Predicting early risk of chronic kidney disease in cats using routine clinical laboratory tests and machine learning
Abstract Background Advanced machine learning methods combined with large sets of health screening data provide opportunities for diagnostic value in human and veterinary medicine. Hypothesis/Objectives To derive a model to predict the risk of cats developing chronic kidney disease (CKD) using data...
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2019-11-01
|
| Series: | Journal of Veterinary Internal Medicine |
| Subjects: | |
| Online Access: | https://doi.org/10.1111/jvim.15623 |
| _version_ | 1819172566426714112 |
|---|---|
| author | Richard Bradley Ilias Tagkopoulos Minseung Kim Yiannis Kokkinos Theodoros Panagiotakos James Kennedy Geert De Meyer Phillip Watson Jonathan Elliott |
| author_facet | Richard Bradley Ilias Tagkopoulos Minseung Kim Yiannis Kokkinos Theodoros Panagiotakos James Kennedy Geert De Meyer Phillip Watson Jonathan Elliott |
| author_sort | Richard Bradley |
| collection | DOAJ |
| description | Abstract Background Advanced machine learning methods combined with large sets of health screening data provide opportunities for diagnostic value in human and veterinary medicine. Hypothesis/Objectives To derive a model to predict the risk of cats developing chronic kidney disease (CKD) using data from electronic health records (EHRs) collected during routine veterinary practice. Animals A total of 106 251 cats that attended Banfield Pet Hospitals between January 1, 1995, and December 31, 2017. Methods Longitudinal EHRs from Banfield Pet Hospitals were extracted and randomly split into 2 parts. The first 67% of the data were used to build a prediction model, which included feature selection and identification of the optimal neural network type and architecture. The remaining unseen EHRs were used to evaluate the model performance. Results The final model was a recurrent neural network (RNN) with 4 features (creatinine, blood urea nitrogen, urine specific gravity, and age). When predicting CKD near the point of diagnosis, the model displayed a sensitivity of 90.7% and a specificity of 98.9%. Model sensitivity decreased when predicting the risk of CKD with a longer horizon, having 63.0% sensitivity 1 year before diagnosis and 44.2% 2 years before diagnosis, but with specificity remaining around 99%. Conclusions and clinical importance The use of models based on machine learning can support veterinary decision making by improving early identification of CKD. |
| first_indexed | 2024-12-22T20:09:13Z |
| format | Article |
| id | doaj.art-f4f8f76e776146099467a078cb594ead |
| institution | Directory Open Access Journal |
| issn | 0891-6640 1939-1676 |
| language | English |
| last_indexed | 2024-12-22T20:09:13Z |
| publishDate | 2019-11-01 |
| publisher | Wiley |
| record_format | Article |
| series | Journal of Veterinary Internal Medicine |
| spelling | doaj.art-f4f8f76e776146099467a078cb594ead2022-12-21T18:14:05ZengWileyJournal of Veterinary Internal Medicine0891-66401939-16762019-11-013362644265610.1111/jvim.15623Predicting early risk of chronic kidney disease in cats using routine clinical laboratory tests and machine learningRichard Bradley0Ilias Tagkopoulos1Minseung Kim2Yiannis Kokkinos3Theodoros Panagiotakos4James Kennedy5Geert De Meyer6Phillip Watson7Jonathan Elliott8WALTHAM® Centre for Pet Nutrition, Freeby Lane, Waltham on the Wolds Leicestershire United KingdomDepartment of Computer Science and Genome Center University of California Davis CaliforniaProcess Integration and Predictive Analytics, PIPA LLC Davis CaliforniaProcess Integration and Predictive Analytics, PIPA LLC Davis CaliforniaProcess Integration and Predictive Analytics, PIPA LLC Davis CaliforniaMars, Incorporated McLean VirginiaWALTHAM® Centre for Pet Nutrition, Freeby Lane, Waltham on the Wolds Leicestershire United KingdomWALTHAM® Centre for Pet Nutrition, Freeby Lane, Waltham on the Wolds Leicestershire United KingdomDepartment of Comparative Biomedical Sciences Royal Veterinary College London United KingdomAbstract Background Advanced machine learning methods combined with large sets of health screening data provide opportunities for diagnostic value in human and veterinary medicine. Hypothesis/Objectives To derive a model to predict the risk of cats developing chronic kidney disease (CKD) using data from electronic health records (EHRs) collected during routine veterinary practice. Animals A total of 106 251 cats that attended Banfield Pet Hospitals between January 1, 1995, and December 31, 2017. Methods Longitudinal EHRs from Banfield Pet Hospitals were extracted and randomly split into 2 parts. The first 67% of the data were used to build a prediction model, which included feature selection and identification of the optimal neural network type and architecture. The remaining unseen EHRs were used to evaluate the model performance. Results The final model was a recurrent neural network (RNN) with 4 features (creatinine, blood urea nitrogen, urine specific gravity, and age). When predicting CKD near the point of diagnosis, the model displayed a sensitivity of 90.7% and a specificity of 98.9%. Model sensitivity decreased when predicting the risk of CKD with a longer horizon, having 63.0% sensitivity 1 year before diagnosis and 44.2% 2 years before diagnosis, but with specificity remaining around 99%. Conclusions and clinical importance The use of models based on machine learning can support veterinary decision making by improving early identification of CKD.https://doi.org/10.1111/jvim.15623artificial neural networkcomputer modelfelinemachine learningrenal |
| spellingShingle | Richard Bradley Ilias Tagkopoulos Minseung Kim Yiannis Kokkinos Theodoros Panagiotakos James Kennedy Geert De Meyer Phillip Watson Jonathan Elliott Predicting early risk of chronic kidney disease in cats using routine clinical laboratory tests and machine learning Journal of Veterinary Internal Medicine artificial neural network computer model feline machine learning renal |
| title | Predicting early risk of chronic kidney disease in cats using routine clinical laboratory tests and machine learning |
| title_full | Predicting early risk of chronic kidney disease in cats using routine clinical laboratory tests and machine learning |
| title_fullStr | Predicting early risk of chronic kidney disease in cats using routine clinical laboratory tests and machine learning |
| title_full_unstemmed | Predicting early risk of chronic kidney disease in cats using routine clinical laboratory tests and machine learning |
| title_short | Predicting early risk of chronic kidney disease in cats using routine clinical laboratory tests and machine learning |
| title_sort | predicting early risk of chronic kidney disease in cats using routine clinical laboratory tests and machine learning |
| topic | artificial neural network computer model feline machine learning renal |
| url | https://doi.org/10.1111/jvim.15623 |
| work_keys_str_mv | AT richardbradley predictingearlyriskofchronickidneydiseaseincatsusingroutineclinicallaboratorytestsandmachinelearning AT iliastagkopoulos predictingearlyriskofchronickidneydiseaseincatsusingroutineclinicallaboratorytestsandmachinelearning AT minseungkim predictingearlyriskofchronickidneydiseaseincatsusingroutineclinicallaboratorytestsandmachinelearning AT yianniskokkinos predictingearlyriskofchronickidneydiseaseincatsusingroutineclinicallaboratorytestsandmachinelearning AT theodorospanagiotakos predictingearlyriskofchronickidneydiseaseincatsusingroutineclinicallaboratorytestsandmachinelearning AT jameskennedy predictingearlyriskofchronickidneydiseaseincatsusingroutineclinicallaboratorytestsandmachinelearning AT geertdemeyer predictingearlyriskofchronickidneydiseaseincatsusingroutineclinicallaboratorytestsandmachinelearning AT phillipwatson predictingearlyriskofchronickidneydiseaseincatsusingroutineclinicallaboratorytestsandmachinelearning AT jonathanelliott predictingearlyriskofchronickidneydiseaseincatsusingroutineclinicallaboratorytestsandmachinelearning |