Robotic vs. video-assisted thoracoscopic surgery for resectable non-small cell lung cancer: a systematic review and meta-analysis of randomized controlled trials

Introduction

Lung cancer continues to be the primary cause of cancer-related deaths worldwide, with non-small cell lung cancer (NSCLC) representing about 85% of all cases (1). For early-stage and selected locally advanced NSCLC, surgical resection remains the cornerstone of curative treatment (2). Over the past two decades, minimally invasive thoracic surgery has transformed the management of resectable disease (3). Video-assisted thoracoscopic surgery (VATS) has become the established standard minimally invasive approach, supported by robust evidence demonstrating its advantages over open thoracotomy, including reduced postoperative morbidity, shorter hospital stays, and improved quality of life (4,5).

In recent years, robotic-assisted thoracoscopic surgery (RATS) has emerged as a promising alternative, introducing a potential paradigm shift (6). The robotic platform offers three-dimensional high-definition visualization, wristed instruments with tremor filtration, and enhanced maneuverability, which address some technical limitations of VATS, particularly in complex anatomical dissections and lymph node sampling (7,8). While earlier studies and meta-analyses reported conflicting results regarding the clinical superiority of RATS, recent high-quality evidence has begun to clarify this debate (9-11).

Crucially, the publication of landmark multicenter randomized controlled trials (RCTs), such as the RAVAL and RVlob trials, has significantly shifted the evidence landscape (12-15). These large-scale studies provide robust Level I evidence that was previously lacking in meta-analyses. As a result, existing reviews which predated these pivotal trials are no longer reflective of the current state of knowledge. Therefore, we conducted a systematic review and meta-analysis incorporating all available RCTs (12-18) to provide a comprehensive and up-to-date comparison of RATS vs. VATS in the management of resectable NSCLC, with the aim of guiding clinical decision-making and surgical practice in the management of resectable NSCLC. We present this article in accordance with the PRISMA reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0917/rc) (19).

Methods

Protocol registration

This systematic review and meta-analysis were conducted following the guidelines outlined in the Cochrane Handbook for Systematic Reviews. The study protocol was registered with the International Prospective Register of Systematic Reviews (PROSPERO) under the registration number CRD420251168709.

Literature search strategy

A comprehensive search was conducted across PubMed, Embase, Cochrane Library, and Web of Science databases from their inception to February 25, 2026 to identify RCTs comparing RATS vs. VATS for resectable NSCLC. Reference lists of relevant reviews and meta-analyses were also manually screened to identify potential additional studies. No restrictions were imposed on publication year, and only studies published in English were included. The detailed search strategy is presented in Appendix 1.

Inclusion and exclusion criteria

Studies were included if they met the following criteria: (I) RCTs; (II) involving adult patients (≥18 years) pathologically diagnosed with NSCLC; (III) comparing VATS with RATS for NSCLC; (IV) at least one of the following: perioperative outcomes (operative time, blood loss, complication rate, conversion rate, length of hospital stay), oncological outcomes [lymph node dissection metrics, R0 resection rate, overall survival (OS), disease-free survival (DFS)], health-related quality of life (HRQoL), or cost-related outcomes.

Exclusion criteria were as follows: (I) non-RCTs; (II) studies with incomplete or insufficient data for meta-analysis; (III) studies involving patients with severe comorbidities that could affect surgical safety or contraindicate general anesthesia. Specifically, to align with the eligibility criteria of the included RCTs (such as the ROMAN trial (12) which excluded patients with severe heart disease), we defined “severe comorbidities” as: cardiovascular: New York Heart Association (NYHA) class III or IV heart failure, unstable angina, or left ventricular ejection fraction (LVEF) <50%. Pulmonary: severe chronic obstructive pulmonary disease (COPD) or other conditions with predicted postoperative FEV1 <40%. Other: active immunodeficiency or uncontrolled systemic infection.

Data extraction and quality assessment

Data extraction was performed independently by two reviewers using a pre-designed data extraction form. Discrepancies were resolved by discussion with a third reviewer. Initially, titles and abstracts were examined to exclude studies that did not comply with the inclusion criteria. Subsequently, the remaining articles were read in their entirety to finalize the selection of studies for inclusion. The extracted information included author, publication year, trial design, sample size (RATS/VATS), study population, follow-up duration, primary outcome, key baseline characteristics, perioperative outcomes, oncological outcomes, cost/HRQoL outcomes, main risk of bias (Tables 1,2).

The RVlob trial was reported in three separate publications (13-15). To address this, we treated the RVlob trial as a single trial unit in our analysis. Specifically, we extracted short-term surgical outcomes, long-term survival data, and HRQoL metrics from their respective publications but ensured that patient-level data were not double-counted by analyzing the cohorts separately per outcome domain. This approach aligns with the recommendations for handling multiple publications from a single RCT in Cochrane guidelines.

The Cochrane risk-of-bias tool (RoB 2) from the Cochrane RevMan 5.4.1 software was used to assess the methodological quality of the included RCTs. This tool evaluates seven domains, including random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other biases. Each domain was rated as “low risk”, “high risk”, or “unclear risk” based on the information provided in the study. The quality assessment process was carried out in accordance with the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions, and similar methods have been applied in previous meta-analyses (20,21).

Statistical analysis

Meta-analysis was conducted using RevMan 5.4.1 software. All statistical analyses, including sensitivity analyses and data preparation calculations, were verified using SPSS 26.0, with a significance threshold established at P<0.05.

Synthesis methods and model selection

The decision to pool studies was based on clinical and methodological similarity. For dichotomous variables, results were pooled as risk ratios (RR) with 95% confidence intervals (CIs). For continuous variables, results were pooled as mean differences (MD) with 95% CI.

To synthesize results, both fixed-effect and random-effects models were employed. The Inverse-Variance method was used for data synthesis. The choice between models was guided by the assessment of statistical heterogeneity: the fixed-effect model was used when heterogeneity was low, and the random-effects model (Mantel-Haenszel method) was used when heterogeneity was significant.

Heterogeneity was evaluated using the I² statistic and the Chi² test (Cochran’s Q). Statistical heterogeneity was defined as: I² <25% (low heterogeneity), 25–50% (moderate heterogeneity), and >50% (high heterogeneity). A significance level of α=0.10 was used for the Chi² test to determine the presence of heterogeneity.

Data preparation and handling of missing statistics

For continuous outcomes reported as medians and interquartile ranges (IQRs) rather than means and standard deviations (SDs), we estimated the sample means and SDs according to the Cochrane Handbook and using the methods described by Wan et al. (22) and Luo et al. (23). Specifically, when the sample size was large (n>25), the median was approximated as the mean, and the SD was estimated as IQR/1.35. For studies providing minimum and maximum values, more precise estimation formulas incorporating the range were applied. For studies that only reported sample sizes, mean values, and P values for two groups without providing SD, the pooled SD was estimated in the present study by first deriving the t-statistic from the P value, and then calculating SD using the two-sample t-test formula. The estimated SD was then used for data entry and statistical analysis in the meta-analysis.

Exploration of heterogeneity and sensitivity analysis

Subgroup analyses were conducted a priori to explore possible causes of heterogeneity based on study scale (large-scale vs. small-scale RCTs). We defined studies with a sample size exceeding 100 as large-scale and those with 100 or fewer participants as small-scale.

Sensitivity analyses were conducted to assess the robustness of the synthesized results. This included re-analyzing the data to ensure that the estimations for missing summary statistics (as described above) did not significantly alter the overall findings.

Assessment of publication bias

Assessment of publication bias using funnel plots or Egger’s regression test was not performed because the meta-analysis included fewer than 10 RCTs for the primary outcomes, which is generally considered insufficient for reliable detection of asymmetry.

Evidence quality assessment

The certainty of evidence for each critical outcome was evaluated using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework. Two independent reviewers assessed the following domains: risk of bias, inconsistency, indirectness, imprecision, and publication bias. Disagreements were resolved through consensus or adjudication by a third reviewer. Evidence was initially rated as high for RCTs, then downgraded (by one or two levels per domain) or upgraded (for large effect sizes, dose-response gradients, or plausible confounding bias) based on predefined criteria. Final evidence certainty was categorized as high, moderate, low, or very low according to GRADE guidelines.

Results

Literature search results

The initial literature search yielded 400 potentially relevant studies. After removing duplicates (n=241), 159 studies were screened based on titles and abstracts, of which 113 were excluded due to irrelevant study design, population, or interventions. Full-text evaluation of the remaining 46 studies resulted in the exclusion of 39 studies. Finally, 7 publications (12-18) involving 5 RCTs and 712 patients (338 in the RATS group and 374 in the VATS group) were included in the meta-analysis. Notably, the RVlob trial contributed 3 publications (13-15) but was counted as one RCT with 320 patients to avoid sample size duplication. The literature screening process and results are shown in Figure 1.

Figure 1 PRISMA chart of the literature screening process and results. NSCLC, non-small cell lung cancer; RCT, randomized controlled trial.

Basic characteristics of the included studies

The baseline characteristics of the included studies are summarized in Tables 1,2. These 5 RCTs (7 publications) were published between 2022 and 2024, with a total of 712 patients enrolled (range: 50 to 363 per study). Among them, 338 patients were allocated to the RATS group and 374 to the VATS group. The included studies were conducted across multiple countries, including Brazil, China, Canada, the United States, France, and Italy, with most being single-center trials except for 2 multicenter studies [RAVAL trial (16) and ROMAN trial (12)]. The study populations were predominantly patients with resectable NSCLC (clinical stages I–IIIa), with tumor diameters ranging from 2.0 cm to 2.45 cm, encompassing both early-stage and locally advanced disease. Key baseline characteristics (age, gender, BMI, comorbidities, tumor stage, and histopathological type) were generally balanced between the RATS and VATS groups in all included studies, except for more severe pleural adhesions in RATS in Catelli et al. [2023] (17).

Risk of bias assessment results

Risk of bias assessment was performed using the Cochrane Collaboration’s Risk of Bias Tool 2.0. The results indicated that all 7 publications were rated as low risk in random sequence generation and allocation concealment, while all 7 publications were rated as high risk in blinding of participants and personnel due to inherent differences between RATS and VATS (e.g., surgical incisions, instruments used) that precluded blinding. For blinding of outcome assessment, 6 publications were rated as unknown risk, and 1 as low risk. No significant risk of selective reporting or incomplete outcome data was identified in any included study, and other biases were considered low risk in all trials. The specific outcomes of the quality assessment are illustrated in Figure 2.

Figure 2 Assessment of bias risk. The risk of bias was assessed utilizing RevMan V5.4.1. (A) The risk of bias for each study was evaluated utilizing the Cochrane tool designed for the assessment of bias risk. (B) Presents a summary graph of bias risks, displayed as the percentage distribution across all included studies.

Meta-analysis results

Surgical safety and complications

Conversion to thoracotomy: regarding conversion to thoracotomy, a total of 712 patients from 5 RCTs (338 in the RATS group and 374 in the VATS group) were analyzed. The pooled results indicated a non-significant trend toward fewer conversions in the RATS group compared to the VATS group (RR =0.67, 95% CI: 0.37 to 1.21, P=0.19), with no significant heterogeneity (P=0.69, I²=0%) (Figure 3A). Subgroup analysis by sample size showed a similar non-significant trend in large-scale RCTs (RR =0.61, 95% CI: 0.31 to 1.19), whereas no difference was observed in small-scale RCTs (RR =0.95, 95% CI: 0.27 to 3.38). The test for subgroup differences was non-significant (P=0.54), indicating consistent findings across study scales (Figure 3B). Overall, these analyses suggest that RATS does not significantly differ from VATS in terms of the rate of conversion to thoracotomy, regardless of study scale.

Figure 3 Forest plots for surgical safety and complications outcomes. (A) Overall pooled analysis of conversion to thoracotomy, (B) subgroup analysis of conversion to thoracotomy by study scale, (C) overall pooled analysis of overall complications, (D) subgroup analysis of overall complications by study scale. CI, confidence interval; df, degrees of freedom; M-H, Mantel-Haenszel; RATS, robotic-assisted thoracic surgery; RCT, randomized controlled trial; VATS, video-assisted thoracoscopic surgery.

Overall complications: 5 RCTs reported postoperative complications, the pooled analysis indicated no statistically significant difference between RATS and VATS (RR =0.91, 95% CI: 0.68 to 1.21, P=0.52, I²=36%) (Figure 3C). Subgroup analysis stratified by sample size revealed comparable complication rates in both large-scale RCTs (RR =0.99, 95% CI: 0.76 to 1.29) and small-scale RCTs (RR =0.82, 95% CI: 0.46 to 1.48), with no significant interaction observed between subgroups (P=0.57, I²=0%); In the large-scale RCTs subgroup (n=2 RCTs, total n=338), the pooled RR was 0.99 (95% CI: 0.76 to 1.29; P=0.95), with low heterogeneity (I²=22%); In the small-scale RCTs subgroup (n=3 studies, total n=374), the RR was 0.82 (95% CI: 0.46 to 1.48; P=0.51), with moderate heterogeneity (I²=50%) (Figure 3D). The incidence of specific complications showed no statistically significant differences between RATS and VATS (Figure S1A-S1F). Collectively, these findings suggest that RATS and VATS have comparable safety profiles regarding overall complications, irrespective of study scale.

Perioperative clinical outcomes

Operative metrics

Operative time: a total of 5 RCTs comprising 712 patients (338 in the RATS group and 374 in the VATS group) were included in the meta-analysis of operative time. The pooled results indicated no statistically significant difference between the RATS and VATS groups (MD =4.92, 95% CI: −7.82 to 17.66, P=0.45). However, significant heterogeneity was observed (I²=69%, P=0.01) (Figure 4A). Subgroup analyses were performed to explore the sources of heterogeneity. When stratified by sample size, large-scale studies showed a trend toward shorter operative time in the RATS group (MD =−3.75, 95% CI: −16.48 to 8.98), whereas small-scale studies showed a trend toward longer operative time (MD =13.79, 95% CI: −4.08 to 31.67) (Figure 4B).

Figure 4 Forest plots for perioperative clinical outcomes. (A) Overall pooled analysis of operative time, (B) subgroup analysis of operative time by study scale, (C) overall pooled analysis of intraoperative blood loss, (D) subgroup analysis of intraoperative blood loss by study scale, (E) overall pooled analysis of postoperative hospital stay, (F) subgroup analysis of postoperative hospital stay by study scale, (G) chest tube duration, (H) readmission rates. CI, confidence interval; df, degrees of freedom; IV, inverse variance; M-H, Mantel-Haenszel; RATS, robotic-assisted thoracic surgery; RCT, randomized controlled trial; SD, standard deviation; VATS, video-assisted thoracoscopic surgery.

Intraoperative blood loss: data regarding intraoperative blood loss were available from 4 RCTs, comprising a total of 635 patients (300 in the RATS group and 335 in the VATS group). The overall pooled analysis revealed that RATS was associated with a statistically significant reduction in blood loss compared to VATS (MD =−62.27 mL; 95% CI: −119.25 to −5.30; P=0.03, Figure 4C). However, substantial heterogeneity was observed among the included studies (I²=96%), necessitating the use of a random-effects model. To explore potential sources of this heterogeneity, subgroup analyses were conducted based on sample size. Neither the large-scale subgroup (MD =−49.69 mL, P=0.32) nor the small-scale subgroup (MD =−197.86 mL, P=0.28) showed a statistically significant difference (Figure 4D). Furthermore, the test for subgroup differences was not significant (P=0.43, I²=0%), suggesting that study scale does not account for the observed variability. Sensitivity analyses using a leave-one-out approach were performed to assess the robustness of the results. Excluding individual studies one by one did not substantially alter the direction of the pooled estimate, nor did it reduce the heterogeneity (I² remained ≥90% in all iterations, as shown in Figure S2). This indicates that the high heterogeneity is likely intrinsic to the clinical diversity across centers rather than driven by a single outlier study. Despite this variability, the overall conclusion favoring RATS for reduced blood loss remained consistent.

Recovery indicators

Postoperative hospital stay: 5 RCTs reported data on the length of hospital stay, involving a total of 712 patients (338 in the RATS group and 374 in the VATS group). The overall pooled analysis indicated that there was no statistically significant difference between the RATS and VATS groups (MD =−0.38, 95% CI: −1.07 to 0.30, P=0.28). However, significant heterogeneity was observed among the studies (I²=79%) (Figure 4E). To explore the source of heterogeneity, subgroup analyses were performed based on trial scale. In large-scale RCTs (2 RCTs), RATS was associated with a trend toward a shorter hospital stay, although this did not reach statistical significance (MD =−0.56, 95% CI: −1.53 to 0.42). In small-scale RCTs (3 RCTs), no significant difference was observed (MD =−0.18, 95% CI: −1.39 to 1.04). The test for subgroup differences was not significant (P=0.63) (Figure 4F).

Chest tube duration: data were available from 4 RCTs comprising 637 participants (313 in the RATS group and 324 in the VATS group). No statistically significant difference was observed between the RATS and VATS groups (MD =0.00; 95% CI: −0.14 to 0.14; P>0.99). The absence of heterogeneity (I²=0%) suggests that high consistency in measurement and outcome across included studies (Figure 4G).

Readmission rates: 3 RCTs reported data on hospital readmission rates, involving a total of 473 patients (232 in the RATS group and 241 in the VATS group). The overall pooled analysis showed no statistically significant difference between the RATS and VATS groups (RR =0.90, 95% CI: 0.11 to 7.24, P=0.92), with moderate to substantial heterogeneity observed (I²=66%) (Figure 4H).

Oncological efficacy

Lymph nodes harvested: 3 RCTs reported the number of lymph nodes harvested during surgery, involving a total of 561 patients (276 in the RATS group and 285 in the VATS group). The overall pooled analysis demonstrated that the RATS group retrieved significantly more lymph nodes than the VATS group (MD =2.56, 95% CI: 0.53 to 4.58, P=0.01), with high heterogeneity (I²=88%) (Figure 5A).

Figure 5 Forest plots for perioperative clinical outcomes. (A) Overall pooled analysis of total lymph nodes harvested, (B) nodal upstaging, (C) 3-year overall survival, (D) 3-year disease-free survival, (E) need for adjuvant therapy. CI, confidence interval; df, degrees of freedom; IV, inverse variance; M-H, Mantel-Haenszel; RATS, robotic-assisted thoracic surgery; SD, standard deviation; VATS, video-assisted thoracoscopic surgery.

For nodal upstaging: 4 RCTs reported the rate of nodal upstaging (defined as pathological N stage higher than clinical N stage), involving a total of 637 patients (313 in the RATS group and 324 in the VATS group). The meta-analysis showed no statistically significant difference between the two groups (RR =0.66; 95% CI: 0.40 to 1.09; P=0.11). There was no evidence of heterogeneity among the studies (I²=0%), indicating consistent findings across trials (Figure 5B).

Three-year OS: 2 RCTs reported OS data, comprising a total of 379 patients (174 in the RATS group and 205 in the VATS group). The meta-analysis revealed no statistically significant difference in OS between the RATS and VATS groups (RR =1.02; 95% CI: 0.97 to 1.08; P=0.38). There was no evidence of heterogeneity among the included studies (I²=0%), indicating consistent findings across trials (Figure 5C).

Three-year DFS: 2 RCTs reported the 3-year DFS rates, involving a total of 370 patients (167 in the RATS group and 203 in the VATS group). The meta-analysis showed no statistically significant difference between the two groups (RR =1.05; 95% CI: 0.97 to 1.12; P=0.21). There was no evidence of heterogeneity among the included RCTs (I²=0%), indicating highly consistent oncological outcomes across trials (Figure 5D).

Need for adjuvant therapy: three RCTs reported the proportion of patients requiring adjuvant therapy after surgery, involving a total of 530 patients (259 in the RATS group and 271 in the VATS group). The meta-analysis showed no statistically significant difference between the two groups (RR =0.84; 95% CI: 0.58 to 1.22; P=0.37). There was no evidence of heterogeneity among the included RCTs (I²=0%), indicating consistent findings across trials (Figure 5E).

HRQoL and cost outcomes

Postoperative pain: 3 RCTs reported postoperative pain scores using Visual Analog Scale (VAS), involving a total of 559 patients (263 in the RATS group and 296 in the VATS group). Although pain assessments were conducted at different postoperative time points across RCTs (ranging from postoperative day 1 to week 4), the meta-analysis demonstrated that RATS was associated with significantly lower pain scores compared to VATS (MD =−0.33; 95% CI: −0.36 to −0.31; P<0.001). Notably, there was no statistical heterogeneity (I²=0%), suggesting a highly consistent effect across the included RCTs despite variations in assessment timing (Figure 6A).

Figure 6 Forest plots for HRQoL and cost outcomes. (A) Postoperative pain, (B) EQ-5D scores. CI, confidence interval; df, degrees of freedom; EQ-5D, EuroQol 5 Dimensions; HRQoL, health-related quality of life; IV, inverse variance; RATS, robotic-assisted thoracic surgery; SD, standard deviation; VATS, video-assisted thoracoscopic surgery.

Quality of life: 4 RCTs assessed HRQoL, primarily utilizing the EQ-5D and EORTC QLQ-C30 scales. As Catelli et al.(17) and the ROMAN study (12) did not report specific numerical values, describing the results only as “no statistically significant difference”, the quantitative synthesis was mainly derived from RAVAL(16) and RVlob (13) trials, with 446 patients were included in the meta-analysis of quality of life, as measured by the EQ-5D score. The pooled analysis demonstrated a non-significant trend favoring the RATS group (MD =0.02, 95% CI: −0.03 to 0.007, P=0.48). Notably, considerable statistical heterogeneity was observed across the RCTs (I²=80%, Figure 6B), thereby preventing a conclusive determination of the superiority of either surgical approach. This pronounced variability stemmed from discordant findings: whereas the RAVAL trial documented significantly elevated EQ-5D scores for the RATS cohort at 12 weeks, the RVlob study detected no significant disparity at the 4-week mark. Consequently, although current evidence indicates that RATS and VATS yield comparable HRQoL outcomes, discrepancies in assessment timelines and reported metrics constrain the validity of pooling these estimates.

Cost analysis: 3 RCTs reported heterogeneous cost data precluding standardized pooling. Catelli et al. (17) qualitatively noted higher surgical costs for RATS. The RAVAL trial (16) showed significantly higher direct medical costs for RATS but deemed it cost-effective based on ICER and comparable EQ-5D scores. The RVlob trial (14) detailed significantly higher total costs for RATS (RATS vs. VATS: $12,821 vs. $8,009; P<0.001), primarily attributable to indirect costs ($5,197 vs. $453; P<0.001) while direct medical costs were similar ($7,624 vs. $7,572; P=0.15). These studies suggest that RATS’s elevated costs stem from equipment-related expenses, not procedural costs, implying potential future cost-effectiveness improvements.

GRADE assessment

The results of the GRADE assessment for the primary outcomes are summarized in Table 3 (21). Overall, the certainty of evidence for most outcomes was rated as “High”. However, we downgraded the certainty to “Moderate” for Operative Time and Total Lymph Nodes Yield due to serious statistical heterogeneity (I²>50%), despite the use of random-effects models. The evidence suggests that RATS is associated with a similar risk of conversion and overall complications compared to VATS, with high certainty. For intraoperative blood loss, RATS showed a significant advantage with high-certainty evidence.

Discussion

Analysis of results

This meta-analysis synthesizes the current high-level evidence from RCTs comparing RATS and VATS for the surgical treatment of NSCLC. By incorporating recent landmark trials such as the RAVAL and RVlob studies, our analysis provides a comprehensive assessment of perioperative, oncological, and patient-reported outcomes. The results indicate that RATS demonstrates comparable safety and efficacy to VATS. However, the higher economic burden associated with robotic technology remains a significant consideration.

Surgical safety and perioperative outcomes

One of the primary concerns in adopting new surgical technology is the potential for increased complications or conversion rates. Regarding surgical safety, our analysis demonstrated equivalence between RATS and VATS in terms of overall complications (RR =0.91) and conversion rates (RR =0.67). This finding is consistent with subgroup analyses by study scale, suggesting that the safety profile of RATS is robust and not influenced by the learning curve. Furthermore, the two techniques showed no significant difference in the incidence of specific complications such as prolonged air leak, atrial fibrillation, or pneumonia. This indicates that the adoption of robotic technology does not compromise patient safety when compared to the established VATS approach.

Notably, RATS was associated with a statistically significant reduction in intraoperative blood loss (MD =−62.27 mL) compared to VATS; however, this analysis was characterized by substantial statistical heterogeneity (I²≥96%) across all sensitivity and subgroup analyses. Unlike typical meta-analyses where heterogeneity can be resolved by excluding outliers, our leave-one-out analysis suggests that this variability stems from fundamental differences in study design or surgical execution across different centers. Several factors may contribute to this inconsistency: first, variability in surgical techniques: “RATS” and “VATS” are broad terms. The specific extent of lymph node dissection, the type of energy devices used, and the definition of “blood loss” vary significantly between institutions. For instance, the BRAVO trial reported extremely high SDs, potentially reflecting a more complex patient cohort or less standardized perioperative management. Second, learning curve effects: the proficiency of surgeons with robotic systems varies. Centers with a steep learning curve might show different hemostatic outcomes compared to high-volume expert centers, contributing to the wide CIs seen in smaller trials like Catelli et al. Third, patient selection bias: differences in inclusion criteria, such as tumor size, location, or prior chest surgery history, could inherently affect bleeding risks regardless of the surgical approach. Despite the high heterogeneity, the consistency of the point estimates favoring RATS across sensitivity analyses provides reassurance regarding the directional benefit of the robotic approach in minimizing bleeding. The mechanism underlying the reduced blood loss in RATS, may be attributed to the technical advantages of the robotic platform in vascular dissection. The tremor filtration and motion scaling features of the robotic system facilitate precise dissection and coagulation around pulmonary vessels, potentially minimizing the risk of minor vascular injuries. Furthermore, the 3D high-definition magnification offers superior visualization of the vascular anatomy compared to standard 2D VATS, aiding in the safer identification and sealing of vascular branches.

The primary finding regarding operative time was the lack of a significant overall difference between RATS and VATS, which aligns with previous meta-analyses (9,20,21,24-26). However, the substantial heterogeneity observed (I²=69%), indicating that operative efficiency varies considerably across studies. To explore potential sources of this heterogeneity, we conducted a subgroup analysis based on sample size. While this analysis revealed a trend where smaller-scale studies favored VATS and larger-scale studies showed comparable outcomes, we acknowledge that sample size alone cannot serve as a direct surrogate for surgical volume or the learning curve. Large multicenter trials often include sites at various stages of robotic adoption, potentially diluting the effect of high-volume expertise, whereas smaller studies may reflect specific protocols or highly selected patient cohorts in specialized settings. Therefore, the observed heterogeneity is likely driven by unreported clinical and technical factors rather than study size per se. For instance, variations in robotic platform generation (e.g., multi-port vs. single-port systems) and institutional workflows (e.g., parallel docking, standardized instrument exchange) are critical determinants of operative efficiency that are rarely stratified in current literature (26-28). The adoption of newer technologies, such as single-port robotic systems, has been shown to reduce port placement and docking times, potentially explaining the superior efficiency seen in certain cohorts regardless of sample size (29,30). Furthermore, differences in case complexity and surgeon experience levels across studies introduce clinical heterogeneity that statistical adjustments cannot fully capture. Future studies should aim to standardize the reporting of specific robotic models, surgeon experience levels, and procedural details to better elucidate these technical determinants. In summary, while RATS demonstrates comparable operative efficiency to VATS in the aggregate, realizing this potential requires optimized workflows and technological maturity, highlighting the need for nuanced interpretation beyond simple trial size categorization.

Oncological efficacy: technical advantage vs. clinical impact

The comparative analysis of oncological outcomes presents a nuanced picture. Our meta-analysis demonstrated a statistically significant advantage for RATS in the total number of lymph nodes retrieved (MD =2.56 nodes). However, this finding was associated with substantial heterogeneity (I²=88%), suggesting variability in reporting standards or pathological processing protocols between institutions, such as the differences observed between the RAVAL and RVlob trials. We acknowledge that this variability may be influenced by operator-dependent factors (such as the surgeon’s experience with mediastinal dissection) or institutional pathology protocols (differences in the submission and examination of fatty tissue). Therefore, while the technical capability of RATS for meticulous dissection is evident, the variability in reporting standards should be considered when interpreting the pooled numerical estimate.

From a mechanistic perspective, this increased nodal yield can be attributed to the enhanced dexterity afforded by wristed instrumentation. Unlike the rigid, non-articulating instruments used in conventional VATS, the robotic “wrist” allows for articulation within the chest cavity, enabling the surgeon to manipulate tissue in a manner analogous to open surgery (28). This facilitates more meticulous mediastinal lymphadenectomy, particularly in anatomically challenging stations, where the ability to retract and dissect simultaneously improves the completeness of fat clearance. Furthermore, the improved ergonomics reduce surgeon fatigue during lengthy dissection, potentially contributing to a more thorough harvest (30).

Despite this technical superiority, the clinical significance of retrieving an average of 2.56 additional nodes requires careful consideration. While statistically significant, an MD of less than three nodes may not be clinically decisive in the majority of individual cases, this quantitative advantage did not translate into a higher rate of nodal upstaging (RR =0.66) or a reduced need for adjuvant therapy (RR =0.84), neither in OS (RR =0.12) or DFS (RR =1.05), compared to VATS. This is further contextualized by the established literature, such as ACOSOG Z0030 trial (31), showing that in NSCLC, systematic lymphadenectomy (complete removal of lymph node stations) has demonstrated equivalent oncologic outcomes to lymph node sampling (targeted removal of suspected nodes) in terms of survival and recurrence, particularly in early-stage disease. Thus, the impact of these additional nodes on actual patient management appears limited (32).

HRQoL and pain

Patient-reported outcomes are increasingly recognized as vital metrics in surgical evaluation. This analysis presents a statistically significant difference in postoperative pain favoring the RATS approach compared to conventional VATS, with an MD of −0.33 (95% CI: −0.36 to −0.31). While this result is biologically plausible given the theoretical biomechanical advantages of the robotic platform (e.g., tremor filtration and precise tissue handling), its interpretation requires considerable nuance. Although the statistical significance is clear (Z=27.36, P<0.001), its clinical significance remains questionable. The marginal reduction of 0.33 points on the VAS is generally considered to be below the threshold for a clinically meaningful change perceptible to an individual patient (typically 1–2 points) (33). Therefore, despite the statistical advantage, the magnitude of this benefit is likely too small to represent a major patient-centered advantage or to substantially alter perioperative pain management strategies, such as reducing opioid consumption. Therefore, the pooled result, though highly significant, suggests that the magnitude of benefit in the current era of enhanced recovery might be too small to represent a major patient-centered advantage. Further complicating the interpretation is the heterogeneity in measurement methodology across the included RCTs. Pain was assessed at widely varying time points, from postoperative day 1 up to week 4, and in different contexts such as at rest vs. during cough. While the lack of statistical heterogeneity (I²=0%) indicates a consistent directional effect favoring RATS, this methodological variability prevents a precise understanding of when the analgesic benefit occurs. Consequently, it remains uncertain whether the technical features of robotic surgery reliably translate into tangible clinical outcomes like reduced analgesic consumption or faster functional recovery (34). In summary, although RATS is associated with a statistically significant but marginal reduction in postoperative pain, the clinical relevance of this small-magnitude effect is questionable (35). Future research should employ standardized pain assessment protocols to determine if this statistical advantage can be translated into a meaningful improvement in patient experience and recovery.

When evaluating overall health status using the EQ-5D instrument, no statistically significant difference was observed between RATS and VATS, with an MD of 0.02. However, the analysis revealed substantial heterogeneity (I²=80%), which is not merely a statistical artifact but reflects the inherent complexity of real-world clinical outcomes and patient-reported experiences. This high level of heterogeneity is primarily driven by the conflicting results from two major trials: the RAVAL trial, which reported a short-term utility gain associated with RATS, and the RVlob trial, which found no such difference. The discrepancy between these trial outcomes suggests that the factors influencing patient-reported quality of life extend well beyond the technical aspects of the surgical procedure itself (32). While RATS may offer specific intraoperative advantages, such as reduced blood loss or improved lymph node dissection, these potential benefits might be offset or masked in broad-spectrum quality of life assessments by various non-surgical factors (36). These include the speed of returning to normal daily activities, patient anxiety related to the perception of robotic technology, and differences in postoperative care pathways or regional healthcare practices. The influence of such factors on subjective well-being highlights how the “noise” from the postoperative period and contextual healthcare differences can obscure the “signal” attributable to the surgical technique in generic HRQoL instruments like the EQ-5D (37).

Collectively, these findings indicate that while RATS provides certain technical advantages, it does not currently demonstrate definitive superiority over VATS in generic, broad-spectrum quality of life metrics. This underscores the complexity of measuring subjective well-being and the limitations of relying solely on global utility scores. Therefore, future research should prioritize the concurrent use of disease-specific quality of life instruments, such as the EORTC QLQ-LC13 for lung cancer, alongside generic measures (38,39). This approach is essential to more accurately capture the specific domains and potential unique benefits of RATS that may not be apparent in overall health utility assessments.

Cost implications

Our analysis reveals a complex economic landscape for RATS. A critical barrier to its widespread adoption remains the significantly higher total hospitalization costs compared to VATS. Detailed cost breakdowns from major trials, such as RVlob and RAVAL, indicate that this disparity is almost entirely attributable to indirect costs rather than direct procedural expenses. Specifically, while direct medical costs—including items like staplers and disposables—are comparable between RATS and VATS groups, the robotic approach incurs substantially higher indirect costs (40). These are primarily driven by the high capital expenditure required for robotic systems and the subsequent costs of equipment depreciation and maintenance (41).

The current financial burden on healthcare systems is therefore undeniably higher with RATS. However, this cost structure is not necessarily static. Our findings prompt a critical reflection on the future trajectory of robotic surgery economics. We hypothesize that potential shifts in healthcare financing models—such as a move toward equipment leasing rather than capital purchase—could alter the economic equation. This hypothesis is increasingly relevant as some healthcare systems in emerging markets are adopting operational leasing models to access high-cost surgical robotics without heavy capital investment. Furthermore, the increasing market penetration of domestically manufactured robotic systems, particularly in regions like China, may lead to a substantial reduction in hardware depreciation costs over time (42).

In such a future scenario, RATS would retain its documented clinical advantages, such as reduced intraoperative blood loss and improved lymph node dissection yield, without imposing a higher financial burden on the healthcare system. While the RAVAL trial suggested that RATS might already approach cost-effectiveness within certain willingness-to-pay thresholds due to comparable health utility gains, a broader paradigm shift in cost structures could solidify its position. This potential evolution suggests that RATS may represent not only a significant surgical innovation but also a future cost-effective strategy. For healthcare policymakers, this underscores the importance of adopting a long-term, dynamic perspective when evaluating robotic technology. Strategic investments in infrastructure today may yield significant clinical and economic dividends in the future as the underlying costs of the technology inevitably decline (43).

While the RAVAL and RVlob trials provided the primary weight for long-term oncological and quality-of-life outcomes due to their large sample sizes, it is important to recognize the complementary role of the other included RCTs. The collective evidence from all 5 trials strengthens the robustness of our conclusions. Specifically, the ROMAN trial (12), despite its early termination, was the first multicenter RCT to compare RATS and VATS in a Western population. Its findings, consistent with our pooled analysis, support the safety of the robotic approach, particularly regarding conversion rates and major complications. The BRAVO trial (18) provided unique insights into postoperative care, reporting a significantly lower readmission rate in the RATS group, which aligns with the trend toward reduced morbidity observed in our analysis. Moreover, the single-center RCT by Catelli et al. (17), although smaller in scale, contributed valuable data on intraoperative efficiency, demonstrating that RATS can achieve comparable operative times to VATS in experienced hands, alongside reduced drainage volumes in the immediate postoperative period.

The convergence of evidence across these diverse studies, from single-center experiences to large multicenter trials, reinforces the conclusion that RATS is a reproducible and safe technique across different healthcare settings.

Comparison with previous studies

The findings of this study are largely consistent with previous meta-analyses and clinical studies comparing VATS and RATS for NSCLC treatment. For instance, a recent meta-analysis by Khan W et al. (24) reported similar results regarding intraoperative blood loss and postoperative complications, supporting the safety and feasibility of both techniques. However, our study differs from some previous reports in terms of operative time. While several studies have suggested that RATS is associated with longer operative times, possibly due to setup requirements and technical complexities (3-5), our analysis found no significant difference between the two techniques. This discrepancy may reflect improvements in surgical techniques and operator proficiency over time, as well as advancements in robotic system design.

One notable innovation of this study is its focus exclusively on clinical RCTs, which enhances the robustness of the evidence base compared to previous analyses that included non-randomized studies. By restricting inclusion criteria to RCTs, we minimized the risk of selection bias and confounding factors, thereby improving the internal validity of our results. Additionally, this study provides a comprehensive comparison of surgical-related indicators and postoperative outcomes, offering a more holistic assessment of the two techniques than some earlier investigations, which often focused on a limited subset of outcomes (44,45).

The value of this study lies in its ability to provide clinicians with up-to-date evidence to inform surgical decision-making for NSCLC patients. Specifically, our findings support the use of RATS as a viable alternative to VATS, particularly in cases where complex anatomical dissection is required. At the same time, we acknowledge the importance of considering resource availability and cost-effectiveness when choosing between the two techniques, as highlighted by previous research (24,44).

Limitations and future directions

While this meta-analysis provides valuable insights into the comparative outcomes of VATS and RATS for NSCLC, several limitations warrant careful consideration. First, the number of included studies, comprising five RCTs, is relatively small. Consequently, the total sample size for specific outcomes, particularly long-term survival and rare complications such as major adverse events, remains modest. This limitation constrains the statistical power of the analysis and increases the risk of a type II error, potentially precluding the detection of meaningful differences between the two surgical techniques for infrequent outcomes.

Second, inherent methodological constraints in surgical trials must be acknowledged. The blinding of participants and surgeons to the intervention was not feasible, which may introduce performance bias. However, the reliance on objective, pre-defined outcome measures helps to mitigate this concern. Third, significant heterogeneity was observed among the included studies. Variations in patient demographics, surgical protocols, and technical execution may have introduced bias into the pooled results. Furthermore, the heterogeneity in healthcare cost structures and economic systems between Eastern and Western countries limits the generalizability of the economic findings to specific regional contexts.

Fourth, although all included studies reported costs in US dollars, we did not adjust the data for inflation to a common reference year. While this limits the comparability of absolute cost values, the significant disparities in total costs and the specific drivers of these costs (e.g., capital investment) remain evident. Finally, the potential for publication bias cannot be ruled out, as the exclusion of unpublished studies or negative results could skew the findings. Moreover, while 3-year survival data are promising, longer-term follow-up (e.g., 5-year data) is required to definitively confirm the oncological non-inferiority of RATS, given the possibility of late recurrences in lung cancer. Collectively, these limitations affect the robustness and generalizability of our conclusions, particularly in resource-limited settings.

To address these gaps, future research should prioritize the conduct of large-scale, multicenter RCTs with extended follow-up periods. Such studies should aim to enroll a diverse patient population and employ standardized surgical and perioperative protocols to minimize clinical and methodological heterogeneity. Concurrently, efforts should be made to identify and incorporate data from unpublished studies or registries to reduce publication bias. Finally, comprehensive and region-specific cost-effectiveness analyses are urgently needed to evaluate the long-term economic implications of adopting RATS on a broader scale. By addressing these priorities, subsequent research can build upon the foundation established by this meta-analysis and further advance the field of minimally invasive thoracic surgery.

Conclusions

In conclusion, this meta-analysis confirms that RATS is a viable and safe alternative to VATS, with equivalent efficacy in major clinical and oncological outcomes. RATS demonstrates specific technical merits, such as reduced blood loss and improved lymph node dissection; however, these do not currently translate into superior patient-centered outcomes like survival or complication rates. The choice between RATS and VATS should be guided by institutional expertise, patient factors, and cost-effectiveness considerations.

Acknowledgments

During the preparation of this manuscript, the authors used Qwen (a Large Language Model developed by Alibaba Cloud) for the purpose of language polishing and grammar checking to improve the readability of the text. The authors reviewed and edited the content as needed and take full responsibility for the accuracy and integrity of the published article.

Footnote

Reporting Checklist: The authors have completed the PRISMA reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0917/rc

Peer Review File: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0917/prf

Funding: This work was supported by Luzhou Science and Technology Bureau Key R&D Project: Development and Validation of an Intervention Protocol for Operating Room Isolation Techniques in Lung Cancer Surgery (grant No. 2022-SYF-90); and Key Research and Development Program of Sichuan Provincial Department of Science and Technology (grant No. 24ZDYF0573): Identification of Prognosis-Related Genes Mediated by Cancer-Associated Fibroblasts in Lung Squamous Cell Carcinoma and Computer-Aided Targeted Drug Screening.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-0917/coif). All authors report that this work was supported by Luzhou Science and Technology Bureau Key R&D Project: Development and Validation of an Intervention Protocol for Operating Room Isolation Techniques in Lung Cancer Surgery (grant No. 2022-SYF-90); and Key Research and Development Program of Sichuan Provincial Department of Science and Technology (grant No. 24ZDYF0573): Identification of Prognosis-Related Genes Mediated by Cancer-Associated Fibroblasts in Lung Squamous Cell Carcinoma and Computer-Aided Targeted Drug Screening. The authors have no other conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Nemeth A, Friedel G, Groß T, et al. Non-small cell lung cancer - surgical treatment options in the early stages. Dtsch Med Wochenschr 2026;151:315-9. [Crossref] [PubMed]
2. Yan HJ, Zheng XY, Zeng YT, et al. Salvage surgery and conversion surgery for patients with nonsmall cell lung cancer: a narrative review. Int J Surg 2025;111:1032-41. [Crossref] [PubMed]
3. D'Ambrosio PD, da Nobrega Oliveira REN, Martins MAB, et al. Minimally invasive vs. open chest wall resection in non-small cell lung cancer: a systematic review and meta-analysis. Transl Lung Cancer Res 2025;14:2626-35.
4. Baboolal I, Lau K, Alvarez Gallesio J, et al. 3DCT reconstruction-does 3DCT improve anatomical lung resection?-a narrative review of the literature. J Thorac Dis 2025;17:11389-401. [Crossref] [PubMed]
5. Ni Y, Zhang J. Comparison of Robotic-Assisted and Uniportal Video-Assisted Thoracoscopic Lobectomy for Early-Stage Non-Small Cell Lung Cancer: A Retrospective Cohort Study. Ann Ital Chir 2026;97:352-63. [Crossref] [PubMed]
6. Li X, Li Q, Zhao E, et al. Perioperative Outcomes of Robotic- vs Video-Assisted Thoracoscopic Surgery in Non-Small Cell Lung Cancer After Neoadjuvant Therapy. Ann Thorac Surg 2026;121:1184-93. [Crossref] [PubMed]
7. Adamica D, Tulinský L, Hrubovčák J, et al. Robot-assisted vs. video-assisted thoracoscopic surgery in the question of radicality of mediastinal lymphadenectomy. Rozhl Chir 2025;104:526-32.
8. Oliveira RENDN, Delgado LM, Pimienta Ibarra AS, et al. Superiority of single-photon emission computed tomography for predicting postoperative lung function after anatomical lung resection: a systematic review and meta-analysis. J Thorac Dis 2026;18:402. [Crossref] [PubMed]
9. Ma J, Li X, Zhao S, et al. Robot-assisted thoracic surgery versus video-assisted thoracic surgery for lung lobectomy or segmentectomy in patients with non-small cell lung cancer: a meta-analysis. BMC Cancer 2021;21:498. [Crossref] [PubMed]
10. Patel AJ, Yasufuku K, Bille A. Robotic Thoracic Surgery: Current Landscape and Future Directions. Interdiscip Cardiovasc Thorac Surg 2026;41:ivag009. [Crossref] [PubMed]
11. Lechien JR. Learning Curves in Robotic Thyroid Surgery: A Systematic Review Comparing Transoral, Axillary, and Facelift Approaches. J Otolaryngol Head Neck Surg 2026;55:19160216261421877. [Crossref] [PubMed]
12. Veronesi G, Abbas AE, Muriana P, et al. Perioperative Outcome of Robotic Approach Versus Manual Videothoracoscopic Major Resection in Patients Affected by Early Lung Cancer: Results of a Randomized Multicentric Study (ROMAN Study). Front Oncol 2021;11:726408. [Crossref] [PubMed]
13. Jin R, Zheng Y, Yuan Y, et al. Robotic-assisted Versus Video-assisted Thoracoscopic Lobectomy: Short-term Results of a Randomized Clinical Trial (RVlob Trial). Ann Surg 2022;275:295-302. [Crossref] [PubMed]
14. Jin R, Zhang Z, Zheng Y, et al. Health-Related Quality of Life Following Robotic-Assisted or Video-Assisted Lobectomy in Patients With Non-Small Cell Lung Cancer: Results From the RVlob Randomized Clinical Trial. Chest 2023;163:1576-88. [Crossref] [PubMed]
15. Niu Z, Cao Y, Du M, et al. Robotic-assisted versus video-assisted lobectomy for resectable non-small-cell lung cancer: the RVlob randomized controlled trial. EClinicalMedicine 2024;74:102707. [Crossref] [PubMed]
16. Patel YS, Hanna WC, Fahim C, et al. RAVAL trial: Protocol of an international, multi-centered, blinded, randomized controlled trial comparing robotic-assisted versus video-assisted lobectomy for early-stage lung cancer. PLoS One 2022;17:e0261767. [Crossref] [PubMed]
17. Catelli C, Corzani R, Zanfrini E, et al. RoboticAssisted (RATS) versus Video-Assisted (VATS) lobectomy: A monocentric prospective randomized trial. Eur J Surg Oncol 2023;49:107256. [Crossref] [PubMed]
18. Terra RM, Araujo PHXN, Lauricella LL, et al. A Brazilian randomized study: Robotic-Assisted vs. Video-assisted lung lobectomy Outcomes (BRAVO trial). J Bras Pneumol 2022;48:e20210464.
19. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372: [Crossref] [PubMed]
20. Lamas FM, Lech GE, Maboni LR, et al. Robotic-assisted thoracic surgery versus video-assisted thoracic surgery for patients undergoing lung resection: a systematic review and meta-analysis of randomized controlled trials. Gen Thorac Cardiovasc Surg 2026;74:1-10. [Crossref] [PubMed]
21. Mirza W, Javid A, Khan ME, et al. Robotic-assisted vs video-assisted thoracoscopic lobectomy for non-small cell lung cancer: a GRADE-assessed systematic review and meta-analysis of randomized controlled trials on surgical quality, perioperative, and oncologic outcomes. J Robot Surg 2025;19:651. [Crossref] [PubMed]
22. Wan X, Wang W, Liu J, et al. Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range. BMC Med Res Methodol 2014;14:135. [Crossref] [PubMed]
23. Luo D, Wan X, Liu J, et al. Optimally estimating the sample mean from the sample size, median, mid-range, and/or mid-quartile range. Stat Methods Med Res 2018;27:1785-805. [Crossref] [PubMed]
24. Khan W, Ali T, Bilal M, et al. Evaluating Surgical Approaches in Early NSCLC: A Meta-Analysis of Robot-Assisted and Video-Assisted Techniques. Ann Surg Oncol 2026;33:2123-36. [Crossref] [PubMed]
25. Wang YC, Diao HX, Xu L, et al. Comparison of Early Functional Recovery Following Triportal Robot-Assisted and Uniportal Video-Assisted Segmentectomy in Patients With Early-Stage Non-Small Cell Lung Cancer: A Propensity Score-Matched Analysis. Thorac Cancer 2025;16:e70041. [Crossref] [PubMed]
26. Noleto da Nobrega Oliveira RE, Salvador ICMC, Passos FS, et al. Comparison of Video-assisted Surgery and Open Surgery for Mediastinal Tumor Resection in Pediatric Population: A Systematic Review and Meta-analysis. J Pediatr Surg 2025;60:162271. [Crossref] [PubMed]
27. Oishi H, Onodera K, Notsuda H, et al. Postoperative Lung Function after Robotic- versus Video-Assisted Thoracoscopic Lobectomy for Lung Cancer: A Propensity Score-Matched Study. Ann Thorac Cardiovasc Surg 2025;31:25-00123. [Crossref] [PubMed]
28. Leão EI, Rocha A, Pinto ACPN, et al. Robotic-assisted thoracoscopic surgery lobectomy versus open or video-assisted thoracoscopic surgery lobectomy for early-stage non-small cell lung cancer. Cochrane Database Syst Rev 2025;12:CD016158. [Crossref] [PubMed]
29. Dela Vega AJM, Terra RM, Lauricella LL, et al. Lymph Node Upstaging in Robotic, Video-Assisted Thoracic Surgery, and Open Lung Resections: A Matched Analysis From the Brazilian Lung Cancer Registry. Eur J Cardiothorac Surg 2026;68:ezag051. [Crossref] [PubMed]
30. Syrigos N, Fyta E, Goumas G, et al. Robotic-Assisted Thoracoscopic Surgery Versus Video-Assisted Thoracoscopic Surgery: Which Is the Preferred Approach for Early-Stage NSCLC? J Clin Med 2025;14:3032. [Crossref] [PubMed]
31. Osarogiagbon RU, Decker PA, Ballman K, et al. Survival Implications of Variation in the Thoroughness of Pathologic Lymph Node Examination in American College of Surgeons Oncology Group Z0030 (Alliance). Ann Thorac Surg 2016;102:363-9. [Crossref] [PubMed]
32. Patel AJ, Bille A. Lymph node dissection in lung cancer surgery. Front Surg 2024;11:1389943. [Crossref] [PubMed]
33. Moran TE, Hu EY, Cervantes JE, et al. The Effect of Body Mass Index on Outcomes After Hip Arthroscopy for Femoroacetabular Impingement Syndrome: A Matched Analysis With 10-Year Follow-up. Am J Sports Med 2026;54:397-406. [Crossref] [PubMed]
34. Mirza W, Orakzai FK, Ali RN, et al. Da Vinci platform robotic adrenalectomy for adrenal neoplasms: a GRADE-assessed systematic review and meta-analysis of prospective randomized trials comparing perioperative, postoperative, and economic outcomes with laparoscopic adrenalectomy. J Robot Surg 2026;20:207. [Crossref] [PubMed]
35. Mamun MA, Saleem Y, El-Tahlawy Y, et al. Intravenous Lidocaine in Video-Assisted Thoracoscopic Surgery: A Systematic Review and Meta-analysis of Randomized Controlled Trials. J Cardiothorac Vasc Anesth 2026;S1053-0770(26)00108-4.
36. Wei X, Yu H, Dai W, et al. Long-Term Patient-Reported Outcomes after Multiport Robot-Assisted Surgery versus Video-Assisted Surgery for Lung Cancer: An Observational Cohort Study. Ann Surg Oncol 2026;33:4288-300. [Crossref] [PubMed]
37. Higashi RT, Lee J, Hernandez Y, et al. How Does Traumatic Hand Injury Impact Patients in a Safety-net Healthcare System? A Mixed Methods Study. Clin Orthop Relat Res 2026;484:1470-82. [Crossref] [PubMed]
38. Cheng Y, Qin T, Chen C, et al. Tislelizumab plus platinum and etoposide versus placebo plus platinum and etoposide as first-line treatment for extensive-stage small-cell lung cancer: patient-reported outcomes in the RATIONALE-312 trial. Curr Med Res Opin 2025;41:2357-67. [Crossref] [PubMed]
39. Gandara D, Gogishvili M, Sezer A, et al. Composite patient-reported outcomes and risk prediction for overall survival in advanced non-small cell lung cancer with first-line cemiplimab. Front Oncol 2026;16:1676687. [Crossref] [PubMed]
40. Hsieh MJ, Wu CY, Wu YF, et al. Early clinical and economic outcomes of uniportal robotic- and video-assisted thoracoscopic surgery for lung anatomic resection: a retrospective study. J Robot Surg 2026;20:256. [Crossref] [PubMed]
41. Shen G, Wang R, Xu J, et al. Comparative effectiveness of robotic and video assisted thoracic surgery for lung cancer across key outcomes. iScience 2026;29:114898. [Crossref] [PubMed]
42. Reis LO, Moretti TBC. Data-Driven Culture in Medicine and Surgery: Policy Pathways to Learning Health Systems. Int J Gen Med 2026;19:583804. [Crossref] [PubMed]
43. Zhang K, Liu W, Zhao Y, et al. Comparison of early postoperative patient-reported outcomes after multiportal robotic-assisted thoracoscopic surgery and uniportal video-assisted thoracoscopic surgery for non-small cell lung cancer. Eur J Surg Oncol 2024;50:108481. [Crossref] [PubMed]
44. Lan Z, Zeng C, Li Z, et al. Early patient-reported outcomes after robotic-assisted versus video-assisted thoracoscopic lobectomy. Thorac Cancer 2024;15:1563-71. [Crossref] [PubMed]
45. Martin JL, Mack SJ, Rshaidat H, et al. Wedge Resection Outcomes: A Comparison of Video-Assisted and Robot-Assisted Wedge Resections. Ann Thorac Surg 2024;118:683-90. [Crossref] [PubMed]