Differentiate common subtypes of renal cell carcinoma with iodine quantitative analysis technology of energy spectral CT

Objective: To assess the utility of iodine quantitative multi-parameter analysis of energy spectral CT in distinguishing common subtypes of renal cell carcinoma (RCC). Methods: All of 102 patients who had underwent energy spectral CT scan and had pathologically proven single RCC, involving 75 cases of clear cell renal cell carcinoma (ccRCC), 27 cases of nonclear cell renal cell carcinoma (non-ccRCC) [15 chromophoberenal cell carcinoma (cRCC) and 12 papillary renal cell carcinoma (pRCC)], were retrospectively assessed. All patients underwent plain CT scan and dual-phase dynamic enhanced scan, and the tumor maximum diameter, absolute CT values and relative CT values, iodine concentration values and normalized iodine concentration (NIC) values were measured and calculated. Stage and classification of each tumor were determined according to TNM stage and Furhman grade. The Chi-square test or Fisher's exact test was used to compare the differences in pathology stage and classification among different RCC subtypes. Independent sample t test was used for the comparison of the maximum diameter and other iodine quantitative parameters. The diagnostic efficiency and optimal diagnostic threshold of each parameter value were evaluated by receiver operating characteristic (ROC) curve, and area under the curve (AUC), the corresponding sensitivity and specificity of each parameter were determined. Results: Significant difference in pathological findings were neither found between ccRCC and non-ccRCC, nor between cRCC and pRCC (P > 0.05). The absolute enhancement CT value, relative enhancement CT value, iodine concentration and NIC were all higher in the ccRCC than in the non-ccRCC (P < 0.05). These values were also higher in the pRCC than in the cRCC (P < 0.05). The AUC of NIC for identifying ccRCC between non-ccRCC was 0.992, with optimal threshold of 0.35, it achieved a sensitivity and specificity of 98.67% and 92.59%, respectively. The AUC of NIC for identifying cRCC between pRCC was 0.978, with optimal threshold of 0.17, it achieved a sensitivity and specificity of 93.33% and 91.67%, respectively. Conclusion: ccRCC and non-ccRCC, cRCC and pRCC can be identified according to iodine quantitative multi-parameter analysis of energy spectral CT, and NIC value helps distinguish some different RCC subtypes. (authors)