Laparoscopic Completion total Gastrectomy as an Alternative Procedure for Gastric Stump Cancer: A Case Control Study

Background Complete gastrectomy for gastric stump cancer can be challenging due to severe adhesions; therefore, advanced techniques are required to perform laparoscopic surgery. This study aimed to evaluate the clinical outcomes of laparoscopic completion total gastrectomy for the treatment of gastric stump cancer. Methods gastrectomy,


Abstract Background
Complete gastrectomy for gastric stump cancer can be challenging due to severe adhesions; therefore, advanced techniques are required to perform laparoscopic surgery. This study aimed to evaluate the clinical outcomes of laparoscopic completion total gastrectomy for the treatment of gastric stump cancer.

Methods
Patient records from January 2010 to October 2018 were retrospectively evaluated. The patients were divided into two groups depending on whether they underwent open or laparoscopic gastrectomy. We

Conclusions
Laparoscopic completion total gastrectomy was safely conducted without complications and mortality implicating the oncological validity for the treatment of gastric stump cancer. With sophistication of laparoscopic skills and advanced technologies, laparoscopic completion total gastrectomy may be the best way to perform less invasive surgery in terms of decreased blood loss and earlier recovery of intestinal peristalsis.

Background
The prognosis of gastric cancer after gastrectomy has improved, however, the incidence of cancer in the remnant stomach is increasing [1,2]. Newly developed gastric cancer after partial gastrectomy for benign disease or gastric cancer is de ned as remnant gastric cancer or gastric stump cancer (GSC), which is found in 1.1-6% of patients [2][3][4][5][6]. Complete resection of the carcinoma combined with a radical lymph node dissection is the only way to secure curability and improve the prognosis in patients who have no other complications [7]. Mesenteric lymph node metastasis around the gastrojejunostomy may worsen the prognosis of GSC [5]. A reported 5-year disease-speci c survival rate for GSC was 7-20% due to the advanced tumor stage [2]. However, recent reports have stated that the overall 5-year survival for GSC has improved to approximately 53-56%, which remains at a lower level than that of proximal gastrectomy for primary gastric cancer (PGC) [4,5,[8][9][10]. Therefore, the diagnosis of GSC at an early stage in patients who have undergone gastrectomy is important to reduce complications. However, complete gastrectomy for GSC has been di cult and invasive due to the severe adhesions that can occur fromthe previous procedures.
Laparoscopic gastrectomy has been con rmed to be safe with improved postoperative pain and earlier recovery than open gastrectomy [11][12][13]. However, only a few studies have reported the feasibility of laparoscopic completion total gastrectomy (LCTG) for GSC [14][15][16][17][18][19][20][21]. This study aimed to evaluate the feasibility, safety, and clinical outcomes of LCTG compared with those of open completion total gastrectomy (OCTG) to prove the oncological validity of LCTG.

Study design and patient characteristics
We performed a database search and identi ed 40 patients who had undergone surgery for GSC at Ishikawa Prefectural Central Hospital in Japan from January 2010 to December 2018. One patient who underwent bypass surgery, one patient who underwent staging laparoscopy, and one patient who underwent robot-assisted completion total gastrectomy were excluded from this study. The remaining 37 patients underwent gastrectomy for GSC and were further divided into two groups according to the initial approach of the operation: the LCTG group (n = 17) and OCTG group (n = 20). The mean age of patients in the LCTG and OCTG groups was 71.9 ± 8.1 and 68.9 ± 9.1 years, respectively. The male-to-female ratios in the LCTG and OCTG groups were 12 to 5 and 16 to 4, respectively. The medical records of all patients were retrospectively evaluated to compare the short-term surgical and long-term oncological outcomes.
The patient characteristics are shown in Table 1.

Surgical Procedures
A ve-port surgical approach was used. The rst 12-mm trocar was inserted at the umbilical area using the Hassan method. The other trocars were subsequently inserted carefully under laparoscopic viewing as in conventional laparoscopy [22]. A 10-mm, 30-degree oblique viewing laparoscope was used, and the CO 2 pressure was maintained at 10-12 mmHg. The Harmonic Scalpel (Ethicon EndoSurgery Inc., Cincinnati, OH), which is an ultrasonic-activated device, and the LigaSure (Medtronic, Minneapolis, MN) were used for adhesiolysis and radical lymphadenectomy. The extent of lymph node dissection was based on the recommendations for total gastrectomy from the Japanese gastric cancer treatment guidelines [23]. In cases of benign disease at the initial surgery, the intact gastric vessels were dissected from the root, and the resected specimen was extracted through the umbilical incision, which was enlarged to the minimum size required for extraction.
Patients who underwent Billroth I reconstruction during the initial surgery The gastrosplenic ligament was resected, starting with the opening of the omental bursa, and adhesions between the posterior gastric wall and pancreas were carefully divided. The dissection was carefully performed because the inferior surface of the left lateral segment of the liver was often severely adhesive to the remnant gastric wall. Moreover, the pneumoperitoneum yielded bloodless dissection. After encirclement of the gastro-duodenal anastomosis, the duodenum was transected using a linear stapler. Next, the remaining dissection around the remnant stomach including the vessels was completed, and the lower esophagus was transected in the same manner.
Patients who underwent Billroth II or Roux-en Y reconstruction during the initial surgery The afferent and efferent loops of the jejunum or Roux limb were resected, securing su cient distance from the anastomosis. The mesenteric lymph nodes were dissected depending on the tumor size or invasion.
Reconstruction after the complete removal of the GSC was performed with the Roux-en-Y method. The jejunum was transected 25 cm from the ligament of Treitz. Approximately 20 cm of the jejunum on the anal side was sacri ced, and the Roux limb was prepared. Jejunojejunostomy was performed with the Y limb. The Roux limb was ascended through the antecolic route, and esophagojejunostomy was performed using the overlap procedure. The mesenteric gap at the Y limb and Petersen's mesenteric defect were closed by continuous suturing using barbed string.

Statistical analysis
Patient ages are presented as the mean ± standard deviation, and all other values are expressed as the median with range. All statistical analyses were completed using R statistical software, version 3.5.0 (R Foundation for Statistical Computing, Vienna, Austria). Chi-squared, Fisher's exact, and Mann-Whitney U tests were performed for comparisons between the two groups. The cumulative 5-year survival rates were calculated according to the Kaplan-Meier method, and survival curves were compared using the log-rank test. The statistical signi cance level was set at p < 0.05.

Results
No signi cant differences between the two groups in sex distribution, body mass index, and comorbidity incidence were observed. A total of 5 and 12 initial gastrectomies were performed in the LCTG group for benign and malignant diseases, respectively; a total of 4 and 16 initial gastrectomies were performed in the OCTG group for benign and malignant diseases, respectively. However, the difference observed between the groups was not statistically signi cant. The median time from the initial gastrectomy to the development of GSC was comparable between the LCTG and OCTG groups (11 vs 15.5 years, respectively; p = 0.385). The most common tumor location in the LCTG and OCTG groups was the nonanastomotic site (12 [70.6%] and 14 [70%] cases, respectively). However, only three cases in each group had the tumor detected at the anastomotic site (17.6% and 15% in the LCTG and OCTG groups, respectively).
Two (11.8%) and seven (35%) patients in the LCTG and OCTG groups, respectively, were diagnosed with tumor depth invasion greater than clinical T stage 3. However, no signi cant difference in the distribution of the clinical T stage between the groups was observed. No patients in the LCTG group had preoperative lymph node metastasis. Therefore, patients in the LCTG group, except for those with clinical T stage 3, were considered to be at clinical stage I. The clinical stages in the OCTG group were diversely distributed; however, no signi cant difference in the stage distribution between the groups was observed. In the initial gastrectomy, several patients in both groups were treated with laparoscopy. The type and reconstruction of the initial gastrectomies were diverse; however, no signi cant difference in the distributions for both groups was observed. The most common reconstruction method in the initial gastrectomy was Billroth I anastomosis in seven (41.2%) and eight (40%) patients in the LCTG and OCTG groups, respectively.
Operative And Postoperative Short-term Outcomes The surgical outcomes are depicted in

Pathological Outcomes
The pathological ndings of the resected specimens are shown in Table 3. The median tumor size in the LCTG group was signi cantly smaller than that in the OCTG group (26 vs. 40 mm; p = 0.0457). No signi cant difference between groups in the median number of retrieved lymph nodes (11 vs. 9.5; p = 0.437), depth of the tumor invasion (p = 0.12), extent of lymphatic metastasis (p = 0.0509), distant metastasis ratio (p = 1), and variation of histological type (p = 1) were observed. However, the median number of metastatic lymph nodes was signi cantly lower in the LCTG group than that in the OCTG group (0 vs. 0.5; p = 0.0108). Additionally, the pathological stage distribution in the LCTG group was lower than that in the OCTG group (p = 0.0346). Postoperative Long-term Outcomes The postoperative long-term outcomes are shown in Table 4. The median follow-up duration was 41 and 31 months for the LCTG and OCTG groups, respectively (p = 0.427). In the LCTG group, one patient died from recurrence, and one patient died from pneumonia. In the OCTG group, six patients died from recurrence, and three patients died from other diseases. Speci cally, one patient in the LCTG group (5.88%) and 11 patients in the OCTG group (55%) developed recurrence, representing a signi cant difference (p = 0.004679). Recurrence in the patient in the LCTG group was due to metastasis to the mediastinal lymph nodes. In the OCTG group, seven cases of metastasis in the liver, three in the peritoneum, two in the pleura or regional lymph nodes, and one in the lung were reported. The 5-year overall survival rate was signi cantly higher in the LCTG group than that in the OCTG group (84.4% vs. 48.5%; p = 0.0373) (Fig. 1). The 5-year disease-free survival of the LCTG group was signi cantly higher than that of the OCTG group (93.3% vs. 41.9%; p = 0.00274) (Fig. 2).  (1) Pneumonia (1) Original (6) Sepsis (1) Pneumonia ( (7) Peritoneum (3) Pleura (2) LYM (2) Lung (1) LCTG laparoscopic completion total gastrectomy, OCTG open completion total gastrectomy, MST mean survival time, DFS disease free survival, LYM lymph node Discussion Our results con rm the feasibility of LCTG for the treatment of GSC. Patients in the LCTG group had signi cantly longer operation times but signi cantly less blood loss and earlier atus passage than the OCTG group. Furthermore, no conversion to open surgery and no higher morbidity than Clavien Dindo class III were reported in the LCTG group.
Additionally, the number of retrieved lymph nodes was equal to that reported in other studies (Table 6) [14][15][16][17][18][19][20][21]. The OCTG group had a higher number of recurrences and deaths than the LCTG group, because the OCTG group had more advanced cases, leading to a worse 5-year overall survival rate than the LCTG group. Recently, the number of lower stages of GSC has been increasing due to the strict postoperative surveillance for PGC; this surveillance combined with the feasibility and validity of LCTG can improve patient survival [7].
The most di cult part of the operative procedure for GSC is the adhesiolysis, which is the key factor to safely performing LCTG [18]. A precise and sharp dissection between the adjacent organ and remnant stomach is necessary to avoid organ injury, and less bowel manipulation leads to early recovery [15,18].
We consider laparoscopy an effective solution to overcome this di culty in the treatment of GSC. The advantages of laparoscopic surgery are pneumoperitoneum, which widens the dissectible layer between the adhered organs, and a magni ed view that enables detection of the loose and dissectible layer.
Moreover, progressive high de nition (HD) imaging signi cantly contributes to the bene ts from such magni ed views. We have been using the HD scope system (Karl Storz SE & Co. KG, Tuttlingen, DE) since the introduction of LCTG in our institution. Advanced energy devices and forceps also contribute to re ning the quality of surgery, reducing bleeding, reducing the trauma to organs, and re ning the precision of lymphadenectomy. Our sophisticated dissection techniques combined with these advanced developments enable us to perform LCTG with an extremely reduced blood loss compared to previous case-controlled studies (Table 6), leading to earlier recovery of digestive peristalsis.
Robotic gastrectomy could be a future advancement for the treatment of GSC in terms of its visual improvement in the surgical eld, which is referred to as robotically enhanced surgical anatomy [34]. The re ned anatomical view of robotic gastrectomy could achieve precise movement of forceps without hand tremors, which could increase operative accuracy. In fact, robotic gastrectomy for PGC has decreased the complication rate despite longer operative time and higher cost than laparoscopic gastrectomy [35,36]. Robotic surgery has already been applied to GSC and reviewed retrospectively, which has shown a lower conversion rate and comparative short-term outcomes to LCTG [37]. We have also introduced robotic surgery to GSC and expect superior results.
We acknowledge some limitations in our study. First, this study had a retrospective design, which could have led to potential selection biases. Therefore, a randomized, controlled study should be completed.
Second, due to the low incidence of GSC, the sample size was too small to elucidate the universal results and superiority of LCTG over OCTG for GSC. A multicenter study is necessary to validate our results.

Conclusion
LCTG was safely conducted without complications and mortality, implicating the oncological validity of LCTG for the treatment of GSC. With sophistication of laparoscopic skills and advanced technologies, LCTG is less invasive, results in reduced blood loss, and leads to earlier recovery of intestinal peristalsis.  Figure 1 Overall survival rate Kaplan-Meier estimates of overall survival probability. The straight and dotted lines indicate the laparoscopic (LCTG) and open completion total gastrectomy (OCTG) groups, respectively.

Abbreviations
There was a signi cant difference between the two groups in the log-rank test (p = 0.0373)