Carolina Parada Villavicencio1, Robert J Mc Carthy2, Frank H Miller3. 1. Department of Radiology, Northwestern Memorial Hospital, Northwestern University Feinberg School of Medicine, 676 North Saint Clair St. Suite 800, Chicago, IL, USA. 2. Department of Anesthesiology, Northwestern Memorial Hospital, Northwestern University Feinberg School of Medicine, 676 North Saint Clair St. Suite 1050, Chicago, IL, USA. 3. Department of Radiology, Northwestern Memorial Hospital, Northwestern University Feinberg School of Medicine, 676 North Saint Clair St. Suite 800, Chicago, IL, USA. fmiller@northwestern.edu.
Abstract
OBJECTIVE: To assess the diagnostic performance of the apparent diffusion coefficient (ADC) in predicting the Fuhrman nuclear grading of clear cell renal cell carcinomas (ccRCC). MATERIALS AND METHODS: A total of 129 patients who underwent partial and radical nephrectomies with pathology-proven ccRCC were retrospectively evaluated. Histopathological characteristics and nuclear grades were analyzed. In addition, conventional magnetic resonance imaging (MRI) features were assessed in consensus by two radiologists to discriminate nuclear grading. ADC values were obtained from a region of interest (ROI) measurement in the ADC maps calculated from diffusion-weighted imaging (DWI) using b values of 50, 500, and 800 s/mm2. The threshold values for predicting and differentiating low-grade cancers (Fuhrman I-II) from high grade (Fuhrman III-IV) was obtained using binary logistic regression. The ADC cut-off value for differentiating low- and high-grade tumors was determined using classification analysis. RESULTS: Significant associations (P < 0.001) were found between nuclear grading, conventional MR features, and DWI. Hemorrhage, necrosis, perirenal fat invasion, enhancement homogeneity, and cystic component were identified as independent predictors of tumor grade. High-grade ccRCC had significantly lower mean ADC values compared to low-grade tumors. An ADC cut-off value of 1.6 × 10-3 mm2/s had an optimal predictive percentage of 65.5% for low-grade tumors above this threshold and 81% for high-grade ccRCC below this threshold. Overall predictive accuracy was 70.5%. CONCLUSION: The addition of ADC values to a model based on MRI conventional features demonstrates increased sensitivity and high specificity improving the distinguishing accuracy between both high-grade and low-grade ccRCC.
OBJECTIVE: To assess the diagnostic performance of the apparent diffusion coefficient (ADC) in predicting the Fuhrman nuclear grading of clear cell renal cell carcinomas (ccRCC). MATERIALS AND METHODS: A total of 129 patients who underwent partial and radical nephrectomies with pathology-proven ccRCC were retrospectively evaluated. Histopathological characteristics and nuclear grades were analyzed. In addition, conventional magnetic resonance imaging (MRI) features were assessed in consensus by two radiologists to discriminate nuclear grading. ADC values were obtained from a region of interest (ROI) measurement in the ADC maps calculated from diffusion-weighted imaging (DWI) using b values of 50, 500, and 800 s/mm2. The threshold values for predicting and differentiating low-grade cancers (Fuhrman I-II) from high grade (Fuhrman III-IV) was obtained using binary logistic regression. The ADC cut-off value for differentiating low- and high-grade tumors was determined using classification analysis. RESULTS: Significant associations (P < 0.001) were found between nuclear grading, conventional MR features, and DWI. Hemorrhage, necrosis, perirenal fat invasion, enhancement homogeneity, and cystic component were identified as independent predictors of tumor grade. High-grade ccRCC had significantly lower mean ADC values compared to low-grade tumors. An ADC cut-off value of 1.6 × 10-3 mm2/s had an optimal predictive percentage of 65.5% for low-grade tumors above this threshold and 81% for high-grade ccRCC below this threshold. Overall predictive accuracy was 70.5%. CONCLUSION: The addition of ADC values to a model based on MRI conventional features demonstrates increased sensitivity and high specificity improving the distinguishing accuracy between both high-grade and low-grade ccRCC.