Svyat Vergun1, Josh I Suhonen2, Veena A Nair2, J S Kuo3, M K Baskaya3, Camille Garcia-Ramos1, Elizabeth E Meyerand4,5, Vivek Prabhakaran6. 1. Department of Medical Physics, University of Wisconsin-Madison, School of Medicine and Public Health, 1111 Highland Avenue, Madison, WI53792-3252, USA. 2. Department of Radiology, University of Wisconsin-Madison, School of Medicine and Public Health, 600 Highland Avenue, Wisconsin Institutes for Medical Research (WIMR), Madison, WI 53705, USA. 3. Department of Neurosurgery, University of Wisconsin-Madison, School of Medicine and Public Health, University of Wisconsin, Box 8660 Clinical Science Center, 600 Highland Ave, Madison, WI 53792, USA. 4. Departments of Biomedical Engineering University of Wisconsin-Madison, 1550 Engineering Dr, Madison, WI 53706, USA. 5. Medical Physics, University of Wisconsin-Madison, 1111 Highland Ave., Suite 1129, Wisconsin Institutes for Medical Research (WIMR), Madison, WI 53705, USA. 6. Department of Radiology, Director of Functional Neuroimaging in Radiology, University of Wisconsin Madison, School of Medicine and Public Health, 600 Highland Avenue, Madison, WI 53792-3252, USA.
Abstract
BACKGROUND: Advanced neuroimaging measures along with clinical variables acquired during standard imaging protocols provide a rich source of information for brain tumor patient treatment and management. Machine learning analysis has had much recent success in neuroimaging applications for normal and patient populations and has potential, specifically for brain tumor patient outcome prediction. The purpose of this work was to construct, using the current patient population distribution, a high accuracy predictor for brain tumor patient outcomes of mortality and morbidity (i.e., transient and persistent language and motor deficits). The clinical value offered is a statistical tool to help guide treatment and planning as well as an investigation of the influential factors of the disease process. METHODS: Resting state fMRI, diffusion tensor imaging, and task fMRI data in combination with clinical and demographic variables were used to represent the tumor patient population (n = 62; mean age = 51.2 yrs.) in a machine learning analysis in order to predict outcomes. RESULTS: A support vector machine classifier with a t-test filter and recursive feature elimination predicted patient mortality (18-month interval) with 80.7% accuracy, language deficits (transient) with 74.2%, motor deficits with 71.0%, language outcomes (persistent) with 80.7% and motor outcomes with 83.9%. The most influential features of the predictors were resting fMRI connectivity, and fractional anisotropy and mean diffusivity measures in the internal capsule, brain stem and superior and inferior longitudinal fasciculi. CONCLUSIONS: This study showed that advanced neuroimaging data with machine learning methods can potentially predict patient outcomes and reveal influential factors driving the predictions.
BACKGROUND: Advanced neuroimaging measures along with clinical variables acquired during standard imaging protocols provide a rich source of information for brain tumor patient treatment and management. Machine learning analysis has had much recent success in neuroimaging applications for normal and patient populations and has potential, specifically for brain tumor patient outcome prediction. The purpose of this work was to construct, using the current patient population distribution, a high accuracy predictor for brain tumor patient outcomes of mortality and morbidity (i.e., transient and persistent language and motor deficits). The clinical value offered is a statistical tool to help guide treatment and planning as well as an investigation of the influential factors of the disease process. METHODS: Resting state fMRI, diffusion tensor imaging, and task fMRI data in combination with clinical and demographic variables were used to represent the tumor patient population (n = 62; mean age = 51.2 yrs.) in a machine learning analysis in order to predict outcomes. RESULTS: A support vector machine classifier with a t-test filter and recursive feature elimination predicted patient mortality (18-month interval) with 80.7% accuracy, language deficits (transient) with 74.2%, motor deficits with 71.0%, language outcomes (persistent) with 80.7% and motor outcomes with 83.9%. The most influential features of the predictors were resting fMRI connectivity, and fractional anisotropy and mean diffusivity measures in the internal capsule, brain stem and superior and inferior longitudinal fasciculi. CONCLUSIONS: This study showed that advanced neuroimaging data with machine learning methods can potentially predict patient outcomes and reveal influential factors driving the predictions.
Authors: Bradford A Moffat; Thomas L Chenevert; Theodore S Lawrence; Charles R Meyer; Timothy D Johnson; Qian Dong; Christina Tsien; Suresh Mukherji; Douglas J Quint; Stephen S Gebarski; Patricia L Robertson; Larry R Junck; Alnawaz Rehemtulla; Brian D Ross Journal: Proc Natl Acad Sci U S A Date: 2005-04-01 Impact factor: 11.205
Authors: Timothy B Meier; Alok S Desphande; Svyatoslav Vergun; Veena A Nair; Jie Song; Bharat B Biswal; Mary E Meyerand; Rasmus M Birn; Vivek Prabhakaran Journal: Neuroimage Date: 2011-12-28 Impact factor: 6.556
Authors: Nico U F Dosenbach; Binyam Nardos; Alexander L Cohen; Damien A Fair; Jonathan D Power; Jessica A Church; Steven M Nelson; Gagan S Wig; Alecia C Vogel; Christina N Lessov-Schlaggar; Kelly Anne Barnes; Joseph W Dubis; Eric Feczko; Rebecca S Coalson; John R Pruett; Deanna M Barch; Steven E Petersen; Bradley L Schlaggar Journal: Science Date: 2010-09-10 Impact factor: 47.728
Authors: Michael H Lev; Yelda Ozsunar; John W Henson; Amjad A Rasheed; Glenn D Barest; Griffith R Harsh; Markus M Fitzek; E Antonio Chiocca; James D Rabinov; Andrew N Csavoy; Bruce R Rosen; Fred H Hochberg; Pamela W Schaefer; R Gilberto Gonzalez Journal: AJNR Am J Neuroradiol Date: 2004-02 Impact factor: 3.825
Authors: Elena A Allen; Erik B Erhardt; Eswar Damaraju; William Gruner; Judith M Segall; Rogers F Silva; Martin Havlicek; Srinivas Rachakonda; Jill Fries; Ravi Kalyanam; Andrew M Michael; Arvind Caprihan; Jessica A Turner; Tom Eichele; Steven Adelsheim; Angela D Bryan; Juan Bustillo; Vincent P Clark; Sarah W Feldstein Ewing; Francesca Filbey; Corey C Ford; Kent Hutchison; Rex E Jung; Kent A Kiehl; Piyadasa Kodituwakku; Yuko M Komesu; Andrew R Mayer; Godfrey D Pearlson; John P Phillips; Joseph R Sadek; Michael Stevens; Ursina Teuscher; Robert J Thoma; Vince D Calhoun Journal: Front Syst Neurosci Date: 2011-02-04
Authors: Maxwell Lewis Neal; Andrew D Trister; Tyler Cloke; Rita Sodt; Sunyoung Ahn; Anne L Baldock; Carly A Bridge; Albert Lai; Timothy F Cloughesy; Maciej M Mrugala; Jason K Rockhill; Russell C Rockne; Kristin R Swanson Journal: PLoS One Date: 2013-01-23 Impact factor: 3.240