Transformer-based multiclass segmentation pipeline for basic kidney histology

Scientific Reports, vol. 15, pp. 38957

Abstract

Current applications of deep learning in renal pathology focused on anatomical structures with morphology, yet little research has focused on the performance of models, such as versatility, in regions with severe kidney damage. In this study, we explored the difference in modal/domain shift capabilities between CNN-based and Transformer-based models. Firstly, we adopted two splitting strategies-WSI-level and patch-level-to stimulate sampling on multiple modal data distribution (i.e., renal WSIs collected from multi-centers). Then, we trained multiple CNN- and Transformer-based models on each splitting scheme respectively. We compared cross-splitting performance and analyzed the effective factors of results. For further validation, all models were tested on an independent external dataset for sensitivity analysis on the degree of fibrosis and inflammation. In conclusion, at both the patch- and WSI-level, M2F-Swin-B substantially outperformed UNet-ResNet18 with an average Intersection over Union (A-IoU) and per-class IoU. Notably, M2F-Swin-B outperformed UNet-ResNet18 in areas of a higher degree of fibrosis and inflammation and, a higher IoU score of arteries. In this study, we developed a robust multi-class segmentation pipeline for kidney histology. Moreover, we showed that the attention mechanism in Mask2Former enables visibly crisper and more uniform segmentation, particularly when the data is inadequate.

Authors 14

1. Junling He Aachen Department of Pathology, LUMC

   Army Medical University · Leiden University Medical Center · Daping Hospital

   Affiliation as printed

   Department of Nephrology, Daping Hospital, Army Medical Center, Army Medical University, Chongqing, China

   Department of Pathology, LUMC, Leiden, The Netherlands

2. P. Valkema Aachen Department of Pathology, LUMC

   Leiden University Medical Center · Amsterdam University Medical Centers

   Affiliation as printed

   Department of Pathology, Amsterdam UMC, Amsterdam, The Netherlands

   Department of Pathology, LUMC, Leiden, The Netherlands

3. Jingmin Long Aachen Leiden Institute of Advanced Computer Science

   Leiden University

   Affiliation as printed

   Leiden Institute of Advanced Computer Science, Leiden University, Leiden, The Netherlands

4. Jia Li Aachen Leiden Institute of Advanced Computer Science

   Leiden University

   Affiliation as printed

   Leiden Institute of Advanced Computer Science, Leiden University, Leiden, The Netherlands

5. Sandrine Florquin

   Amsterdam University Medical Centers

   Affiliation as printed

   Department of Pathology, Amsterdam UMC, Amsterdam, The Netherlands

6. Maarten Naesens

   KU Leuven

   Affiliation as printed

   Department of Nephrology and Transplantation, KU Leuven, Leuven, Belgium

7. Priyanka Joseph Koshy

   KU Leuven

   Affiliation as printed

   Department of Microbiology, Immunology and Transplantation, KU Leuven, Leuven, Belgium

8. Tri Quang Nguyen

   University Medical Center Utrecht

   Affiliation as printed

   Department of Pathology, UMC Utrecht, Utrecht, The Netherlands

9. Soufian Meziyerh Aachen Division of Nephrology Department of Medicine, LUMC Leiden Transplant Center

   Leiden University · Leiden University Medical Center

   Affiliation as printed

   Division of Nephrology, Department of Medicine, LUMC, Leiden, The Netherlands

   Leiden Transplant Center, LUMC and Leiden University, Leiden, The Netherlands

10. Aiko P. J. de Vries Aachen Division of Nephrology Department of Medicine, LUMC Leiden Transplant Center

    Leiden University · Leiden University Medical Center

    Affiliation as printed

    Division of Nephrology, Department of Medicine, LUMC, Leiden, The Netherlands

    Leiden Transplant Center, LUMC and Leiden University, Leiden, The Netherlands

11. Onno J. de Boer

    Amsterdam University Medical Centers

    Affiliation as printed

    Department of Pathology, Amsterdam UMC, Amsterdam, The Netherlands

12. Fons J. Verbeek Aachen Leiden Institute of Advanced Computer Science

    Leiden University

    Affiliation as printed

    Leiden Institute of Advanced Computer Science, Leiden University, Leiden, The Netherlands

13. Zhan Xiong Aachen Leiden Institute of Advanced Computer Science

    Leiden University

    Affiliation as printed

    Leiden Institute of Advanced Computer Science, Leiden University, Leiden, The Netherlands

14. Jesper Kers corresponding Aachen Department of Pathology, LUMC Leiden Transplant Center

    Leiden University · Leiden University Medical Center · Erasmus MC · Amsterdam University Medical Centers

    Affiliation as printed

    Department of Pathology, Amsterdam UMC, Amsterdam, The Netherlands. j.kers@lumc.nl

    Department of Pathology, Erasmus Medical Center, Rotterdam, The Netherlands. j.kers@lumc.nl

    Department of Pathology, LUMC, Leiden, The Netherlands. j.kers@lumc.nl

    Leiden Transplant Center, LUMC and Leiden University, Leiden, The Netherlands. j.kers@lumc.nl

    Department of Pathology, Erasmus Medical Center, Rotterdam, The Netherlands