Automated Computational Detection of Interstitial Fibrosis, Tubular Atrophy, and Glomerulosclerosis.
convolutional neural network
diabetes
eGFR
glomerulosclerosis
interstitial fibrosis
machine learning
prognostication
transplant
tubular atrophy
whole slide segmentation
Journal
Journal of the American Society of Nephrology : JASN
ISSN: 1533-3450
Titre abrégé: J Am Soc Nephrol
Pays: United States
ID NLM: 9013836
Informations de publication
Date de publication:
Apr 2021
Apr 2021
Historique:
received:
13
05
2020
accepted:
14
12
2020
pubmed:
25
2
2021
medline:
25
2
2021
entrez:
24
2
2021
Statut:
ppublish
Résumé
Interstitial fibrosis, tubular atrophy (IFTA), and glomerulosclerosis are indicators of irrecoverable kidney injury. Modern machine learning (ML) tools have enabled robust, automated identification of image structures that can be comparable with analysis by human experts. ML algorithms were developed and tested for the ability to replicate the detection and quantification of IFTA and glomerulosclerosis that renal pathologists perform. A renal pathologist annotated renal biopsy specimens from 116 whole-slide images (WSIs) for IFTA and glomerulosclerosis. A total of 79 WSIs were used for training different configurations of a convolutional neural network (CNN), and 17 and 20 WSIs were used as internal and external testing cases, respectively. The best model was compared against the input of four renal pathologists on 20 new testing slides. Further, for 87 testing biopsy specimens, IFTA and glomerulosclerosis measurements made by pathologists and the CNN were correlated to patient outcome using classic statistical tools. The best average performance across all image classes came from a DeepLab version 2 network trained at 40× magnification. IFTA and glomerulosclerosis percentages derived from this CNN achieved high levels of agreement with four renal pathologists. The pathologist- and CNN-based analyses of IFTA and glomerulosclerosis showed statistically significant and equivalent correlation with all patient-outcome variables. ML algorithms can be trained to replicate the IFTA and glomerulosclerosis assessment performed by renal pathologists. This suggests computational methods may be able to provide a standardized approach to evaluate the extent of chronic kidney injury in situations in which renal-pathologist time is restricted or unavailable.
Sections du résumé
BACKGROUND
BACKGROUND
Interstitial fibrosis, tubular atrophy (IFTA), and glomerulosclerosis are indicators of irrecoverable kidney injury. Modern machine learning (ML) tools have enabled robust, automated identification of image structures that can be comparable with analysis by human experts. ML algorithms were developed and tested for the ability to replicate the detection and quantification of IFTA and glomerulosclerosis that renal pathologists perform.
METHODS
METHODS
A renal pathologist annotated renal biopsy specimens from 116 whole-slide images (WSIs) for IFTA and glomerulosclerosis. A total of 79 WSIs were used for training different configurations of a convolutional neural network (CNN), and 17 and 20 WSIs were used as internal and external testing cases, respectively. The best model was compared against the input of four renal pathologists on 20 new testing slides. Further, for 87 testing biopsy specimens, IFTA and glomerulosclerosis measurements made by pathologists and the CNN were correlated to patient outcome using classic statistical tools.
RESULTS
RESULTS
The best average performance across all image classes came from a DeepLab version 2 network trained at 40× magnification. IFTA and glomerulosclerosis percentages derived from this CNN achieved high levels of agreement with four renal pathologists. The pathologist- and CNN-based analyses of IFTA and glomerulosclerosis showed statistically significant and equivalent correlation with all patient-outcome variables.
CONCLUSIONS
CONCLUSIONS
ML algorithms can be trained to replicate the IFTA and glomerulosclerosis assessment performed by renal pathologists. This suggests computational methods may be able to provide a standardized approach to evaluate the extent of chronic kidney injury in situations in which renal-pathologist time is restricted or unavailable.
Identifiants
pubmed: 33622976
pii: 00001751-202104000-00011
doi: 10.1681/ASN.2020050652
pmc: PMC8017538
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
837-850Subventions
Organisme : NCATS NIH HHS
ID : UL1 TR001863
Pays : United States
Organisme : NIDDK NIH HHS
ID : R01 DK093770
Pays : United States
Organisme : NIDDK NIH HHS
ID : R01 DK114485
Pays : United States
Organisme : NIDDK NIH HHS
ID : U2C DK114886
Pays : United States
Organisme : NIH HHS
ID : S10 OD024973
Pays : United States
Organisme : NIDDK NIH HHS
ID : R01 DK056942
Pays : United States
Commentaires et corrections
Type : CommentIn
Informations de copyright
Copyright © 2021 by the American Society of Nephrology.
Références
Okada T, Nagao T, Matsumoto H, Nagaoka Y, Wada T, Nakao T: Histological predictors for renal prognosis in diabetic nephropathy in diabetes mellitus type 2 patients with overt proteinuria. Nephrology (Carlton) 17: 68–75, 2012 21933307
Tervaert TW, Mooyaart AL, Amann K, Cohen AH, Cook HT, Drachenberg CB, et al.; Renal Pathology Society: Pathologic classification of diabetic nephropathy. J Am Soc Nephrol 21: 556–563, 2010 20167701
Weening JJ, D’Agati VD, Schwartz MM, Seshan SV, Alpers CE, Appel GB, et al.; International Society of Nephrology Working Group on the Classification of Lupus Nephritis; Renal Pathology Society Working Group on the Classification of Lupus Nephritis: The classification of glomerulonephritis in systemic lupus erythematosus revisited [published correction appears in Kidney Int 65: 1132, 2004]. Kidney Int 65: 521–530, 2004 14717922
Roberts IS, Cook HT, Troyanov S, Alpers CE, Amore A, Barratt J, et al.; Working Group of the International IgA Nephropathy Network and the Renal Pathology Society: The Oxford classification of IgA nephropathy: Pathology definitions, correlations, and reproducibility. Kidney Int 76: 546–556, 2009 19571790
Sethi S, D’Agati VD, Nast CC, Fogo AB, De Vriese AS, Markowitz GS, et al.: A proposal for standardized grading of chronic changes in native kidney biopsy specimens. Kidney Int 91: 787–789, 2017 28314581
Srivastava A, Palsson R, Kaze AD, Chen ME, Palacios P, Sabbisetti V, et al.: The prognostic value of histopathologic lesions in native kidney biopsy specimens: Results from the Boston kidney biopsy cohort study. J Am Soc Nephrol 29: 2213–2224, 2018 29866798
Cosio FG, El Ters M, Cornell LD, Schinstock CA, Stegall MD: Changing kidney allograft histology early posttransplant: Prognostic implications of 1-year protocol biopsies. Am J Transplant 16: 194–203, 2016 26274817
Serón D, Moreso F: Protocol biopsies in renal transplantation: Prognostic value of structural monitoring. Kidney Int 72: 690–697, 2007 17597702
Myllymäki J, Saha H, Mustonen J, Helin H, Pasternack A: IgM nephropathy: Clinical picture and long-term prognosis. Am J Kidney Dis 41: 343–350, 2003 12552495
Bohle A, Wehrmann M, Bogenschütz O, Batz C, Müller CA, Müller GA: The pathogenesis of chronic renal failure in diabetic nephropathy. Investigation of 488 cases of diabetic glomerulosclerosis. Pathol Res Pract 187: 251–259, 1991 2068008
Farris AB, Chan S, Climenhaga J, Adam B, Bellamy CO, Serón D, et al.: Banff fibrosis study: Multicenter visual assessment and computerized analysis of interstitial fibrosis in kidney biopsies. Am J Transplant 14: 897–907, 2014 24712330
Snoeijs MG, Boonstra LA, Buurman WA, Goldschmeding R, van Suylen RJ, van Heurn LW, et al.: Histological assessment of pre-transplant kidney biopsies is reproducible and representative. Histopathology 56: 198–202, 2010 20102398
Grootscholten C, Bajema IM, Florquin S, Steenbergen EJ, Peutz-Kootstra CJ, Goldschmeding R, et al.: Interobserver agreement of scoring of histopathological characteristics and classification of lupus nephritis. Nephrol Dial Transplant 23: 223–230, 2008 17981886
Gough J, Rush D, Jeffery J, Nickerson P, McKenna R, Solez K, et al.: Reproducibility of the Banff schema in reporting protocol biopsies of stable renal allografts. Nephrol Dial Transplant 17: 1081–1084, 2002 12032200
Furness PN, Taub N, Assmann KJ, Banfi G, Cosyns JP, Dorman AM, et al.: International variation in histologic grading is large, and persistent feedback does not improve reproducibility. Am J Surg Pathol 27: 805–810, 2003 12766585
Farris AB, Adams CD, Brousaides N, Della Pelle PA, Collins AB, Moradi E, et al.: Morphometric and visual evaluation of fibrosis in renal biopsies. J Am Soc Nephrol 22: 176–186, 2011 21115619
Servais A, Meas-Yedid V, Noël LH, Martinez F, Panterne C, Kreis H, et al.: Interstitial fibrosis evolution on early sequential screening renal allograft biopsies using quantitative image analysis. Am J Transplant 11: 1456–1463, 2011 21672152
Hermsen M, de Bel T, den Boer M, Steenbergen EJ, Kers J, Florquin S, et al.: Deep learning-based histopathologic assessment of kidney tissue. J Am Soc Nephrol 30: 1968–1979, 2019 31488607
Kolachalama VB, Singh P, Lin CQ, Mun D, Belghasem ME, Henderson JM, et al.: Association of pathological fibrosis with renal survival using deep neural networks. Kidney Int Rep 3: 464–475, 2018 29725651
Bukowy JD, Dayton A, Cloutier D, Manis AD, Staruschenko A, Lombard JH, et al.: Region-based convolutional neural nets for localization of glomeruli in trichrome-stained whole kidney sections. J Am Soc Nephrol 29: 2081–2088, 2018 29921718
Pedraza A, Gallego J, Lopez S, Gonzalez L, Laurinavicius A, Bueno G: Glomerulus classification with convolutional neural networks. In: Medical Image Understanding and Analysis: 21st Annual Conference, MIUA 2017, Edinburgh, United Kingdom, July 11–13, 2017, Proceedings, edited by Valdés Hernández M, González-Castro V. Cham, Switzerland, Springer International Publishing, 2017, pp 839–849
Lutnick B, Ginley B, Govind D, McGarry SD, LaViolette PS, Yacoub R, et al.: An integrated iterative annotation technique for easing neural network training in medical image analysis. Nat Mach Intell 1: 112–119, 2019 31187088
Ginley B, Lutnick B, Jen KY, Fogo AB, Jain S, Rosenberg A, et al.: Computational segmentation and classification of diabetic glomerulosclerosis. J Am Soc Nephrol 30: 1953–1967, 2019 31488606
Wang Z, Ji S: Smoothed dilated convolutions for improved dense prediction. Presented at the 2018 24th Association for Computing Machinery Special Interest Group on Knowledge Discovery in Data International Conference on Knowledge Discovery and Data Mining, London, August 19–23, 2018
Abadi M, Agarwal A, Barham P, Brevdo E, Chen Z, Citro C, et al.: Tensorflow: Large-scale machine learning on heterogeneous distributed systems. arXiv. 1603.04467 (Preprint posted March 14, 2016)
Lin TY, Maire M, Belongie S, Hays J, Perona P, Ramanan D, et al.: Microsoft COCO: Common objects in context. Lect Notes Comput Sci 8693: 740–755, 2014
Russakovsky O, Deng J, Su H, Krause J, Satheesh S, Ma S, et al.: ImageNet large scale visual recognition challenge. Int J Comput Vis 115: 211–252, 2015
Matthews BW: Comparison of the predicted and observed secondary structure of T4 phage lysozyme. Biochim Biophys Acta 405: 442–451, 1975 1180967
Powers DMW: Evaluation: From precision, recall and F-measure to ROC, informedness, markedness and correlation. J Mach Learn Technol 2: 37–63, 2011
Chicco D, Jurman G: The advantages of the Matthews correlation coefficient (MCC) over F1 score and accuracy in binary classification evaluation. BMC Genomics 21: 6, 2020 31898477
Chan YH: Biostatistics 104: Correlational analysis. Singapore Med J 44: 614–619, 2003 14770254
Akoglu H: User’s guide to correlation coefficients. Turk J Emerg Med 18: 91–93, 2018 30191186
Gutman DA, Khalilia M, Lee S, Nalisnik M, Mullen Z, Beezley J, et al.: The digital slide archive: A software platform for management, integration, and analysis of histology for cancer research. Cancer Res 77: e75–e78, 2017 29092945
Meyer F, Beucher S: Morphological segmentation. J Vis Commun Image Represent 1: 21–46, 1990
Serge Beucher FM: The morphological approach to segmentation: The watershed transformation. In: Mathematical Morphology in Image Processing, edited by Dougherty ED, Boca Raton, FL, CRC Press, 1993, pp 433–481
Zaiontz C: Real statistics resource pack software (release 7.2), 2020. Available at: https://www.real-statistics.com/ . Accessed January 7, 2020
Landis JR, Koch GG: The measurement of observer agreement for categorical data. Biometrics 33: 159–174, 1977 843571
Meng XL, Rosenthal R, Rubin DB: Comparing correlated correlation-coefficients. Psychol Bull 111: 172–175, 1992
Diedenhofen B, Musch J: cocor: A comprehensive solution for the statistical comparison of correlations [published correction appears in PLoS One 10: e0131499, 2015 10.1371/journal.pone.0131499]. PLoS One 10: e0121945, 2015
R Core Team: A language and environment for statistical computing, 2020. Available at: https://www.R-project.org/ . Accessed January 7, 2020
Hosmer DW, Lemeshow S, Sturdivant RX: Applied Logistic Regression, 3rd Ed., Hoboken, NJ, Wiley, 2013
Jen KY, Olson JL, Brodsky S, Zhou XJ, Nadasdy T, Laszik ZG: Reliability of whole slide images as a diagnostic modality for renal allograft biopsies. Hum Pathol 44: 888–894, 2013 23199528
Ozluk Y, Blanco PL, Mengel M, Solez K, Halloran PF, Sis B: Superiority of virtual microscopy versus light microscopy in transplantation pathology. Clin Transplant 26: 336–344, 2012 21955102
Austin HA 3rd, Boumpas DT, Vaughan EM, Balow JE: Predicting renal outcomes in severe lupus nephritis: Contributions of clinical and histologic data. Kidney Int 45: 544–550, 1994 8164443
Roufosse C, Simmonds N, Clahsen-van Groningen M, Haas M, Henriksen KJ, Horsfield C, et al.: A 2018 reference guide to the Banff classification of renal allograft pathology. Transplantation 102: 1795–1814, 2018 30028786
Farris AB, Alpers CE. What is the best way to measure renal fibrosis?: A pathologist’s perspective. Kidney Int Suppl (2011) 4: 9–15, 2014
University at Buffalo: Center for Computational Research. Available at: http://www.buffalo.edu/ccr.html