Deep learning segmentation of glomeruli on kidney donor frozen sections

Xiang Li, Richard Davis, Yuemei Xu, Zehan Wang, Nao Souma, Gina Sotolongo, Jonathan Bell, Matthew J. Ellis, David N. Howell, Xiling Shen, Kyle J. Lafata, Laura BarisoniJournal of Medical Imaging 2021 · DOI

Abstract

<b>Purpose:</b> Recent advances in computational image analysis offer the opportunity to develop automatic quantification of histologic parameters as aid tools for practicing pathologists. We aim to develop deep learning (DL) models to quantify nonsclerotic and sclerotic glomeruli on frozen sections from donor kidney biopsies. <b>Approach:</b> A total of 258 whole slide images (WSI) from cadaveric donor kidney biopsies performed at our institution ( <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mi>n</mml:mi> <mml:mo>=</mml:mo> <mml:mn>123</mml:mn></mml:mrow> </mml:math> ) and at external institutions ( <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mi>n</mml:mi> <mml:mo>=</mml:mo> <mml:mn>135</mml:mn></mml:mrow> </mml:math> ) were used in this study. WSIs from our institution were divided at the patient level into training and validation datasets (ratio: 0.8:0.2), and external WSIs were used as an independent testing dataset. Nonsclerotic ( <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mi>n</mml:mi> <mml:mo>=</mml:mo> <mml:mn>22767</mml:mn></mml:mrow> </mml:math> ) and sclerotic ( <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mi>n</mml:mi> <mml:mo>=</mml:mo> <mml:mn>1366</mml:mn></mml:mrow> </mml:math> ) glomeruli were manually annotated by study pathologists on all WSIs. A nine-layer convolutional neural network based on the common U-Net architecture was developed and tested for the segmentation of nonsclerotic and sclerotic glomeruli. DL-derived, manual segmentation, and reported glomerular count (standard of care) were compared. <b>Results:</b> The average Dice similarity coefficient testing was 0.90 and 0.83. And the <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"><mml:mrow><mml:mi>F</mml:mi> <mml:mn>1</mml:mn></mml:mrow> </mml:math> , recall, and precision scores were 0.93, 0.96, and 0.90, and 0.87, 0.93, and 0.81, for nonsclerotic and sclerotic glomeruli, respectively. DL-derived and manual segmentation-derived glomerular counts were comparable, but statistically different from reported glomerular count. <b>Conclusions:</b> DL segmentation is a feasible and robust approach for automatic quantification of glomeruli. We represent the first step toward new protocols for the evaluation of donor kidney biopsies.
BibTeX
@article{li2021deep,
  author = {Xiang Li and Richard Davis and Yuemei Xu and Zehan Wang and Nao Souma and Gina Sotolongo and Jonathan Bell and Matthew J. Ellis and David N. Howell and Xiling Shen and Kyle J. Lafata and Laura Barisoni},
  title = {{Deep learning segmentation of glomeruli on kidney donor frozen sections}},
  journal = {Journal of Medical Imaging},
  year = {2021},
  doi = {10.1117/1.jmi.8.6.067501},
}

Loading related content...

Figure