Robotic partial nephrectomy versus percutaneous cryoablation for small renal masses: A propensity score matched comparative analysis

Background Robotic partial nephrectomy (RPN) and percutaneous cryoablation (CA) are established nephron-sparing treatment options for small renal masses (sRMs); however, direct comparison remains challenging because these modalities are typically offered to clinically distinct patient populations, resulting in substantial selection bias in existing observational studies. We aimed to compare the perioperative, functional, and oncological outcomes of RPN and CA using propensity score matching (PSM) to minimize baseline differences and provide a more balanced comparative analysis.Methods A retrospective cohort study of 337 consecutive patients treated with RPN (n = 271) or CA (n = 66) for clinical T1a renal masses between January 2012 and March 2026. A 1:1 nearest-neighbour PSM algorithm incorporating eleven pre-treatment covariates yielded 50 matched pairs (C-statistic 0.862; caliper 0.383 logit units). Primary endpoints were local disease progression and eGFR percent change.Results After PSM, with a median follow-up of 20 months in the CA group and 28 months in the RPN group, local disease recurrence was significantly higher in the CA group (16.0% vs. 0.0%; p = 0.010). eGFR percent change was equivalent (CA 0.0% vs. RPN −4.4%; p = 0.184). CA was associated with shorter operative time (125 vs. 155 min; p < 0.001) and hospital stay (1.0 vs. 5.0 days; p < 0.001). Overall complication rates were not significantly different (10.0% vs. 22.0%; p = 0.173). No cancer-specific deaths occurred in either group.Conclusions RPN was associated with superior local oncological control, whereas renal functional preservation, cancer-specific mortality, and distant metastasis were comparable to percutaneous CA. The comparable functional outcomes after propensity score matching suggest that previously reported renal functional advantages of CA may largely reflect baseline differences in patient characteristics.