Prediction of Malignant and Benign Lung Tumors Using a Quantitative Radiomic Method

Lung cancer is a leading cause of cancer mortality around the world. Accurate diagnosis of lung cancer is significant for treatment regimen selection. Radiomics refers to comprehensively quantifying the tumor phenotypes by applying a large number of quantitative image features. Here we analyze a computed tomography (CT) data set of 619 patients with lung cancer on the lung image database consortium (LIDC) by radiomic method. Combining with the medical character and clinical recognition of lung tumor, we present a radiomic analysis of 60 features. Then, we use SVM to build a prediction model and find radiomic features which have predictive value for discrimination of malignant and benign lung tumors. Nowadays, as CT imaging is routinely used in lung cancer clinical diagnosis, there is an increase in data set size. We consider that our radiomic prediction model will be developed a valuable medical software and an auxiliary tool which can provide malignant and benign information of lung tumors efficiently.