Radiomic features analysis in computed tomography images of lung nodule classification

• DOI: 10.1371/journal.pone.0192002
• Publication Date: 2018-02-05T13:31:26Z
• Authors: Chia-Hung Chen; Chih-Kun Chang; Chih-Yen Tu; Wei-Chih Liao; Bing-Ru Wu; Kuei-Ting Chou; Yu-Rou Chiou; Shih-Neng Yang; Geoffrey Zhang; Tzung-Chi Huang
• Keywords: Nodule (geology); Solitary pulmonary nodule

Purpose Radiomics, which extract large amount of quantification image features from diagnostic medical images had been widely used for prognostication, treatment response prediction and cancer detection. The options lung nodules depend on their diagnosis, benign or malignant. Conventionally, nodule diagnosis is based invasive biopsy. Recently, radiomics features, a non-invasive method clinical images, have shown high potential in lesion classification, outcome prediction. Methods Lung classification using Computed Tomography (CT) data was investigated 4-feature signature introduced classification. Retrospectively, 72 patients with 75 pulmonary were collected. Radiomics feature extraction performed non-enhanced CT contours delineated by an experienced radiation oncologist. Result Among the 750 each case, 76 found to significant differences between malignant lesions. A composed best 4 included Laws_LSL_min, Laws_SLL_energy, Laws_SSL_skewness Laws_EEL_uniformity. accuracy 84% sensitivity 92.85% specificity 72.73%. Conclusion demonstrated very good application.