Robot-assisted Surgery for Esophageal Cancer: Analysis of Short-and Long-term Outcomes

s reviewed: 146 Excluded = 101 references Comparison: 66 Intervention: 1 Systematic review: 7 Review/editorial: 19 Protocol: 8 Total title screened: 390 Excluded: 244 Excluded = 23 references Intervention: 6 Comparison: 3 Small sample size: 1 Not original research: 1 Duplicate: 11 Unavailable: 1 Clinical outcomes: 20 Clinical outcomes and cost: 1 Cost only: 1 Full-text review: 45 Robot-assisted Esophagectomy for Cancer Evidence Synthesis Program 14 KEY QUESTION 1 – What is the clinical effectiveness of robotassisted esophagectomy compared to thoracoscopic/laparoscopic or open esophagectomy for cancer? Intra-operative Outcomes Intra-operative: RAMIE compared with VAMIE Figure 2 presents 3 intra-operative outcomes: OR time, LN harvest, and EBL. For the RCT,15 OR time was not longer for RAMIE as compared to VAMIE. The number of LNs harvested was greater for RAMIE, and EBL was not different. For the observational studies, OR time was reported as longer for RAMIE in 4 of the matched studies18,20,23,26 and as no difference in 3 of the other matched studies.16,17,22 One matched study reported shorter OR time24 for RAMIE. For the unmatched observational studies, 1 reported longer OR time with RAMIE27 and 4 reported no difference.29-31,34 Two of the matched studies reported a greater number of LNs harvested for RAMIE,18,19 whereas 6 reported no difference.16,17,20,23,24,26 For unmatched observational studies, 2 reported greater number of LNs harvested with RAMIE,27,31 and 3 reported no difference.29,30,34 None of the matched or unmatched observational studies reported differences in EBL for RAMIE as compared to VAMIE.16-18,20,23,24,26,27,30,31,34 Intra-operative: RAMIE compared with Open Esophagectomy For the RCT,14 OR time was significantly longer for RAMIE compared to the open approach. The number of lymph nodes harvested was not different in the RAMIE arm. EBL was less for RAMIE. For the observational studies, there was a signal of longer OR time for RAMIE (4 reported significantly longer;25,27,32,34 3 no difference21,28,33). Seven studies reported on the number of LNs harvested and of these, 3 reported higher numbers removed by the RAMIE approach.19,27,32 More than half of the observational studies reported less EBL with RAMIE27,28,32,33 whereas 3 reported no difference.21,25,34 Robot-assisted Esophagectomy for Cancer Evidence Synthesis Program 15 Figure 2. Intra-operative Outcomes 1Dotted line separates RCT from the observational studies 2Solid line separates the studies comparing RAMIE with VAMIE or RAMIE with open esophagectomy Robot-assisted Esophagectomy for Cancer Evidence Synthesis Program 16 Short-term Post-operative Outcomes Figure 3 presents 6 short-term post-operative outcomes: anastomotic leak, recurrent laryngeal nerve palsy/hoarseness, pulmonary complications, duration of hospitalization, total complications, and mortality. Twelve studies compared RAMIE vs VAMIE,15-18,20,22-24,26,29-31 6 studies compared RAMIE with open esophagectomy,14,21,25,28,32,33 and 3 studies compared RAMIE with both VAMIE and open esophagectomy.19,27,34 Short-Term: RAMIE compared with VAMIE Of the studies comparing RAMIE with VAMIE, 14 assessed anastomotic leak, and there was no significant difference between study arms or trend favoring RAMIE or VAMIE in either the RCT or the 8 matched and 5 unmatched observational studies.15-18,20,22-24,26,27,29-31,34 Cervical anastomoses were used in 9 studies, including 8 studies primarily or exclusively utilizing the McKeown approach15-18,20,24,31,34 and 1 utilizing a transhiatal approach.29 Three studies directly compared intrathoracic anastomoses with an Ivor-Lewis approach in both study arms.22,23,26 One study was from a large database and compared robot-assisted Ivor-Lewis with an unspecified “transthoracic” MIE, suggesting at least 1 study arm had an intrathoracic anastomosis.27 Another study reported a transthoracic approach for both study arms but did not specify whether an intrathoracic or cervical anastomosis was performed.30 There was no clear difference favoring RAMIE or VAMIE when evaluating studies with a cervical or intrathoracic anastomosis separately. The RCT found no difference in recurrent laryngeal nerve (RLN) palsy between study arms.15 Of the 6 matched observational studies assessing RLN injury, 1 reported lower RLN palsy with RAMIE,17 1 reported lower RLN palsy with VAMIE,24 and 4 reported no difference.16,19,20,26 For the 3 unmatched studies, 1 reported lower rate of RLN palsy as compared to RAMIE,30 and 2 reported no difference.31,34 The RCT15 did not report a difference in pulmonary complications for RAMIE as compared to VAMIE. One propensity matched study reported fewer pulmonary complications24; however, the other 7 studies did not.16-18,20,22,23,26 None of the 4 unmatched observational studies reported a difference between RAMIE and VAMIE approaches.27,30,31,34 Most of the studies had a point estimate of effect falling within the 95% confidence interval of the RCT, which may suggest a possible signal that there may be fewer pulmonary complications in RAMIE compared with VAMIE. None of the 4 US observational studies assessing LOS found a significant difference between RAMIE as compared to VAMIE; 2 were matched and 2 were unmatched studies.19,22,27,29 One of these studies compared robot-assisted and laparoscopic transhiatal esophagectomy, which had no difference in LOS.29 Nine non-US studies evaluated LOS, of which none demonstrated differences between RAMIE and VAMIE (see Appendix G. Evidence table).15-18,20,23,24,26,34 All but 1 of the 9 non-US studies had a LOS with a central tendency (mean or median) greater than 10 days in both study arms,26 whereas all US studies had a measure of central tendency of 10 days or less. Ten studies assessed outcomes for total complications.15,18,20,22,24,26,27,29,31,34 One study compared robot-assisted transhiatal and laparoscopic transhiatal esophagectomy.29 The remaining studies Robot-assisted Esophagectomy for Cancer Evidence Synthesis Program 17 compared a robot-assisted transthoracic approach to a thoracoscopic approach. Neither the RCT nor the matched and unmatched observational studies found a difference in complications. Mortality was assessed in 14 studies.15-20,22-24,26,27,29,31,34 Mortality was not different in the RCT or the matched and unmatched observational studies. In general, mortality rate was low across all studies. Short-Term: RAMIE compared with Open Esophagectomy Eight studies comparing RAMIE and open esophagectomy assessed anastomotic leak rate.14,21,25,27,28,32-34 The RCT, 2 matched observational studies, and 5 unmatched observational studies reported no difference in leak rate. One observational study utilized the robot for the abdominal portion combined with thoracotomy, which did not demonstrate a difference in anastomotic leak rates, as the technique for creating the anastomosis was the same in both arms of the study.33 Of the 2 matched21,25 and 3 unmatched observational studies28,32,34 assessing RLN palsy, none found a difference between RAMIE and open esophagectomy. Eight studies assessed pulmonary complications.14,21,25,27,28,32-34 The rate of pulmonary complications was lower for RAMIE and open esophagectomy in the RCT.14 One matched observation study and 2 unmatched observational studies also reported a lower rate for RAMIE.25,27,28 The largest difference was seen in the RCT14 but significance was also achieved in the 1 matched observational study25 and 2 unmatched observational studies.27,28 One matched observational study and 3 unmatched observational studies did not report a difference in pulmonary complications.21,32-34 Three US studies evaluated LOS.19,27,28 One matched observational study19 and an unmatched observational study28 demonstrated a shorter time to discharge with RAMIE. The third study (unmatched observational) did not find any differences in LOS between the study arms.27 Of the 6 non-US studies that assessed LOS,14,21,25,32-34 2 demonstrated a shorter hospital stay for RAMIE compared with open esophagectomy (see Appendix G. Evidence Table).32,33 The central tendency for LOS was greater than 10 days in both arms of the non-US studies except for one.33 One out of the 3 US studies had LOS with a central tendency greater than 10 days.28 Six studies assessed total complication rate.14,21,27,28,33,34 The RCT demonstrated a lower total complication rate with RAMIE.14 Additionally, 1 matched observational study21 and 1 unmatched study33 showed reduced rates of total complications with RAMIE. Of note, the unmatched study compared the utilization of the robot for the abdominal portion with laparotomy (thoracic portion was performed via thoracotomy in both study arms). Three additional unmatched studies reported no difference total complications with RAMIE as compared to open esophagectomy.27,28,34 Mortality was assessed in 9 studies.14,19,21,25,27,28,32-34 The RCT reported no different in mortality for RAMIE compared with open esophagectomy.14 One matched observational study found that Robot-assisted Esophagectomy for Cancer Evidence Synthesis Program 18 RAMIE was associated with a lower mortality compared with open esophagectomy.25 The remaining studies did not show a difference in mortality between study arms. Figure 3. Short-term Post-operative Outcomes 1Dotted line separates RCT from the observational studies 2Solid line separates the studies comparing RAMIE with VAMIE or RAMIE with open esophagectomy Robot-assisted Esophagectomy for Cancer Evidence Synthesis Program 19 Long-term Outcomes Long-term: RAMIE compared with VAMIE or Open Esophagectomy Figure 4 presents graphically the results of long-term outcomes for recurrence and cancer-free survival. These outcomes were less frequently reported than the intra-operative and short-term post-operative outcomes. These were evaluated in 2 RCTs14,15 and 3 observational studies.24,25,31 One study reported overall survival instead of cancer-f