A systematic review of robotic versus open and video assisted thoracoscopic surgery (VATS) approaches for thymectomy

Abstract

Background: Median sternotomy has been the most commonly used approach for thymectomy to date. Recent advances in video-assisted thoracoscopic surgery (VATS) and robotic access with CO₂ insufflation techniques have allowed more minimally invasive approaches. However, prior reviews have not compared robotic to both open and VATS thymectomy.

Methods: A systematic review was conducted in accordance with the PRISMA guidelines using PubMed, Embase and Scopus databases. Original research articles comparing robotic to VATS or to open thymectomy for myasthenia gravis, anterior mediastinal masses, or thymomas were included. Meta-analyses were performed for mortality, operative time, blood loss, transfusions, length of stay, conversion to open, intraoperative and postoperative complication rates, and positive/negative margin rates.

Results: Robotic thymectomy is a valid alternative to the open approach; advantages include: reduced blood loss [weighted mean difference (WMD): -173.03, 95% confidence interval (95% CI): -305.90, -40.17, P=0.01], fewer postoperative complications (odds ratio: 0.37, 95% CI: 0.22, 0.60, P<0.00001), a shorter hospital stay (WMD: -2.78, 95% CI: -3.22, -2.33, P<0.00001), and a lower positive margin rate (relative difference: -0.04, 95% CI: -0.07, -0.01, P=0.01), with comparable operative times (WMD: 6.73, 95% CI: -21.20, 34.66, P=0.64). Robotic thymectomy was comparable with the VATS approach; both have the advantage of avoiding median sternotomy.

Conclusions: While randomized controlled studies are required to make definitive conclusions, current data suggests that robotic thymectomy is superior to open surgery and comparable to a VATS approach. Long-term follow-up is required to further delineate oncological outcomes.

Keywords: Thymectomy; da Vinci; robotic; video-assisted thoracoscopic surgery (VATS).

Figure 1

PRISMA flowchart of search.

Figure 2

Risk of bias assessment. Graph lists each paper by first author and year of publication and shows high (red circle with a minus sign), low (green circle with a plus sign), and unclear (yellow circle with a question mark) risk of: selection (systematic differences between groups in baseline characteristics, comparability of groups), performance (systematic differences between groups in the care that is provided), detection (systematic differences between groups in how outcomes are determined), attrition (systematic differences between groups in withdrawals from a study, completeness of sample, follow-up, or data), reporting (systematic differences between reported and unreported findings, selective reporting of results), and other (learning curve, conflicts of interest, funding) bias. Summarized criteria for low risk determinations were: selection bias: the cohorts were contemporary and comparable, matched for patient characteristics, or adjusted for confounding factors. Performance bias: the cohorts were matched on operative and/or hospital characteristics, such as surgical technique, care pathways, and length of follow up to make combining them reasonable, or differences were addressed. Detection bias: data capture and entry was standardized/performed by trained personnel and precise definitions of outcomes of interest were provided. Attrition bias: there was no missing data, or missing data not an issue, no (or few) patients were lost to follow up, and length of follow up comparable and sufficient. Reporting bias: all pre-specified outcomes of interest (and meta-analyses) were reported in the pre-specified way regardless of significance and complete enough for inclusion in a meta-analysis. Other bias: there were no funding or industry conflicts of interest that were deemed an issue, they authors accounted for experience/volume of surgeons and/or hospital. No other obvious bias.

Figure 3

Robotic vs. open thymectomy forest plots. Forest plots showing comparisons between robotic and open cohorts for outcomes of interest. For operative time (A), [1] is operative room time, [2-6,8] are matched, and [7] does not include robotic set up time. For estimated blood loss (EBL) (B), [1-3] are matched. For length of hospital stay (LOS) (C), [1] LOS in Austria is prolonged due to less pressure from insurance companies, [2-7] are matched. For intraoperative complications (D), [1,2] are matched. For postoperative complications (E), [1-6] are matched, [7] does not specify if postoperative, [8] is the perioperative complication rate. For mortality (F), [1] is the intraoperative rate, [2-5] are matched. For positive margin rate (G), [1-4] are matched, n was based on the number of thymoma cases.

Figure 4

Robotic vs. VATS thymectomy forest plots. Forest plots showing comparisons between robotic and VATS cohorts for outcomes of interest. For operative time (A), [1] included robotic set up time, [2] did not include robotic set up time, [3] authors only completed 11 robotic cases in 20 years. For estimated blood loss (B), [1] was reported as median (interquartile range) and standard deviation was calculated as zero and estimated at 0.001 to allow for calculation of a mean difference and 95% confidence interval. Blood transfusions were reported in two studies (C). For length of hospital stay (D), [1] standard deviation was extrapolated from Figure 3 in paper. For conversions to open (E), only papers reporting both a robotic and a VATS rate were included in the analysis. Intraoperative complications (F), postoperative complications (G), mortality (H), and positive surgical margin rate (I) were also reported.

Figure S1

Funnel plots. Graphs showing funnel plots for any outcome with at least ten studies, including operative time for robotic vs. open (A), postoperative complication rate for robotic vs. open (B), mortality rate for robotic vs. open (C), and length of hospital stay for robotic vs. open (D).