Introduction

Proximal gastrectomy (PG) has recently been adopted as a function-preserving surgery for select patients with early proximal gastric cancer or adenocarcinoma of the esophagogastric junction [1–3]. PG has numerous potential advantages in preservation of the physiological function of the remnant stomach; in addition to the conservation of the gastric reservoir, the secretion of several critical factors such as gastric acid, intrinsic factor critical for vitamin B12 absorption, and various digestive hormones including the appetite hormone ghrelin is maintained [4, 5]. However, the lower esophageal sphincter and the acute angle of His are lost after PG, increasing the postoperative risk of reflux esophagitis [6, 7]. Therefore, reconstruction with preservation of the anti-reflux mechanism should be emphasized.

Although many reconstructive procedures after PG, including esophagogastrostomy, jejunal interposition, jejunal pouch interposition, and double tract reconstruction, have been reported [8], optimal reconstructive procedures have not yet been established. Esophagogastrostomy has been commonly performed after PG due to its simplicity [9], whereas several anti-reflux procedures, including fundoplication, narrow gastric conduit, fixation of the esophagus into the diaphragm to create a new cardiac notch (angle of His), and lower esophageal sphincter preservation, have been reported [10]. Recently, laparoscopic valvuloplastic esophagogastrostomy by double flap technique (VEG-DFT) has been proposed as an ideal reconstruction method that could potentially reproduce a nearly physiological anti-reflux mechanism [11]; however, the laparoscopic approach, in which the degree of freedom is quite limited, may not suit VEG-DFT which requires complicated suturing and ligation.

Recently, surgical robots have been developed to overcome some of the disadvantages of standard minimally invasive surgery [12–15]. The da Vinci Surgical System (DVSS, Intuitive, Sunnyvale, CA, USA) provides surgeons with a three-dimensional, ten-fold magnified vivid view of the operating field. DVSS filters the physiological hand tremor of the surgeon, restores the natural hand-eye coordination axis as a result of the ergonomically designed surgeon’s console, and offers more degrees of freedom through its articulating surgical instruments. As a result, this robotic system facilitates precise dissection in a confined surgical field with impressive dexterity [12–15].

We previously reported the technique for robotic radical gastrectomy for gastric cancer and its promising short-term outcomes with a focus on reduction in local complications [14, 15]. We also showed that the long-term oncological outcomes of this robotic approach were comparable to those achieved with laparoscopic gastrectomy [16]. Based on our abundant experience in robotic gastrectomy, we hypothesized that DVSS would allow surgeons to perform reproducible VEG-DFT procedures in a safer environment. Thus, we launched robotic VEG-DFT in January 2014 at our institution. The objective of this single-institution retrospective study was to determine the feasibility and safety as well as the learning effect of robotic VEG-DFT.

Materials and methods

Patients

In this retrospective single-institution study, the medical records of patients who underwent gastrectomy between January 2014 and December 2015 were reviewed. The patient selection process is summarized in Fig. 1. In brief, during the study period, a total of 362 consecutive patients underwent gastrectomy for primary gastric malignancies, including gastric cancer, gastric gastrointestinal stromal tumor, and adenocarcinoma of the esophagogastric junction at our institute. Among these, 80 consecutive patients who agreed to the uninsured use of DVSS underwent robotic gastrectomy, and 12 of these patients who underwent robotic PG with VEG-DFT were enrolled in the present study. VEG-DFT is indicated for early gastric cancer localized in the upper third of the stomach, adenocarcinoma of the esophagogastric junction according to the Nishi classification, and gastric gastrointestinal stromal tumor (GIST) near the cardia [17]. For gastric cancer and adenocarcinoma of the esophagogastric junction, clinical tumor staging was based on the 14th edition of the Japanese Classification of Gastric Carcinoma [17], and the extent of systemic lymph node dissection was determined according to the Japanese Gastric Cancer Treatment Guidelines 2014 [3]. For gastric GIST, clinical tumor staging was determined according to the 7th edition of the International Union Against Cancer (UICC) Tumor Node Metastasis Classification [18], and surgical treatment including the extent of lymph node dissection was determined according to the guideline of the European Society for Medical Oncology (ESMO) [19]. Patients were fully involved in the decision-making process, and informed consent was obtained from all patients. All patients were charged 2,200,000 JPY during perioperative admission [14, 20]. All procedures were performed by the same gastric surgeon (I.U.), who had performed more than 1500 laparoscopic and 250 robotic gastrectomy procedures. This study was approved by the institutional review board of Fujita Health University.

Fig. 1
figure 1

Flow diagram of the study selection process. DG distal gastrectomy, TG total gastrectomy, PD pancreaticoduodenectomy, PG proximal gastrectomy

Full size image

Operative procedure

Setup for robotic proximal gastrectomy

For robotic PG, we used a 5-port system with Nathanson hook liver retractors (Yufu Itonaga, Tokyo, Japan), as previously described [15]. We preferred using the robotic third arm with our right hand; thus, it was placed on the left side of the patient. In detail, under general anesthesia, a longitudinal 12-mm incision was created on the umbilicus, and a camera trocar was put in place. Robotic third, first, and second arms were docked on 8-mm left upper, left lower, and right upper trocars, respectively. The 8-mm left lower trocar was placed through the 12/75-mm trocar (trocar-in-trocar technique). The assistant surgeon used the 12/100-mm right lower trocar (Fig. 2). To prevent collision of the robotic arms, the patient cart was docked in accordance with the “da Vinci’s plane theory,” and intracorporeal positioning of the forceps was determined based on the “monitor quadrisection theory” as previously described [14].

Fig. 2
figure 2

Trocar arrangement and rolls of each trocar in robotic proximal gastrectomy

Full size image

Proximal gastrectomy and lymph node dissection

Omentum was mobilized approximately 4 cm inferior to the gastroepiploic arcade. Root of the gastric branch of the left gastroepiploic artery was subsequently divided. Short gastric vessels were divided at their root for No. 4sa lymph node dissection. In patients with clinical tumor stage ≥2 (cases 2 and 6), adipose tissue bearing station 11d and 10 lymph nodes without spleen were dissected along the distant splenic artery up to the splenic hilum, as previously described [21]. Then, the communicating vessel between the right and left gastroepiploic arteries was divided, and adipose tissue bearing station 4sb lymph nodes were dissected along the greater curvature of the stomach. The right gastroepiploic vessels were carefully preserved. Next, lesser omentum was mobilized, and station 3a and 1 lymph nodes were dissected along the lesser curvature of the stomach up to the esophagogastric junction, with careful preservation of the right gastric vessels and the hepatic branch of the anterior vagal trunk. Celiac branches of the posterior vagal trunk were not routinely preserved. Suprapancreatic nodal dissection for No. 7, 8a, 9, and 11p lymph nodes was performed using the medial approach along the outermost layer of the autonomic nerve, as previously described [15, 22–25]. Finally, the esophagogastric junction was mobilized with dissecting station 2 lymph nodes, and the abdominal esophagus was transected using a linear stapler via the assistant trocar. If necessary, lower mediastinal dissection was conducted as previously described [26]. After the patient cart of DVSS was temporarily undocked, the camera port at the umbilicus was extended to 4 cm and protected using a Multi Flap Gate™ (Sumitomo Bakelite, Tokyo, Japan). Next, the stomach was extracted, and the upper body of the stomach was transected using linear staplers. A typical procedure of robotic PG with D1+ dissection for gastric cancer is shown in Supplemental video clip #1. In the only patient with GIST in this study (case 9), only station 7 among suprapancreatic lymph nodes was dissected.

Reconstruction with valvuloplastic esophagogastrostomy by double flap technique

After the conclusion of robotic PG, transverse H-shaped seromuscular flaps were extracorporeally created at the anterior wall of the remnant stomach according to previous reports with minor modifications [11, 27, 28]. The dimensions of the door were 2.5 cm × 3.5 cm (width × height) situated at 5 cm below the tip of the greater curvature of the remnant stomach (Fig. 3A). This residue formed a pseudo-fornix. Using a sharp knife and an electric cautery, the submucosal layer was cautiously detached and the mucosa was exposed, resulting in the creation of the seromuscular flaps (Fig. 3B). After the seromuscular flap was created, a mucosal window was opened 1 cm above the caudal end of the seromuscular flap (Fig. 3C). Then, stomach was inserted into the abdomen, and the patient cart was redocked.

Fig. 3
figure 3

Creation of the seromuscular flaps. A Dimensions of the door are 2.5 cm × 3.5 cm (width × height) at 5 cm below the tip of the greater curvature of the proximal margin. B Using a sharp knife and an electric cautery, the submucosal layer is carefully detached to expose the mucosa, resulting in the creation of seromuscular flaps. C After the seromuscular flaps are created, a mucosal window is opened 1 cm above the inferior end of the seromuscular flaps

Full size image

First, the posterior aspect of the esophagus 4 cm above the cut end was marked with a solid line using methylrosanilinium chloride (pyoktanin) (Fig. 4A) and fixed to the cranial end of the mucosal window using five stitches (Fig. 4B). Next, the proximal esophagus was clamped with a clamping clip (Fig. 4C), and the staple line at the cut end was removed (Fig. 4D). Continuous suturing using an absorbable unidirectional barbed suture (4-0 V-Loc™; Covidien, Mansfield, MA, USA) was performed between the full thickness of the posterior wall of the esophagus and the cranial opening of the mucosa of the remnant stomach (Fig. 5A). Then, continuous or interrupted layer-to-layer sutures were placed between the anterior wall of the esophagus and the caudal opening of the mucosa (5B, C). Finally, the anastomosis was covered by the seromuscular flaps (Fig. 5D). The cranial ends of both sides of the seromuscular flaps were fixed in position slightly below the caudal end of the mucosal window to allow for the anastomosis to be fully covered by the flaps. Both side ends of the flaps were upwardly sutured in the midst of the window. Near the cranial end of the window, it was not necessary to join both ends of the flaps when the tension between the flaps appeared exceptionally strong. After completion of the anastomosis, the esophagus was fixed firmly to the hiatus. When the anastomosis was created in the mediastinum, the diaphragm was closed with the body of the stomach fixed circumferentially to the hiatus. A closed drain was routinely inserted near the anastomosis. Additional pyloroplasty was not performed in any of the cases in the present study. The entire VEG-DFT is summarized in Supplemental video clip #2.

Fig. 4
figure 4

Fixation of the esophagus and preparation for anastomosis. A Esophagus at 4 cm above the cut end is marked with a solid line using methylrosanilinium chloride (pyoktanin). B The posterior wall of the esophagus is fixed to the superior end of the mucosal window using five stitches. C The proximal esophagus is clamped by a clamping clip. D The stapler line at the cut end of esophagus is cut by scissors

Full size image
Fig. 5
figure 5

Robotic esophagogastrostomy by layer-to-layer handsewn anastomosis, followed with coverage by the seromuscular flap. A Continuous suturing using a unidirectional absorbable barbed 4-0 V-Loc™ suture is performed between the posterior wall of the esophagus and the superior opening of the mucosa of the remnant stomach. B Continuous muco-mucosal anastomosis using a unidirectional absorbable barbed 4-0 V-Loc™ suture is placed between the anterior wall of the esophagus and the inferior opening of the mucosa. C Continuous or interrupted muscular-seromuscular anastomosis using a unidirectional absorbable barbed 4-0 V-Loc™ suture is placed between the anterior wall of the esophagus and the inferior opening of the mucosa. D Anastomosis is covered by the seromuscular flaps. Both side ends of flaps are upwardly sutured in the midst of the window

Full size image

Postoperative management

None of the patients underwent nasogastric tube placement after surgery. Postoperative care was provided to all patients according to the same clinical protocol: Walking and drinking water were resumed on postoperative day (POD) 1, soft meals were resumed on POD 3 after the examination of anastomosis by a water-soluble contrast (50 ml of meglumine sodium amidotrizoate), a closed drain was removed on POD 4, and hospital discharge was permitted on POD 7 for patients with a favorable postoperative course. All patients who underwent VEG-DFT were routinely administered an oral proton pump inhibitor (PPI) for several months after surgery.

Postoperative follow-up protocols

Postoperative chemotherapy and oncological follow-up protocols were clearly demonstrated in our previous publications [14]. Briefly, the discharged patients visited our outpatient clinic at least 1 month, 3 months, and every 6 months until 5 years after surgery. At the clinic, routine laboratory tests with carcinoembryonic antigen and carbohydrate antigen 19-9 and physical examination were performed. The patients were examined with chest and abdominopelvic computed tomographic scans every 6 months to detect local recurrence and systemic metastasis. Upper-gastrointestinal endoscopy was performed every year to rule out local recurrence and metachronous multicentric or multiple cancers. In addition to these routine follow-up protocols, upper-gastrointestinal endoscopy was performed routinely every 6 months in patients who underwent PG after VEG-DFT to evaluate the anastomotic passage and grade of reflux esophagitis. The severity of the reflux esophagitis was determined according to the modified Los Angeles classification.

Assessment of learning curve by cumulative summation method

The learning curve for VEG-DFT was analyzed using the cumulative summation (CUSUM) method, which is uniquely suited to learning curve evaluation as previously reported [29]. The CUSUM graph represents the accumulation of differences between the individual data and the mean of all data. CUSUM of the anastomotic time (AT; CUSUMAT) for each patient was determined. First, cases were organized in chronological order, and the recursive process was started according to the surgical data as follows: CUSUMAT for the first case was the difference between the AT of the first case and the mean AT for all cases (μAT). CUSUMAT for the second case was the former CUSUMAT added to the difference between the AT of the second case and μAT. The same procedure was repeated for all patients except for the last one, for whom the CUSUMAT was calculated as 0.

Assessment of technical risk factors contributing to anastomotic stenosis and the impact of learning effect on VEG-DFT

To explore the impact of learning effect on VEG-DFT and to determine the technical risk factors contributing to anastomotic stenosis, AT and number of stitches used for anastomosis were assessed within the following five steps that comprised VEG-DFT by reviewing non-edited videos that recorded the entire intra-abdominal operative procedure in all cases: (1) fixation of the esophagus on the cranial end of the mucosal window of the stomach (Fig. 4B), (2) anastomosis between the posterior wall of the esophagus and the cranial opening of the mucosa of the remnant stomach (Fig. 5A), (3) mucosal anastomosis between the anterior wall of the esophagus and the caudal opening of the mucosa (Fig. 5B), (4) seromuscular suture between the anterior wall of the esophagus and the caudal opening of the mucosa (Fig. 5C), and (5) semi-closure of the seromuscular flaps (Fig. 5D).

Measurements

All patients were observed for 30 days following surgery, and short-term surgical outcomes, including anastomosis-related and non-anastomosis-related complications as well as operative time, surgeon console time, AT, total number of stitches used to complete VEG-DFT, estimated blood loss, length of postoperative hospital stay, and clinicopathological characteristics, were assessed. Postoperative complications that were ≥grade II or III based on the Clavien–Dindo (CD) classification were recorded [30] and classified in accordance with the Japan Clinical Oncology Group Postoperative Complication Criteria according to CD ver. 2.0 [31]. Total operative time was defined as the time from the start of the abdominal incision until the end of complete closure of the wounds. Surgeon console time was defined as the time of DVSS operation during surgery. AT was defined as the time from the start of the fixation of the posterior aspect of the esophagus to the cranial side of the mucosal window through until the completion of the semi-closure of the seromuscular flaps. Blood loss was estimated by weighing suctioned blood and gauze pieces with absorbed blood. Body weight and serum levels of hemoglobin and albumin were measured at preoperation and at 6 months after surgery. The relative rate of change in body weight was calculated as (postoperative body weight) / (preoperative body weight). Ideal body weight was determined as the weight with a body mass index (BMI) of 22 kg/m2, and the ratio of ideal body weight was calculated as (postoperative body weight) / (ideal body weight).

Statistical analysis

Data were presented as medians with range. Independent continuous variables were compared by Student’s t test, and categorical variables were compared by the Chi-square test with Yates’ correction where necessary. Statistical analyses were performed using StatMate IV for Windows (Advanced Technology for Medicine & Science, Tokyo, Japan), and p values < 0.05 were considered statistically significant.

Results

Clinical and oncological characteristics of the patients

Background and tumor characteristics of the 12 patients included in this study are presented in Table 1. There were nine males, and median age and BMI were 67 (45–87) years and 23.1 (18.1–26.7) kg/m2, respectively. The American Society of Anesthesiologists’ (ASA) score was 1 in three patients and 2 in the remaining nine patients. Preoperative diagnoses were gastric cancer in five patients, adenocarcinoma of the esophagogastric junction in six patients, and gastric GIST in one patient. Clinical tumor stages were I, II, and III in 9, 2, and 1 patient, respectively.

Table 1 Patient characteristics and surgical outcomes of robotic proximal gastrectomy followed by valvuloplastic esophagogastrostomy using double flap technique
Full size table

Surgical and short-term outcomes

Surgical and short-term outcomes are presented in Table 1. Intra-abdominal anastomosis was performed in eight patients, whereas the remaining four patients underwent intramediastinal anastomosis. No patients suffered from intraoperative adverse events. Median operative time, console time, and AT were 406 (324–613), 267 (214–483), 104 (76–186) min, respectively, and median estimated blood loss was 31 (5–130) ml. None of the patients required conversion to laparoscopic or open surgery. Mortality was 0%. Postoperative complications of CD grade II occurred in 1 (8%) patient who was diagnosed with intra-abdominal abscess around the pancreatic body via computed tomography and recovered with antibiotic administration for one week. There were no postoperative complications of CD grade III or higher. Median length of postoperative hospital stay was 10 (9–30) days.

Late anastomotic complications after postoperative day 30 within 6 months after operation

Late anastomotic complications after postoperative day 30 are also presented in Table 1. Before the routine postsurgical evaluation 6 months after surgery, three patients (25%; cases 4, 5, and 8) complained of dysphagia at the outpatient clinic. Endoscopic examination revealed a diagnosis of anastomotic stenosis. Then, they underwent endoscopic balloon dilation.

Risk factors for anastomotic stenosis

Assessment of technical risk factors for anastomotic stenosis showed that the total number of stitches in patients who developed anastomotic stenosis was significantly more than that in patients without anastomotic stenosis (p < 0.001, Table 2). Meanwhile, there was no difference in total AT irrespective of the presence of anastomotic stenosis (p = 0.098, Table 2).

Table 2 Analysis of the association between total number of stitches and postoperative anastomotic stenosis. Student’s t test was used for between-group comparisons of total number of stitches and anastomotic time (AT)
Full size table

Learning effect

Raw ATs and the CUSUMAT graph were plotted in chronological order of cases (Fig. 5). The CUSUMAT learning curve was best modeled as a third-order polynomial with the following equation: CUSUMAT (min) = 0.1085 × case number3 − 7.2061 × case number2 + 77.475 × case number − 78.313, which had a high R 2 value of 0.9137 (Fig. 6). According to the change in the slope of Fig. 5, the CUSUMAT learning curve was divided into two phases: Phase 1 included the first six cases, and the remaining six cases constituted phase 2. Figure 7 displays the curves with best fit for both phases. Phase 1 represented the initial phase with the learning curve, which was the best fit as a straight line with a positive slope (R 2 > 0.9). Conversely, phase 2 was the best fit as a straight line with a negative slope (R 2 > 0.9). These findings indicated that the first six cases were required to reach a learning plateau.

Fig. 6
figure 6

Anastomotic time (AT) and cumulative sum (CUSUM)AT plotted against case numbers (gray dotted line), which represents the curve of best fit for the plot (third-order polynomial with the following equation: CUSUMAT = 0.1085 × case number3 −7.2061 x case number2 + 77.475 × case number − 78.313 (R 2 = 0.9137)

Full size image
Fig. 7
figure 7

Lines of best fit for each phase of the CUSUMAT learning curve. A Phase 1 represents the initial learning curve. B Phase 2 represents the well-developed phase

Full size image

To determine which specific step was vulnerable to the learning effect, the time in each step of the VEG-DFT was compared between the initial six and the latter six cases, as shown in Table 3. The median time of all steps except for step 5 (semi-closure of the seromuscular flaps) of the latter six cases was significantly shorter than that of the initial six cases. As a result, total AT in the latter six cases was also significantly shorter than that of the initial six cases (p = 0.003). The mean number of stitches in step 1 was significantly fewer in the latter 6 cases than in the initial 6 cases, but those of the remaining steps were not significantly different between the two groups of cases. As a result, the mean total number of stitches used for VEG-DFT was not significantly different between the two groups (p = 0.327).

Table 3 Anastomotic time (AT) and the number of stitches used in each step of the anastomosis during valvuloplastic esophagogastrostomy using double flap technique. Student’s t test was used for between-group comparisons
Full size table

Oncologic outcomes 6 months after surgery

The pathological diagnosis, surgical curativity, and status of recurrence 6 months after surgery are shown in Table 4. The R0 resection was completed successfully in all patients. Although the follow-up period may be too short to evaluate oncological outcomes, no tumor recurrence was identified in any patient within 6 months after surgery.

Table 4 Oncologic outcomes and endoscopic evaluation of reflux esophagitis 6 months after surgery
Full size table

Endoscopic evaluation for reflux esophagitis 6 months after surgery

The endoscopic findings 6 months after surgery are shown in Table 4. Nine patients without anastomotic stenosis had no endoscopic findings of reflux esophagitis of grade A or higher according to the modified Los Angeles classification. Among them, four patients (cases 1, 7, 10, and 12) successfully discontinued PPI therapy within 1 month according to their preferences in the outpatient clinic. Among patients who underwent endoscopic balloon dilation, two (cases 4 and 5) developed moderate reflux esophagitis of grades A and B, respectively, whereas the remaining patient (case 8) did not develop reflux esophagitis within 6 months after surgery.

Nutritional status at six months after operation

To assess the nutritional status at six months after VEG-DFT, the relative rate of change in body weight, ratio of ideal body weight to postoperative body weight, and serum levels of hemoglobin and albumin were investigated. The relative rates of change in body weight and ratio of ideal body weight to postoperative body weight at six months after surgery were 91.5% (78.7–96.4%) and 92.7% (76.8–106.5%), respectively. The serum levels of hemoglobin and albumin were not significantly different between the preoperative period and at six months after surgery (hemoglobin, 13.9 (10.1–16.1) and 13.5 (9.7–16.1) g/dL, p = 0.279; albumin, 4.5 (3.4–5.1) and 4.5 (3.9–4.8) g/dL, p = 0.396, respectively).

Discussion

VEG-DFT was first reported as an efficient reconstruction approach after open PG to prevent reflux esophagitis by Kamikawa et al. [27]. Distal esophagus and the site of anastomosis are implanted in the submucosal layer, and the anterior side of anastomosis is fully covered by the seromuscular double flap. These structures and intragastric pressure create a pressure gradient between the esophagus and stomach, acting as a one-way valve. There were several reports demonstrating the technical feasibility of VEG-DFT via open, laparoscopy-assisted, or laparoscopic approaches and the associated favorable short-term outcomes; however, most of these studies enrolled a small number of patients [11, 27, 28], and only one report included more than ten patients [12]. To the best of our knowledge, this is the first study reporting the feasibility and short-term outcomes of robotic VEG-DFT. Four major findings emerged in the present study.

First, robotic PG, followed by VEG-DFT, was performed with minimal intraoperative blood loss and without severe intraoperative or early major postoperative complications. Although one patient (8.3%) developed intra-abdominal abscess (CD grade II) and his hospital stay was prolonged to 30 days, the remaining 11 patients were discharged within 2 weeks after VEG-DFT. These data clearly demonstrated the feasibility and safety of our method. In agreement with our experience, the only report that evaluated the short-term outcomes of VEG-DFT via laparoscopic approach clearly showed that the incidence of postoperative early complications was 3% [11], whereas open or laparoscopic PG followed by esophagogastrostomy, excluding VEG-DFT, induced leakage and morbidity in 0–8.8% and 0–35.7% of the cases, respectively [10].

Second, none of the patients developed reflux esophagitis of grade A or higher immediately after surgery; however, three patients did develop secondary reflux esophagitis after endoscopic balloon dilation to treat anastomotic stenosis, and all patients received prophylactic PPI therapy. Kuroda et al. also reported that none of the patients who underwent VEG-DFT experienced grade B or higher reflux esophagitis [11]. Esophagogastrostomy after PG without additional anti-reflux maneuvers is usually considered to increase the risk of reflux esophagitis, with a reported incidence of 0–33% [10]. Accordingly, the anti-reflux mechanism of VEG-DFT has great potential in prevention of reflux esophagitis. Given that robotic surgery offers great advantages in intracorporeal suturing due to several technical properties, including wristed instruments, tremor filtration, ability to scale motion, and stereoscopic vision [12–15], we believe that the use of DVSS might be superior to laparoscopic or open approaches in the reproducible application of VEG-DFT to prevent reflux esophagitis.

Third, the learning curve of robotic VEG-DFT was quite short in the present study. This outcome was in agreement with the previous reports demonstrating that one of the major advantages of the DVSS was its short learning curve [32–35]. In fact, robotic gastrectomy required about 20 or fewer cases to overcome learning curves, whereas forty to sixty cases were required for surgical experience in laparoscopic gastrectomy [32]. In the present study, only six patients were needed to reach the learning plateau. After the initial experience with the first 6 cases, the AT was significantly shortened from 140 to 90 min. The median AT in the latter group of cases was comparable to the previously reported median reconstruction times by laparoscopic or mini-laparotomy procedures [11]. Since the total number of stitches used for anastomosis did not differ between these two groups, the improvement in AT might be achieved at least partly by increasing the efficiency of the anastomotic process including the creation of a good surgical field and suturing maneuvers.

Fourth, attention should be paid to prevent anastomotic stenosis after VEG-DFT. In the present study, postoperative stenosis occurred in three patients at two months after surgery. The previous study by Kuroda et al. showed that anastomotic stricture occurred in 5% of the patients that underwent reconstruction via mini-laparotomy and in 15% of those that underwent total laparoscopic reconstruction [11]. Therefore, anastomotic stenosis should be recognized as an important surgical complication after VEG-DFT. In the present study, a greater number of stitches were used in patients with anastomotic stenosis than in those who did not develop anastomotic stenosis, suggesting that the anastomotic site might be too tightly closed with an excessive number of stitches. Intriguingly, in contrast to the observed risk of associated anastomotic stenosis in the present study, VEG-DFT is recognized as a reconstruction procedure with decreased risk of anastomotic leakage, as the anterior aspect of the esophagogastrostomy is fully wrapped with a seromuscular flap, which results in reinforcing the anastomosis [11]. In fact, no anastomotic leakage occurred in either Kuroda’s or the present study [11]; however, the number of cases was limited in both studies. Therefore, it might be important to ensure that the sutures are somewhat loose to prevent stenosis of the anastomotic site but are tight enough to prevent anastomotic leakage.

There are some limitations to the present study. First, this study was conducted by the same highly skilled surgeon (I. U.). Therefore, it is unclear whether this learning curve can be equally applied to other surgeons, including trainees and skilled laparoscopic surgeons with minimal experience with robotic surgery. Second, this was a single-institution, single-arm retrospective cohort study and included a small number of patients. Although the total number of stitches was identified as a risk factor for postoperative anastomotic stenosis, further analyses, including multivariate analysis to identify other potential risk factors were not possible due to the small sample size. In addition, we did not compare robotic VEG-DFT with laparoscopic or open VEG-DFT. Furthermore, we did not compare robotic VEG-DFT with esophagogastrostomy followed by other anti-reflux procedures, such as the Toupet-like partial fundoplication [36], narrow gastric conduit [37], fixation of esophagus using a knifeless endoscopic linear stapler [38, 39], recreation of a sharp angle of His [40], and lower esophageal sphincter preservation [41]. Thus, further studies, including a larger number of patients and prospective randomized controlled studies comparing robotic VEG-DFT with other procedures, are necessary to clarify the anti-reflux efficacy of this procedure. Third, the postoperative observation period was limited to only six months, and PPI therapy was prophylactically administrated in all patients for several months in the present study. Since the primary intent of this procedure is to prevent reflux esophagitis without the use of antiulcer agents [11, 27], we will assess the long-term effects of robotic VEG-DFT on reflux esophagitis after tapering or discontinuing PPIs in future studies. Additionally, longer observation periods are required to more extensively evaluate the efficacy of robotic VEG-DFT in maintaining nutritional status and preventing anemia, although the present study showed that there was a 10% decrease in mean body weight and that there were no changes in the levels of albumin and hemoglobin at six months after robotic VEG-DFT. In addition, long-term oncological outcomes deserve careful attention. Because our indication for PG is within the recommendation of the guideline, we believe that cancer curability is not impaired by this procedure. Further study is warranted to determine the oncological safety of this procedure.

In conclusion, robotic PG followed by VEG-DFT is a promising procedure to prevent reflux esophagitis, based not only on its feasibility but also its shorter learning curve, requiring only six patients. Attention is required to prevent postoperative anastomotic stenosis, possibly caused by an excessive number of stitches for esophagogastrostomy.