This study adhered to the standards specifed in PRISMA30((Preferred Reporting Items for Systematic Reviews and Meta-Analysis) and was prospectively registered in the PROSPERO database (CRD42023 411427). Articles included in the systematic review were investigated independently by two reviewers (WZ and WL). The data obtained from the literature were before April 1, 2023. Extensive literature searches were conducted using MEDLINE, EMBASE, and Google Scholar databases. The search was limited to English language papers. Medical Subject Headings (MeSH) terms and keywords, such as " Solitary Kidney*", " Ablation*" OR "Cryoablation*", " Partial Nephrectomy". In addition, we manually searched and reviewed relevant references to avoid any omission.
Selection criteria
The inclusion and exclusion criteria for the studies were determined using the PICOS method. The inclusion criteria consisted of solitary renal tumors undergoing ablation in the experimental group and partial nephrectomy in the control group. The primary outcome measures of this study were changes in creatinine and estimated glomerular filtration rate, new onset of chronic kidney disease after surgery, postoperative dialysis rate (percentage of patients requiring dialysis after treatment), local recurrence, and postoperative metastasis. Secondary outcome measures included length of hospital stay, operation time, complications (intraoperative complications, postoperative complications, major complications), intraoperative blood loss, blood transfusion, overall survival, and cancer-specific survival. Cohort studies, case-control studies, and randomized controlled trials (RCT) were eligible for inclusion. Non-comparative studies, editorial comments, conference abstracts, case reports, unpublished studies or comments, articles published in non-English language, and studies with no data available were excluded. Data extraction was performed independently by two reviewers and included general information such as the first author, publication year, and demographic characteristics such as age, sex, follow-up time, tumor size, and perioperative outcomes such as operative time, length of hospital stay, and intraoperative complications.
Study screening and selection
Two independent authors (WZ and WL) manually screened all retrieved records. When consensus could not be reached between the two authors, it was resolved by consultation with a third author (LY). Retrospective, prospective, non-randomized, and randomized studies were included according to PICOS criteria. Reviews, conference abstracts, case reports, letters to editors, and editorials were excluded. Papers were selected for screening by reading the full text if found to be relevant to the objectives of this study.
Statistical analysis
This study used Review Manager V5.4.1 software (Cochrane Collaboration, Oxford, UK) for statistical analysis. Results are presented as 95% confidence intervals (CI) and odds ratios (OR) for dichotomous variables and weighted mean difference (WMD) for continuous variables. Data from some studies reporting only medians, quartiles, or extreme value ranges were converted to means and standard deviations (SDs) using data conversion tables provided by McGrath31. Meta-analysis was performed using the Mantel-Haenszel method for dichotomous variables and the inverse variance method for continuous variables. For survival data, because some articles do not give survival curves, we analyze survival data in the form of binary variables, the simplest way is to collect them directly from the original article, and if the data are provided only in the form of survival curves, we extract survival rates from some designated times. We used a random-effects model for all analyses due to the predictable significance of heterogeneity across trials. Study heterogeneity was calculated using the I2 statistic, with 0–40% defined as mild heterogeneity; 40% – 60% as moderate heterogeneity; 60–75% as large heterogeneity; and 75–100% as high heterogeneity32. Values of p < 0.05 were considered statistically significant.
Bias risk assessment
We included articles that were all cohort studies and no randomized cohort studies were identified. The risk of bias was assessed using the Newcastle-Ottawa Scale (NOS) and the quality of the literaturetative evaluation was evaluated using a semiquantistar system, which consisted of 9 stars.
Sensitivity Analysis
We used the leave-one-out method to exclude studies from the pooled effect one at a time to assess the robustness of the estimates. Furthermore, we evaluated the robustness based on the study cohort size (excluding studies with < 100 patients), which may contribute to heterogeneity. However, we cannot perform sensitivity analyses comparing three or fewer studies.
Publication Bias
Funnel plots were used to screen for potential publication bias.