This single-center, single-arm, prospective, open-label study was conducted between October 2020 and April 2022, with approval from the Institutional Review Board (approval number: RIN2009-005) and in accordance with the Declaration of Helsinki. All participants provided written informed consent. This study was registered in the University Hospital Medical Information Network Clinical Trials Registry (study ID: UMIN000041358).

The primary end point was the diagnostic yield of single-core and multiple-core biopsies. The secondary end points were i) feasibility of multiple-core biopsy, ii) appropriateness of specimens obtained by single-core and multiple-core biopsies, iii) histological diagnosis, iv) renal cell carcinoma (RCC) subtype, and v) Fuhrman grade of RCC.

The inclusion criteria were i) renal tumor with an indication for image-guided biopsy, ii) tumor > 1.5 cm, and iii) age ≥ 20 years. The exclusion criteria were i) cystic tumor; ii) tumor with an obvious fat component (i.e., typical angiomyolipoma); iii) abnormal laboratory data, including white blood cell count < 2,000/μL, hemoglobin level < 6.0 g/dL, platelet count < 50,000/μL, and prothrombin time > 1.5; iv) pregnancy; v) mental illness (e.g., dementia); and vi) ineligibility determined by the responding physician.

All biopsies were performed in the interventional radiology suite under US (Aplio 500; Canon Medical Systems, Otawara, Japan) or CT fluoroscopy (Aquilion; Canon Medical Systems) guidance. An operator completed the biopsies with the co-axial method using one image guidance modality per procedure. The co-axial system comprised a 17-gauge introducer and an 18-gauge semi-automatic cutting biopsy needle (Temno Evolution, Care Fusion, IL, USA). A board-certified interventional radiologist (staff) or non-staff member, under a staff member’s direct supervision, performed the biopsy. Operators included four staff with 15 years of median experience (range, 10–16 years) and two non-staff with 6 and 3 years of experience, respectively.

First, the patient was placed on a CT table under conscious sedation, and an US examination was performed to detect the target renal tumor. If possible, biopsy was performed under US guidance. If the target was undetected with US or CT guidance appeared safer and/or more reliable, the biopsy was performed under CT fluoroscopy guidance. Before CT fluoroscopy-guided biopsy, a conventional abdominal CT examination with a 5-mm slice thickness was performed to precisely locate the tumor and plan the needle insertion route. If the target was not visible on plain CT images, a contrast medium was administered intravenously for tumor visualization before obtaining the first specimen. After administration of local anesthesia with lidocaine, the introducer needle was advanced until its tip was in front of the tumor. The internal stylet of the introducer needle was replaced with a biopsy needle. Furthermore, multiple specimens were obtained until the quantity was sufficient (e.g., a total core length > 2 cm). When operators lacked certainty regarding needle placement within the targeted tumor, additional biopsies were performed until they were certain that the target had been successfully biopsied. The tumor target area (i.e., central or peripheral) for the biopsy was determined at the operator’s discretion. Biopsy specimens were divided into first and subsequent specimens and were diagnosed histopathologically. Complications were evaluated using US or plain conventional CT with a ≤ 5-mm slice thickness. All complications were graded according to the Clavien–Dindo classification of surgical complications [10].

The specimens were evaluated for their appropriateness for diagnosis. The pathologists evaluated and diagnosed the benign or malignant tumors histologically. In RCC tumors, the subtype and Fuhrman grade were evaluated. The first and subsequent specimens were paraffin-embedded separately, and tissue specimens were prepared. Evaluations of “the first specimen” and “all specimens” were blindly and independently performed by two board-certified pathologists with 31 and 16 years of experience (other and T.T.); the diagnosis was made via consensus (Fig. 1). If a benign lesion was histologically diagnosed, a 3-month imaging follow-up was performed to confirm whether the tumor size was unchanged.

Flowchart of histological evaluation

According a previous study’s results [11] (59% diagnostic yield with one core, 77–80% with two cores [vs. one core, P < 0.01], and 85% with three cores [vs. one core, P = 0.001]), we set the diagnostic yield of the first specimen = 0.59 and that of all specimens (i.e., first and subsequent specimens) = 0.80. Furthermore, we set α = 0.05 and β = 0.2. Because this study compares the diagnostic yield between the first specimen and all specimens for the same patient, we calculated a sample size sufficient to detect differences using McNemar's test.

The probabilities of a 2 × 2 table under these conditions are as follows (Table 1). The probability of “b” = 0 because diagnosis by all specimens was always possible when diagnosis by the first specimen was possible. The formula had to be assigned a small value instead of b = 0 for the calculations. Using b = 0.01 to 10–10 as a small number, 35–41 pairs were needed. Since the pairs are “the first specimen” and “all specimens,” the required tumor number was between 35 and 41. Calculating where the actual number in cell b is 0 pairs and the actual number in cell c is six pairs or more, McNemar’s test showed significance (“a” and “d” do not contribute to the test results). This number increased by 10% in anticipation of protocol deviations. Therefore, we enrolled 45 patients with renal tumors.

The diagnostic yields of the first specimen and all specimens were calculated and compared using McNemar’s test. The core number was compared between the two groups (i.e., biopsies with and without complications) using the Mann–Whitney U test.

The histological results of the first specimen were classified as a diagnostic success or failure. Patient-, tumor-, and biopsy-related variables were evaluated to assess risk factors for diagnostic failure. Age and sex were included as patient-related variables. Tumor-related variables included size, laterality (left or right kidney), anteroposterior location (ventral or dorsal), longitudinal location (upper or lower pole), position (exophytic or non-exophytic), and diagnosis (benign lesion or malignancy). Biopsy-related variables included operator experience (staff or non-staff), guided image modality (CT or US), and tumor puncture site (central or non-central). Tumor position was classified as exophytic or non-exophytic (parenchymal, mixed, and central) according to Gervais et al.’s definition [12]. Tumor locations were classified according to the transverse and longitudinal kidney axes as follows: ventral (tumor arises from the anterior half), dorsal (tumor arises from the posterior half), upper pole (tumor arises from the upper half), and lower pole (tumor arises from the lower half).

The variables were compared between the two groups using Fisher’s exact test for categorical values and the Mann–Whitney U test for numerical values. Two authors (J.S. and K.T.) performed analyses using R version 3.6.1 and SPSS version 26 (IBM Corp., Armonk, NY, USA). Statistical significance was set at P < 0.05.