Guillaume Madelin1, Frederick Poidevin2, Antonios Makrymallis3, Ravinder R Regatte1. 1. Department of Radiology, Center for Biomedical Imaging, New York University School of Medicine, New York, New York, USA. 2. Departamento de Astrofísica, Instituto de Astrofísica de Canarias, La Laguna, Tenerife, Spain; Universidad de La Laguna, La Laguna, Tenerife, Spain. 3. Department of Physics & Astronomy, University College London, Kathleen Lonsdale Building, Gower Place, London, UK.
Abstract
PURPOSE: To assess the possible utility of machine learning for classifying subjects with and subjects without osteoarthritis using sodium magnetic resonance imaging data. Theory: Support vector machine, k-nearest neighbors, naïve Bayes, discriminant analysis, linear regression, logistic regression, neural networks, decision tree, and tree bagging were tested. METHODS: Sodium magnetic resonance imaging with and without fluid suppression by inversion recovery was acquired on the knee cartilage of 19 controls and 28 osteoarthritis patients. Sodium concentrations were measured in regions of interests in the knee for both acquisitions. Mean (MEAN) and standard deviation (STD) of these concentrations were measured in each regions of interest, and the minimum, maximum, and mean of these two measurements were calculated over all regions of interests for each subject. The resulting 12 variables per subject were used as predictors for classification. RESULTS: Either Min [STD] alone, or in combination with Mean [MEAN] or Min [MEAN], all from fluid suppressed data, were the best predictors with an accuracy >74%, mainly with linear logistic regression and linear support vector machine. Other good classifiers include discriminant analysis, linear regression, and naïve Bayes. CONCLUSION: Machine learning is a promising technique for classifying osteoarthritis patients and controls from sodium magnetic resonance imaging data.
PURPOSE: To assess the possible utility of machine learning for classifying subjects with and subjects without osteoarthritis using sodium magnetic resonance imaging data. Theory: Support vector machine, k-nearest neighbors, naïve Bayes, discriminant analysis, linear regression, logistic regression, neural networks, decision tree, and tree bagging were tested. METHODS:Sodium magnetic resonance imaging with and without fluid suppression by inversion recovery was acquired on the knee cartilage of 19 controls and 28 osteoarthritispatients. Sodium concentrations were measured in regions of interests in the knee for both acquisitions. Mean (MEAN) and standard deviation (STD) of these concentrations were measured in each regions of interest, and the minimum, maximum, and mean of these two measurements were calculated over all regions of interests for each subject. The resulting 12 variables per subject were used as predictors for classification. RESULTS: Either Min [STD] alone, or in combination with Mean [MEAN] or Min [MEAN], all from fluid suppressed data, were the best predictors with an accuracy >74%, mainly with linear logistic regression and linear support vector machine. Other good classifiers include discriminant analysis, linear regression, and naïve Bayes. CONCLUSION: Machine learning is a promising technique for classifying osteoarthritispatients and controls from sodium magnetic resonance imaging data.
Authors: Arijitt Borthakur; Erik M Shapiro; Sarma V S Akella; Alexander Gougoutas; J Bruce Kneeland; Ravinder Reddy Journal: Radiology Date: 2002-08 Impact factor: 11.105
Authors: Andrew J Wheaton; Arijitt Borthakur; Erik M Shapiro; Ravinder R Regatte; Sarma V S Akella; J Bruce Kneeland; Ravinder Reddy Journal: Radiology Date: 2004-06 Impact factor: 11.105
Authors: H E Smith; T J Mosher; B J Dardzinski; B G Collins; C M Collins; Q X Yang; V J Schmithorst; M B Smith Journal: J Magn Reson Imaging Date: 2001-07 Impact factor: 4.813
Authors: R Altman; E Asch; D Bloch; G Bole; D Borenstein; K Brandt; W Christy; T D Cooke; R Greenwald; M Hochberg Journal: Arthritis Rheum Date: 1986-08
Authors: S V Akella; R R Regatte; A J Gougoutas; A Borthakur; E M Shapiro; J B Kneeland; J S Leigh; R Reddy Journal: Magn Reson Med Date: 2001-09 Impact factor: 4.668
Authors: Ryan Brown; Guillaume Madelin; Riccardo Lattanzi; Gregory Chang; Ravinder R Regatte; Daniel K Sodickson; Graham C Wiggins Journal: Magn Reson Med Date: 2012-08-08 Impact factor: 4.668
Authors: Muhammad H Khan; John J Walsh; Jelena M Mihailović; Sandeep K Mishra; Daniel Coman; Fahmeed Hyder Journal: Sci Rep Date: 2021-03-23 Impact factor: 4.379
Authors: Benedikt Kamp; Miriam Frenken; Jan M Henke; Daniel B Abrar; Armin M Nagel; Lena V Gast; Georg Oeltzschner; Lena M Wilms; Sven Nebelung; Gerald Antoch; Hans-Jörg Wittsack; Anja Müller-Lutz Journal: Diagnostics (Basel) Date: 2021-12-08