Lan Bowen1, Li Xiaojing2. 1. Huizhou Central People's Hospital, No.41, Elingbei Road, 516000 Huizhou, China. Electronic address: lanbowenbw@163.com. 2. Huizhou Central People's Hospital, No.41, Elingbei Road, 516000 Huizhou, China.
Abstract
PURPOSE: To investigate associations between clear-cell renal cell carcinoma mRNA-based subtyping and CT features. MATERIALS AND METHODS: The CT data from 177 patients generated with The Cancer Imaging Archive were reviewed. The correlation was analyzed using chi-square test and univariate regression analysis. RESULTS: Identified were 124 (53.2%) m1, 67 (28.8%) m2, 17 (7.3%) m3, and 14 (8.7%) m4 subtypes. m1-subtype rates were significantly higher in well-defined margin lesions (p = 0.041). m3-subtype rates were significantly higher in ill-defined margin lesions (p = 0.012), in collecting system invasion lesions (p = 0.028) and collecting system invasion lesions (p = 0.026).On univariate logistic regression analysis, tumor margin (well-defined margin vs ill-defined margin, OR: 2.104; p = 0.041; 95% CI: 1.024-4.322) was associated with m1-subtype. Tumor margin (well-defined margin vs ill-defined margin, OR: 2.104; p = 0.012; 95% CI: 0.212-0.834) and collecting system invasion (yes vs no, OR: 0.421; p = 0.028; 95% CI: 0.212-0.834) and renal vein invasion (yes vs no, OR: 2.164; p = 0.026; 95% CI: 1.090-4.294) were associated with m3-subtype. There was no significant difference between mRNA-based subtyping (m2 vs other; m4 vs other) and the CT features. CONCLUSIONS: This preliminary radiogenomics analysis of clear-cell renal cell carcinoma revealed associations between CT features and mRNA-based subtyping which warrant further investigation and validation.
PURPOSE: To investigate associations between clear-cell renal cell carcinoma mRNA-based subtyping and CT features. MATERIALS AND METHODS: The CT data from 177 patients generated with The Cancer Imaging Archive were reviewed. The correlation was analyzed using chi-square test and univariate regression analysis. RESULTS: Identified were 124 (53.2%) m1, 67 (28.8%) m2, 17 (7.3%) m3, and 14 (8.7%) m4 subtypes. m1-subtype rates were significantly higher in well-defined margin lesions (p = 0.041). m3-subtype rates were significantly higher in ill-defined margin lesions (p = 0.012), in collecting system invasion lesions (p = 0.028) and collecting system invasion lesions (p = 0.026).On univariate logistic regression analysis, tumor margin (well-defined margin vs ill-defined margin, OR: 2.104; p = 0.041; 95% CI: 1.024-4.322) was associated with m1-subtype. Tumor margin (well-defined margin vs ill-defined margin, OR: 2.104; p = 0.012; 95% CI: 0.212-0.834) and collecting system invasion (yes vs no, OR: 0.421; p = 0.028; 95% CI: 0.212-0.834) and renal vein invasion (yes vs no, OR: 2.164; p = 0.026; 95% CI: 1.090-4.294) were associated with m3-subtype. There was no significant difference between mRNA-based subtyping (m2 vs other; m4 vs other) and the CT features. CONCLUSIONS: This preliminary radiogenomics analysis of clear-cell renal cell carcinoma revealed associations between CT features and mRNA-based subtyping which warrant further investigation and validation.