Esophagectomy with extended lymphadenectomy remains the primary treatment option for resectable esophageal cancer [1] [2] . However, this procedure is highly invasive and is associated with high rate of morbidity and mortality rate [3] -[5] . Since the first report by Cushieri et al., minimally invasive surgery for esophageal cancer has been performed with increasing frequency, because it reduces the operative invasiveness and provides the rapid postoperative recovery [6] -[8] . On the other hand, the comparison of clinical outcomes between minimally invasive (MIE) and open esophagectomy (OE), the operation time is significantly longer. Moreover, the incidence of anastomotic leakage and reoperation is significantly higher in MIE [9] . Although thoracoscopic esophagectomy is potentially feasible and less invasive alternative, a large multi-center clinical trialis need for confirming the clinical benefits of this procedure.

Thoracoscopic esophagectomy in the prone position is performed under artificial pneumothorax [10] [11] . The advantages of this technique include enhanced visualization of operative fields and surgeons’ ergonomics than left lateral position. However, thoracoscopic esophagectomy in prone position takes a significant longer operation time and is not established for emergent thoracotomy [12] . Therefore, the patients’ position during thoracoscopic esophagectomy remains unclear and does not allow for significant conclusions.

We have developed a new technique in thoracoscopic esophagectomy combining left lateral position with artificial pneumothorax. The aim of this retrospective study is to evaluate the technical feasibility and effect of this technique on terms of operation time and intraoperative blood loss.

We retrospectively analyzed all the patients who underwent thoracoscopic esophagectomy with artificial pnuemothorax in the left lateral decubitus position followed by laparoscopic gastric mobilization between April 2010 and November 2012 at our institution. The detailed of our operative techniques have been described in previous report [13] . During this period, thoracoscopic esopagectomy was indicated for 60 patients with thoracic esophageal cancer. All thoracoscopic esophagectomy were performed by a single surgeon. He has previously experienced esophagectomy under thoracotomy in 200 cases with esophageal cancer in other institutions.

All the patients underwent preoperative esophagogastroscopy with biopsy, computed tomography (CT) from the neck to abdomen, barium swallow, and positron emission tomography/CT. Preoperative and postoperative staging, and treatment criteria followed the Japanese Guidelines for Diagnosis and Treatment of Carcinoma of the Esophagus [14] . Patients with diagnosed as clinical stage II or III esophageal cancer by preoperative examination received 2 courses of neoadjuvant chemotherapy with 5-fluorouracil and cisplatin. These patients underwent surgery 4 weeks after the termination of neoadjuvant chemotherapy. Patients with stage I esophageal cancer underwent esophagectomy without neoadjuvant chemotherapy. Chemoradiation or chemotherapy was indicated for patients with clinical stage IV (including T4) cancer.

This study measures included the clinicopathological factors and short-term surgical outcomes of the patients and improvements in operation time and intraoperative blood loss with an increased number of operated cases. To define the learning curve of our procedure in terms of operation time and estimated blood loss, the 60 patients were divided into 6 sequential groups (n = 10 each). Other learning indicators such as requirement for blood transfusion, rate to conversion to conventional open surgery, rate of postoperative complications, and number of harvested lymph nodes, were evaluated for statistical significance.

All data were analyzed using SPSS version 19 (SPSS, Inc., Chicago, II, USA). Statistical analyses were performed using Kruskal-Wallis test for continuous variables and chi-squared test for categorical variables. Data are expressed as median and range. Probability (P) values of <0.05 were considered statistically significant in all tests.

The clinicopathological and operative findings of the patients are shown in Table 1. The tumor location was in the upper third of the thoracic esophagus in 9, middle third in 36, and lower third in15, respectively. For 38 out