Differential diagnostic ability of 18F-FDG PET/CT radiomics features between renal cell carcinoma and renal lymphoma.

BACKGROUND: To determine the differential diagnostic value of texture parameters of PET/CT on renal cell carcinoma and renal lymphoma. METHODS: Twenty renal lymphoma and 18 renal cell carcinoma (RCC) patients were analyzed in this study. The pathological information and basic characteristics were extracted from the electronic medical record system of our hospital. We used LIFEx package to extract data from the radiomics images. Receiver operating characteristic analysis and binary logistic regression analysis was applied in determining the diagnostic accuracy of texture parameters as well as the synthetic parameter, of which the sensitivity and specificity was improved. RESULTS: There were 14 (2 in HISTO, 2 in GLCM, 5 in GLRLM, 5 in GLZLM) out of the texture parameters showing AUC > 0.7 and p<0.05. Synthesized parameters of each section showed even higher differentiation ability, with AUC varying from 0.725 to 1.000. CONCLUSIONS: Texture analysis of 18F-FDG PET/CT could effectively differentiate between RCCs and renal lymphomas.