Computer-Aided Diagnosis of Lung Cancer in Magnetic Resonance Imaging Exams

Lung cancer is the type of cancer that most makes victims around the world and often presents a late diagnosis. Computed tomography (CT) is currently the reference imaging test for the diagnosis and staging of lung tumors. Recent studies have shown relevance in the characterization of lung tumors by different sequences obtained with magnetic resonance imaging (MRI). MRI also has the advantage of not exposing the patient to ionizing radiation, as occurs in CT scans. This paper presents an investigation about the applicability of pattern recognition methods to computer-aided diagnosis of lung cancer in MRI exams. A set of 21 T1-weighted contrast-enhanced MR images associated with lung lesions (14 malignant and 7 benign) was retrospectively constructed and semi-automatically segmented. Quantitative features were obtained from tumor 2D and 3D segmentation, totaling 150 features. Unbalancing problems were solved synthetically oversampling the dataset. Tumor classification was based on five machine learning classifiers and leave-one-out cross-validation. Relevant feature selection was performed for all classifiers. Results showed significant performance on balanced dataset, presenting area under the receiver operating characteristic (ROC) curve of 0.885 during the validation, and 0.938 during the test process. The investigated approach demonstrates potential for computer-aided diagnosis of lung cancer in MRI.