Patients
This retrospective single-institution study was approved by the institutional review board of our facility. Written informed consent for the use of individual patient data in the analysis was waived because of the retrospective nature of the investigation. A total of 281partial nephrectomies were performed for renal tumors between January 2017 and February 2018. Of these, 75 patients were excluded for no preoperative contrast-enhanced CT at our institution (n = 61), rare neoplasms (n = 4), non-neoplastic lesions (n = 7), and AML with apparent fat on CT (n = 3) because differentiation is not difficult. Thus, the final study group consisted of 206 patients. The clinical characteristics of the patients are summarized in Table 1 (Figure 5). Along with histological types of tumors, the presence or absence of peritumoral pseudocapsule was evaluated in 171 (90.0%) of the 190 RCCs.
CT examinations
CT examination was performed using 64 or 320-row detector scanners. The median interval period from CT to surgery was 28.5 days, range 1–390 days. Images were obtained at the following settings: pitch, 0.81–0.83; collimation, 0.5 mm; reconstruction thickness/interval, 1.0 mm/1.0 mm; and 120 kVp with automatic exposure control. Images at the corticomedullary phase were obtained using a bolus-tracking contrast monitoring system after injecting contrast materials for 30 s. The nephrographic and excretory phase images were subsequently obtained at 90 and 300 s after injecting contrast medium. The amount of contrast material, 600 mgI/kg with a maximum 150 mL, and injection rate of maximum 5.0 mL/s depended on body weight and concentration of iodine (300, 350, or 370 mgI/mL) used.
CT evaluation
Images were independently evaluated by two radiologists (S.M and Y.O, each with more than 18 and five years of experience, respectively) using a viewer with a three-dimensional workstation. For the evaluation of the EDCB, trans-axial and multi-planar reconstruction images at any angle with 1.0-mm thickness and interval were used. All images were randomized, and the reviewers were blinded to information regarding the pathological findings. The reviewers evaluated the presence or absence of EDCB, which was an unenhanced thin line between the tumor and renal cortex on the corticomedullary phase images. If it was observed in only one part, it was recognized as being present. Where discrepancies occurred, final judgment was obtained by consensus. The presence or absence of high-attenuation rims around the tumor on the nephrographic and excretory phase images were also evaluated. Schemas of these findings are shown in Figure 1. Other previously reported parameters for differentiating ccRCC from Fp-AML on CT were also evaluated according to past reports [4–8].
Statistical analysis
Statistical analyses were performed using JMP 15 software (SAS Institute Inc). Reader agreement for recording EDCB, nephrographic, and excretory phase high-attenuation rim were assessed with k statistics. The significant findings of ccRCCs were extracted based on the univariate analysis using Fisher’s exact test and Mann-Whitney's U test. Independent predictors for ccRCCs were identified among the extracted findings using multivariate analysis with the logistic regression model. The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of EDCB for differentiating ccRCC from Fp-AML were calculated. These values, with Fisher’s exact test of EDCB, nephrographic, and excretory phase high-attenuation rim for detecting peritumoral pseudocapsules in RCC, were calculated with histopathologic data as the standard of reference.