Three-Port Versus Five-Port Laparoscopic Distal Gastrectomy for Early Gastric Cancer Patients: A Propensity Score Matched Case-Control Study

ABSTRACT

Aim: The aim of this study was to evaluate the feasibility and safety of three-port laparoscopic distal gastrectomy (TP-LDG) as a reduced port laparoscopic gastrectomy. Materials and Methods: We retrospectively identified 146 patients preoperatively diagnosed with early gastric cancer who underwent five-port laparoscopic distal gastrectomy (FP-LDG) or TP-LDG between May 2013 and July 2016. A propensity score matching analysis was used to create patient groups (48 patients in each group) matched for sex, age, body mass index, previous abdominal surgery history, and American Society of Anesthesiologist score. The short-term surgical outcomes between TP-LDG and FP-LDG were compared. Results: The TP-LDG group had a statistically shorter umbilical wound length [3.4 (range, 3.0–4.0) cm vs. 3.9 (range, 3.7–4.0) cm, p = .000], shorter operative time [230 (range, 190–310) min vs. 250 (range, 180–320) min, p = .036], and lower estimated blood loss [68 (range, 20–180) mL vs. 80 (range, 40–150) mL, p = .005] compared to that in the FP-LDG group. However, there was no specific superiority regarding inflammatory profiles. Complication rates were also similar (8.4% TP-LDG vs. 12.6% FP-LDG, p = .504). Conclusions: TP-LDG is a feasible and safe surgical procedure for the patients with early gastric cancer and provides the benefit of better cosmesis.

Keywords::

• reduced port laparoscopic gastrectomy
• three-port laparoscopic gastrectomy
• propensity score matching analysis
• early gastric cancer
• outcomes

INTRODUCTION

In recent years, reduced port laparoscopic gastrectomy (RPLG) has been increasingly used in the treatment of early gastric cancer (EGC). However, five-port laparoscopic distal gastrectomy (FP-LDG) has been the primary consideration for the treatment of EGC since the procedure was first performed in Japan in 2002 [1]. The safety, reliability and good cosmetic result of FP-LDG contributed to its standing [2, 3]. With the accumulation of experience and the advancement of equipment for FP-LDG, the number of incisions for ports has been reduced, resulting in improved postoperative quality of life in terms of scarring [4].

RPLG has been conducted in the Department of Surgery at Haeundae Paik Hospital since November 2015. At first, both two- and three-port laparoscopic distal gastrectomy were performed; however, with two-port laparoscopic distal gastrectomy, techniques using straight and inflexible equipment were complex and difficult because of interference between the operator's right hand instrument and other equipment. Accordingly, we have performed mainly three-port laparoscopic distal gastrectomy (TP-LDG), which involves two operator's trocars located independently in the right side of the abdomen and one multi-access port in the umbilicus. This surgical approach enables the operator to freely perform the laparoscopic maneuver. The appropriate retraction and triangulation are still maintained despite the decreased total number of ports as the operator apply the packing gauze to the appropriate area and the assistant can pass instruments through the multi-access port in the umbilicus.

The aim of this study was to evaluate the feasibility and safety of TP-LDG by comparing the short-term surgical outcomes for TP-LDG and FP-LDG.

MATERIALS AND METHODS

Patient Selection

Since March 2010, 1,115 patients preoperatively diagnosed with gastric cancer underwent gastrectomy at the Department of Surgery at Haeundae Paik Hospital, Inje University College of Medicine, Korea. For EGC, we performed laparoscopic assisted distal gastrectomy (LADG) until March 2013. We then performed FP-LDG using five ports until October 2015, when TP-LDG was introduced. The indications of operative methods, either FP-LDG or TP-LDG, were not different. The difference for the selection of procedure was just dependent on when the operation was conducted. All patients who underwent Billroth-I and Roux-en-Y gastrojejunostomy reconstruction were excluded, because the number of these patients were small compared to the number of the patients who underwent Billroth-II reconstruction during study period. There were no cases of conversion to FP-LDG or open surgery, so that no one was excluded for conversion.

As a result, a retrospective review of 146 patients preoperatively diagnosed with EGC who underwent FP-LDG (94 patients) or TP-LDG (52 patients) between May 2013 and July 2016 was conducted. All surgeries were performed by a single surgeon (S.J. OH) who had performed more than 300 cases of LADG and FP-LDG before starting TP-LDG. The final patient selection was based on a propensity score matching (PSM) analysis, which was used to create patient groups matched for sex, age, body mass index (BMI), previous abdominal surgery history, and American Society of Anesthesiologist (ASA) score. Finally, 48 patients in each group were enrolled.

This study was approved by the Institutional Review Board of the Ethics Committee (Inje University Haeundae Paik Hospital IRB).

Surgical Techniques

The patient was placed supine in the lithotomy and reverse Trendelenburg position under general anesthesia. The operator stood at the patient's right side, the assistant stood at the patient's left side, and the endoscopist stood between the patient's legs. During both FP-LDG and TP-LDG, the instruments and endoscope used were straight and inflexible, and the endoscope was 10-mm diameter with 30° direction of view. The extent of lymph nodes (LNs) dissection was D1+ or greater, based on the recommendations of the Japanese Gastric Cancer Association Guidelines [5]. Before closing the incision sites, one drainage tube was applied through the right upper trocar insertion site.

Five-Port Laparoscopic Distal Gastrectomy

During FP-LDG, five ports were introduced. One 12-mm port was inserted into the umbilicus for the endoscope. Another 12-mm port was inserted at the right mid-clavicular line for the operator's right hand, and one 5-mm port was inserted at the right mid-axillary line below the costal margin for the operator's left hand. Another two 5-mm ports were inserted at the left side of the abdomen in the same manner for the assistant's maneuver.

Liver Retraction

Liver retraction is an important procedure to ensure a good visual field and adequate working space. During laparoscopic gastrectomy, the operator used a technique involving a Penrose drain to suspend the liver and avoid injury.

Dissection of the Greater Omentum

Dissection of the greater omentum was started from the mid-portion of the transverse colon, continuing toward the lower pole of the spleen. After the left gastroepiploic vessels were ligated and the LN station 4sb was dissected, right-sided omentectomy was continued toward the hepatic flexure of the colon. As the dissection of the avascular plane among the stomach, the duodenum and the colon progressed, the gastrocolic trunk of Henle and the anterior superior pancreaticoduodenal vein appeared. After the identification of these vessels, the right gastroepiploic vessels could be identified and ligated, and LN station 6 could be dissected. Then, the right gastric vessels were ligated at their roots and LN station 5 was dissected.

Transection of the Duodenum

The iDrive™ Ultra Powered Stapling Device (Covidien, Mansfield, MA, USA) was inserted through the operator's 12-mm port. This linear stapler was rotated and articulated as appropriate, and the duodenum was transected 2 cm distal to the pylorus.

Dissection in the Suprapancreatic Area

After the duodenum was transected, the common hepatic, proper hepatic, and celiac arteries were identified. Along these arteries, LN station 8, 12, and 9 were dissected. The assistant pushed the pancreas gently downward using a grasper for traction, and the operator dissected the remaining suprapancreatic tissues including LN station 11, along the splenic vein. After the dissection of LN station 9, the left gastric vessels could be identified and ligated, and LN station 7 could be dissected. In addition, LN station 1 was dissected and the remaining lesser curvature side was cleared before transecting the stomach.

Transection of the Stomach and Specimen Extraction

On the day before the surgery, endoscopic tumor localization was performed by clipping proximal to the tumor. Before gastric transection, metal clips were applied to the greater and lesser curvatures at the endpoint of the planned proximal resection margin. In addition, portable C-arm was used to confirm the safe resection line. After confirmation of margin safety, the stomach was transected intracorporeally using two linear staplers. The stapled line was intracorporeally reinforced using interrupted sutures with silk 3-0 thread.

After gastric transection, a plastic bag was introduced through the umbilical 12-mm port, and the endoscopist introduced the endoscope through the right-sided 12-mm port. The operator placed the specimen into a plastic bag and extended the umbilical port insertion site to 3–4 cm. The plastic bag containing the specimen was removed from the abdominal cavity, and the extended umbilical trocar site was closed, leaving a hole large enough to re-introduce the 12-mm port.

Reconstruction

Reconstruction was performed using the Billroth-II method. The jejunotomy was intracorporeally performed using an energy device. Gastrotomy was also intracorporeally performed using an energy device 2-cm distal from the endpoint of the greater curvature and on the posterior wall of the remnant stomach. The operator introduced the thinner fork of a linear stapler into the jejunotomy, delivering it near the remnant stomach. In addition, the operator introduced the thicker fork into the gastrotomy. After performing the gastrojejunostomy, the common entry hole was closed using running sutures with V-LOC™ (Covidien, Mansfield, MA, USA) thread, reinforced by interrupted sutures using silk 3-0 thread.

Three-Port Laparoscopic Distal Gastrectomy Protocol

During TP-LDG, a 3-cm sized incision was made at the umbilicus to apply both the wound retractor and OCTO™ Port (DalimsurgNET, Seoul, Korea). The OCTO™ Port has four accesses, one 5–12-mm port, one 5/10-mm port, and two 5-mm ports. The endoscopist usually used 5–12-mm port and assistant used other ports. In addition, one 12-mm port and one 5-mm port were inserted in the right side of the abdomen in the same manner as with FP-LDG for the operator's instruments (Figure 1). The operator performed most of procedures from the right-sided two ports, except for gastric transection procedure. During gastric transection, it was more convenient for the operator to adjust^{TABLE 1} the rotation and angulation of the linear stapler by introducing the linear stapler through the OCTO™ Port. At this time, the endoscope was introduced through the 12-mm port.

FIGURE 1 (A) The port locations for TP-LDG are shown. The assistant's instruments and endoscope are introduced through an OCTO™ Port (DalimsurgNET, Seoul, Korea) in the umbilicus. The Operator's laparoscopic instruments are introduced through the ports in the right side of the abdomen. (B) The appearance of the wound 14 days after TP-LDG is shown.

The procedures for gastric mobilization, LNs dissections, gastric transection, and reconstruction were similar to those during FP-LDG. Generally, the procedures requiring the assistant's role could be performed without much trouble because the navigation paths of the assistant's instruments were almost parallel with those of the endoscope. Therefore, there was little interference between instruments and endoscope. The procedure for specimen extraction was different compared to that during FP-LDG. After gastric transection, the OCTO™ Port was removed and the specimen was then easily extracted through the wound retractor. Subsequently, the umbilical port was recovered simply by re-attaching the OCTO™ Port.

Postoperative Management

For the postoperative management of patients, the same established protocol was applied to both groups. All patients received Patient-controlled analgesia (PCA). If a patient required additional pain control, nonsteroidal anti-inflammatory drugs were administered. The visual analogue scale (VAS) of postoperative pain was measured on postoperative days 1, 3, and 5, was collected at 08:00 each day. The inflammatory profile, described in terms of white blood cell (WBC) count and C-reactive protein (CRP) level, was assessed on postoperative days 1, 2, 4, and 7. Water was provided after the first flatus had been passed, followed by a liquid diet and then soft diet, introduced gradually. When the diet was acceptable and no adverse postoperative events were noted, the discharge of patient was recommended.

Statistical Analysis

The MatchIt package in the statistical program R (https://www.r-project.org) with nearest-neighbor 1-to-1 matching was used to conduct a PSM analysis to reduce selection biases between the two patient groups. Sex, age, BMI, previous abdominal surgery history, and ASA score were used as covariates. In addition, the options were set as follows: method = “nearest,” discard = “both,” caliper = 0.2, and ratio = 2. To measure the balance of the individual covariates before and after PSM, the standardized mean difference of each covariate was calculated and standardization was confirmed. Finally, we compared 48 TP-LDG patients with 48 matched FP-LDG patients in terms of short-term surgical outcomes.

Continuous variables were analyzed using t tests or Wilcoxon rank sum tests, and categorical variables were analyzed using χ² tests or Fisher's exact tests, as appropriate. p-Values <.05 were considered statistically significant. Statistical analyses were performed using SPSS version 18.0 (SPSS Inc., Chicago, IL, USA) and the free statistical program R (ۙThe R Foundation).

RESULTS

Demographics

The demographics of the selected patients are shown in (Table 1). There were no significant group differences for sex, age, BMI, previous abdominal surgery history, or ASA score, as expected.

TABLE 1 Patient demographics

Variable	FP-LDG (n = 48)	TP-LDG (n = 48)	p Value
Sex			.681
Male	28 (58.3)	26 (54.2)
Female	20 (41.7)	22 (45.8)
Age (year)	57 (34–76)	57 (38–75)	.950
BMI^a (kg/m^2)	24.5 (16.8–31.2)	24.3 (17.3–31.1)	.378
Comorbidity	24 (50)	20 (41.7)	.413
ASA^b classification			.892
I	22 (45.8)	23 (47.9)
II	22 (45.8)	20 (41.7)
III	4 (8.4)	5 (10.4)
Previous abdominal surgery	6 (12.5)	3 (6.3)	.294

Data are expressed as n (%) or median (range).

^a Body mass index.

^b American Society of Anesthesiologist.

Pathologic Characteristics

(Table 2) shows the pathologic characteristics of the patients. No group differences were observed for tumor size, tumor location, distance to the proximal and distal resection margins, Lauren classification, histopathological scores (T and N stages), or the number of total retrieved lymph nodes. Resection margins were always tumor-free.

TABLE 2 Pathologic characteristics

Variable	FP-LDG (n = 48)	TP-LDG (n = 48)	p Value
Tumor size (cm)	2.0 (0.4–7.5)	2.0 (0.8–9.0)	.457
Tumor location			.235
Middle third	12 (25)	22 (45.8)
Lower third	36 (75)	26 (54.2)
Proximal resection margin (cm)	3.0 (0.4–15.5)	3.5 (0.7–12.5)	.736
Distal resection margin (cm)	6.2 (2.0–12.5)	6.0 (0.5–14.0)	.487
Lauren classification			.814
Intestinal	25 (52.1)	26 (54.2)
Diffuse	21 (43.8)	20 (41.6)
Mixed	2 (4.2)	2 (4.2)
Depth of invasion			.985
T1a	37 (77.1)	27 (56.3)
T1b	5 (10.4)	17 (35.4)
T2	2 (4.2)	4 (8.3)
T3	4 (8.3)	0 (0)
Total retrieved lymph nodes	48 (23–116)	49 (24–82)	.714
N stage			.143
N0	41 (85.4)	43 (89.6)
N1	3 (6.2)	5 (10.4)
N2	2 (4.2)	0 (0)
N3a	2 (4.2)	0 (0)

Data are expressed as n (%) or median (range).

Surgical Outcomes and Postoperative Courses

The short-term surgical outcomes and postoperative courses are described in (Tables 3–5). Statistical group differences were found for umbilical wound length, operative time, and estimated blood loss (Table 3). The umbilical wound length and operative time were shorter in the TP-LDG group compared to those in the FP-LDG group [3.4 (range, 3.0–4.0) cm vs. 3.9 (range, 3.7–4.0) cm, p = .000; 230 (range, 190–310) min vs. 250 (range, 180–320) min, p = .036, respectively]. In addition, the amount of estimated blood loss was lower in the TP-LDG group compared to that in FP-LDG group [68 (range, 20–180) ml vs. 80 (range, 40–150) ml, p = .005]. In contrast, there were no significant group differences in the time to first flatus after surgery, day on which a soft diet was started, VAS of postoperative pain, or the duration of the postoperative hospital stay (Table 3). Except for the WBC at postoperative days 1 and 4, there were no significant group differences in the inflammatory profiles (Table 4).

TABLE 3 Postoperative outcomes

Variable	FP-LDG (n = 48)	TP-LDG (n = 48)	p Value
Umbilical wound length (cm)	3.9 (3.7–4.0)	3.4 (3.0–4.0)	.000
Operative time (min)	250 (180–320)	230 (190–310)	.036
Estimated blood loss (mL)	80 (40–150)	68 (20–180)	.005
Time to first flatus (day)	3 (2–5)	3 (2–5)	.393
Time to soft diet feeding (day)	6 (5–10)	6 (5–12)	.789
VAS^a of postoperative pain
POD^b 1	3 (2–7)	3 (2–7)	.167
POD 3	3 (2–7)	3 (2–8)	.094
POD 5	3 (1–4)	3 (2–7)	1.000
Postoperative hospital stay (day)	13 (8–21)	13 (8–30)	.472

Data are expressed as median (range).

^a Visual analogue scale.

^b Postoperative day.

TABLE 4 Inflammatory profiles

Variable	FP-LDG (n = 48)	TP-LDG (n = 48)	p Value
Inflammatory profile
WBC^a count (× 10^3/mm^3)
POD^c 1	10.5 (4.9–21.1)	9.0 (3.8–16.5)	.012
POD 2	10.7 (1.3–21.5)	9.1 (5.5–21.0)	.357
POD 4	6.8 (3.5–13.2)	6.1 (3.4–13.0)	.027
POD 7	6.3 (3.2–16.2)	6.5 (3.0–14.9)	.938
CRP^b (mg/dL)
POD 1	3.4 (0.5–10.0)	3.0 (1.1–8.1)	.991
POD 2	10.6 (1.5–28.1)	11.5 (3.0–25.0)	.758
POD 4	7.1 (1.1–26.8)	8.7 (0.8–23.7)	.608
POD 7	2.5 (0.5–19.8)	4.0 (0.4–17.7)	.691

Data are expressed as median (range).

^a White blood cell.

^b C-reactive protein.

^c Postoperative day.

TABLE 5 Complications by Clavien-Dindo classification

Variable	FP-LDG (n = 48)	TP-LDG (n = 48)	p Value
Complications	6 (12.6)	4 (8.4)	.504
Grade I
Wound infection	1 (2.1)	0
Grade II
Pneumonia	1 (2.1)	0
Pancreatitis	0	1 (2.1)
Intraabdominal bleeding	1 (2.1)	0
Leakage	2 (4.2)	0
Fluid collection	0	3 (6.3)
Stenosis	1 (2.1)	0

Data are expressed as n (%).

Complications were defined as morbidities occurring within 30 days after discharge, and were classified according to the Clavien-Dindo classification system. There were 6 (12.6%) complications noted in the FP-LDG group, including wound infection, pneumonia, intraabdominal bleeding, leakage and stenosis (Table 5). Similarly, there were 4 (8.4%) complications noted in the TP-LDG group, including pancreatitis and fluid collection (Table 5). All morbidities were well managed with medication or intervention, and no deaths were noted in either group over the course of the follow-up.

DISCUSSION

In its early years, laparoscopic gastrectomy was very difficult and was performed restrictively because of the limited visual field and restricted working space relative to that in open gastrectomy. Thus, additional ports were needed over recent laparoscopic gastrectomy techniques, and mini-laparotomy was applied for gastric transection and extracorporeal anastomosis. However, with accumulating experiences, innovations in equipment and refinements in techniques, mini-laparotomy is no longer needed and the number of ports has decreased. Furthermore, surgical outcomes and oncologic results in laparoscopic gastrectomy are comparable to those for open gastrectomy, stimulating unceasing efforts to reduce the number of incisions. Thus, the concept of RPLG was introduced and performed; however, the concept of RPLG has not been clearly established. Many surgeons have attempted to reduce the number of ports for cosmetic superiority without affecting the oncologic safety. Some operators use a two-port system with one port in the right side of the abdomen and one multi-access port in the umbilicus [6]. Other operators use a three-port system with one port in the umbilicus and two ports in the two flanks, or one multi-access port in the umbilicus and two ports in the right flank [7, 8]. Occasionally, single port incision surgery (SILS) has been used in the concept of RPLG [4, 9].

Several studies regarding the surgical outcomes for RPLG have reported attractive results, such as better cosmesis, shorter operative time, less estimated blood loss, and less pain [9–11]. However, it is difficult to perform RPLG using the equipment and techniques of conventional laparoscopic gastrectomy. Especially during SILS, depth perception is lost because the endoscope lines up with the instruments of the operator and the assistant; collision and conflict between the instruments and endoscope are well generated [12]. Furthermore, some institutions are still equipped with straight and inflexible instruments and endoscope. However, flexible, bent, or thinner instruments and endoscope are effective in facilitating dissection, ligation, and intracorporeal suturing while guaranteeing safety and reliability [13]. In following the trends of RPLG, we have performed several cases of two-port laparoscopic distal gastrectomy and found it stressful to perform the surgical procedure with rigid instruments and endoscope. As a result, we have performed mainly TP-LDG for the treatment of EGC, as a precursor to two-port laparoscopic distal gastrectomy and SILS. This surgical technique uses one multi-access port at the umbilicus and other two ports for the operator's maneuvers in the right side of the abdomen, which improves the freedom. In this study, the outcomes for TP-LDG were comparable to those of FP-LDG, suggesting the safety and reliability of this surgical approach.

Oncologic outcome is a criterion of success in cancer surgery. Both the distance from the proximal resection margin and the number of harvested LNs correspond with the oncologic outcome during the short postoperative period of gastric cancer surgery [5, 14]. In this study, these two variables were not different significantly between the TP-LDG and FP-LDG groups. In surgical techniques using less than two ports, relatively many instruments are inserted through an umbilical multi-access port at the same time. As a result, lymph node dissection and gastric transection during two-port laparoscopic distal gastrectomy or SILS are more difficult than those during FP-LDG and the operator is required to have more extraordinary skills. However, the role of the operator in the TP-LDG was not greater compared to that in FP-LDG because the location of the operator's ports in TP-LDG were independent, as in FP-LDG. In addition, the interference between the assistant's instruments and endoscope in the umbilical multi-access port was not generated well because the heading directions of these devices were ordinarily parallel during TP-LDG. Therefore, lymph node dissection and gastric transection in TP-LDG was conducted with the same level of difficulty as in FP-LDG, which supported the safeness of TP-LDG.

The wound length as well as the number of incisions affects postoperative cosmetic result. During specimen extraction in the FP-LDG, the operator drew a plastic bag containing the specimen through the extended umbilical trocar site. At this moment, the bottlenecks occurred frequently, and the operator often needed to extend the umbilical trocar site additionally. Meanwhile, during TP-LDG, the operator did not have to pull the specimen in a mass, because the applied wound retractor had self-retaining and self-extending effect. Therefore, the umbilical wound length was shorter in the TP-LDG group compared to that in FP-LDG group.

In this study, the operative time was statistically shorter in TP-LDG group compared to that in FP-LDG group, similar to the result of previous studies comparing various types of RPLG and FP-LDG [15, 16]. The operative time is influenced by many factors, such as operator's proficiency, BMI, previous abdominal surgery history, anatomical characteristics, bleeding risk, tumor characteristics, extent of LNs dissection and reconstruction method. In most cases during study period, FP-LDG was performed earlier than RPLG, and the operator could overcome the learning curve for FP-LDG before performing mainly TP-LDG. We believe that advancement of the operator's skill over time and no difference in location of the operator's ports between TP-LDG and FP-LDG account for the difference in operative time. Therefore, we could conclude that TP-LDG was acceptable operative method in the aspect of the operative time, even the factors that overcoming the learning curve influenced the results and the difference in the operative time was not clinically much. In contrast, some studies have reported that the operative time in RPLG was similar to or longer than that in FP-LDG [11, 17]. However, the operative time in both approaches could be shortened by applying a simple method for determining the safe proximal resection margin rather than an intraoperative portable C-arm [18].

The amount of intraoperative blood loss is considered as one of the parameters of intraoperative morbidity [19]. In this study, the estimated blood loss in the two groups was statistically different, but clinically not much different. Most of the existing studies comparing RPLG with FP-LDG have shown a similar result regarding the amount of estimated blood loss, prompting the conclusion that RPLG is as safe as FP-LDG [15, 17, 20, 21]. As with the cause of difference in the operative time between the two groups, difference in the estimated blood loss might be understood.

The time to first flatus, time to the introduction of soft diet, postoperative pain, and length of the postoperative hospital stay are criteria typically applied in the evaluation of the postoperative course. Some studies have reported that pain after RPLG is similar to that after FP-LDG [13,15]. These studies concluded that pain might be dependent on various factors, such as the number of ports, length of incision, and individual characteristics. In contrast, other studies have reported differences between approaches, with a decreased number of incisions associated with reduced pain [22]. In this study, pain scoring was compared between the two groups using VAS of postoperative pain. The VAS on postoperative day 1 was not as high as expected, and VAS decrement over time was accordingly small in both groups. Therefore, no significant differences in VAS were noted at any of the assessment points. For a more detailed clarification of whether pain is less or not after TP-LDG than after FP-LDG, it will be necessary to conduct a well-designed randomized controlled study with a standard measurement method of postoperative pain. In addition, this study was limited in determining the true length of the postoperative hospital stay. In Korea, the discharge timing was often adjusted between the surgeon's recommendation and the patient's individual circumstances, such as their expectations, distance between the hospital and home, reimbursement system, high subscription rate of personal health insurance, and other cultural issues. Generally, the discharge date was often delayed.

The inflammatory responses, including WBC and CRP levels, were evaluated to compare the invasiveness after surgery between the two types of laparoscopic gastrectomy. The normalization of WBC and CRP levels certainly comes earlier after laparoscopic gastrectomy compared to that after open gastrectomy, because of the definite difference in the invasiveness between these two operative methods [23–26]. However, there was no definite evidence suggesting TP-LDG was superior to FP-LDG in terms of the inflammatory responses in this study, similar to that reported in other studies on RPLG [11,21]. Therefore, the degree of invasiveness between TP-LDG and FP-LDG does not appear to differ.

The operative safety is a long-standing concern for surgeons performing TLDG. First, the postoperative complications are a highly important issue, especially when introducing new technology. Previous studies have reported varied morbidity rates in FP-LDG and RPLG [3, 6, 9, 27, 28]. In this study, there was no specific trend in the complication rates between the two methods. During the operation using less than two ports, the operator and assistant have to share the multi-access port, and the collision and conflict between instruments occur frequently. However, as mentioned previously, the operator's maneuver during TP-LDG was not different from that during FP-LDG. Therefore, the chance for the postoperative complications in TP-LDG was not increased relative to that in FP-LDG, and the safeness of TP-LDG could be guaranteed. Second, the comfort during laparoscopic maneuver is also related with the operative safety. During FP-LDG, both operator and assistant had independent two ports each, so they could perform laparoscopic maneuver freely. However, during TP-LDG, the assistant could use just umbilical multi-access port. Therefore, sometimes the assistant faced the ergonomic problems associated with stand position, because the distance between the assistant and umbilical port during TP-LDG was further than the distance between the assistant and left-sided two ports during FP-LDG. This problem could be relieved by applying the packing gauze in the appropriate area to make the tissue retraction, and the assistant could save the effort to overcome the ergonomic problems.

This study has several limitations. First, this study was conducted retrospectively and was not based on a random assignment of operative technique. However, we performed a propensity score matching analysis to reduce the biases. Second, the surgeries were conducted by one surgeon, which might undermine the generalizability of the current results.

In conclusion, TP-LDG can be performed feasibly and safely without adverse influence on the oncologic outcomes for the treatment of early gastric cancer, despite the decreased number of ports compared to FP-LDG. It also has the benefit of better cosmesis in terms of the decreased number of incisions and length of umbilical incision site. Finally, well-designed prospective randomized trials would allow a more robust evaluation of the benefits of TP-LDG.

Declaration of Interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the article.

ETHICS APPROVAL AND CONSENT TO PARTICIPATE

This study was approved by the Institutional Review Board of the Ethics Committee (Inje University Haeundae Paik Hospital IRB), the reference number is HPIRB 2017-03-003-001.

ACKNOWLEDGMENTS

The authors appreciate the help provided by Professor Kyungha Seok in the Department of Statistics at Inje University in conducting the statistical analyses.