Usefulness of three‐dimensional thoracoscope for prone position thoracoscopic esophagectomy improves mediastinal lymph node dissection and prognosis for esophageal cancer

Abstract Objectives This study aimed to assess the superiority of 3D flexible thoracoscope against 2D thoracoscope for lymph node dissection (LND) and prognosis for prone‐position thoracoscopic esophagectomy (TE) in esophageal cancer. Methods Three hundred and sixty‐seven esophageal cancer patients who underwent prone‐position TE with 3‐field LND between 2009 and 2018 were evaluated. 2D and 3D thoracoscope was used in 182 (2D group) and 185 cases (3D group), respectively. Short‐term surgical outcomes, numbers of retrieved mediastinal lymph node (LN), and rates of LN recurrence were compared. Risk factors for mediastinal LN recurrence and long‐time prognosis were also evaluated. Results No differences in postoperative complications were observed between the groups. The numbers of retrieved mediastinal LN were significantly higher, and the rates of LN recurrence were significantly lower in the 3D group compared to 2D group. Use of 2D thoracoscope was a significant independent factor of middle mediastinal LN recurrence by multivariable analysis. Survival was compared by cox regression analysis, and the 3D group had a significantly better prognosis than the 2D group. Conclusions Prone position TE using 3D thoracoscope may improve the accuracy of mediastinal LND and prognosis without increasing postoperative complications for esophageal cancer.

applied in Japan. The most frequent metastasis in thoracic esophageal cancer is mediastinal LN, therefore, an adequate dissection of mediastinal LNs is expected to improve outcomes. 7,8 Development of endoscopic surgery has shifted the esophagectomy procedures from open thoracotomy to thoracoscopy approach. 9,10 The magnification provided by the thoracoscope increased the accuracy of LND. 11 Three-dimensional (3D) endoscope had been developed recently to further improve surgical procedures. 12 In Japan, the clinical use of 3D endoscope began around 2013, and then have been widely introduced for thoracoscopic esophagectomy (TE). The stereoscopic visibility and depth perception of the 3D endoscope are considered suitable for mediastinal LND for esophageal cancer. Although several studies had reported the utility of 3D endoscope in treating esophageal cancer, [13][14][15] the clinical impact of 3D endoscope on LND accuracy, prevention of LN recurrence, and the contribution to patients' prognosis has not yet been clarified.
The aim of this study was to assess the superiority of 3D flexible thoracoscope over 2D thoracoscope in prone-position TE to investigate whether 3D thoracoscope can improve short-and long-term surgical outcomes for patients with esophageal cancer. This study will provide important insights into the potential benefits of 3D endoscopes in the surgical treatment of esophageal cancer, which could have a significant impact on patient care and improve overall prognosis.

| Approval for human experiments
Written informed consent was obtained for the surgical procedures prior to the thoracoscopic esophagectomy. All methods of this study were carried out in accordance with relevant guidelines and regulations. The present study protocol was approved by the Ethics Committee of Tokai University School of Medicine (Approval No. 18R236) and performed with an opt-out option, as explained in instructions posted on the website of the hospital.

| Mediastinal lymph nodes
Esophageal cancer was staged according to TNM classification by the Union for International Cancer Control (UICC) 8th edition. 16 The Japanese classification of esophageal cancer, 11th edition, 17 was used for analyzing the LN station. In addition, we classified the retrieved LNs into five zones of origin: cervical, upper mediastinal, middle mediastinal, lower mediastinal, and abdominal. 18 The middle mediastinal zone and the lower mediastinal zone were divided by the caudal margin of the inferior pulmonary vein. Supraclavicular LN metastasis was not considered as distant metastasis and still judged to be resectable according to the present Japanese esophageal cancer treatment guideline. 19

| Surgical procedures
All the patients enrolled underwent TE in a prone-position, and the surgery was initiated using thoracoscopy. Briefly, patients were placed under general anesthesia in a semi-supine position with the right arm raised, followed by full prone position with bed rotation. The patients underwent esophageal mobilization and mediastinal LN dissection under thoracoscopy. 20 After cervical LND, including bilateral supraclavicular LND, an anastomosis was performed on the left side of the neck, typically using a gastric conduit via retrosternal route. 3D thoracoscope was introduced at our institution in 2015, since then, all patients undergone thoracoscopic surgery using the 3D thoracoscope.
Both 2D and 3D scopes were flexible types and there was no F I G U R E 1 Patient flow of the study. Of the 399 patients who received TE, 32 were excluded and 367 were enrolled; they were divided into a 2D group of 182 patients and a 3D group of 185 patients. LND lymph nodes dissection; TE, thoracoscopic esophagectomy. difference in resolution. Other than the change from 2D to 3D thoracoscope, all surgical procedures in the study period was performed by the same surgical team and also pre/postoperative management remained the same throughout the study period.

| Statistical analysis
Surgical outcomes, the numbers of retrieved LNs, and patient outcomes were compared between the 3D and 2D groups. In particular, the presence of regional LN recurrence was investigated. The numbers of retrieved LNs and LN recurrences were examined separately for each zone.
Correlations with categorizable variables were evaluated using a chi-square test and the Fisher exact test, while those with continuous variables were evaluated using the Mann-Whitney U test. The Student's t test or Welch t test was used to compare the numbers of dissected LNs after checking for equal variances using the Levene test. To assess factors capable of predicting regional LN metastasis, a multivariable logistic regression analysis was performed after checking the multicollinearity of each factor. The survival times were assessed using cox regression analysis. The statistical examinations were performed using SPSS 26.0 (IBM SPSS, New York). All tests were two-sided, and p values <.05 were considered to indicate statistical significance.

| Short-term surgical outcomes
The clinical and oncological characteristics for each group are shown in Table 1. Sex, tumor location, and histology did not differ significantly between the 3D group and the 2D group except for age was significantly higher in the 3D group ( p = .008). Although clinical T category and Stage did not differ between the groups, clinical N category was significantly advanced in the 3D group compared to 2D group ( p = .03). As a result, preoperative treatment was also significantly more frequent in the 3D group than the 2D group ( p = .001).
Surgical outcomes for the patients in both groups are shown in Table 2. The thoracoscopic operation time was significantly shorter and blood loss was significantly lesser in the 3D group compared to 2D group (p < .001). The incidence of postoperative complications did not differ significantly, nor did the incidences of recurrent laryngeal nerve paralysis, chylothorax, or pneumonia. The R1 resection rate was significantly higher in the 2D group (p = .003). No significant difference in postoperative hospital stay was observed between the groups. Table 3 Table 4 shows the recurrences in both groups, especially regional LN recurrences, according to the LN regions, zones, or stations. The incidence of overall recurrence and regional LN recurrence was lower in the 3D group compared with the 2D group (p = .004 and p = .01, respectively). The incidence of mediastinal LN recurrence was also significantly lower in the 3D group than the 2D group ( p = .005). The rate of cervical LN recurrence was not significantly different between the groups. The numbers of recurrences in the middle mediastinal zone and at the No. 109L station were significantly lower in the 3D group compared than the 2D group ( p = .001 and p = .004, respectively).

| Prediction of lymph node recurrence and survival
A multiple logistic regression analysis was performed to assess the risk factors for mediastinal LN recurrence after esophagectomy. The   First, the safety and ease of the surgical procedure is discussed in terms of operative time, blood loss, and postoperative complications.
A simulation-based study reported that the use of 3D scope provided an advantage on depth perception, compared with 2D scope, enabling good results regarding accuracy and speed of work. 21 15 In the present study, no reduction in postoperative hospital stay was observed, but reductions in operative time and blood loss were observed in the 3D group, as shown in Table 2.
Next, the accuracy of the surgical technique is discussed in terms of the number of dissected LNs. While two of the three aforementioned previous studies reported no differences in the numbers of dissected LNs, 14,15 we demonstrated that the numbers of dissected mediastinal LNs were significantly higher in the 3D group than in the 2D group, as shown in Table 3. A detailed analysis of the LN stations showed that the numbers of dissected LNs significantly increased in areas where a field of view was difficult to obtain, such as 109L and 112, or in areas requiring precise manipulation, for example, around the recurrent laryngeal nerve (106recL/recR/tbL). For LN station No. 101, which was dissected from the cervical incision, the number of retrieved LNs was significantly lower in the 3D group. This could be attributable to the fact that LNs at the cervicothoracic border were sufficiently dissected by thoracoscopic manipulation using the 3D thoracoscope. These very interesting results were likely caused by stereovision, which was the great advantage of using 3D endoscopes.
Specifically, stereovision enables the surgeon to understand the depth of space available and to perform more reliable operations.
Finally, we discuss the differences in long-term prognosis resulting from different surgical techniques. The impact of use of 3D thoracoscope during surgery on postoperative recurrence and prognosis for esophageal cancer patients had not yet been reported before. To the best of our knowledge, for the first time our current study showed significantly fewer mediastinal LN recurrences, especially LN recurrences in the middle mediastinal zone with the assistance of 3D thoracoscope for esophageal cancer patients, as shown in Table 4.
Multivariate analysis, as shown in Consequently, the present results might not be applicable to robotassisted surgery. However, robot-assisted surgery also provides a sense of depth and stereoscopic vision, so that, robot-assisted surgery is expected to be capable of performing precise mediastinal LND, similar to the presently reported results.
The results of this study suggest that the use of 3D thoracoscopes in supine-position TE may provide better results in clinical practice, and therefore the use of 3D thoracoscopes is recommended, although some problems remain regarding the supply and cost of medical equipment.
In conclusion, prone-position TE performed using a 3D thoraco-

ACKNOWLEDGMENTS
Not applicable.

FUNDING INFORMATION
No funding.

CONFLICT OF INTEREST STATEMENT
The authors declare no conflict of interest.

DATA AVAILABILITY STATEMENT
The data set used to conduct this research will be made available on request.

ETHICS STATEMENT
The protocol for this research project has been approved by a suitably