OBJECTIVETo explore the significance of multi-detector CT (MDCT) in differential diagnosis of papillary renal cell carcinoma and chromophobe renal cell carcinoma.

METHODSClinical data of forty-one cases of renal cancers confirmed pathologically were collected, including 21 cases of papillary renal cell carcinoma (PRCC) (14 type I, 7 type II) and 20 cases of chromophobe renal cell carcinoma (ChRCC). Their morphological and MDCT characteristics were retrospectively analyzed. Receiver operator characteristic curve (ROC) was used to analyze the value of MDCT in differential diagnosis of PRCC and ChRCC. Two senior radiologists analyzed the morphological and the dynamic enhancement characteristics of the images. The attenuation of the lesions and the adjacent renal parenchyma were measured. The morphological indexes were compared with chi-square test and the quantitative indexes were compared with independent sample T-test. Receiver operator characteristic curve (ROC) was used to analyze the sensitivity, specificity and accuracy of diagnosis of PRCC and ChRCC.

RESULTSAngioid enhancement and filled enhancement were more common in ChRCC than in PRCC, while delayed enhancement was more often seen in PRCC than in ChRCC. Calcification was more common in type I than type II PRCC. The enhancement value (ΔCT value) in corticomedullary phase was (29.08 ± 20.12) Hu for PRCC, significantly lower than the (48.29 ± 26.70) Hu for ChRCC (t = -2.611, P = 0.013). The ΔCT value of type I PRCC in corticomedullary phase was (26.36 ± 18.16) Hu, showing a significant difference from that of ChRCC (t = -2.666, P = 0.012). The lesion to kidney ratio (LKR) in corticomedullary phase was 0.44 ± 0.19 for PRCC and 0.58 ± 0.15 for ChRCC, with a significant difference between them (t = -2.587, P = 0.014). The LKR of type I PRCC in corticomedullary phase was 0.39 ± 0.15, showing a significant difference from that of ChRCC (t = -3.628, P = 0.001). The difference value (D-value) of the attenuation of lesion between corticomedullary and nephrographic phases was (-3.69 ± 8.90) Hu for PRCC and (8.39 ± 21.98) Hu for ChRCC, with a significant difference between them (t = -2.285, P = 0.031). The D-value of type I PRCC was (-4.55 ± 9.82) Hu, showing a significant difference from that of ChRCC (t = -2.323, P = 0.028). There was no significant difference between the ΔCT, LKR and D-value of the type II PRCC and ChRCC (P > 0.05 for all). The area under the curve (AUC) for ΔCT value, LKR value in corticomedullary phase, and D-value were 0.718, 0.751 and 0.668, respectively, and there were no significant differences among them (z values were 0.896, 0.683 and 0.559, respectively, and P values were 0.370, 0.495 and 0.576, respectively). Using 49.350 Hu as the cutoff value for ΔCT value in corticomedullary phase, resulted in a sensitivity, specificity and accuracy of 50.0%, 90.5% and 70.7%, respectively. Corresponding values were 65.0%, 81.0% and 73.2%, when using a cutoff value of 0.532 for LKR in corticomedullary phase, and were 60.0%, 76.2% and 68.3%, when using a D-value of 0.400 Hu.

CONCLUSIONSThe ΔCT value, LKR value in corticomedullary phase, and the D-value are all useful indexes for the differentiation of PRCC and ChRCC.