The study included 52 kidney percutaneous needle biopsies collected at the National Cancer Institute in Vilnius, Lithuania between 2018 and 2021. The biopsies were taken during routine procedure for confirmation of diagnosis of RCC, immediately flash-frozen after the procedure, and stored until DNA extraction at -80 °C.

The patients were selected if (1) older than 18 years; (2) had a solid renal mass less than 4 cm in diameter diagnosed by imaging (CT, MRI, or ultrasonography); (3) patient’s agreement for active surveillance of SMR was obtained and the informed consent form was signed. The patients that had an uninformative biopsy, a life expectancy of less than a year, or had undergone systematic therapy for malignancy were excluded. The cases with suspected hereditary cancer were also excluded. The study was approved by the regional bioethics committee (No. 158200-17-952-457).

Of the 52 biopsies, 38 were confirmed RCC (5 chRCC, 26 ccRCC, 6 pRCC, and 1 p/ccRCC case), and 14 were benign (10 OCT, 3 AML, and 1 other). Clinical data is provided in Table 1.

DNA extraction from kidney needle biopsy samples was performed using standard phenol-chloroform extraction and ethanol precipitation protocols. In brief, the needle biopsy tissue samples digested for 18 h at 55 °C with Proteinase K (Thermo Scientific, Wilmington, DE, USA) and 500 µL lysis solution containing (50 mM Tris-HCl, pH 8.5; 1 mM EDTA; 0.5% Tween-20, Carl Roth, Karlsruhe, Germany). After incubation, Phenol/Chloroform/Isoamyl alcohol (25:24:1, Carl Roth, Karlsruhe, Germany) was used for DNA extraction following the chloroform (Carl Roth, Karlsruhe, Germany) step. The final DNA is precipitated using 40 µL 5 M ammonium acetate (Thermo Fisher, Kandel, Germany), 1 µL glycogen (Thermo Scientific, Vilnius, Lithuania), and 1 mL 96% ethanol. The DNA was then washed twice using 70% ethanol and dissolved in nuclease-free water. DNA quantity was measured using Qubit™ dsDNA BR Assay Kit on a Qubit™ 2.0 Fluorimeter (Invitrogen, TFS, Eugene, OR, USA). The DNA samples were stored at -80 °C until further experiments.

All 52 biopsy samples were analyzed for predominant mutations in CHEK2 using TaqMan SNP Genotyping assays (rs17879961 c.470T > C; rs555607708 c.1100delC; rs121908698 c.444 + 1G > A). The qPCR reactions were conducted using 2X TaqMan Universal Master Mix II (Applied Biosystems (AB), Thermo Fisher Scientific Baltics, Vilnius, Lithuania) on QuantStudio 5 Real-Time PCR System (AB, Singapore) following the manufacturer’s instructions.

A custom 50 gene panel consisting of ABL1, AKT1, ALK, APC, ATM, BRAF, CDH1, CDKN2A, CSF1R, CTNNB1, EGFR, ERBB2, ERBB4, EZH2, FBXW7, FGFR1, FGFR2, FGFR3, FLT3, GNA11, GNAQ, GNAS, HNF1A, HRAS, IDH1, IDH2, JAK2, JAK3, KDR, KIT, KRAS, MET, MLH1, MPL, NOTCH1, NPM1, NRAS, PDGFRA, PIK3A, PTEN, PTPN11, RB1, RET, SMAD4, SMARCB1, SMO, SRC, STK11, TP53, and VHL was used for targeted next generation sequencing of the 52 DNA samples from kidney percutaneous needle biopsies.

The sequencing libraries were prepared using Ion AmpliSeq™ Library Kit 2.0 and custom On-Demand Panel (Life Technologies (LT), Carlsbad, CA, USA) following the manufacturer’s protocols. Library quantification was conducted using Ion Library TaqMan™ Quantification Kit (AB, TFS, Vilnius, Lithuania). The final libraries were sequenced on the Ion 520™ chip using the Ion Torrent™ Ion S5™ system (LT, Singapore). The NGS data analysis was conducted on the Ion Reporter 5.18 tool (LT, Carlsbad, CA, USA).

Each variant was first classified using ClinVar database (Landrum et al. 2018) according to the American College of Medical Genetics and Genomics (ACMG) / the College of American Pathologist (AMP) guidelines (Richards et al. 2015) into benign/likely benign, pathogenic/likely pathogenic or uncertain significance (VUS) groups. Variants not listed or without interpretation on ClinVar were regarded as VUS. Every VUS was then evaluated using five publicly available interpretation knowledgebases: Varsome (Saphetor SA) (Kopanos et al. 2019), Franklin (Genoox) (2024), InterVar (Li and Wang 2017), Cancer Genome Interpreter (CGI) (Tamborero et al. 2018), and CancerVar (Li et al. 2022) for mutation pathogenicity interpretation. The VarSome, Franklin, CancerVar, and InterVar databases provide predictions for ACMG/AMP classification using unique data aggregation algorithms and use the available data to automatically classify each variant according to ACMG/AMP criteria. Franklin database separately classify each variant according to both ACMG (classifying each variant into benign/likely benign, VUS, low or moderate oncogenic support, and pathogenic/likely pathogenic variants) and AMP classification (using tier system: Tier 1 – strong clinical significance, Tier 2 – potential clinical significance, Tier 3 – unknown clinical significance, Tier 4 – benign/likely benign) (Li et al. 2017). CancerVar OPAI algorithm takes CancerVar interpretation and combines the 23 in silico scores for semi-supervised deep learning oncogenicity prediction (Li et al. 2022). CGI base provides a rule based OncodriveMUT algorithm to designate variants as passenger or driver mutations (Tamborero et al. 2018). All knowledgebases accessed in January 2024.

For statistical analysis, the Shapiro-Wilk’s W test was used for testing the normal distribution. Mann-Whitney U test or Welch two sample t test was used for testing associations between two independent samples as appropriate, while Fisher’s exact test was used for categorical variable associations. Univariable and multivariable odds ratios calculated using logistic regression models. The data was analyzed using R x64 (version 4.3.1, R Foundation for statistical computing, Vienna, Austria) on the RStudio (version 2023.06.0, Posit, PBC, Boston, MA, USA) software. ComplexHeatmap package (version 2.16.0) was used for oncoprint visualization. Receiver operating characteristic (ROC) analysis was performed using pROC package (version 1.18.5). Statistical significance was considered when p-value was < 0.050.