Investigation of the support vector machine algorithm to predict lung radiation-induced pneumonitis

• Published in: Medical physics (Lancaster) Vol. 34; no. 10; pp. 3808 - 3814
• Main Authors: Chen, Shifeng, Zhou, Sumin, Yin, Fang-Fang, Marks, Lawrence B., Das, Shiva K.
• Format: Journal Article
• Language: English
• Published: United States American Association of Physicists in Medicine 01.10.2007
• Subjects: Algorithms; biological effects of radiation; Biomedical modeling; cancer; Chemotherapy; Computer‐aided diagnosis; Diseases; Dose-Response Relationship, Radiation; Dosimetry; Effects of electromagnetic and acoustic fields on biological systems; Humans; Internet; Laser Doppler velocimetry; lung; Lung - pathology; Lung - radiation effects; Lungs; medical diagnostic computing; modeling; Models, Statistical; Multivariate analysis; prediction; radiation pneumonitis; Radiation Pneumonitis - diagnosis; Radiation Pneumonitis - etiology; radiation therapy; Radiotherapy Dosage; Radiotherapy Planning, Computer-Assisted; ROC Curve; sensitivity analysis; Sensitivity and Specificity; Software; support vector machines; Testing procedures; prediction; support vector machines; modeling; radiation pneumonitis
• ISSN: 0094-2405; 2473-4209; 1522-8541; 2473-4209
• DOI: 10.1118/1.2776669

The purpose of this study is to build and test a support vector machine (SVM) model to predict for the occurrence of lung radiation-induced Grade 2+ pneumonitis. SVM is a sophisticated statistical technique capable of separating the two categories of patients (with/without pneumonitis) using a boundary defined by a complex hypersurface. Despite the complexity, the SVM boundary is only minimally influenced by outliers that are difficult to separate. By contrast, the simple hyperplane boundary computed by the more commonly used and related linear discriminant analysis method is heavily influenced by outliers. Two SVM models were built using data from 219 patients with lung cancer treated using radiotherapy (34 diagnosed with pneumonitis). One model ( SVM all ) selected input features from all dose and non-dose factors. For comparison, the other model ( SVM dose ) selected input features only from lung dose-volume factors. Model predictive ability was evaluated using ten-fold cross-validation and receiver operating characteristics (ROC) analysis. For the model SVM all , the area under the cross-validated ROC curve was 0.76 ( sensitivity ∕ specificity = 74 % ∕ 75 % ) . Compared to the corresponding SVM dose area of 0.71 ( sensitivity ∕ specificity = 68 % ∕ 68 % ) , the predictive ability of SVM all was improved, indicating that non-dose features are important contributors to separating patients with and without pneumonitis. Among the input features selected by model SVM all , the two with highest importance for predicting lung pneumonitis were: (a) generalized equivalent uniform doses close to the mean lung dose, and (b) chemotherapy prior to radiotherapy. The model SVM all is publicly available via internet access.