An Artificial Intelligence Generated Automated Algorithm to Measure Total Kidney Volume in ADPKD

Introduction: Accurate tools to inform individual prognosis in patients with autosomal dominant polycystic kidney disease (ADPKD) are lacking. Here, we report an artificial intelligence (AI)-generated method for routinely measuring total kidney volume (TKV). Methods: An ensemble U-net algorithm was...

Full description

Saved in:
Bibliographic Details
Main Authors: Jonathan Taylor, Richard Thomas, Peter Metherall, Marieke van Gastel, Emilie Cornec-Le Gall, Anna Caroli, Monica Furlano, Nathalie Demoulin, Olivier Devuyst, Jean Winterbottom, Roser Torra, Norberto Perico, Yannick Le Meur, Sebastian Schoenherr, Lukas Forer, Ron T. Gansevoort, Roslyn J. Simms, Albert C.M. Ong
Format: Article
Language:English
Published: Elsevier 2024-02-01
Series:Kidney International Reports
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2468024923015711
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849693165391446016
author Jonathan Taylor
Richard Thomas
Peter Metherall
Marieke van Gastel
Emilie Cornec-Le Gall
Anna Caroli
Monica Furlano
Nathalie Demoulin
Olivier Devuyst
Jean Winterbottom
Roser Torra
Norberto Perico
Yannick Le Meur
Sebastian Schoenherr
Lukas Forer
Ron T. Gansevoort
Roslyn J. Simms
Albert C.M. Ong
author_facet Jonathan Taylor
Richard Thomas
Peter Metherall
Marieke van Gastel
Emilie Cornec-Le Gall
Anna Caroli
Monica Furlano
Nathalie Demoulin
Olivier Devuyst
Jean Winterbottom
Roser Torra
Norberto Perico
Yannick Le Meur
Sebastian Schoenherr
Lukas Forer
Ron T. Gansevoort
Roslyn J. Simms
Albert C.M. Ong
author_sort Jonathan Taylor
collection DOAJ
description Introduction: Accurate tools to inform individual prognosis in patients with autosomal dominant polycystic kidney disease (ADPKD) are lacking. Here, we report an artificial intelligence (AI)-generated method for routinely measuring total kidney volume (TKV). Methods: An ensemble U-net algorithm was created using the nnUNet approach. The training and internal cross-validation cohort consisted of all 1.5T magnetic resonance imaging (MRI) data acquired using 5 different MRI scanners (454 kidneys, 227 scans) in the CYSTic consortium, which was first manually segmented by a single human operator. As an independent validation cohort, we utilized 48 sequential clinical MRI scans with reference results of manual segmentation acquired by 6 individual analysts at a single center. The tool was then implemented for clinical use and its performance analyzed. Results: The training or internal validation cohort was younger (mean age 44.0 vs. 51.5 years) and the female-to-male ratio higher (1.2 vs. 0.94) compared to the clinical validation cohort. The majority of CYSTic patients had PKD1 mutations (79%) and typical disease (Mayo Imaging class 1, 86%). The median DICE score on the clinical validation data set between the algorithm and human analysts was 0.96 for left and right kidneys with a median TKV error of −1.8%. The time taken to manually segment kidneys in the CYSTic data set was 56 (±28) minutes, whereas manual corrections of the algorithm output took 8.5 (±9.2) minutes per scan. Conclusion: Our AI-based algorithm demonstrates performance comparable to manual segmentation. Its rapidity and precision in real-world clinical cases demonstrate its suitability for clinical application.
format Article
id doaj-art-8d207eef50c940349d6429e450b9a25e
institution DOAJ
issn 2468-0249
language English
publishDate 2024-02-01
publisher Elsevier
record_format Article
series Kidney International Reports
spelling doaj-art-8d207eef50c940349d6429e450b9a25e2025-08-20T03:20:30ZengElsevierKidney International Reports2468-02492024-02-019224925610.1016/j.ekir.2023.10.029An Artificial Intelligence Generated Automated Algorithm to Measure Total Kidney Volume in ADPKDJonathan Taylor0Richard Thomas1Peter Metherall2Marieke van Gastel3Emilie Cornec-Le Gall4Anna Caroli5Monica Furlano6Nathalie Demoulin7Olivier Devuyst8Jean Winterbottom9Roser Torra10Norberto Perico11Yannick Le Meur12Sebastian Schoenherr13Lukas Forer14Ron T. Gansevoort15Roslyn J. Simms16Albert C.M. Ong173DLab, Medical Imaging Medical Physics, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK3DLab, Medical Imaging Medical Physics, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK3DLab, Medical Imaging Medical Physics, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UKDepartment of Nephrology, University Medical Centre Groningen, Groningen, The NetherlandsUniversity Brest, Inserm, UMR 1078, GGB, CHU Brest, F-29200 Brest, FranceIstituto di Ricerche Farmacologiche Mario Negri IRCCS, Bergamo, ItalyInherited Kidney Disorders, Nephrology Department, Fundació Puigvert, IIB Sant Pau, Universitat Autònoma de Barcelona, Barcelona, SpainCliniques Universitaires Saint-Luc, UCLouvain Medical School, Brussels, BelgiumCliniques Universitaires Saint-Luc, UCLouvain Medical School, Brussels, BelgiumAcademic Nephrology, Division of Clinical Medicine, School of Medicine and Population Health, Faculty of Health, University of Sheffield, Sheffield, UK; Sheffield Kidney Institute, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UKInherited Kidney Disorders, Nephrology Department, Fundació Puigvert, IIB Sant Pau, Universitat Autònoma de Barcelona, Barcelona, SpainIstituto di Ricerche Farmacologiche Mario Negri IRCCS, Bergamo, ItalyUniversity Brest, Inserm, UMR 1227, LBAI, CHU Brest, F-29200 Brest, FranceInstitute of Genetic Epidemiology, Department of Genetics and Pharmacology, Medical University of Innsbruck, AustriaInstitute of Genetic Epidemiology, Department of Genetics and Pharmacology, Medical University of Innsbruck, AustriaDepartment of Nephrology, University Medical Centre Groningen, Groningen, The NetherlandsAcademic Nephrology, Division of Clinical Medicine, School of Medicine and Population Health, Faculty of Health, University of Sheffield, Sheffield, UK; Sheffield Kidney Institute, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK; Correspondence: Roslyn J. Simms or Albert C.M. Ong, Academic Nephrology, Division of Clinical Medicine, School of Medicine and Population Health, University of Sheffield, Beech Hill Road, Sheffield S10 2RX, UK.Academic Nephrology, Division of Clinical Medicine, School of Medicine and Population Health, Faculty of Health, University of Sheffield, Sheffield, UK; Sheffield Kidney Institute, Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK; Correspondence: Roslyn J. Simms or Albert C.M. Ong, Academic Nephrology, Division of Clinical Medicine, School of Medicine and Population Health, University of Sheffield, Beech Hill Road, Sheffield S10 2RX, UK.Introduction: Accurate tools to inform individual prognosis in patients with autosomal dominant polycystic kidney disease (ADPKD) are lacking. Here, we report an artificial intelligence (AI)-generated method for routinely measuring total kidney volume (TKV). Methods: An ensemble U-net algorithm was created using the nnUNet approach. The training and internal cross-validation cohort consisted of all 1.5T magnetic resonance imaging (MRI) data acquired using 5 different MRI scanners (454 kidneys, 227 scans) in the CYSTic consortium, which was first manually segmented by a single human operator. As an independent validation cohort, we utilized 48 sequential clinical MRI scans with reference results of manual segmentation acquired by 6 individual analysts at a single center. The tool was then implemented for clinical use and its performance analyzed. Results: The training or internal validation cohort was younger (mean age 44.0 vs. 51.5 years) and the female-to-male ratio higher (1.2 vs. 0.94) compared to the clinical validation cohort. The majority of CYSTic patients had PKD1 mutations (79%) and typical disease (Mayo Imaging class 1, 86%). The median DICE score on the clinical validation data set between the algorithm and human analysts was 0.96 for left and right kidneys with a median TKV error of −1.8%. The time taken to manually segment kidneys in the CYSTic data set was 56 (±28) minutes, whereas manual corrections of the algorithm output took 8.5 (±9.2) minutes per scan. Conclusion: Our AI-based algorithm demonstrates performance comparable to manual segmentation. Its rapidity and precision in real-world clinical cases demonstrate its suitability for clinical application.http://www.sciencedirect.com/science/article/pii/S2468024923015711ADPKDartificial intelligencemachine learningmagnetic resonance imagingtotal kidney volume
spellingShingle Jonathan Taylor
Richard Thomas
Peter Metherall
Marieke van Gastel
Emilie Cornec-Le Gall
Anna Caroli
Monica Furlano
Nathalie Demoulin
Olivier Devuyst
Jean Winterbottom
Roser Torra
Norberto Perico
Yannick Le Meur
Sebastian Schoenherr
Lukas Forer
Ron T. Gansevoort
Roslyn J. Simms
Albert C.M. Ong
An Artificial Intelligence Generated Automated Algorithm to Measure Total Kidney Volume in ADPKD
Kidney International Reports
ADPKD
artificial intelligence
machine learning
magnetic resonance imaging
total kidney volume
title An Artificial Intelligence Generated Automated Algorithm to Measure Total Kidney Volume in ADPKD
title_full An Artificial Intelligence Generated Automated Algorithm to Measure Total Kidney Volume in ADPKD
title_fullStr An Artificial Intelligence Generated Automated Algorithm to Measure Total Kidney Volume in ADPKD
title_full_unstemmed An Artificial Intelligence Generated Automated Algorithm to Measure Total Kidney Volume in ADPKD
title_short An Artificial Intelligence Generated Automated Algorithm to Measure Total Kidney Volume in ADPKD
title_sort artificial intelligence generated automated algorithm to measure total kidney volume in adpkd
topic ADPKD
artificial intelligence
machine learning
magnetic resonance imaging
total kidney volume
url http://www.sciencedirect.com/science/article/pii/S2468024923015711
work_keys_str_mv AT jonathantaylor anartificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT richardthomas anartificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT petermetherall anartificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT mariekevangastel anartificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT emiliecorneclegall anartificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT annacaroli anartificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT monicafurlano anartificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT nathaliedemoulin anartificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT olivierdevuyst anartificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT jeanwinterbottom anartificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT rosertorra anartificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT norbertoperico anartificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT yannicklemeur anartificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT sebastianschoenherr anartificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT lukasforer anartificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT rontgansevoort anartificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT roslynjsimms anartificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT albertcmong anartificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT jonathantaylor artificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT richardthomas artificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT petermetherall artificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT mariekevangastel artificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT emiliecorneclegall artificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT annacaroli artificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT monicafurlano artificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT nathaliedemoulin artificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT olivierdevuyst artificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT jeanwinterbottom artificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT rosertorra artificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT norbertoperico artificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT yannicklemeur artificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT sebastianschoenherr artificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT lukasforer artificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT rontgansevoort artificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT roslynjsimms artificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd
AT albertcmong artificialintelligencegeneratedautomatedalgorithmtomeasuretotalkidneyvolumeinadpkd