Abstract
Background
Robot-assisted minimally invasive oesophagectomy and conventional minimally invasive oesophagectomy are superior to open techniques. However, few studies have directly compared the outcomes of the two minimally invasive approaches.
Methods
A retrospective study of patients from six medical centres with oesophageal squamous cell carcinoma who underwent minimally invasive oesophagectomy between 2015 and 2022. Perioperative outcomes were compared after applying inverse probability of treatment weighting.
Results
The study included 577 patients (robot-assisted minimally invasive oesophagectomy: 206; conventional minimally invasive oesophagectomy: 371). After applying inverse probability of treatment weighting, robot-assisted minimally invasive oesophagectomy was found to yield a higher number of mediastinal nodes compared with conventional minimally invasive oesophagectomy (14.86 versus 12.66, P = 0.017). Robot-assisted minimally invasive oesophagectomy was notably effective in retrieving upper mediastinal left recurrent laryngeal nerve nodes, averaging 1.97 nodes versus 1.14 nodes harvested by conventional minimally invasive oesophagectomy (P < 0.001). This was coupled by a significant decrease in nerve palsy rates (13.9% versus 22.8%, P = 0.020). A significantly larger percentage of patients in the robot-assisted minimally invasive oesophagectomy group had an uncomplicated postoperative course (51.8% versus 34%, P < 0.001). Robot-assisted minimally invasive oesophagectomy also led to a reduction in pneumonia rates (8.6% versus 15.2%, P = 0.041) and was linked to a shorter length of stay (length of stay; 16.64 versus 21.14 days, P = 0.007). The advantage of robot-assisted minimally invasive oesophagectomy in reducing the length of stay was especially pronounced in patients with a high Charlson co-morbidity index (≥2, mean difference 8.46 days; P = 0.0069) and those who underwent neoadjuvant therapy (mean difference 5.63 days; P < 0.001).
Conclusion
In oesophageal squamous cell carcinoma, the use of robot-assisted minimally invasive oesophagectomy led to fewer cases of pneumonia and faster recovery compared with conventional minimally invasive oesophagectomy. Additionally, robot-assisted minimally invasive oesophagectomy significantly improved the feasibility and safety of performing lymph node dissection along the recurrent laryngeal nerve.
In oesophageal squamous cell carcinoma, the use of robot-assisted minimally invasive oesophagectomy led to fewer cases of pneumonia and faster recovery compared with conventional minimally invasive oesophagectomy. Additionally, robot-assisted minimally invasive oesophagectomy significantly improved the feasibility and safety of performing lymph node dissection along the recurrent laryngeal nerve.
Introduction
The use of minimally invasive oesophagectomy (MIO) has emerged as the primary surgical approach for the management of oesophageal cancer1–3. This procedure, which can be performed using video-assisted thoracoscopic surgery or robotic surgery, offers a range of advantages over traditional open surgery including reduced wound trauma, less postoperative pain and lower morbidity rates. While both robot-assisted MIO (RAMIO) and conventional MIO (CMIO) have shown superior short-term outcomes compared with the open approach, the specific benefits and differences between the two techniques remain incompletely understood4–8.
To address this knowledge gap, the authors undertook a comprehensive retrospective multicentre study in Taiwan to compare the clinical outcomes of RAMIO and CMIO for oesophageal squamous cell carcinoma (OSCC). To ensure the validity of the findings, the authors employed inverse probability of treatment weighting (IPTW) utilizing propensity scores (PS). This approach allowed them to effectively address any selection bias stemming from baseline differences between the two groups.
Methods
Design, setting and participants
A retrospective observational study was conducted to examine patients with OSCC who had undergone MIO for oesophageal cancer at six medical centres between January 2015 and December 2022. Patients were divided into two groups according to the type of oesophagectomy they underwent: transthoracic oesophagectomy by conventional thoracoscopic resection (CMIO group) or robotic resection (RAMIO group). The collection and recording of data were conducted by investigators affiliated with the participating institutions, utilizing a standardized clinical research form. The individual data collections were subsequently merged to create a comprehensive, multi-institutional data set. In order to ensure data security, the research data were securely stored on an external hard drive with password protection. The principal investigator closely supervised the storage process. The study data were only accessible to participating researchers, and their access was subject to strict monitoring and control measures. To comply with the Personal Data Protection Act, every effort was made to safeguard confidentiality. This involved anonymizing patient identification and using pseudocodes for analysis. The institutional review board of each participating institution granted a waiver for study-specific informed consent, as the data collection was retrospective in nature.
Data collection and variable definitions
A comprehensive data set was collected, including information about demographics, clinical variables, details of surgery and any complications occurring during the perioperative period. The Charlson co-morbidity index (CCI) was utilized to evaluate the severity of co-morbidities9. The CCI assigns a total score ranging between 0 and 37, with a score of zero indicating the absence of co-morbidities. The total operating time for RAMIO included the time required for docking. The criteria for perioperative complications were established according to the Esophagectomy Complications Consensus Group (ECCG) system10. The Clavien–Dindo classification system was used to categorize the severity of complications, with grade IIIa or higher being considered major complications requiring surgical, endoscopic or radiologic interventions. The classification of recurrent laryngeal nerve (RLN) injury was based on its site, distinguishing between unilateral and bilateral cases. Thirty- and 90-day mortality rates were defined as any death that occurred within 30 or 90 days after surgery respectively. An unplanned readmission was defined as any hospitalization not scheduled in advance that occurred within 30 days of a patient being discharged after undergoing an oesophagectomy. The reasons for readmission were categorized into the following groups: anastomotic complications, pulmonary complications, chylothorax, gastrointestinal complications, infections other than pneumonia, wound infections and other miscellaneous causes11.
Pretreatment staging and indications for neoadjuvant therapy
Staging was conducted according to the eighth edition of the American Joint Committee on Cancer (AJCC) Staging Manual. The results were based on the findings from patients undergoing computed tomography of the chest and abdomen, PET imaging and optionally, endoscopic ultrasound. Patients who were clinically staged as T2N0M0 or higher were presented with the option of neoadjuvant chemoradiotherapy (nCRT). The most commonly used chemotherapy regimens across the six study centres were either taxane-based (typically paclitaxel or docetaxel combined with carboplatin or cisplatin) or fluoropyrimidine plus platinum (FP) regimens (typically cisplatin plus 5-FU)12. Radiotherapy was administered at doses varying from 41.4 to 50.4 Gy. Surgical resection was performed 6–12 weeks after completing nCRT, with the McKeown oesophagectomy being the preferred approach for all six centres. The Ivor Lewis procedure was specifically employed for lower third or junctional cancers. In terms of MIO approaches, patients who consented to a partially insured operation with the da Vinci® robotic system (Intuitive Surgical, Sunnyvale, CA, USA) were offered RAMIO. All other patients underwent CMIO fully covered by health insurance. Regarding RLN lymph node dissection (LND), three surgeons routinely performed nodal dissection along both RLNs. The other three surgeons routinely performed right RLN LND, but left RLN LND was only selectively performed when there was clinical evidence of suspicious nodal involvement. Except for one surgeon, radical neck dissection was not a standard procedure.
Statistical analysis
To account for potential selection bias, the IPTW method was applied. PS, calculated with a multivariable logistic regression model, represented the probability of patients receiving treatment by either RAMIO or CMIO based on observed baseline characteristics. Modelling variables for the logistic regression analysis of the RAMIO and CMIO groups included age, sex, body mass index (BMI), smoking history, CCI, tumour histology type, clinical stage, tumour length, tumour location, the use of neoadjuvant therapy, the type of resection (McKeown or Ivor Lewis), the type of abdominal procedure and radical neck dissection. IPTW derived from the PS were used to construct a pseudopopulation. For patients in the RAMIO group, weights were assigned based on the inverse of the corresponding PS, while for patients in the CMIO group, the authors assigned weights based on the inverse of 1 minus the corresponding PS. By employing the IPTW method, stabilized weights were obtained without diminishing the matched sample. Covariate balance was assessed using the standardized mean difference (SMD) before and after applying IPTW. An SMD below 0.1 indicated satisfactory balance. Statistical comparisons were made using the Student’s t-test for continuous variables and the Fisher’s exact test for categorical data. Analyses were performed using SAS, version 9.4 (SAS Institute, Cary, NC, USA), and two-tailed P values of less than 0.05 were considered statistically significant.
Results
Patient characteristics
Figure 1 presents a flow chart of the study. The analysis included a total of 577 patients with OSCC. Of these, 206 underwent RAMIO, whereas the remaining 371 patients received CMIO. The number of patients enrolled per centre is provided in Table S1. Table 1 displays the baseline characteristics of the two study groups before and after the application of IPTW. In the original cohort, patients who underwent RAMIO exhibited a lower mean age, higher mean BMI and a greater prevalence of smoking compared with those who were treated with CMIO. Significant differences were also observed in other variables, including clinical stage distribution, pretreatment tumour length and the use of neoadjuvant therapy. In terms of surgical approach, apart from the thoracic phase, the CMIO group received a higher number of laparotomies and included a greater proportion of patients who underwent neck nodal dissection. The application of IPTW led to the attainment of a satisfactory covariate balance, as indicated by all SMD values being below 0.025.
Fig. 1.
Flow chart depicting patient progression through the study
OSCC, oesophageal squamous cell carcinoma; RAMIO, robot-assisted minimally invasive oesophagectomy; CMIO, conventional minimally invasive oesophagectomy.
Table 1.
General characteristics of the two study groups before and after the application of inverse probability of treatment weighting
| Characteristic | Original cohort | IPTW cohort | ||||
|---|---|---|---|---|---|---|
| RAMIO (n = 206) | CMIO (n = 371) | SMD | RAMIO | CMIO | SMD | |
| Sex | ||||||
| Male | 189 (91.7) | 344 (92.7) | 0.036 | 92.3% | 92.4% | 0.002 |
| Female | 17 (8.3) | 27 (7.3) | 7.7% | 7.6% | ||
| Age (years), mean(s.d.) | 58.28(8.83) | 57.22(8.63) | 0.121 | 57.63(8.65) | 57.54(8.77) | 0.011 |
| BMI (kg/m2), mean(s.d.) | 22.49(3.28) | 21.84(3.38) | 0.193 | 22.16(3.20) | 22.14(3.36) | 0.007 |
| Smoking | 167 (82.3) | 318 (86.9) | 0.128 | 85.5% | 85.4% | 0.004 |
| CCI | 0.83(1.38) | 0.90(1.31) | 0.049 | 0.89(1.27) | 0.85(1.28) | 0.025 |
| Clinical AJCC stage | ||||||
| I | 20 (9.7) | 34 (9.2) | 0.140 | 8.8% | 8.4% | 0.017 |
| II | 45 (21.8) | 80 (21.6) | 21.9% | 21.8% | ||
| III | 103 (50.0) | 168 (45.3) | 46.5% | 46.5% | ||
| IV | 38 (18.4) | 89 (24.0) | 22.8% | 23.3% | ||
| Tumour length (cm), mean(s.d.) | 4.70(2.54) | 4.89(2.59) | 0.073 | 4.86(2.56) | 4.87(2.56) | 0.002 |
| Abdominal approach | ||||||
| Laparoscopy | 190 (92.2) | 311 (83.8) | 0.261 | 86.7% | 86.3% | 0.013 |
| Laparotomy | 16 (7.8) | 60 (16.2) | 13.3% | 13.7% | ||
| Type of resection | ||||||
| Ivor Lewis | 13 (6.3) | 28 (7.5) | 0.049 | 7.7% | 7.3% | 0.015 |
| McKeown | 193 (93.7) | 343 (92.5) | 92.3% | 92.7% | ||
| Radical neck dissection | 38 (18.4) | 89 (24.0) | 0.136 | 20.8% | 20.4% | 0.010 |
| Tumour location | ||||||
| Upper third | 49 (23.8) | 68 (18.3) | 0.216 | 20.7% | 20.2% | 0.021 |
| Middle third | 100 (48.5) | 172 (46.4) | 45.6% | 46.2% | ||
| Lower third | 55 (26.7) | 120 (32.3) | 31.5% | 31.3% | ||
| Gastro-oesophageal junction | 2 (1.0) | 11 (3.0) | 2.1% | 2.3% | ||
| Preoperative treatment | 157 (76.2) | 308 (83.0) | 0.170 | 80.8% | 81.0% | 0.003 |
| Taxane-based regimen | 96 (61.1) | 163 (52.9) | 60.6% | 53.8% | ||
| FP regimen | 58 (36.9) | 138 (44.8) | 37.6% | 44.7% | ||
| Other | 3 (2.0) | 7 (2.3) | 1.7% | 1.6% | ||
Values are n (%) unless otherwise stated. IPTW, inverse probability of treatment weighting; RAMIO, robot-assisted minimally invasive oesophagectomy; CMIO, conventional minimally invasive oesophagectomy; SMD, standardized mean difference; CCI, Charlson co-morbidity index; AJCC, American Joint Committee on Cancer; FP, 5-fluorouracil plus cisplatin.
Surgical and perioperative outcomes after IPTW
Table 2 provides a summary of perioperative outcomes and variables associated with surgical quality. No significant difference was found between the two groups in terms of intraoperative complications or conversion rates. However, patients who underwent RAMIO had a longer operating time compared with CMIO (505.03 min versus 468.75 min respectively; P < 0.001). Conversely, RAMIO was associated with lower blood loss (121.54 ml versus 165.73 ml respectively; P = 0.007).
Table 2.
Surgical and pathological outcomes in the two study groups before and after the application of inverse probability of treatment weighting
| Outcome | Original cohort | IPTW cohort | ||||
|---|---|---|---|---|---|---|
| RAMIO (n = 206) | CMIO (n = 371) | P | RAMIO | CMIO | P | |
| Total operating time (min), mean(s.d.) | 500.05(99.89) | 466.75(110.69) | <0.001 | 505.03(101.96) | 468.75(110.75) | <0.001 |
| Blood loss (ml), mean(s.d.) | 130.44(214.49) | 169.3(209.16) | 0.034 | 121.54(155.45) | 165.73(198.47) | 0.007 |
| Intraoperative complication rate | 2 (1.0) | 10 (2.7) | 0.277 | 0.5% | 2.4% | 0.093 |
| Conversion required for the chest phase | 0 (0.0) | 3 (0.8) | 0.490 | 0% | 0.9% | 0.089 |
| Total number of excised nodes, mean(s.d.) | 29.95(15.53) | 28.50(15.91) | 0.327 | 30.01(16.11) | 28.03(15.48) | 0.194 |
| Total number of excised mediastinal nodes | 14.51(9.84) | 12.51(9.17) | 0.015 | 14.86(9.95) | 12.66(9.13) | 0.017 |
| Total number of excised right RLN nodes | 1.94(1.99) | 1.99(2.58) | 0.819 | 1.94(2.01) | 1.89(2.51) | 0.813 |
| Total number of excised left RLN nodes | 2.05(2.48) | 1.11(1.92) | <0.001 | 1.97(2.46) | 1.14(1.95) | <0.001 |
| Total number of excised abdominal nodes | 13.06(9.28) | 13.25(13.29) | 0.857 | 12.19(8.29) | 13.08(13.17) | 0.348 |
| R0 resection rate | ||||||
| R0 | 184 (89.3) | 328 (88.4) | 0.944 | 88.8% | 88.2% | 0.912 |
| R1 | 10 (4.9) | 20 (5.4) | 5.3% | 5.0% | ||
| R2 | 12 (5.8) | 23 (6.2) | 5.9% | 6.8% | ||
| ypTNM AJCC stage | ||||||
| I | 89 (56.7) | 140 (45.5) | 0.075 | 54.5% | 46.2% | 0.262 |
| II | 31 (19.7) | 67 (21.8) | 20.7% | 22.4% | ||
| IIIa | 20 (12.7) | 34 (11.0) | 13.2% | 10.2% | ||
| IIIb | 13 (8.3) | 50 (16.2) | 8.9% | 15.7% | ||
| IVa | 2 (1.3) | 9 (2.9) | 1.5% | 2.8% | ||
| IVb | 2 (1.3) | 8 (2.6) | 1.3% | 2.7% | ||
| pTNM AJCC stage | ||||||
| I | 3 (6.1) | 2 (3.2) | 0.476 | 3.7% | 3.9% | 0.596 |
| Ia | 11 (22.4) | 15 (23.8) | 26.7% | 22.0% | ||
| Ib | 17 (34.7) | 17 (27.0) | 32.7% | 27.2% | ||
| IIa | 3 (6.1) | 12 (19.0) | 7.0% | 18.7% | ||
| IIb | 3 (6.1) | 7 (11.1) | 6.8% | 12.1% | ||
| IIIa | 2 (4.1) | 1 (1.6) | 2.7% | 1.5% | ||
| IIIb | 9 (18.4) | 8 (12.7) | 20.4% | 13.5% | ||
| IVa | 1 (2.0) | 1 (1.6) | 0.0% | 1.1% | ||
Values are n (%) unless otherwise stated. IPTW, inverse probability of treatment weighting; RAMIO, robot-assisted minimally invasive oesophagectomy; CMIO, conventional minimally invasive oesophagectomy; RLN, recurrent laryngeal nerve; AJCC, American Joint Committee on Cancer; ypTNM, postneoadjuvant treatment tumour-node-metastasis staging; pTNM, pathological tumour-node-metastasis staging.
Regarding pathological parameters, both groups showed a similar distribution of pathology stages, and there were no significant intergroup differences in terms of positive margins. However, RAMIO demonstrated a significant benefit over CMIO regarding the number of mediastinal lymph nodes harvested. On average, RAMIO yielded 14.86 mediastinal nodes, whereas CMIO resulted in 12.66 mediastinal nodes (P = 0.017).
Table 3 presents a detailed comparison of the incidence of specific complications between the two groups, as defined by the ECCG criteria. Specifically, RAMIO was associated with significantly lower rates of pneumonia (8.6% versus 15.2%, P = 0.041). Additionally, RAMIO resulted in a shorter postoperative mechanical ventilation support time and intensive care unit (ICU) stay. Although there was no difference in the in-hospital mortality rate, patients who underwent RAMIO had a significantly shorter postoperative length of stay (LOS) compared with CMIO (16.64 days versus 21.14 days respectively; P = 0.007).
Table 3.
Postoperative complications in the two study groups before and after the application of inverse probability of treatment weighting
| Postoperative complication | Original cohort | IPTW cohort | ||||
|---|---|---|---|---|---|---|
| RAMIO (n = 206) | CMIO (n = 371) | P | RAMIO | CMIO | P | |
| Complication | ||||||
| Pneumonia | 18 (8.7) | 55 (14.8) | 0.048 | 8.6% | 15.2% | 0.041 |
| Pleural effusion | 33 (16.0) | 90 (24.3) | 0.027 | 17.0% | 22.9% | 0.131 |
| Respiratory failure | 2 (1.0) | 7 (1.9) | 0.617 | 0.5% | 1.7% | 0.242 |
| Anastomotic leakage | 18 (8.7) | 45 (12.1) | 0.266 | 10.3% | 12.3% | 0.548 |
| Neurological and psychiatric complications | 2 (1.0) | 2 (0.5) | 0.940 | 1.1% | 0.6% | 0.515 |
| Cardiovascular complications | 13 (6.3) | 23 (6.2) | 1.000 | 6.5% | 6.6% | 0.958 |
| Urologic complications | 9 (4.4) | 13 (3.5) | 0.770 | 3.8% | 2.8% | 0.511 |
| Infections | 16 (7.8) | 50 (13.5) | 0.054 | 8.5% | 12.7% | 0.195 |
| Thromboembolic complications | 1 (0.5) | 2 (0.5) | 1.000 | 0.4% | 0.6% | 0.680 |
| Chylothorax | 3 (1.5) | 11 (3.0) | 0.398 | 1.5% | 3.4% | 0.203 |
| RLN injury | 28 (13.6) | 83 (22.4) | 0.014 | 13.9% | 22.8% | 0.020 |
| Right RLN injury | 1 (0.5) | 14 (3.8) | 0.035 | 0.4% | 4.0% | 0.005 |
| Left RLN injury | 28 (13.6) | 80 (21.6) | 0.025 | 13.9% | 21.8% | 0.038 |
| Postoperative ICU duration (hours), mean(s.d.) | 48.48(72.27) | 68.86(130.85) | 0.039 | 49.30(68.87) | 68.33(135.76) | 0.033 |
| Postoperative intubation duration (hours), mean(s.d.) | 18.74(53.27) | 31.83(104.43) | 0.093 | 18.67(47.14) | 32.66(108.48) | 0.037 |
| Postoperative LOS (days), mean(s.d.) | 16.46(19.60) | 21.15(16.84) | 0.003 | 16.64(19.18) | 21.14(16.99) | 0.007 |
| In-hospital mortality rate | 3 (1.5) | 9 (2.4) | 0.633 | 0.8% | 2.3% | 0.192 |
| 30-day mortality rate | 2 (1.0) | 4 (1.1) | 1.000 | 0.5% | 0.9% | 0.636 |
| 90-day mortality rate | 6 (2.9) | 9 (2.4) | 0.937 | 2.6% | 2.3% | 0.871 |
| Reoperation within 30 days | 13 (6.3) | 39 (10.5) | 0.124 | 6.4% | 11.0% | 0.111 |
| Unplanned readmission within 30 days | 26 (12.6) | 43 (11.6) | 0.817 | 12.0% | 12.1% | 0.979 |
| Anastomotic complications | 5 (19.2) | 11 (25.6) | 11.1% | 26.1% | ||
| Pulmonary complications | 5 (19.2) | 10 (23.3) | 19.4% | 24.3% | ||
| Chylothorax | 1 (3.8) | 2 (4.7) | 3.4% | 5.0% | ||
| Gastrointestinal complications | 5 (19.2) | 6 (14.0) | 26.0% | 13.6% | ||
| Infections other than pneumonia | 1 (3.8) | 4 (9.3) | 0.0% | 7.5% | ||
| Wound infections | 8 (30.8) | 3 (7.0) | 31.5% | 7.2% | ||
| Other miscellaneous causes | 1 (3.8) | 7 (16.3) | 8.7% | 16.3% | ||
Values are n (%) unless otherwise stated. IPTW, inverse probability of treatment weighting; RAMIO, robot-assisted minimally invasive oesophagectomy; CMIO, conventional minimally invasive oesophagectomy; RLN, recurrent laryngeal nerve; ICU, intensive care unit; LOS, length of stay.
Table 4 illustrates the overall incidence and frequency of complications. RAMIO was associated with a significantly higher rate of regular postoperative course (Clavien–Dindo grade 0; 51.8% versus 33.9% respectively; P < 0.001) and a lower proportion of patients experiencing at least two complications (14.9% versus 24.2% respectively; P = 0.003). Furthermore, the incidence of major complications, classified as Clavien–Dindo grade IIIa or higher, was significantly lower in the RAMIO group compared with the CMIO group (23.3% versus 36.3% respectively; P = 0.003).
Table 4.
Overall incidence and severity of complications before and after the application of inverse probability of treatment weighting
| Before IPTW | P | After IPTW | P | |||
|---|---|---|---|---|---|---|
| RAMIO | CMIO | RAMIO | CMIO | |||
| Complications | <0.001 | <0.001 | ||||
| No | 104 (50.5) | 129 (34.8) | 51.8% | 33.9% | ||
| Yes | 102 (49.5) | 242 (65.2) | 48.2% | 66.1% | ||
| Number of complications | 0.001 | 0.003 | ||||
| 0 | 104 (50.5) | 129 (34.8) | 51.8% | 33.9% | ||
| 1 | 73 (35.4) | 150 (40.4) | 33.4% | 41.9% | ||
| 2 | 17 (8.3) | 50 (13.5) | 8.0% | 13.7% | ||
| ≥3 | 12 (5.8) | 42 (11.3) | 6.9% | 10.5% | ||
| C–D grade | 0.034 | 0.028 | ||||
| Grade I | 25 (12.1) | 45 (12.1) | 10.1% | 12.5% | ||
| Grade II | 24 (11.7) | 51 (13.7) | 12.5% | 13.6% | ||
| Grade IIIa | 30 (14.6) | 81 (21.8) | 14.2% | 21.2% | ||
| Grade IIIb | 10 (4.9) | 31 (8.4) | 4.7% | 9.4% | ||
| Grade IVa | 5 (2.4) | 12 (3.2) | 2.9% | 3.4% | ||
| Grade IVb | 1 (0.5) | 0 (0.0) | 0.5% | 0.0% | ||
| Grade V | 3 (1.5) | 9 (2.4) | 0.9% | 2.3% | ||
| Grade ≥ IIIa | 49 (23.8) | 133 (35.8) | 0.004 | 23.3% | 36.3% | 0.003 |
Values are n (%) unless otherwise stated. IPTW, inverse probability of treatment weighting; RAMIO, robot-assisted minimally invasive oesophagectomy; CMIO, conventional minimally invasive oesophagectomy; C–D, Clavien–Dindo.
Since centre A enrolled more than 50% of the patients, the authors performed subgroup analysis based on centre A and centre-A excluded cohorts. The surgical and complication data are provided in Tables S2, S3. RAMIO was associated with fewer postoperative complications and faster recovery compared with CMIO in both cohorts. Regarding surgical approach, a similar pattern in both the McKeown and Ivor Lewis procedures was noted (Table S4). However, due to the limited number of Ivor Lewis cases (n = 13 in RAMIO and n = 28 in CMIO), results did not reach statistical significance.
Recurrent laryngeal nerve palsy following lymph node dissection
RAMIO was particularly effective in retrieving upper mediastinal left RLN nodes, averaging 1.97 nodes compared with the 1.14 nodes harvested by CMIO (P < 0.001). However, there was no significant difference in the retrieval of mediastinal right RLN nodes (Table 2). Importantly, the application of RAMIO enhanced the safety of dissection, resulting in a reduced overall nerve palsy rate (13.9% versus 22.8%, P = 0.020), right nerve palsy rate (0.4% versus 4%, P = 0.005) and left nerve palsy rate (13.9% versus 21.8%, P = 0.038; Table 3).
Subgroup analysis
Figure 2 presents forest plots depicting the subgroup analyses of LOS in patients who underwent RAMIO or CMIO, considering the influence of five distinct covariates. Notably, RAMIO showcased significant benefits in reducing hospital stay compared with CMIO, especially in the subgroups of patients with a high CCI (≥2, mean difference, 8.46 days, P = 0.0069) and those who underwent neoadjuvant therapy (mean difference, 5.63 days, P < 0.001).
Fig. 2.
Subgroup analysis of the duration of hospital stay, with lines indicating the 95% confidence intervals for the mean difference in stay duration MD, mean difference; CCI, Charlson co-morbidity index; RAMIO, robot-assisted minimally invasive oesophagectomy; CMIO, conventional minimally invasive oesophagectomy.
Discussion
This multicentre study, using data gathered across six medical centres in Taiwan, showcased the significant short-term advantages of deploying robotic surgery in the treatment of OSCC. Specifically, it was found that the use of RAMIO was associated with fewer postoperative complications and faster recovery compared with CMIO. These advantages were evident through a reduced incidence of pneumonia, as well as shorter postoperative ICU and hospital stays. Additionally, it was demonstrated that RAMIO was significantly associated with increased efficacy and safety of performing LND along the RLN.
A review of the literature identified several studies that compared RAMIO with CMIO for OSCC (Table 5)13–18. Despite this, the current investigation possesses several distinctive strengths. Firstly, by utilizing a multi-institutional study design, the authors were able to include a substantial number of patients. Furthermore, this approach allowed them to gather highly granular data, such as the operating time, specific locations of RLN palsy and the number of lymph nodes harvested at each location. Secondly, by utilizing PS-derived IPTW as opposed to matching, the authors effectively addressed potential selection bias issues, while also preserving crucial data from the majority of patients. These two advantages allowed the authors to conduct comprehensive subgroup analyses with substantial statistical power, facilitating the identification of specific patient subsets that experienced the most significant benefits from the RAMIO approach (Fig. 2). Notably, this encompassed individuals with higher CCI scores and those who had undergone neoadjuvant therapy. The findings presented here offer valuable insights into the patient selection process within the constraints of limited medical resources, particularly in light of the high costs associated with robotic surgery.
Table 5.
Summary of studies conducted in the past 5 years on the comparative efficacy of robot-assisted minimally invasive oesophagectomy versus conventional minimally invasive oesophagectomy for oesophageal squamous cell carcinoma
| Authors | Number of patients | Study design | PS model | Number of dissected nodes | Surgical time | RLN palsy | Pneumonia | LOS |
|---|---|---|---|---|---|---|---|---|
| Deng et al., 201813 | RAMIO = 42 | Retrospective, single-centre | Matching | RAMIO↑ | RAMIO↑ | Equal | Equal | Equal |
| CMIO = 42 | ||||||||
| Xu et al., 202115 | RAMIO = 310 | Retrospective, single-centre | Matching | RAMIO↑ | N/A | Equal | Equal | N/A |
| CMIO = 411 | ||||||||
| Yang et al., 202017 | RAMIO = 280 | Retrospective, single-centre | Matching | RAMIO↑ | RAMIO↓ | RAMIO↑ | Equal | Equal |
| CMIO = 372 | ||||||||
| Chen et al., 201912 | RAMIO = 68 | Retrospective, single-centre | Matching | N/A | N/A | RAMIO↓ | Equal | Equal |
| CMIO = 74 | ||||||||
| Yang et al., 202216 | RAMIO = 181 | Prospective, multicentre | None | RAMIO↑ | RAMIO↓ | Equal | Equal | Equal |
| CMIO = 177 | ||||||||
| Present study | RAMIO = 206 | Retrospective, multicentre | Weighting | RAMIO↑ | RAMIO↑ | RAMIO↓ | RAMIO↓ | RAMIO↓ |
| CMIO = 373 |
PS, propensity score; RLN, recurrent laryngeal nerve; LOS, length of stay; RAMIO, robot-assisted minimally invasive oesophagectomy; CMIO, conventional minimally invasive oesophagectomy; N/A, not available.
Performing LND along the upper mediastinum in patients with OSCC presents a challenging dilemma for surgeons. On the one hand, the lymph nodes located in the upper mediastinum have shown the highest efficacy index after removal19. On the other hand, the close proximity of these lymph nodes to the RLN poses a significant risk of nerve injury during dissection15,20. The clinical consequences of RLN damage extend beyond phonatory issues and can lead to aspiration-induced pulmonary complications, which are associated with an increased mortality rate15,20,21. Considering these factors, it is imperative to devise strategies to maximize surgical radicality while concurrently mitigating the risk of RLN palsy during oesophagectomy. Notably, the authors found that RAMIO significantly improved the effectiveness and safety of LND along the RLN. The adoption of robotic technology for this procedure not only increased the lymph node yield but also resulted in a marked decrease in the incidence of nerve palsy. These improvements were instrumental in lowering the pneumonia rates observed in patients undergoing RAMIO.
In the present study, the overall complication rate was 59.6%. This rate is higher than those reported in some comprehensive national audits of oesophagectomy, but aligns with studies that employed the same outcome definition proposed by the ECCG22,23. The short-term clinical outcomes observed in this study also met the highest current standards22, with an in-hospital mortality rate of 2% and a 90-day mortality rate of 3%.
In the present analysis of various efficacy parameters, the operating time was the only metric that did not significantly favour the RAMIO group. This result diverges from findings of a recent randomized trial17, warranting further discussion. It is crucial to note that this study focused on patients who underwent MIO after January 2015. By this time, CMIO had already been established as the standard of care across all participating centres, whereas RAMIO was still in its early stages of adoption24. Therefore, the longer operating time observed in the RAMIO group could potentially be attributed to the learning curve associated with the new technique.
When interpreting these findings, it is crucial to acknowledge several limitations. Firstly, despite the efforts to account for numerous variables using IPTW, there remains the possibility of unmeasured and residual confounding factors that could influence the results. Specifically, the use of RAMIO was determined by the patient’s willingness to undergo a partially insured robot-assisted operation (additional 5500~6500 Euros out of their own pocket), which may introduce selection bias related to their financial status. The authors also recognize that the initial learning curve may have adversely influenced the results of the RAMIO group. Nonetheless, this does not alter the primary conclusions of this study. Secondly, due to the retrospective nature of this analysis, the thoroughness of RLN LND via video review could not be verified. Instead, the authors had to rely solely on the count of retrieved lymph nodes. To comprehensively understand the advantages provided by robotic systems in RLN LND, a RCT (NCT03713749) is currently underway25. Thirdly, the majority of the RAMIO procedures were limited to the thoracic phase, with the abdominal phase being predominantly performed using laparoscopy. Additional research is needed to explore the potential benefits of applying total RAMIO, as it may lead to further improvements in patient outcomes. Lastly, the present study was limited by a short follow-up period, which restricted the ability to provide long-term survival data. Future research should aim to overcome these caveats to enhance the robustness and applicability of the current findings.
Supplementary Material
Acknowledgements
This study was carried out on behalf of the Taiwan Robotic Thoracic Surgery Task Force Group, with the support of the Taiwan Association of Thoracic and Cardiovascular Surgery.
Contributor Information
Yin-Kai Chao, Department of Thoracic and Cardiovascular Surgery, Chang Gung Memorial Hospital-Linkou, Chang Gung University, Tao-Yuan, Taiwan.
Jui-Ying Lee, Division of Thoracic Surgery, Department of Surgery, Kaohsiung Medical University Hospital, Kaohsiung Medical University, Kaohsiung, Taiwan; Graduate Institute of Clinical Medicine, College of Medicine, Kaohsiung Medical University, Kaohsiung, Taiwan.
Wen-Chien Huang, Department of Thoracic Surgery, Mackay Memorial Hospital, Taipei, Taiwan.
Jang-Ming Lee, Department of Surgery, National Taiwan University Hospital, National Taiwan University, Taipei, Taiwan.
Yau-Lin Tseng, Division of Thoracic Surgery, Department of Surgery, National Cheng Kung University Hospital, College of Medicine, National Cheng Kung University, Tainan, Taiwan.
Hung-I Lu, Department of Cardiovascular and Thoracic Surgery, Chang Gung Memorial Hospital-Kaohsiung, Chang Gung University, Kaohsiung, Taiwan.
Funding
This research received financial support from a grant (CORPG3N0521) provided by the Chang Gung Memorial Hospital, Linkou.
Disclosure
Y-K.C. has received honoraria from Intuitive Surgical Sarl (Taiwan Branch) for his speaking engagements and educational contributions. The authors declare no other conflicts of interest.
Supplementary material
Supplementary material is available at BJS Open online.
Data availability
The data that support the findings of this study are available from the corresponding author (Y-K.C.) upon reasonable request.
Author contributions
Yin-Kai Chao (Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Supervision, Visualization, Writing—original draft, Writing—review & editing), Jui-Ying Lee (Conceptualization, Data curation, Writing—review & editing), Wen-Chien Huang (Conceptualization, Data curation, Resources, Writing—review & editing), Jan-Ming Lee (Conceptualization, Data curation, Formal analysis, Writing—review & editing), Yau-Lin Tseng (Conceptualization, Data curation, Investigation, Methodology, Writing—review & editing) and Hung-I Lu (Conceptualization, Data curation, Methodology, Writing—review & editing)
References
- 1. Khaitan PG, Vekstein AM, Thibault D, Kosinski A, Hartwig MG, Block M et al. Robotic esophagectomy trends and early surgical outcomes: the US experience. Ann Thorac Surg 2023;115:710–717 [DOI] [PubMed] [Google Scholar]
- 2. Kalff MC, van Berge Henegouwen MI, Baas PC, Bahadoer RR, Belt EJ, Brattinga B et al. Trends in distal esophageal and gastroesophageal junction cancer care: the Dutch nationwide IVORY study. Ann Surg 2023;277:619–628 [DOI] [PubMed] [Google Scholar]
- 3. de Groot E, Goense L, Kingma B, Haverkamp L, Ruurda J, van Hillegersberg R. Trends in surgical techniques for the treatment of esophageal and gastroesophageal junction cancer: the 2022 update. Dis Esophagus 2023;36:doac099. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. van der Sluis PC, van der Horst S, May AM, Schippers C, Brosens LAA, Joore HCA et al. Robot-assisted minimally invasive thoracolaparoscopic esophagectomy versus open transthoracic esophagectomy for resectable esophageal cancer: a randomized controlled trial. Ann Surg 2019;269:621–630 [DOI] [PubMed] [Google Scholar]
- 5. Weindelmayer J, De Pasqual CA, Turolo C, Gervasi MC, Sacco M, Bencivenga M et al. Robotic versus open Ivor–Lewis esophagectomy: a more accurate lymph node dissection without burdening the leak rate. J Surg Oncol 2023;127:1109–1115 [DOI] [PubMed] [Google Scholar]
- 6. Straatman J, Van Der Wielen N, Cuesta MA, Daams F, Garcia JR, Bonavina L et al. Minimally invasive versus open esophageal resection: three-year follow-up of the previously reported randomized controlled trial: the TIME trial. Ann Surg 2017;266:232–236 [DOI] [PubMed] [Google Scholar]
- 7. Nuytens F, Dabakuyo-Yonli TS, Meunier B, Gagnière J, Collet D, D’Journo XB et al. Five-year survival outcomes of hybrid minimally invasive esophagectomy in esophageal cancer: results of the MIRO randomized clinical trial. JAMA Surg 2021;156:323–332 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Yun J, Chong B, Kim H, Lee I, Gong C, Kim B et al. Comparative outcomes of robot-assisted minimally invasive versus open esophagectomy in patients with esophageal squamous cell carcinoma: a propensity score-weighted analysis. Dis Esophagus 2020;33:doz071. [DOI] [PubMed] [Google Scholar]
- 9. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373–383 [DOI] [PubMed] [Google Scholar]
- 10. Kuppusamy MK, Low DE; International Esodata Study Group (IESG) . Evaluation of international contemporary operative outcomes and management trends associated with esophagectomy: a 4-year study of > 6000 patients using ECCG definitions and the online Esodata database. Ann Surg 2022;275:515–525 [DOI] [PubMed] [Google Scholar]
- 11. Chen YT, Wang CT, Chiu CH, Chao YK. Salvage surgery, chylothorax and pneumonia are the main drivers of unplanned readmissions after oesophagectomy for cancer. Eur J Cardio-Thorac Surg 2021;59:1021–1029 [DOI] [PubMed] [Google Scholar]
- 12. Gao X, Tsai PC, Chuang KH, Pai CP, Hsu PK, Li SH et al. Neoadjuvant carboplatin/paclitaxel versus 5-fluorouracil/cisplatin in combination with radiotherapy for locally advanced esophageal squamous cell carcinoma: a multicenter comparative study. Cancers (Basel) 2022;14:2610. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Chen J, Liu Q, Zhang X, Yang H, Tan Z, Lin Y et al. Comparisons of short-term outcomes between robot-assisted and thoraco-laparoscopic esophagectomy with extended two-field lymph node dissection for resectable thoracic esophageal squamous cell carcinoma. J Thorac Dis 2019;11:3874–3880 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Deng HY, Huang WX, Li G, Li SX, Luo J, Alai G et al. Comparison of short-term outcomes between robot-assisted minimally invasive esophagectomy and video-assisted minimally invasive esophagectomy in treating middle thoracic esophageal cancer. Dis Esophagus 2018;31:doy012. [DOI] [PubMed] [Google Scholar]
- 15. Taniyama Y, Miyata G, Kamei T, Nakano T, Abe S, Katsura K et al. Complications following recurrent laryngeal nerve lymph node dissection in oesophageal cancer surgery. Interact Cardiovasc Thorac Surg 2015;20:41–46 [DOI] [PubMed] [Google Scholar]
- 16. Xu Y, Li X-K, Cong Z-Z, Zhou H, Wu W-J, Qiang Y et al. Long-term outcomes of robotic-assisted versus thoraco-laparoscopic McKeown esophagectomy for esophageal cancer: a propensity score-matched study. Dis Esophagus 2021;34:doaa114. [DOI] [PubMed] [Google Scholar]
- 17. Yang Y, Li B, Yi J, Hua R, Chen H, Tan L et al. Robot-assisted versus conventional minimally invasive esophagectomy for resectable esophageal squamous cell carcinoma: early results of a multicenter randomized controlled trial: the RAMIE trial. Ann Surg 2022;275:646–653 [DOI] [PubMed] [Google Scholar]
- 18. Yang Y, Zhang X, Li B, Hua R, Yang Y, He Y et al. Short-and mid-term outcomes of robotic versus thoraco-laparoscopic McKeown esophagectomy for squamous cell esophageal cancer: a propensity score-matched study. Dis Esophagus 2020;33:doz080. [DOI] [PubMed] [Google Scholar]
- 19. Tachimori Y, Ozawa S, Numasaki H, Matsubara H, Shinoda M, Toh Y et al. Efficacy of lymph node dissection for each station based on esophageal tumor location. Esophagus 2016;13:138–145 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Sato Y, Kosugi S-I, Aizawa N, Ishikawa T, Kano Y, Ichikawa H et al. Risk factors and clinical outcomes of recurrent laryngeal nerve paralysis after esophagectomy for thoracic esophageal carcinoma. World J Surg 2016;40:129–136 [DOI] [PubMed] [Google Scholar]
- 21. Oshikiri T, Takiguchi G, Hasegawa H, Yamamoto M, Kanaji S, Yamashita K et al. Postoperative recurrent laryngeal nerve palsy is associated with pneumonia in minimally invasive esophagectomy for esophageal cancer. Surg Endosc 2021;35:837–844 [DOI] [PubMed] [Google Scholar]
- 22. Low DE, Kuppusamy MK, Alderson D, Cecconello I, Chang AC, Darling G et al. Benchmarking complications associated with esophagectomy. Ann Surg 2019;269:291–298 [DOI] [PubMed] [Google Scholar]
- 23. Reynolds JV, Donlon N, Elliott JA, Donohoe C, Ravi N, Kuppusamy MK et al. Comparison of esophagectomy outcomes between a National Center, a National Audit Collaborative, and an international database using the Esophageal Complications Consensus Group (ECCG) standardized definitions. Dis Esophagus 2021;34:doaa060. [DOI] [PubMed] [Google Scholar]
- 24. Chao YK, Wen YW. Cost-effectiveness analysis of thoracoscopic versus open esophagectomy for esophageal cancer: a population-based study. Dis Esophagus 2021;34:doaa116. [DOI] [PubMed] [Google Scholar]
- 25. Chao YK, Li ZG, Wen YW, Kim DJ, Park SY, Chang YL et al. Robotic-assisted Esophagectomy vs Video-Assisted Thoracoscopic Esophagectomy (REVATE): study protocol for a randomized controlled trial. Trials 2019;20:346. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data that support the findings of this study are available from the corresponding author (Y-K.C.) upon reasonable request.