Shuang Wu1,2, Jin Meng1,2, Qi Yu1,2, Ping Li3, Shen Fu4,5,6,7,8. 1. Department of Radiation Oncology, Fudan University Shanghai Cancer Center, 270 Dong'An Road, Xuhui District, Shanghai, 200032, China. 2. Department of Oncology, Shanghai Medical College, Fudan University, Shanghai, 200032, China. 3. Department of Radiation Oncology, Shanghai Proton and Heavy Ion Center, Shanghai, 201321, China. 4. Department of Radiation Oncology, Fudan University Shanghai Cancer Center, 270 Dong'An Road, Xuhui District, Shanghai, 200032, China. shen_fu@hotmail.com. 5. Department of Oncology, Shanghai Medical College, Fudan University, Shanghai, 200032, China. shen_fu@hotmail.com. 6. Department of Radiation Oncology, Shanghai Proton and Heavy Ion Center, Shanghai, 201321, China. shen_fu@hotmail.com. 7. Key Laboratory of Nuclear Physics and Ion-beam Application (MOE), Fudan University, Shanghai, 200433, China. shen_fu@hotmail.com. 8. Department of Radiation Oncology, Shanghai Concord Cancer Hospital, Shanghai, 200020, China. shen_fu@hotmail.com.
Abstract
PURPOSE: Reliable and accurate predictive models are necessary to drive the success of radiomics. Our aim was to identify the optimal radiomics-based machine learning method for isocitrate dehydrogenase (IDH) genotype prediction in diffuse gliomas. METHODS: Eight classical machine learning methods were evaluated in terms of their stability and performance for pre-operative IDH genotype prediction. A total of 126 patients were enrolled for analysis. Overall, 704 radiomic features extracted from the pre-operative MRI images were analyzed. The patients were randomly assigned to either the training set or the validation set at a ratio of 2:1. Feature selection and classification model training were done using the training set, whereas the predictive performance and stability of the model were independently assessed using the validation set. RESULTS: Random Forest (RF) showed high predictive performance (accuracy 0.885 ± 0.041, AUC 0.931 ± 0.036), whereas neural network (NN) (accuracy 0.829 ± 0.064, AUC 0.878 ± 0.052) and flexible discriminant analysis (FDA) (accuracy 0.851 ± 0.049, AUC 0.875 ± 0.057) displayed low predictive performance. With regard to stability, RF also showed high robustness against data perturbation (relative standard deviations, RSD 3.87%). CONCLUSIONS: RF is a promising machine learning method in predicting IDH genotype. Development of an accurate and reliable model can assist in the initial diagnostic evaluation and treatment planning for diffuse glioma patients.
RCT Entities:
PURPOSE: Reliable and accurate predictive models are necessary to drive the success of radiomics. Our aim was to identify the optimal radiomics-based machine learning method for isocitrate dehydrogenase (IDH) genotype prediction in diffuse gliomas. METHODS: Eight classical machine learning methods were evaluated in terms of their stability and performance for pre-operative IDH genotype prediction. A total of 126 patients were enrolled for analysis. Overall, 704 radiomic features extracted from the pre-operative MRI images were analyzed. The patients were randomly assigned to either the training set or the validation set at a ratio of 2:1. Feature selection and classification model training were done using the training set, whereas the predictive performance and stability of the model were independently assessed using the validation set. RESULTS: Random Forest (RF) showed high predictive performance (accuracy 0.885 ± 0.041, AUC 0.931 ± 0.036), whereas neural network (NN) (accuracy 0.829 ± 0.064, AUC 0.878 ± 0.052) and flexible discriminant analysis (FDA) (accuracy 0.851 ± 0.049, AUC 0.875 ± 0.057) displayed low predictive performance. With regard to stability, RF also showed high robustness against data perturbation (relative standard deviations, RSD 3.87%). CONCLUSIONS: RF is a promising machine learning method in predicting IDH genotype. Development of an accurate and reliable model can assist in the initial diagnostic evaluation and treatment planning for diffuse gliomapatients.
Authors: Nicolas De Jay; Simon Papillon-Cavanagh; Catharina Olsen; Nehme El-Hachem; Gianluca Bontempi; Benjamin Haibe-Kains Journal: Bioinformatics Date: 2013-07-03 Impact factor: 6.937
Authors: Christian Hartmann; Bettina Hentschel; Wolfgang Wick; David Capper; Jörg Felsberg; Matthias Simon; Manfred Westphal; Gabriele Schackert; Richard Meyermann; Torsten Pietsch; Guido Reifenberger; Michael Weller; Markus Loeffler; Andreas von Deimling Journal: Acta Neuropathol Date: 2010-11-19 Impact factor: 17.088
Authors: Philippe Lambin; Emmanuel Rios-Velazquez; Ralph Leijenaar; Sara Carvalho; Ruud G P M van Stiphout; Patrick Granton; Catharina M L Zegers; Robert Gillies; Ronald Boellard; André Dekker; Hugo J W L Aerts Journal: Eur J Cancer Date: 2012-01-16 Impact factor: 9.162
Authors: J A Carrillo; A Lai; P L Nghiemphu; H J Kim; H S Phillips; S Kharbanda; P Moftakhar; S Lalaezari; W Yong; B M Ellingson; T F Cloughesy; W B Pope Journal: AJNR Am J Neuroradiol Date: 2012-02-09 Impact factor: 3.825
Authors: D Williams Parsons; Siân Jones; Xiaosong Zhang; Jimmy Cheng-Ho Lin; Rebecca J Leary; Philipp Angenendt; Parminder Mankoo; Hannah Carter; I-Mei Siu; Gary L Gallia; Alessandro Olivi; Roger McLendon; B Ahmed Rasheed; Stephen Keir; Tatiana Nikolskaya; Yuri Nikolsky; Dana A Busam; Hanna Tekleab; Luis A Diaz; James Hartigan; Doug R Smith; Robert L Strausberg; Suely Kazue Nagahashi Marie; Sueli Mieko Oba Shinjo; Hai Yan; Gregory J Riggins; Darell D Bigner; Rachel Karchin; Nick Papadopoulos; Giovanni Parmigiani; Bert Vogelstein; Victor E Velculescu; Kenneth W Kinzler Journal: Science Date: 2008-09-04 Impact factor: 47.728
Authors: Hai Yan; D Williams Parsons; Genglin Jin; Roger McLendon; B Ahmed Rasheed; Weishi Yuan; Ivan Kos; Ines Batinic-Haberle; Siân Jones; Gregory J Riggins; Henry Friedman; Allan Friedman; David Reardon; James Herndon; Kenneth W Kinzler; Victor E Velculescu; Bert Vogelstein; Darell D Bigner Journal: N Engl J Med Date: 2009-02-19 Impact factor: 176.079
Authors: Evi J van Kempen; Max Post; Manoj Mannil; Benno Kusters; Mark Ter Laan; Frederick J A Meijer; Dylan J H A Henssen Journal: Cancers (Basel) Date: 2021-05-26 Impact factor: 6.639