Introduction

Surgical procedure for esophageal cancer can be challenging as it involves radical resection of the esophagus, local and regional lymphadenectomy, and restoration of gastrointestinal continuity.

Though conventional open procedures like thoracotomy and laparotomy are effective in achieving oncological clearance, they are associated with major complications and poor postoperative quality of life.123

With the advancement of videoscopic technology, minimally invasive surgery using thoracoscopic and laparoscopic approaches for esophageal cancers became an attractive option both for the surgeons and for the patients4567 with an equally good oncological clearance as compared with open procedures.8

Traditionally, minimally invasive esophageal dissection has been performed in either left lateral position8 or prone position.910 Left lateral position has the advantage of familiar surgical field as open surgery, and easy and quick conversion to thoracotomy. But it has some disadvantages like the need for an extra port for lung retraction, pooling of blood in the operative field, and limited exposure to the posterior mediastinum, especially left recurrent laryngeal nerve group of lymph nodes. The prone position has the advantage of a good access to the posterior mediastinal structures without any need for lung retraction. However, patient positioning is difficult and in case of need for emergency thoracotomy, changing to lateral position can be time consuming.

At our cancer center we have been doing minimally invasive esophagectomy (MIE) using a modified technique of a semi-prone position since the year 2007. This technique aims to combine the advantages and overcome the disadvantages of the two conventional positions. In this report we describe our technique and analyze our data for feasibility of performing MIE in semi-prone position in terms of intraoperative and postoperative performances and final oncological outcome.

Methods

Technique of Surgery

Patient Position

We developed an innovative “semi-prone position,” which is a left lateral position with an anterior inclination of 45 degrees with the horizontal (Fig. 1). A double-lumen endotracheal tube for single-lung ventilation is preferable, although the procedure can also be done with routine dual-lung ventilation. An epidural analgesia is used in all cases.

Close

Fig. 1: A sketch showing patient position: semi-prone with 45 degrees angle to the horizontal.

Export to PPT

Port Positions

In a three-port approach, ports are placed in fifth, seventh, and ninth intercostal spaces in the posterior, mid, and anterior axillary lines, respectively, in the right chest (Fig. 2). Camera port (10 mm) is passed through the seventh intercostal space whereas two 5 mm operating ports are passed through the fifth and ninth intercostal spaces. Pneumothorax is created with CO₂ pressure of 8 to 10 mm Hg. This helps to keep the lungs away without requiring an extra port. The thoracoscopic surgery is performed in three steps, viz., infra-azygous dissection, retro-azygous dissection, and supra-azygous dissection.

Close

Fig. 2: Port positions: at fifth, seventh, and ninth intercostal spaces in the posterior(p), mid(m), and anterior(a) axillary lines, respectively. Sc: angle of scapula.

Export to PPT

Infra-Azygous Dissection

Dissection is started in infra-azygous region after opening the mediastinal pleura overlying the esophagus. The dissection begins in the posterior plane between the esophagus and the aorta, starting from the azygous vein up to the diaphragmatic crura (Fig. 3A, B). All the paraesophageal lymph nodes and fibro-fatty tissues are dissected circumferentially in a centripetal fashion toward the esophagus.

Close

Fig. 3: Thoracoscopic view of infra-azygous dissection. (A) View showing lower paraesophageal lymph nodes: aorta A; left crura of diaphragm B; right diaphragmatic crura C; esophagus D; lower paraesophageal lymph nodes 8L; pulmonary ligament nodes 9; and diaphragmatic nodes 15. (B) View after lymph node clearance: aorta A; pericardium E; inferior pulmonary vein F; and esophagus retracted upward (not shown).

Export to PPT

Then the dissection continues anteriorly clearing the lower paraesophageal and pericardial nodes and proceeds superiorly to clear the right and left bronchial, subcarinal, and aorto-pulmonary group of lymph nodes. The vagus nerves are divided well below the carina after saving the tracheobronchial branches. Thus, circumferential mobilization of the esophagus along with all the surrounding lymph nodes, periesophageal tissue, and fat is performed from the diaphragmatic reflection up to the azygous vein.

Retro-Azygous Dissection

Azygous vein is lifted and dissection is done to release the esophagus from surrounding structures. The aorto-pulmonary lymph node dissection is performed at this stage (Fig. 4). The azygous vein and right bronchial artery running underneath it are preserved to maintain vascularity of right bronchus. This is expected to reduce the pulmonary morbidity by maintaining a good blood supply to the right bronchus. The azygous vein acts as a support to the thin and delicate right bronchial artery. Also, the vein helps in subsequently keeping the gastric conduit in the same anatomical position as the esophagus. However, in case of T4 lesions at the level of the azygous vein, both these structures can be sacrificed for radical clearance.

Close

Fig. 4: Thoracoscopic view of retro-azygous dissection: arch of aorta G; left main bronchus H; right main bronchus I; azygous vein J; subcarinal nodes 7; left bronchial nodes 10L; and right bronchial nodes 10R.

Export to PPT

Supra-Azygous Dissection

This includes the dissection of paraesophageal, lower paratracheal, and right and left recurrent laryngeal group of lymph nodes along with esophageal mobilization right up to the thoracic inlet. The semi-prone position facilitates clear visualization of both the recurrent laryngeal nerves and surrounding lymph nodes (Fig. 5A, B). Dissection under vision also minimizes traction injury to the recurrent nerves as they are carefully dissected away from the esophagus.

Close

Fig. 5: Thoracoscopic view of supra-azygous dissection. (5a) Esophagus D; left recurrent laryngeal nerve M; trachea K; and right recurrent laryngeal group of lymph nodes (RLN LN). (5b) Right vagus nerve N; right recurrent laryngeal nerve P; trachea K; and RLN LN.

Export to PPT

An intercostal underwater drain tube is placed through the ninth intercostal port site. This step completes the esophageal mobilization and thoracic part of the extended two-field lymphadenectomy (Fig. 6).

Close

Fig. 6: Thoracoscopic view of completely mobilized esophagus and the lymph nodal dissection: esophagus D; trachea K; azygous vein J; and lower paratracheal nodes—right 4R and left 4L.

Export to PPT

Laparotomy

Following this, the patient is turned supine and the anesthesia team changes the double-lumen endotracheal tube to a single-lumen tube. A single-lumen tube is less bulky and facilitates working in the neck. In addition, in case the patient is not ready for extubation at the end of surgery, a single-lumen tube is easier to manage. It also allows toilet bronchoscopy at the end of the procedure if required before extubation.

After upper midline laparotomy, stomach is mobilized based on right gastroepiploic arcade. Lymphadenectomy is performed along the left gastric, hepatic, celiac, and splenic vessels and along lesser curvature of the stomach.

Neck Incision

A left neck 4 to 5 cm transverse skin crease incision is taken at supraclavicular fossa. The left recurrent laryngeal nerve is identified. The esophagus is dissected and divided in the neck and the specimen is pulled in the abdomen. Gastric tube is prepared and brought in the neck orthotopically for an end-to-side hand sewn anastomosis in the neck. Care is taken to avoid rotation of the gastric tube.

A feeding jejunostomy is done prior to abdominal closure.

Postoperatively, patient is extubated and kept in intensive care unit for a day. Jejunostomy tube feeding is started from the second postoperative day. On postoperative day-10, oral feeds are started after water soluble contrast swallow examination shows no evidence of leak and patent anastomosis.

Results

We obtained data of 224 patients from our hospital database. Fourteen patients required conversion to an open procedure, hence they were excluded from further analysis. Rest 210 patients were considered for analysis with the available follow-up. Median age was 60 years (range: 27–80 years). Patient variables are given in Table 1. Majority of the patients belonged to stage II (80 [38%]) and stage III (126 [60%]).

As can be seen in Table 1 location wise, the incidence of gastroesophageal junction tumors is still low in our patients: 7 (3.3%). Also, squamous cell carcinoma of lower third esophagus and gastroesophageal junction is still the predominant presentation: 146 (70%). Majority of the patients had received neoadjuvant treatment in view of locally advanced but resectable disease.

Table 2 shows intraoperative performance of the surgery. The mean blood loss and operative duration for thoracoscopy part of the surgery were less compared with most other studies of MIE. Median number of lymph nodes dissected were 16 (range: 5–32) with 36% having 20 or more lymph nodes dissected.

Total 18 (8.5%) patients had intraoperative complications (Table 2). Fourteen (6.6%) patients required conversion to thoracotomy for bleeding (3), tracheal injury (1), and difficult dissection mainly due to dense pleural adhesions (10). There was no intraoperative mortality.

Table 3 shows postoperative outcomes. The most common were the pulmonary complications: 63 (30%). All patients with ECCG grade I and II anastomotic leaks recovered with proper wound drainage. Those with vocal cord paresis recovered to normal within 6 to 8 weeks. The median duration of hospital stay was 14 days (range: 8–33 days).

Complete resection (R0) rate was 98%. Final histopathological evaluation showed squamous cell carcinoma in 146 (70%) patients, adenocarcinoma in 62 (29%) patients, and 1 patient having neuroendocrine tumor of esophagus. As per grading of the tumor, well-differentiated carcinoma was seen in 58 (28.5%) patients, moderately differentiated in 85 (39%), and poorly differentiated in 34 (18.5%) patients. The pathological complete response was seen in 23 (11%) cases. Appropriate adjuvant therapy was given wherever required.

The median follow-up period was 36 months (range: 8–108 months). Total 59 (28%) patients had recurrence locally or in distant organs. The median disease-free survival (DFS) was 18 months (range: 3–108 months; 95% confidence interval [CI] =12.26–23.74 months). OS was 22 months (range: 6–108 months; 95% CI = 28.01–43.58). The 1-, 3-, and 5-year OS rate was 82.5, 46, and 15%, respectively.

Discussion

Of the 224 patients of cancer of esophagus and esophago-gastric junction who were treated at our center and operated in semi-prone position using minimally invasive approach, surgery could be successfully completed by this approach in 94% with a median operative time of just over an hour for the thoracic part of the surgery.

MIE was assessed by various authors in terms of feasibility, short-term results, and oncological results.4679 Early reports of MIE demonstrated reduced trauma of surgery and decreased morbidity and mortality as compared with an open approach.10 Smithers et al demonstrated comparable DFS and OS between open and MIE.5 Nguyen et al concluded MIE to be an oncologically acceptable surgical approach.4

There have been variations among surgeons with respect to the approach of MIE. MIE in the left lateral position was initially described in 1992 by Cuschieri et al12 and popularized by Luketich et al.8 Easy conversion to thoracotomy, which can be done quickly without repositioning the patient, was the main advantage of this position. Also, for the operating surgeon, the anatomical orientation is the same as that of open thoracotomy. However, in this position, exposure to upper mediastinal structures is limited, especially the left recurrent laryngeal group of lymph nodes. Lung as well as diaphragm need to be retracted for exposure, which requires an extra assistance port. Also, there is pooling of blood in the surgical field. Since surgeon’s arms remain abducted during surgery, there is more shoulder and arms fatigue.13

In 1994, Cuschieri first described thoracoscopic mobilization of the esophagus in prone position in six patients.9 Subsequently, prone position was suggested by some authors to improve the exposure of the posterior mediastinum.1012 A recently published systematic review demonstrated reduced pulmonary complications and blood loss in MIE done in prone position as compared with MIE done in left lateral position.14 Due to gravity effect, lung remains retracted to dependent position obviating the need for lung retraction. Similarly, blood does not interfere in the operative field. Since operating surgeon’s hands remain at a low level, there is less fatigue.1315 Ergonomically, prone position is better than lateral position as seen by Shen et al.16 However, conversion to a classic thoracotomy is more difficult and time consuming in the prone position. This may prove critical, especially in emergency situations.

To retain the benefits and overcome the disadvantages of the left lateral decubitus and prone positions, a modified semi-prone position has been used by our team. To the best of our knowledge, this patient position for MIE has been first described by us.17 With the patient in semi-prone position, one gets maximum exposure of the esophagus and the entire posterior mediastinum. The thoracic duct can be well visualized along its entire length. Bilateral recurrent laryngeal group of lymph nodes can be easily and safely cleared. There is minimum pooling of blood at the surgical field. This position avoids crowding of instruments and offers better ergonomics. Also, since the surgeon’s elbow remains at the level of the patient, there is less elbow fatigue. The theoretical ease of performing surgery in various patient positions is shown in Table 4.

To our knowledge, there was no published study on performing MIE in semi-prone position before 2007 when we first started performing it at our center. Subsequently, very few reports have described the experience of performing MIE using semi-prone position.181920 Lin et al compared the feasibility and safety of performing MIE in semi-prone and left lateral positions and found statistically insignificant differences in intraoperative and postoperative complications favoring semi-prone position.19 Seesing et al compared the feasibility and safety of performing MIE in semi-prone versus prone position and found semi-prone positioning being safe, feasible, and at least comparable to MIE in prone position in terms of oncological clearance and postoperative complications.20

Consistent with the observations reproduced by Ma et al in their similar trial,18 in our experience, posterior mediastinal and aorto-pulmonary dissection was technically easier to perform in the semi-prone approach as compared with the lateral approach. The diaphragm did not hinder the lower esophageal dissection as in lateral position. Also, bilateral recurrent laryngeal lymph node dissections were performed with technically the same ease as in prone approach. Moreover, semi-prone position kept the lung and blood away from the posterior mediastinum. Also, in cases where single-lung ventilation failed or was contraindicated, the procedure could still be completed with an additional port for lung retraction.

In our study 14 (6.6%) patients needed conversion into an open procedure, most of them due to pleural adhesions leading to difficult dissection. On two occasions we required emergency conversions, one for superior vena cava rent leading to bleeding and another for tracheal tear. On both occasions we could do immediate thoracotomy. When conversion to thoracotomy is needed during MIE in prone position, the patient needs to be repositioned and re-draped, resulting in the loss of valuable time. However, in semi-prone position, we noticed that without losing any time, an emergency thoracotomy could be performed by simply tilting the operation table to convert the patient position into left lateral position.

In our study the postoperative overall complication rate was 50% (Table 3). Most common complication was pulmonary: 63 (30%). The 30-day mortality in our study was 7 (3.3%). The reasons for the deaths were pneumonia with sepsis in two patients, tube necrosis followed by sepsis in two, myocardial infarction in two, and one patient having cerebrovascular accident leading to death.

Perioperative parameters of our study as compared with the other studies in various patient positions are shown in Table 5. The results of our study are comparable to the other studies. As a result of the modified position, convenient port positions, and intraoperative ease of handling of tissues, we suppose our duration of surgery is shorter as compared with other studies.

Fifty-nine (28%) patients had either loco-regional and/or distant metastases. Consistent with the literature, the median DFS in our study was 18 months (range: 3–108 months). The median OS was 22 months (range: 6–108 months) as shown in Fig. 7.

Close

Fig. 7: Kaplan–Meier graph showing (A) disease-free interval (DFI) and (B) overall survival (OS). CI, confidence interval.

Export to PPT

Conclusion

Thoracoscopic esophagectomy with mediastinal lymphadenectomy in semi-prone position (Pawar technique) is a feasible, convenient, and a safe option that can combine the benefits of the two conventional, that is, left lateral and prone, approaches. With this technique, the intraoperative and postoperative outcomes and long-term oncological outcomes were comparable with the conventional approaches.