PodoCount: A Robust, Fully Automated, Whole-Slide Podocyte Quantification Tool

Briana A. Santo; Darshana Govind; Parnaz Daneshpajouhnejad; Xiaoping Yang; Xiaoxin X. Wang; Komuraiah Myakala; Bryce A. Jones; Moshe Levi; Jeffrey B. Kopp; Teruhiko Yoshida; Laura J. Niedernhofer; David Manthey; Kyung Chul Moon; Seung Seok Han; Jarcy Zee; Avi Z. Rosenberg; Pinaki Sarder

doi:10.1016/j.ekir.2022.03.004

PodoCount: A Robust, Fully Automated, Whole-Slide Podocyte Quantification Tool

Briana A. Santo, Darshana Govind, Parnaz Daneshpajouhnejad, Xiaoping Yang, Xiaoxin X. Wang, Komuraiah Myakala, Bryce A. Jones, Moshe Levi, Jeffrey B. Kopp, Teruhiko Yoshida, Laura J. Niedernhofer, David Manthey, Kyung Chul Moon, Seung Seok Han, Jarcy Zee, Avi Z. Rosenberg, Pinaki Sarder

Biochemistry, Molecular Biology, and Biophysics (TMED)

Research output: Contribution to journal › Article › peer-review

3 Scopus citations

Abstract

Introduction: Podocyte depletion is a histomorphologic indicator of glomerular injury and predicts clinical outcomes. Podocyte estimation methods or podometrics are semiquantitative, technically involved, and laborious. Implementation of high-throughput podometrics in experimental and clinical workflows necessitates an automated podometrics pipeline. Recognizing that computational image analysis offers a robust approach to study cell and tissue structure, we developed and validated PodoCount (a computational tool for automated podocyte quantification in immunohistochemically labeled tissues) using a diverse data set. Methods: Whole-slide images (WSIs) of tissues immunostained with a podocyte nuclear marker and periodic acid–Schiff counterstain were acquired. The data set consisted of murine whole kidney sections (n = 135) from 6 disease models and human kidney biopsy specimens from patients with diabetic nephropathy (DN) (n = 45). Within segmented glomeruli, podocytes were extracted and image analysis was applied to compute measures of podocyte depletion and nuclear morphometry. Computational performance evaluation and statistical testing were performed to validate podometric and associated image features. PodoCount was disbursed as an open-source, cloud-based computational tool. Results: PodoCount produced highly accurate podocyte quantification when benchmarked against existing methods. Podocyte nuclear profiles were identified with 0.98 accuracy and segmented with 0.85 sensitivity and 0.99 specificity. Errors in podocyte count were bounded by 1 podocyte per glomerulus. Podocyte-specific image features were found to be significant predictors of disease state, proteinuria, and clinical outcome. Conclusion: PodoCount offers high-performance podocyte quantitation in diverse murine disease models and in human kidney biopsy specimens. Resultant features offer significant correlation with associated metadata and outcome. Our cloud-based tool will provide end users with a standardized approach for automated podometrics from gigapixel-sized WSIs.

Original language	English (US)
Pages (from-to)	1377-1392
Number of pages	16
Journal	Kidney International Reports
Volume	7
Issue number	6
DOIs	https://doi.org/10.1016/j.ekir.2022.03.004
State	Published - Jun 2022

Bibliographical note

Funding Information:
The authors are grateful to Brendon Lutnick, the developer of the Human-AI-Loop tool and the implementer of the Digital Slide Archive tool in the Sarder Laboratory, for his guidance. We acknowledge the assistance of the Reference Histology Core at the Johns Hopkins University School of Medicine, Baltimore, Maryland. We acknowledge the assistance of the Seoul National University Hospital Human Biobank, a member of the National Biobank of Korea, which is supported by the Ministry of Health and Welfare, Republic of Korea, for provision of human biospecimens used. The project was supported by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) grant R01 DK114485 (PS), National Institutes of Health OD grants R01 DK114485 02S1 and 03S1 (PS), NIDDK chronic kidney disease Biomarker Consortium grant U01 DK103225 (PS), NIDDK Kidney Precision Medicine Project glue grant U2C DK114886 (PS; contact: Dr. Jonathan Himmelfarb), a multidisciplinary small team grant RSG201047.2 (PS) from the State University of New York , a pilot grant (PS) from the University of Buffalo’s Clinical and Translational Science Institute grant 3UL1TR00141206 S1 (contact: Dr. Timothy Murphy), a DiaComp Pilot & Feasibility Project 21AU4180 (PS) with support from NIDDK Diabetic Complications Consortium grants U24 DK076169 and U24 DK115255 (contact: Dr. Richard A. McIndoe), National Institutes of Health OD Human Biomolecular Atlas Program grant U54 HL145608 (PS), and NIDDK grant R01 DK131189 (PS; contact: Dr. Farzad Fereidouni); National Institute on Aging grants P01 AG044376 (LJN), R01 AG063543 (LJN), and U19 AG056278 (LJN); National Center for Advancing Translational Sciences grants TL1TR001431 (BAJ), R01 DK127830 (ML), and R01 DK116567 (ML); and by the NIDDK Intramural Research Program (JBK).

Funding Information:
The authors are grateful to Brendon Lutnick, the developer of the Human-AI-Loop tool and the implementer of the Digital Slide Archive tool in the Sarder Laboratory, for his guidance. We acknowledge the assistance of the Reference Histology Core at the Johns Hopkins University School of Medicine, Baltimore, Maryland. We acknowledge the assistance of the Seoul National University Hospital Human Biobank, a member of the National Biobank of Korea, which is supported by the Ministry of Health and Welfare, Republic of Korea, for provision of human biospecimens used. The project was supported by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) grant R01 DK114485 (PS), National Institutes of Health OD grants R01 DK114485 02S1 and 03S1 (PS), NIDDK chronic kidney disease Biomarker Consortium grant U01 DK103225 (PS), NIDDK Kidney Precision Medicine Project glue grant U2C DK114886 (PS; contact: Dr. Jonathan Himmelfarb), a multidisciplinary small team grant RSG201047.2 (PS) from the State University of New York, a pilot grant (PS) from the University of Buffalo's Clinical and Translational Science Institute grant 3UL1TR00141206 S1 (contact: Dr. Timothy Murphy), a DiaComp Pilot & Feasibility Project 21AU4180 (PS) with support from NIDDK Diabetic Complications Consortium grants U24 DK076169 and U24 DK115255 (contact: Dr. Richard A. McIndoe), National Institutes of Health OD Human Biomolecular Atlas Program grant U54 HL145608 (PS), and NIDDK grant R01 DK131189 (PS; contact: Dr. Farzad Fereidouni); National Institute on Aging grants P01 AG044376 (LJN), R01 AG063543 (LJN), and U19 AG056278 (LJN); National Center for Advancing Translational Sciences grants TL1TR001431 (BAJ), R01 DK127830 (ML), and R01 DK116567 (ML); and by the NIDDK Intramural Research Program (JBK). BAS conceptualized and performed the quantitative analyses, designed and conducted the computational methods, completed all statistical analyses, interpreted the results, and wrote the manuscript. DG and DM assisted with the adaptation of PodoCount codes to the cloud format. PD assisted with completion of ground truth podocyte counts. XY assisted in manuscript preparation. db/db, KKAy, and Aging cohorts were generated by the Levi Lab (XXW, KM, BAJ, ML). The FSGS (SAND) and HIV mouse cohorts were generated by the Kopp Lab (JBK, TY). JBK also assisted in manuscript editing. The Ercc1−/Δ Progeroid cohort and littermate controls were generated by the Niedernhofer Lab (LJN). SSH and KCM provided human biopsy and clinical data for patients with diabetes. JZ assisted with statistical design and analytical considerations. AZR conceived and optimized the staining technique that enables this computational method. AZR also assisted in manuscript preparation. PS conceived the idea of automated computational quantification and enumeration of podocytes from digital whole-slide images in the domain of renal pathology and making this tool available for nephrology community via an open-source cloud resource. PS also assisted in manuscript preparation, coordinated with the study team, assisted in study design, supervised the computational implementation, and critically analyzed the results.

Publisher Copyright:
© 2022 International Society of Nephrology

Keywords

chronic kidney disease
digital pathology
gigapixel size images
glomerular disease
podocyte
podometrics

PubMed: MeSH publication types

Journal Article

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access

10.1016/j.ekir.2022.03.004

OpenUrl availability

Full text

Cite this

Santo, B. A., Govind, D., Daneshpajouhnejad, P., Yang, X., Wang, X. X., Myakala, K., Jones, B. A., Levi, M., Kopp, J. B., Yoshida, T., Niedernhofer, L. J., Manthey, D., Moon, K. C., Han, S. S., Zee, J., Rosenberg, A. Z., & Sarder, P. (2022). PodoCount: A Robust, Fully Automated, Whole-Slide Podocyte Quantification Tool. Kidney International Reports, 7(6), 1377-1392. https://doi.org/10.1016/j.ekir.2022.03.004

Santo, BA, Govind, D, Daneshpajouhnejad, P, Yang, X, Wang, XX, Myakala, K, Jones, BA, Levi, M, Kopp, JB, Yoshida, T, Niedernhofer, LJ, Manthey, D, Moon, KC, Han, SS, Zee, J, Rosenberg, AZ & Sarder, P 2022, 'PodoCount: A Robust, Fully Automated, Whole-Slide Podocyte Quantification Tool', Kidney International Reports, vol. 7, no. 6, pp. 1377-1392. https://doi.org/10.1016/j.ekir.2022.03.004

@article{a0d698ac2792492e9dcbbc4871b3fb79,

title = "PodoCount: A Robust, Fully Automated, Whole-Slide Podocyte Quantification Tool",

abstract = "Introduction: Podocyte depletion is a histomorphologic indicator of glomerular injury and predicts clinical outcomes. Podocyte estimation methods or podometrics are semiquantitative, technically involved, and laborious. Implementation of high-throughput podometrics in experimental and clinical workflows necessitates an automated podometrics pipeline. Recognizing that computational image analysis offers a robust approach to study cell and tissue structure, we developed and validated PodoCount (a computational tool for automated podocyte quantification in immunohistochemically labeled tissues) using a diverse data set. Methods: Whole-slide images (WSIs) of tissues immunostained with a podocyte nuclear marker and periodic acid–Schiff counterstain were acquired. The data set consisted of murine whole kidney sections (n = 135) from 6 disease models and human kidney biopsy specimens from patients with diabetic nephropathy (DN) (n = 45). Within segmented glomeruli, podocytes were extracted and image analysis was applied to compute measures of podocyte depletion and nuclear morphometry. Computational performance evaluation and statistical testing were performed to validate podometric and associated image features. PodoCount was disbursed as an open-source, cloud-based computational tool. Results: PodoCount produced highly accurate podocyte quantification when benchmarked against existing methods. Podocyte nuclear profiles were identified with 0.98 accuracy and segmented with 0.85 sensitivity and 0.99 specificity. Errors in podocyte count were bounded by 1 podocyte per glomerulus. Podocyte-specific image features were found to be significant predictors of disease state, proteinuria, and clinical outcome. Conclusion: PodoCount offers high-performance podocyte quantitation in diverse murine disease models and in human kidney biopsy specimens. Resultant features offer significant correlation with associated metadata and outcome. Our cloud-based tool will provide end users with a standardized approach for automated podometrics from gigapixel-sized WSIs.",

keywords = "chronic kidney disease, digital pathology, gigapixel size images, glomerular disease, podocyte, podometrics",

author = "Santo, {Briana A.} and Darshana Govind and Parnaz Daneshpajouhnejad and Xiaoping Yang and Wang, {Xiaoxin X.} and Komuraiah Myakala and Jones, {Bryce A.} and Moshe Levi and Kopp, {Jeffrey B.} and Teruhiko Yoshida and Niedernhofer, {Laura J.} and David Manthey and Moon, {Kyung Chul} and Han, {Seung Seok} and Jarcy Zee and Rosenberg, {Avi Z.} and Pinaki Sarder",

note = "Funding Information: The authors are grateful to Brendon Lutnick, the developer of the Human-AI-Loop tool and the implementer of the Digital Slide Archive tool in the Sarder Laboratory, for his guidance. We acknowledge the assistance of the Reference Histology Core at the Johns Hopkins University School of Medicine, Baltimore, Maryland. We acknowledge the assistance of the Seoul National University Hospital Human Biobank, a member of the National Biobank of Korea, which is supported by the Ministry of Health and Welfare, Republic of Korea, for provision of human biospecimens used. The project was supported by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) grant R01 DK114485 (PS), National Institutes of Health OD grants R01 DK114485 02S1 and 03S1 (PS), NIDDK chronic kidney disease Biomarker Consortium grant U01 DK103225 (PS), NIDDK Kidney Precision Medicine Project glue grant U2C DK114886 (PS; contact: Dr. Jonathan Himmelfarb), a multidisciplinary small team grant RSG201047.2 (PS) from the State University of New York , a pilot grant (PS) from the University of Buffalo{\textquoteright}s Clinical and Translational Science Institute grant 3UL1TR00141206 S1 (contact: Dr. Timothy Murphy), a DiaComp Pilot & Feasibility Project 21AU4180 (PS) with support from NIDDK Diabetic Complications Consortium grants U24 DK076169 and U24 DK115255 (contact: Dr. Richard A. McIndoe), National Institutes of Health OD Human Biomolecular Atlas Program grant U54 HL145608 (PS), and NIDDK grant R01 DK131189 (PS; contact: Dr. Farzad Fereidouni); National Institute on Aging grants P01 AG044376 (LJN), R01 AG063543 (LJN), and U19 AG056278 (LJN); National Center for Advancing Translational Sciences grants TL1TR001431 (BAJ), R01 DK127830 (ML), and R01 DK116567 (ML); and by the NIDDK Intramural Research Program (JBK). Funding Information: The authors are grateful to Brendon Lutnick, the developer of the Human-AI-Loop tool and the implementer of the Digital Slide Archive tool in the Sarder Laboratory, for his guidance. We acknowledge the assistance of the Reference Histology Core at the Johns Hopkins University School of Medicine, Baltimore, Maryland. We acknowledge the assistance of the Seoul National University Hospital Human Biobank, a member of the National Biobank of Korea, which is supported by the Ministry of Health and Welfare, Republic of Korea, for provision of human biospecimens used. The project was supported by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) grant R01 DK114485 (PS), National Institutes of Health OD grants R01 DK114485 02S1 and 03S1 (PS), NIDDK chronic kidney disease Biomarker Consortium grant U01 DK103225 (PS), NIDDK Kidney Precision Medicine Project glue grant U2C DK114886 (PS; contact: Dr. Jonathan Himmelfarb), a multidisciplinary small team grant RSG201047.2 (PS) from the State University of New York, a pilot grant (PS) from the University of Buffalo's Clinical and Translational Science Institute grant 3UL1TR00141206 S1 (contact: Dr. Timothy Murphy), a DiaComp Pilot & Feasibility Project 21AU4180 (PS) with support from NIDDK Diabetic Complications Consortium grants U24 DK076169 and U24 DK115255 (contact: Dr. Richard A. McIndoe), National Institutes of Health OD Human Biomolecular Atlas Program grant U54 HL145608 (PS), and NIDDK grant R01 DK131189 (PS; contact: Dr. Farzad Fereidouni); National Institute on Aging grants P01 AG044376 (LJN), R01 AG063543 (LJN), and U19 AG056278 (LJN); National Center for Advancing Translational Sciences grants TL1TR001431 (BAJ), R01 DK127830 (ML), and R01 DK116567 (ML); and by the NIDDK Intramural Research Program (JBK). BAS conceptualized and performed the quantitative analyses, designed and conducted the computational methods, completed all statistical analyses, interpreted the results, and wrote the manuscript. DG and DM assisted with the adaptation of PodoCount codes to the cloud format. PD assisted with completion of ground truth podocyte counts. XY assisted in manuscript preparation. db/db, KKAy, and Aging cohorts were generated by the Levi Lab (XXW, KM, BAJ, ML). The FSGS (SAND) and HIV mouse cohorts were generated by the Kopp Lab (JBK, TY). JBK also assisted in manuscript editing. The Ercc1−/Δ Progeroid cohort and littermate controls were generated by the Niedernhofer Lab (LJN). SSH and KCM provided human biopsy and clinical data for patients with diabetes. JZ assisted with statistical design and analytical considerations. AZR conceived and optimized the staining technique that enables this computational method. AZR also assisted in manuscript preparation. PS conceived the idea of automated computational quantification and enumeration of podocytes from digital whole-slide images in the domain of renal pathology and making this tool available for nephrology community via an open-source cloud resource. PS also assisted in manuscript preparation, coordinated with the study team, assisted in study design, supervised the computational implementation, and critically analyzed the results. Publisher Copyright: {\textcopyright} 2022 International Society of Nephrology",

year = "2022",

month = jun,

doi = "10.1016/j.ekir.2022.03.004",

language = "English (US)",

volume = "7",

pages = "1377--1392",

journal = "Kidney International Reports",

issn = "2468-0249",

publisher = "Elsevier Inc.",

number = "6",

}

TY - JOUR

T1 - PodoCount

T2 - A Robust, Fully Automated, Whole-Slide Podocyte Quantification Tool

AU - Santo, Briana A.

AU - Govind, Darshana

AU - Daneshpajouhnejad, Parnaz

AU - Yang, Xiaoping

AU - Wang, Xiaoxin X.

AU - Myakala, Komuraiah

AU - Jones, Bryce A.

AU - Levi, Moshe

AU - Kopp, Jeffrey B.

AU - Yoshida, Teruhiko

AU - Niedernhofer, Laura J.

AU - Manthey, David

AU - Moon, Kyung Chul

AU - Han, Seung Seok

AU - Zee, Jarcy

AU - Rosenberg, Avi Z.

AU - Sarder, Pinaki

N1 - Funding Information: The authors are grateful to Brendon Lutnick, the developer of the Human-AI-Loop tool and the implementer of the Digital Slide Archive tool in the Sarder Laboratory, for his guidance. We acknowledge the assistance of the Reference Histology Core at the Johns Hopkins University School of Medicine, Baltimore, Maryland. We acknowledge the assistance of the Seoul National University Hospital Human Biobank, a member of the National Biobank of Korea, which is supported by the Ministry of Health and Welfare, Republic of Korea, for provision of human biospecimens used. The project was supported by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) grant R01 DK114485 (PS), National Institutes of Health OD grants R01 DK114485 02S1 and 03S1 (PS), NIDDK chronic kidney disease Biomarker Consortium grant U01 DK103225 (PS), NIDDK Kidney Precision Medicine Project glue grant U2C DK114886 (PS; contact: Dr. Jonathan Himmelfarb), a multidisciplinary small team grant RSG201047.2 (PS) from the State University of New York , a pilot grant (PS) from the University of Buffalo’s Clinical and Translational Science Institute grant 3UL1TR00141206 S1 (contact: Dr. Timothy Murphy), a DiaComp Pilot & Feasibility Project 21AU4180 (PS) with support from NIDDK Diabetic Complications Consortium grants U24 DK076169 and U24 DK115255 (contact: Dr. Richard A. McIndoe), National Institutes of Health OD Human Biomolecular Atlas Program grant U54 HL145608 (PS), and NIDDK grant R01 DK131189 (PS; contact: Dr. Farzad Fereidouni); National Institute on Aging grants P01 AG044376 (LJN), R01 AG063543 (LJN), and U19 AG056278 (LJN); National Center for Advancing Translational Sciences grants TL1TR001431 (BAJ), R01 DK127830 (ML), and R01 DK116567 (ML); and by the NIDDK Intramural Research Program (JBK). Funding Information: The authors are grateful to Brendon Lutnick, the developer of the Human-AI-Loop tool and the implementer of the Digital Slide Archive tool in the Sarder Laboratory, for his guidance. We acknowledge the assistance of the Reference Histology Core at the Johns Hopkins University School of Medicine, Baltimore, Maryland. We acknowledge the assistance of the Seoul National University Hospital Human Biobank, a member of the National Biobank of Korea, which is supported by the Ministry of Health and Welfare, Republic of Korea, for provision of human biospecimens used. The project was supported by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) grant R01 DK114485 (PS), National Institutes of Health OD grants R01 DK114485 02S1 and 03S1 (PS), NIDDK chronic kidney disease Biomarker Consortium grant U01 DK103225 (PS), NIDDK Kidney Precision Medicine Project glue grant U2C DK114886 (PS; contact: Dr. Jonathan Himmelfarb), a multidisciplinary small team grant RSG201047.2 (PS) from the State University of New York, a pilot grant (PS) from the University of Buffalo's Clinical and Translational Science Institute grant 3UL1TR00141206 S1 (contact: Dr. Timothy Murphy), a DiaComp Pilot & Feasibility Project 21AU4180 (PS) with support from NIDDK Diabetic Complications Consortium grants U24 DK076169 and U24 DK115255 (contact: Dr. Richard A. McIndoe), National Institutes of Health OD Human Biomolecular Atlas Program grant U54 HL145608 (PS), and NIDDK grant R01 DK131189 (PS; contact: Dr. Farzad Fereidouni); National Institute on Aging grants P01 AG044376 (LJN), R01 AG063543 (LJN), and U19 AG056278 (LJN); National Center for Advancing Translational Sciences grants TL1TR001431 (BAJ), R01 DK127830 (ML), and R01 DK116567 (ML); and by the NIDDK Intramural Research Program (JBK). BAS conceptualized and performed the quantitative analyses, designed and conducted the computational methods, completed all statistical analyses, interpreted the results, and wrote the manuscript. DG and DM assisted with the adaptation of PodoCount codes to the cloud format. PD assisted with completion of ground truth podocyte counts. XY assisted in manuscript preparation. db/db, KKAy, and Aging cohorts were generated by the Levi Lab (XXW, KM, BAJ, ML). The FSGS (SAND) and HIV mouse cohorts were generated by the Kopp Lab (JBK, TY). JBK also assisted in manuscript editing. The Ercc1−/Δ Progeroid cohort and littermate controls were generated by the Niedernhofer Lab (LJN). SSH and KCM provided human biopsy and clinical data for patients with diabetes. JZ assisted with statistical design and analytical considerations. AZR conceived and optimized the staining technique that enables this computational method. AZR also assisted in manuscript preparation. PS conceived the idea of automated computational quantification and enumeration of podocytes from digital whole-slide images in the domain of renal pathology and making this tool available for nephrology community via an open-source cloud resource. PS also assisted in manuscript preparation, coordinated with the study team, assisted in study design, supervised the computational implementation, and critically analyzed the results. Publisher Copyright: © 2022 International Society of Nephrology

PY - 2022/6

Y1 - 2022/6

N2 - Introduction: Podocyte depletion is a histomorphologic indicator of glomerular injury and predicts clinical outcomes. Podocyte estimation methods or podometrics are semiquantitative, technically involved, and laborious. Implementation of high-throughput podometrics in experimental and clinical workflows necessitates an automated podometrics pipeline. Recognizing that computational image analysis offers a robust approach to study cell and tissue structure, we developed and validated PodoCount (a computational tool for automated podocyte quantification in immunohistochemically labeled tissues) using a diverse data set. Methods: Whole-slide images (WSIs) of tissues immunostained with a podocyte nuclear marker and periodic acid–Schiff counterstain were acquired. The data set consisted of murine whole kidney sections (n = 135) from 6 disease models and human kidney biopsy specimens from patients with diabetic nephropathy (DN) (n = 45). Within segmented glomeruli, podocytes were extracted and image analysis was applied to compute measures of podocyte depletion and nuclear morphometry. Computational performance evaluation and statistical testing were performed to validate podometric and associated image features. PodoCount was disbursed as an open-source, cloud-based computational tool. Results: PodoCount produced highly accurate podocyte quantification when benchmarked against existing methods. Podocyte nuclear profiles were identified with 0.98 accuracy and segmented with 0.85 sensitivity and 0.99 specificity. Errors in podocyte count were bounded by 1 podocyte per glomerulus. Podocyte-specific image features were found to be significant predictors of disease state, proteinuria, and clinical outcome. Conclusion: PodoCount offers high-performance podocyte quantitation in diverse murine disease models and in human kidney biopsy specimens. Resultant features offer significant correlation with associated metadata and outcome. Our cloud-based tool will provide end users with a standardized approach for automated podometrics from gigapixel-sized WSIs.

AB - Introduction: Podocyte depletion is a histomorphologic indicator of glomerular injury and predicts clinical outcomes. Podocyte estimation methods or podometrics are semiquantitative, technically involved, and laborious. Implementation of high-throughput podometrics in experimental and clinical workflows necessitates an automated podometrics pipeline. Recognizing that computational image analysis offers a robust approach to study cell and tissue structure, we developed and validated PodoCount (a computational tool for automated podocyte quantification in immunohistochemically labeled tissues) using a diverse data set. Methods: Whole-slide images (WSIs) of tissues immunostained with a podocyte nuclear marker and periodic acid–Schiff counterstain were acquired. The data set consisted of murine whole kidney sections (n = 135) from 6 disease models and human kidney biopsy specimens from patients with diabetic nephropathy (DN) (n = 45). Within segmented glomeruli, podocytes were extracted and image analysis was applied to compute measures of podocyte depletion and nuclear morphometry. Computational performance evaluation and statistical testing were performed to validate podometric and associated image features. PodoCount was disbursed as an open-source, cloud-based computational tool. Results: PodoCount produced highly accurate podocyte quantification when benchmarked against existing methods. Podocyte nuclear profiles were identified with 0.98 accuracy and segmented with 0.85 sensitivity and 0.99 specificity. Errors in podocyte count were bounded by 1 podocyte per glomerulus. Podocyte-specific image features were found to be significant predictors of disease state, proteinuria, and clinical outcome. Conclusion: PodoCount offers high-performance podocyte quantitation in diverse murine disease models and in human kidney biopsy specimens. Resultant features offer significant correlation with associated metadata and outcome. Our cloud-based tool will provide end users with a standardized approach for automated podometrics from gigapixel-sized WSIs.

KW - chronic kidney disease

KW - digital pathology

KW - gigapixel size images

KW - glomerular disease

KW - podocyte

KW - podometrics

UR - http://www.scopus.com/inward/record.url?scp=85132411934&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85132411934&partnerID=8YFLogxK

U2 - 10.1016/j.ekir.2022.03.004

DO - 10.1016/j.ekir.2022.03.004

M3 - Article

C2 - 35694561

AN - SCOPUS:85132411934

SN - 2468-0249

VL - 7

SP - 1377

EP - 1392

JO - Kidney International Reports

JF - Kidney International Reports

IS - 6

ER -

PodoCount: A Robust, Fully Automated, Whole-Slide Podocyte Quantification Tool

Abstract

Bibliographical note

Keywords

PubMed: MeSH publication types

UN SDGs

Access

OpenUrl availability

Other files and links

Fingerprint

Cite this