Is a single portal venous phase in contrast-enhanced CT sufficient to detect metastases or recurrence in clear cell renal cell carcinoma? – a single-center retrospective study

This retrospective data evaluation was approved by the institutional review board (approval number 197/2020BO2). Verbal and written informed consent were waived for this retrospective study. The data were collected from the institution’s electronic medical records. We identified a total of 584 patients that were diagnosed with ccRCC between January 2016 and April 2020. Inclusion criteria were as follows: (1) Histologically proven ccRCC, (2) metastatic spread of the ccRCC, proven by histology or interim follow-up of at least 2 years as evidence of metastasis, (3) follow-up CT examination with AP and PVP CECT. Exclusion criteria were defined by missing or incomplete CT-scans or an insufficient follow-up as evidence of new metastatic occurrences. This was most frequently due to the contraindication of contrast agent injection such as a creatinine clearance below 30mL/min/kg resulting in a change of the follow-up protocol (non-enhanced chest CT and abdominal MRI). Of those patients remaining, 43 were included in the study (Fig. 1). Of those, 22 patients had no previous treatment, 11 patients had undergone first-line treatment (6 with checkpoint inhibitor/5 with tyrosine kinase inhibitor) and 10 had undergone second line treatment (6 with checkpoint inhibitor/4 with tyrosine kinase inhibitor).

All CT examinations were performed on a third generation dual-energy scanner consisting of two 192 detector rows (Siemens SOMATOM Force, Siemens Healthineers, Forchheim, Germany). The CT-images of the thorax and abdomen/pelvis were acquired in dual-energy technique using a tube current of 300mAs (CARE Dose4D) for tube A (100 kV) and ref. 232mAs tube current for tube B (Sn150kV), 0.6 mm single collimation width, spiral pitch factor 0.6, matrix 512 × 512 and a convolution kernel Br40 for arterial and Bf40 for PVP image data sets (see Table 1). Image reconstruction was performed with a thickness of 3 mm in 2 planes (axial and coronary) for both contrast phases. Furthermore, high-resolution reconstructions and maximum intensity projections from the PV image datasets were included in the analysis. All patients received Iomeprol as contrast agent (Imerone 400; Bracco, Milan, Italy) intravenously by a dual syringe injector at 2-3 mL/sec (CT Stellant, Medrad, Indianola, PA, USA) followed by a saline chaser bolus. To improve both image quality and parenchymal enhancement the contrast agent was applied taking into account a lean body weight-adapted dosing protocol [15] (body weight in g + 20 = amount of contrast agent in mL.) following the subsequent flow and delay depending on the contrast phase (see Table 1): Based on the findings of Itoh et al., reporting similar detection rates for the late arterial contrast phase after 24.6 to 36.0 Sect.  [16]. The portal venous phase was set based on the findings of Birnbaum et al., reporting nephrographic phase after 60 to 136 Sect.  [17].

Four radiologists with 2, 4, 8 and 12 years of experience in reading oncological CT images analyzed the data sets in two groups and were blinded to the number and location of the metastases: The first group consisted of the less experienced radiologists (2 and 4 years of experience) and read both data sets, the AP and the PVP. The second group of radiologists (8 and 12 years of experience) only had the PVP data sets available for the identification of metastases. As ground truth for the presence of metastatic lesions every patient had a follow-up within a timespan of at least 2 years with interval growth or the metastasis were resected and proven by histology. Suspicious lesions were defined as hypervascularized lesions with distortion of the normal organ structure. Ten potential manifestation regions were defined before reading: either local recurrences, or systemic recurrences, such as manifestation at the opposite kidney, peritoneal metastases, lymph node metastases, pancreatic metastases, hepatic metastases, adrenal metastases, pulmonary metastases, soft-tissue metastases and bone metastases. All lesions were scored individually on the 5-point Likert scale (1 = no metastasis, 2 = unlikely metastasis, 3 = possible metastasis, 4 = most likely metastasis, 5 = definite metastasis). Only lesions classified as “most likely” or “definite” (Likert 4 and 5) were considered as positive metastasis. Furthermore, the lesions were classified depending on their visual aspect and the measured Hounsfield units as being solid, cystic heterogeneous. The presence or absence of calcification was also determined.

Statistical analysis was performed using SPSS (version 27.0.0, IBM, Armonk, NY, USA). Continuous data were expressed as means ± standard deviation (SD). Differences in the number of imaging patterns (solid, cystic, heterogenous) were tested by the Kruskal–Wallis H test. Interrater reliability between the 4 readers was tested with Fleiss’ Kappa (κ). Values from 0.0 to 0.2 indicate slight agreement, 0.21 to 0.40 fair agreement, 0.41 to 0.60 moderate agreement, 0.61 to 0.80 substantial agreement, and values ranging from 0.81 to 1.0 indicate almost perfect or perfect agreement [18]. Sensitivity, specificity as well as accuracy were calculated for every radiologist and plotted on a receiver operating characteristics (ROC)-curve. Determination of the optimal cut-off point was achieved by the Youden index. After verification of the non-Gaussian distribution of every parameter by the Shapiro-Wilk test, we opted for a non-parametric test (Kruskal-Wallis-H test) to analyze differences between the four independent radiologists. A two-tailed p-value of less than 0.05 was considered to indicate statistical significance.