground truth masks, and overlay visualizations for each task. The dataset includes diverse mouse models to highlight morphological and pathological variations.
  • Figure 2: Box plots of quantitative results on different tasks from participants. This figure shows the box plots of the quantitative performance of participants on two tasks. Results from the top 10 teams are reported. Dice similarity coefficient scores are provided for Task 1, while both Dice similarity coefficient scores and F1 scores are presented for Task 2.
  • Figure 3: Patch-level quantitative results - The segmentation results from different participants' methods on glomeruli under various experimental conditions (normal, NEP25, DN, and 5/6Nx).
  • Figure 4: WSI-level quantitative results - This figure shows the segmentation results from different participants' methods on glomeruli under various experimental conditions (normal, NEP25, DN, and 5/6Nx). The yellow bounding box highlights the zoomed-in region on the WSI, with the zoomed-in images below showing the corresponding area.