Enhancing individual glomerular filtration rate assessment: can we trust the equation? Development and validation of machine learning models to assess the trustworthiness of estimated GFR compared to measured GFR
Abstract Background Creatinine-based estimated glomerular filtration rate (eGFR) equations are widely used in clinical practice but exhibit inherent limitations. On the other side, measuring GFR is time consuming and not available in routine clinical practice. We developed and validated machine lear...
Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2025-01-01
|
| Series: | BMC Nephrology |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s12882-025-03972-0 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832571962557202432 |
|---|---|
| author | Antoine Lanot Anna Akesson Felipe Kenji Nakano Celine Vens Jonas Björk Ulf Nyman Anders Grubb Per-Ola Sundin Björn O. Eriksen Toralf Melsom Andrew D. Rule Ulla Berg Karin Littmann Kajsa Åsling-Monemi Magnus Hansson Anders Larsson Marie Courbebaisse Laurence Dubourg Lionel Couzi Francois Gaillard Cyril Garrouste Lola Jacquemont Nassim Kamar Christophe Legendre Lionel Rostaing Natalie Ebert Elke Schaeffner Arend Bökenkamp Christophe Mariat Hans Pottel Pierre Delanaye |
| author_facet | Antoine Lanot Anna Akesson Felipe Kenji Nakano Celine Vens Jonas Björk Ulf Nyman Anders Grubb Per-Ola Sundin Björn O. Eriksen Toralf Melsom Andrew D. Rule Ulla Berg Karin Littmann Kajsa Åsling-Monemi Magnus Hansson Anders Larsson Marie Courbebaisse Laurence Dubourg Lionel Couzi Francois Gaillard Cyril Garrouste Lola Jacquemont Nassim Kamar Christophe Legendre Lionel Rostaing Natalie Ebert Elke Schaeffner Arend Bökenkamp Christophe Mariat Hans Pottel Pierre Delanaye |
| author_sort | Antoine Lanot |
| collection | DOAJ |
| description | Abstract Background Creatinine-based estimated glomerular filtration rate (eGFR) equations are widely used in clinical practice but exhibit inherent limitations. On the other side, measuring GFR is time consuming and not available in routine clinical practice. We developed and validated machine learning models to assess the trustworthiness (i.e. the ability of equations to estimate measured GFR (mGFR) within 10%, 20% or 30%) of the European Kidney Function Consortium (EKFC) equation at the individual level. Methods This observational study used data from European and US cohorts, comprising 22,343 participants of all ages with available mGFR results. Four machine learning and two traditional logistic regression models were trained on a cohort of 9,202 participants to predict the likelihood of the EKFC creatinine-derived eGFR falling within 30% (p30), 20% (p20) or 10% (p10) of the mGFR value. The algorithms were internally and then externally validated on cohorts of respectively 3,034 and 10,107 participants. The predictors included in the models were creatinine, age, sex, height, weight, and EKFC. Results The random forest model was the most robust model. In the external validation cohort, the model achieved an area under the curve of 0.675 (95%CI 0.660;0.690) and an accuracy of 0.716 (95%CI 0.707;0.725) for the P30 criterion. Sensitivity was 0.756 (95%CI 0.747;0.765) and specificity was 0.485 (95%CI 0.460; 0.511) at the 80% probability level that EKFC falls within 30% of mGFR. At the population level, the PPV of this machine learning model was 89.5%, higher than the EKFC P30 of 85.2%. A free web-application was developed to allow the physician to assess the trustworthiness of EKFC at the individual level. Conclusions A strategy using machine learning model marginally improves the trustworthiness of GFR estimation at the population level. An additional value of this approach lies in its ability to provide assessments at the individual level. |
| format | Article |
| id | doaj-art-db9917c0485d4ac8b46dea7eac950a31 |
| institution | Kabale University |
| issn | 1471-2369 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | BMC |
| record_format | Article |
| series | BMC Nephrology |
| spelling | doaj-art-db9917c0485d4ac8b46dea7eac950a312025-02-02T12:12:30ZengBMCBMC Nephrology1471-23692025-01-0126111310.1186/s12882-025-03972-0Enhancing individual glomerular filtration rate assessment: can we trust the equation? Development and validation of machine learning models to assess the trustworthiness of estimated GFR compared to measured GFRAntoine Lanot0Anna Akesson1Felipe Kenji Nakano2Celine Vens3Jonas Björk4Ulf NymanAnders Grubb5Per-Ola Sundin6Björn O. Eriksen7Toralf Melsom8Andrew D. Rule9Ulla Berg10Karin Littmann11Kajsa Åsling-Monemi12Magnus Hansson13Anders Larsson14Marie Courbebaisse15Laurence Dubourg16Lionel Couzi17Francois Gaillard18Cyril Garrouste19Lola Jacquemont20Nassim Kamar21Christophe Legendre22Lionel Rostaing23Natalie Ebert24Elke Schaeffner25Arend Bökenkamp26Christophe Mariat27Hans Pottel28Pierre Delanaye29Normandie Univ, UNICAEN, CHU de Caen Normandie, NéphrologieSkane University Hospital, Clinical Studies Sweden Forum SouthDepartment of Public Health and Primary CareDepartment of Public Health and Primary CareLund UniversityDepartment of Clinical Chemistry and Pharmacology, Laboratory Lund UniversityKarla Healthcare Centre, Faculty of Medicine and Health, Örebro UniversityUniversity Hospital of North Norway (UNN)University Hospital of North Norway (UNN)Division of Nephrology and Hypertension, Mayo ClinicDepartment of Clinical Science, Intervention and Technology, Division of Pediatrics, Karolinska Institutet, Karolinska University. Hospital HuddingeDepartment of Medicine Huddinge, Karolinska Institutet, C2:91 Karolinska University HospitalBarnnjursektionen K 88, Astrid Lindgrens Barnsjukhus, Karolinska University HospitalDepartment of Clinical Chemistry, C1:74 Huddinge, Karolinska University HospitalClinical Chemistry and Pharmacology, Entrance 61, 2Nd Floor, Akademiska HospitalService de Physiologie-Explorations, Fonctionnelles Renales Hopital Europeen Georges PompidouExploration Fonctionnelle Renale Pavillon P, Hopital Edouard HerriotCHU de Bordeaux, Nephrologie-Transplantation-Dialyse, Hopital Pellegrin, Universite de Bordeaux, Place Amelie Raba LeonRenal Transplantation Department, Assistance Publique–Hopitaux de Paris (AP-HP), Hopital BichatDepartment of Nephrology, Clermont-Ferrand University HospitalService de Nephrologie Et Immunologie Clinique, CHU de NantesDepartment of Nephrology and Organ Transplantation, CHU RangueilTransplantation Renale, Hopital NeckerService de Nephrologie, Hemodialyse, Aphereses Et Transplantation Renale, Hopital Michallon, Centre Hospitalier Universitaire Grenoble-AlpesInstitute of Public Health, Charité. Universitätsmedizin BerlinInstitute of Public Health, Charité. Universitätsmedizin BerlinAmsterdam UMC, Vrije UniversiteitService de Nephrologie, Dialyse Et Transplantation Renale, Hopital Nord, CHU de Saint-EtienneDepartment of Public Health and Primary CareDepartment of Nephrology-Dialysis-Transplantation, University of Liège, CHU Sart TilmanAbstract Background Creatinine-based estimated glomerular filtration rate (eGFR) equations are widely used in clinical practice but exhibit inherent limitations. On the other side, measuring GFR is time consuming and not available in routine clinical practice. We developed and validated machine learning models to assess the trustworthiness (i.e. the ability of equations to estimate measured GFR (mGFR) within 10%, 20% or 30%) of the European Kidney Function Consortium (EKFC) equation at the individual level. Methods This observational study used data from European and US cohorts, comprising 22,343 participants of all ages with available mGFR results. Four machine learning and two traditional logistic regression models were trained on a cohort of 9,202 participants to predict the likelihood of the EKFC creatinine-derived eGFR falling within 30% (p30), 20% (p20) or 10% (p10) of the mGFR value. The algorithms were internally and then externally validated on cohorts of respectively 3,034 and 10,107 participants. The predictors included in the models were creatinine, age, sex, height, weight, and EKFC. Results The random forest model was the most robust model. In the external validation cohort, the model achieved an area under the curve of 0.675 (95%CI 0.660;0.690) and an accuracy of 0.716 (95%CI 0.707;0.725) for the P30 criterion. Sensitivity was 0.756 (95%CI 0.747;0.765) and specificity was 0.485 (95%CI 0.460; 0.511) at the 80% probability level that EKFC falls within 30% of mGFR. At the population level, the PPV of this machine learning model was 89.5%, higher than the EKFC P30 of 85.2%. A free web-application was developed to allow the physician to assess the trustworthiness of EKFC at the individual level. Conclusions A strategy using machine learning model marginally improves the trustworthiness of GFR estimation at the population level. An additional value of this approach lies in its ability to provide assessments at the individual level.https://doi.org/10.1186/s12882-025-03972-0Glomerular filtration rateChronic kidney diseaseCreatinineMachine learningRandom forest |
| spellingShingle | Antoine Lanot Anna Akesson Felipe Kenji Nakano Celine Vens Jonas Björk Ulf Nyman Anders Grubb Per-Ola Sundin Björn O. Eriksen Toralf Melsom Andrew D. Rule Ulla Berg Karin Littmann Kajsa Åsling-Monemi Magnus Hansson Anders Larsson Marie Courbebaisse Laurence Dubourg Lionel Couzi Francois Gaillard Cyril Garrouste Lola Jacquemont Nassim Kamar Christophe Legendre Lionel Rostaing Natalie Ebert Elke Schaeffner Arend Bökenkamp Christophe Mariat Hans Pottel Pierre Delanaye Enhancing individual glomerular filtration rate assessment: can we trust the equation? Development and validation of machine learning models to assess the trustworthiness of estimated GFR compared to measured GFR BMC Nephrology Glomerular filtration rate Chronic kidney disease Creatinine Machine learning Random forest |
| title | Enhancing individual glomerular filtration rate assessment: can we trust the equation? Development and validation of machine learning models to assess the trustworthiness of estimated GFR compared to measured GFR |
| title_full | Enhancing individual glomerular filtration rate assessment: can we trust the equation? Development and validation of machine learning models to assess the trustworthiness of estimated GFR compared to measured GFR |
| title_fullStr | Enhancing individual glomerular filtration rate assessment: can we trust the equation? Development and validation of machine learning models to assess the trustworthiness of estimated GFR compared to measured GFR |
| title_full_unstemmed | Enhancing individual glomerular filtration rate assessment: can we trust the equation? Development and validation of machine learning models to assess the trustworthiness of estimated GFR compared to measured GFR |
| title_short | Enhancing individual glomerular filtration rate assessment: can we trust the equation? Development and validation of machine learning models to assess the trustworthiness of estimated GFR compared to measured GFR |
| title_sort | enhancing individual glomerular filtration rate assessment can we trust the equation development and validation of machine learning models to assess the trustworthiness of estimated gfr compared to measured gfr |
| topic | Glomerular filtration rate Chronic kidney disease Creatinine Machine learning Random forest |
| url | https://doi.org/10.1186/s12882-025-03972-0 |
| work_keys_str_mv | AT antoinelanot enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT annaakesson enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT felipekenjinakano enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT celinevens enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT jonasbjork enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT ulfnyman enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT andersgrubb enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT perolasundin enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT bjornoeriksen enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT toralfmelsom enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT andrewdrule enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT ullaberg enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT karinlittmann enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT kajsaaslingmonemi enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT magnushansson enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT anderslarsson enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT mariecourbebaisse enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT laurencedubourg enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT lionelcouzi enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT francoisgaillard enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT cyrilgarrouste enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT lolajacquemont enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT nassimkamar enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT christophelegendre enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT lionelrostaing enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT natalieebert enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT elkeschaeffner enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT arendbokenkamp enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT christophemariat enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT hanspottel enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr AT pierredelanaye enhancingindividualglomerularfiltrationrateassessmentcanwetrusttheequationdevelopmentandvalidationofmachinelearningmodelstoassessthetrustworthinessofestimatedgfrcomparedtomeasuredgfr |