CT Texture Analysis of Mediastinal Lymphadenopathy: Discrimination Between Sarcoidosis and Lymph Node Metastasis by Small Cell Lung Cancer

Abstract

Rationale: To evaluate the potential of texture analysis using computed tomography (CT) to discriminate mediastinal lymphadenopathy caused by sarcoidosis from that caused by metastasis of small cell lung cancer (SCLC). Methods: A total of 30 patients, 16 with sarcoidosis and 14 with SCLC were enrolled. All patients demonstrated mediastinal lymphadenopathy on enhanced chest CT. On CT images showing the largest lymphadenopathy, a fixed-size round region of interest (ROI) was drawn on the lymphadenopathy, and 31 texture features were automatically measured. Among the 30 patients, 19 (12 with sarcoidosis and 7 with SCLC) underwent endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA), and the fat-to-lesion strain ratio (FLR) was recorded. Texture features and the FLR were compared between the two patient groups using the Mann-Whitney test. In addition, logistic regression analyses were performed to evaluate the diagnostic accuracy of the FLR combined with texture features. Results: Of the 31 texture features, 11 showed significant differences between the sarcoidosis and SCLC patients. Among them, the grey-level run length matrix with high gray-level run emphasis (GLRLM-HGRE) showed the greatest difference (P<0.01). The FLR also showed a significant difference between the groups (P<0.05). Logistic regression analysis together with receiver operating characteristic analysis revealed that the FLR combined with the GLRLM-HGRE showed a high diagnostic accuracy (100% sensitivity, 92% specificity, 0.988 area under the curve) for discriminating sarcoidosis from SCLC. Conclusion: Texture analysis, particularly when combined with the FLR, is useful for discriminating mediastinal lymphadenopathy caused by sarcoidosis from that caused by SCLC metastasis.