Since the first reported laparoscopic gastrectomy (LG) performed by Kitano et al1 in 1994, laparoscopic surgery has been increasingly used for treatment of gastric cancer. However, standardization of this minimally invasive approach has not been reached yet because of its technical difficulties and the concern about oncological safety.2 Therefore, its learning curve is known to be longer than that for other laparoscopic procedures.3
In order to analyze the performance improvement of learning curve in LG, several variables such as operation time, absence of microscopic residual tumor, appropriate extent of lymph node (LN) dissection, conversion rate, major complication, and 30-day mortality have been proposed.4
Nowadays, many authors have perfectly analyzed the learning curve of LG, focusing not only on laparoscopic distal gastrectomy,3,5,6 but also on laparoscopic total gastrectomy (TG).7
A learning curve is considered completed when the monitored parameters reach a steady state, and the final results can be compared with published parameters.8 In case of LG, it is estimated that an experience of about 60 to 90 cases is required to achieve a complete training in this operation.5,9–11 However, the decreased incidence of gastric cancer in Western countries compared with Eastern countries might be another cause of variability in the learning curve, probably extending the number of cases required to master this procedure in our countries.
The aim of this study was to evaluate the technical feasibility and short-term surgical outcomes of LG during our learning curve and compare the results obtained between a group of senior gastric surgeons with experience in open surgery and a young surgeon with skills in advanced laparoscopic surgery but not much experience in gastric cancer surgery.
MATERIALS AND METHODS
All first consecutive 100 cases of LG performed at our Department of Surgery in Ramón y Cajal University Hospital (Madrid, Spain) from November 2008 to February 2018 were enrolled. Patients undergoing laparoscopic exploration or laparoscopic gastrojejunostomy for peritoneal carcinomatosis were excluded. Patients requiring conversion to open surgery or those with a history of prior gastrectomy were included.
Learning Curve for LG
In order to define LG learning curve, patients were divided into 2 groups on the basis of the period during which they were operated. The training phase (TP) was considered between 2008 and 2014 (46 cases) and the more developed phase (MDP) between 2015 and 2018 (54 cases) according to the creation of the Division of Esophagogastric and Division Unit.
We also formed 2 groups on the basis of expertise in esophagogastric surgery. Group SS was formed by a group of senior gastric surgeons who were experts in open gastric surgery with previous experience in laparoscopic bariatric surgery (44 cases) and Group YS was formed by a young surgeon with little experience in open gastric surgery, but with good skills in advanced laparoscopic surgery (56 cases). Five surgeons were implicated in the study [4 senior surgeons (SSs) and 1 young surgeon (YSs)].
Conversion, lymphadenectomy dissection and retrieved LNs, hospital length of stay, mean operative time, postoperative complications (graded by Clavien-Dindo classification12), reintervention rate, and 30-day mortality were compared between the 2 phases of the learning curve and between both groups of surgeons.
The surgery was performed with the patient under general anesthesia and placed in a modified lithotomy position. The surgeon stood between the patient’s legs with the assistants on the patient’s left and right side. Pneumoperitoneum was created with Veress needle in the left hypocondrium, and carbon dioxide was insufflated to maintain a 12 to 14 mm Hg intra-abdominal pressure. A 30-degree laparoscope was used. The procedure was performed with three 12-mm trocars, one 5-mm trocar, and the liver Nathanson retractor in the epigastrium, similar to standard laparoscopic antireflux surgery.
Surgical treatment varied from TG to subtotal gastrectomy (SG) with regard to the location and the nature of the tumor. Extent of LN dissection followed the Japanese Gastric Cancer Association guidelines.13 Reconstruction of the gastrointestinal tract was performed laparoscopically with a 45-mm linear stapler for intracorporeal end-to-side esophagojejunostomy (closing the hole with a running barbed suture), or via a minilaparotomy with a circular stapler end-to-side esophagojejunostomy (at the beginning of the experience for TG). The specimen was normally extracted through a Pfannestiel incision. One drain was left in SG and 2 drains in TG. The nasogastric tube was left in cases of TG depending on the surgeon’s criteria.
Postoperative care was different among SG or TG patients. In brief, the nasogastric tube and urinary catheter were removed on postoperative day (POD) 1. Postoperative mobilization was also encouraged from POD1. Recovery of oral intake was initiated on POD3 with water, followed by a dietary progression from liquid to soft food in cases of SG. However, a routine radiologic study using a water-soluble contrast-medium at POD7 was performed on all TG patients in order to check the anastomosis before starting oral diet. Until hydrosoluble contrast test was performed, patients were treated with parenteral nutrition.
Statistical analysis was performed with the statistical software SPSS 23.0 for Windows (IBM SPSS Inc., Chicago, IL). Quantitative variables that followed a normal distribution were defined by the mean and SD. For non-Gaussian variables, the median and range were used. Qualitative variables were defined by number and percentage of cases. Comparison of variables was performed with a Student’s t test (Mann-Whitney test in non-Gaussian variables). Comparison of qualitative variables was performed with the χ2 test; in cases with <5 observations per cell, the Fisher exact probability method was used. P<0.05 was considered statistically significant. Short-term and midterm global and disease-free survival data were determined by Kaplan-Meier analysis.
Among the 100 LG cases, there were 46 male (46%) and 54 female individuals (54%) with a median age of 74 years (range, 66 to 80 y). Eighteen patients (18%) received neoadjuvant treatment with chemotherapy. Table 1 shows clinicopathologic characteristics of the patients.
Forty-five TGs (45%), 56 SGs (56%), and 1 proximal (1%) gastrectomy were performed with a mean operative time of 250 minutes (range, 220 to 305 min). Conversion to open surgery was needed in 18 cases (18%). Causes of conversion are described in Table 2.
Mean hospital length of stay was 12 days (range, 9 to 21 d). Surgical postoperative complications occurred in 29 patients (29%). The most frequent adverse event was the esophagojejunal leak that was described in 8 cases (17.7%). However, complete healing was achieved in 100% of the cases with conservative management (rate of fatal esophagojejunal leaks was 0%). Stump duodenal leak occurred in 9 cases (9%), and reoperation was necessary in 6 cases (66.6%). Five cases of intra-abdominal abscess (5%) were described, requiring a new laparotomy in only 1 case (20%). Finally, haemoperitoneum developed in 3 cases (3%), requiring reintervention in 100% of cases. With respect to Clavien-Dindo classification, the rate of surgical complications requiring endoscopic, radiologic, or surgical therapy (grade III or higher) was 13% (5 cases of grade I, 15 patients of grade II, 4 cases of grade IIIA, 5 patients of grade IIIB, 4 cases of grade IVA, and 4 patients with grade V).
Indeed, reintervention rate was 8% (stump duodenal leak, 3 cases; stump duodenal leak with haemoperitoneum, 3 cases; intra-abdominal abscess, 1 case; and, finally, a patient with a septic shock, multiorganic failure, and evisceration).
Readmission and mortality rates were 6% and 4%, respectively (one patient with stump duodenal leak, 2 cases of simultaneous stump duodenal leak and haemoperitoneum, and the patient with multiorganic failure).
Learning Curve Between TP and MDP
When comparing TP results against MDP, it was observed that there was a statistically higher number of D2 LNs in the second group (P=0.0001). Indeed, the number of retrieved LNs was higher in MDP (17±8.6 vs. 23.3±10.4; P=0.004). Furthermore, it was also observed that there was a statistically higher number of patients with neoadjuvant treatment in the MDP group (1 case vs. 17 cases; P=0.0001). Moreover, there was a significant drop in overall complications (478% vs. 27.8%; P=0.038), reintervention rate (15.2% vs. 1.85%; P=0.023), and overall mortality (8.7% vs. 0%; P=0.003) in MDP. In contrast, there were no significant differences in conversion rate, mean operative time, and hospital length of stay between phases (Table 2).
Learning Curve Between SSs and YS
When comparing results between SS and YS, although patients in the YS group were statistically younger (74.45 vs. 69.84 y; P=0.04), it was observed that there was a statistically higher D1+/D2 LN in the second group (63.6% vs. 87.5%; P=0.0001). Indeed, the number of retrieved LNs was higher in the YS group (18.90±8 vs. 23.52±11; P=0.042). Although the overall complication rate was not statistically reduced, we found less complications in the YS group (40.9% vs. 33.9%). Furthermore, Clavien-Dindo III to V complications were statistically higher in SS (20.4% vs. 14.3%; P=0.028). There was no statistically significant difference between the rest of the parameters; however, we observed a decreased reintervention rate (11.36% vs. 5.35%) and mortality (6.8% vs. 1.8%) in the YS group (Table 3).
In the following figures (Figs. 1–8), we have compared our results obtained with regard to the conversion rate, lymphadenectomy dissection and retrieved LNs, hospital length of stay, mean operative time, postoperative complications (graded by Clavien-Dindo classification), reintervention rate, readmission, and 30-day mortality with the steady state published in the literature. The continuous line expresses the average of our results, and the discontinuous line refers to the steady state published in the literature.
Overall and disease-free survival at 5 years was 40% and 59%, respectively, with mean overall survival being 65.54 months (range, 55 08 to 76.01 mo) and mean disease-free survival being 65.97 months (range, 55.49 to 76.44 mo) (Figs. 9, 10). We have to notice that, in our manuscript, we have mixed malignant patients with gastric adenocarcinoma with other different neoplasms (neuroendocrine tumors, gastrointestinal stromal tumors) or even benign pathologies.
In spite of the great expansion of minimally invasive approach for treatment of several pathologies, the application for gastric cancer treatment has been very slowly introduced, because of its technical difficulties and the concern about oncological safety.2 Recently, a meta-analysis14 has shown that the safety and efficacy of the laparoscopy-assisted distal gastrectomy (LADG) in patients with early gastric cancer, and another multicenter study has reported comparable long-term oncological outcomes of LG and open gastrectomy.10 However, the learning curve of LG is known to be longer than that for other laparoscopic procedures.3,15
Adequate lymphadenectomy for oncological safety was one of our major concerns throughout, from the beginning of the learning curve. There should be no compromise in the principles of oncological surgery for completing a gastrectomy laparoscopically. However, the number of LNs retrieved by laparoscopy varies from West to East publications.16 Previous studies have shown that the number of retrieved LNs differed from 27 to 35,16–19 and the number of LNs retrieved by laparoscopy was significantly lower than that of open surgery.20,21 In our series, comparing TP and MDP, a statistically higher D2 LN in the second group was observed (3 vs. 30; P=0.0001). It is obvious that, with more experience, not only in open gastrectomy, but of course in laparoscopic surgery, the type of lymphadenectomy was more accurate according to the Japanese Gastric Cancer Association guidelines. Indeed, the mean retrieved number of LNs was 20 (range, 14 to 28), with a statistically significant increase from 17+8 to 23.73±10 in MDP. Despite this improvement, we are still not satisfied with the results. However, the role of the pathologist is crucial in this outcome, which we had not considered previously in this review. Actually, the number of retrieved LNs is known to be influenced by factors such as race, sex, age, tumor characteristics, surgeons’ expertise, and pathologist’s interest.22 Therefore, when we analyzed the number of retrieved LNs during 2015 and 2016 (cases 47 to 73 of the series) with permanent pathologists and surgeons, we observed a mean retrieved number of LNs of 28 (range, 21 to 35.5). Although, there is an agreement that at least 15 LNs should be dissected to provide an adequate and accurate postoperative N staging,13,23 our aim was to achieve at least 25 LNs in each LG.24 When we compared SS against the YS group, we observed a statistically higher number of D2 LNs (7 vs. 26; P=0.0001) and a statistically significant increased number of retrieved LNs from 18.9±8 to 23.52±1 (P=0.042) in favor of the YS group.
Other common parameters used to evaluate the learning curve is conversion to open surgery. The rate observed in our study is higher than what has been published in the literature, with a conversion rate of 18%. However, there were no significant differences in the conversion rate between TP and MDP (19.56% vs. 16.67%). Dividing the cases into 4 groups of 25 patients, we observe that the conversion rate was 12% in the first 25 cases, and slightly higher in the next 25 cases (28%), due to an increase in the indications for these approaches (local advanced gastric cancer). In the last 25 cases, the conversion rate was 4% (only one case due to an anesthesiologist problem), which is near to the results published in the literature.7,25 This parameter achieves a statistically significant difference if we compare the YS against the SS group (10.7% vs. 27.3%; P=0.039). Furthermore, if we analyzed the causes of conversion, only 4 cases (4%) were due to intraoperative complications, more of them being due to a wrong indication for the laparoscopic approach.
Another important parameter of the learning curve is surgical postoperative complications. To be honest, our rate of postoperative complications is a little bit higher (29%) compared with other series.16,17,19,20 However, when we used the Clavien-Dindo classification, it is true that only 13% are grades III to V. However, we have noticed a statistically significant decrease in overall complication rates (47.8% vs. 27.8%; P=0.038) from TP to MDP. Furthermore, we have observed a very high rate of esophagojejunal leak (17.7%). As we have described in the Materials and Methods section, we systematically performed the routine radiologic study using a water-soluble contrast-medium at POD7 on all TG patients in order to check the anastomosis before starting oral diet. Therefore, we sometimes obtain radiologic images of leaks in asymptomatic patients; hence, this is probably the reason for finding a higher rate of esophagojejunal leak. A complete healing in 100% of the cases was achieved with conservative management (parenteral nutrition, antibiotics, and drainage), finding no case of fatal esophagojejunal leaks’ (0%) subsidiary of reintervention. Esophagojejunal anastomosis is one of the most challenging steps of laparoscopic TG. At the beginning, we performed the anastomoses with a circular stapler through a minilaparotomy. However, this step was not so easy, and, in most cases, an extended incision was required, losing the benefits of the laparoscopic approach. We did not consider the total laparoscopic circular stapler with OrVil, because it has been associated with a higher rate of stenosis and esophagojejunal leaks in some other studies.26 For this reason, we normally perform it laparoscopically with a 45-mm linear stapler for intracorporeal end-to-side esophagojejunostomy (closing the hole with a running barbed suture).
In contrast, we are concerned with duodenal stump leak, which is around 9%, requiring reintervention in 6 cases, and 3 died. It is true that in 2 patients, duodenal stump leak occurred in those converted to open surgery, but, in order to reduce this life-threatening complication and encouraged by our excellent results in sleeve gastrectomy, we started using peristrip dry (permanent bovine pericardial strips) to reinforce the (gold) cartridge of the stapler. We have described a reduced number of stump duodenal leaks in 6 cases from the first 46 cases (13%) to 3 cases (patient 66, 67, and 94) in the last 54 cases (5.5%). Moreover, in the last cases, only one patient needed reoperation, and mortality associated with this complication was 0%. Comparison between YS and SS, did not reveal a statistically significant drop of overall surgical complication rates (33.9% vs. 40.9%), but it revealed a statistically significant decrease in grades III to V complication rates (14.3% vs. 20.4%; P=0.028).
Finally, we have observed a decrease in reintervention rate (15.2% vs. 1.85%; P=0.023) and overall mortality (8.7% vs. 0%; P=0.003) in MDP, which is next to the recommendations of the Japanese Gastric Cancer Association.
A learning curve is considered completed when the monitored parameters reach a steady state, and the final results can be compared with published parameters.8 The number of procedures that have to be performed by a surgeon to master LG has been debated.
In case of LG, it is estimated that an experience of about 60 to 90 cases is required to achieve a complete training in this operation.3,5,9–11 However, these variations may also arise from differences in the annual volume of the surgeon (East vs. West), patient characteristics, the surgeon’s prior experience and skill with laparoscopic procedures, the assistant surgeon’s experience, and innate ability, probably extending the number of cases required to master this procedure in our western countries.27 Kim et al,28 presented their results in a high-volume gastric center in Korea, and they pointed that the number of procedures in which the learning curve plateau was reached was 30 LADG (7 mo). However, these results should be interpreted with caution, because it was achieved by a few experienced surgeons in a high-volume center in Korea, where a large number of gastric cancer patients are treated by a small number of experts. In contrast, in the study by Moon et al,10 a stable learning curve plateau was not achieved by either of the 2 surgeons after 90 procedures, which was probably owing to low annual LADG volume for both surgeons. They considered that with an average of <2 LADGs per month by each surgeon, it was difficult to master the learning curve in a short period of time. In our study, it is difficult to determine the number of cases required to complete the learning curve, because we have performed 100 cases of surgery, and the surgeon with more experience has operated 56 cases in 10 years. However, with regard to the outcomes of the surgical postoperative parameters, we believe that, in our western hospital, an average of 60-100 cases per surgeon should be required to master LG.
Nowadays, one aspect that can reduce the learning curve is the possibility of recording the procedure and reviewing the operation afterwards, to aid self-examination of their surgical technique. Moreover, young surgeons have higher laparoscopic skills compared with senior surgeons because they have trained in a period in which laparoscopic surgery is very popular and had previously performed a lot of laparoscopic operations like appendectomies, cholecystectomies, antireflux surgery, colectomies, bariatric surgery etc., thus feeling more familiar with laparoscopic instruments compared with surgeons in the past.29
Finally, we have to recognize that there were also several limitations in our study. First of all, its retrospective design and the low number of LGs performed during these 10 years. There was also a selection criteria bias of patients undergoing laparoscopic approach depending on the surgeon’s criteria, and also the LG technique was not completely standardized in all cases because of the lessons learned with the experience acquired in the first cases. Another limitation of this study was the comparative analysis between SS and YS, because the first group had a huge previous experience not only in open gastrectomy but also in bariatric laparoscopic surgery in comparision with YS group.
To conclude and based on our first 100 cases of LG at our hospital, we have observed an improvement on surgical parameters and postoperative course during these 2 phases of learning curve. However, if we consider our results, we believe our learning curve is not yet completed, because the average of monitored parameters has not reached a steady state yet. Anyway, monitoring graphs have showed that our results in the last 2 years were near to the best results published in the literature. It is difficult to identify the exact number of cases required to complete the learning curve in LG, but we considered that, with our outcomes, it seems to be between 60and100 cases individually.
1. Kitano S, Iso Y, Moriyama M, et al. Laparoscopyassisted Billroth I gastrectomy. Surg Laparosc Endosc. 1994;4:146–148.
2. Jeong O, Jung MR, Kim GY, et al. Comparison of short-term surgical outcomes
between laparoscopic and open total gastrectomy for gastric carcinoma: case control study using propensity score matching method. J Am Coll Surg. 2013;216:184–191.
3. Kim MC, Jung GJ, Kim HH. Learning curve
of laparoscopy-assisted distal gastrectomy with systemic lymphadenectomy for early gastric cancer
. World J Gastroenterol. 2005;11:7508–7511.
4. Noshiro H, Nagai E, Shimizu S, et al. Laparoscopically assisted distal gastrectomy with standard radical lymph node dissection for gastric cancer
. Surg Endosc. 2005;19:1592–1596.
5. Jin SH, Kim DY, Kim H, et al. Multidimensional learning curve
in laparoscopy-assisted gastrectomy for early gastric cancer
. Surg Endosc. 2007;21:28–33.
6. Zhang X, Tanigawa N. Learning curve
of laparoscopic surgery for gastric cancer
, a laparoscopic distal gastrectomy-based analysis. Surg Endosc. 2009;23:1259–1264.
7. Jung DH, Son SH, Park YS, et al. The learning curve
associated with laparoscopic total gastrectomy. Gastric Cancer
8. Buchmann P, Dincler S. Learning curve
-calculation and value in laparoscopic surgery. Ther Umsch. 2005;62:69–75.
9. Kunisaki C, Makino H, Yamamoto N, et al. Learning curve
for laparoscopy-assisted distal gastrectomy with regional lymph node dissection for early gastric cancer
. Surg Laparosc Endosc Percutan Tech. 2008;18:236–241.
10. Moon JS, Park MS, Kim JH, et al. Lessons learned from a comparative analysis of surgical outcomes
of and learning curves for laparoscopy-assisted distal gastrectomy. J Gastric Cancer
11. Kang SY, Lee SY, Kim CY, et al. Comparison of learning curves and clinical outcomes
between laparoscopy-assisted distal gastrectomy and open distal gastrectomy. J Gastric Cancer
12. Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg. 2004;240:205–213.
13. Japanese Gastric Cancer
Association. Japanese gastric cancer
treatment guidelines 2010 (version 3). Gastric Cancer
14. Zeng YK, Yang ZL, Peng JS, et al. Laparoscopy-assisted versus open distal gastrectomy for early gastric cancer
: evidence from randomized and nonrandomized clinical trials. Ann Surg. 2012;256:39–52.
15. Kim HH, Han SU, Kim MC, et al. Long-term results of laparoscopic gastrectomy
for gastric cancer
: a large-scale case-control and case-matched Korean multicenter study. J Clin Oncol. 2014;32:627–633.
16. Zhao LY, Zhang WH, Sun Y, et al. Learning curve
for gastric cancer
patients with laparoscopy-assisted distal gastrectomy. 6-year experience from a single institution in western China. Medicine. 2016;95:37.
17. Lee JH, Nam BH, Ryu KW, et al. Comparison of outcomes
after laparoscopy-assisted and open total gastrectomy for early gastric cancer
. Br J Surg. 2015;102:1500–1505.
18. Moisan F, Norero E, Slako M, et al. Completely laparoscopic versus open gastrectomy for early and advanced gastric cancer
: a matched cohort study. Surg Endosc. 2012;26:661–672.
19. Park DJ, Han SU, Hyung WJ, et al. Long-term outcomes
after laparoscopy-assisted gastrectomy for advanced gastric cancer
: a large scale multicenter retrospective study. Surg Endosc. 2012;26:1548–1553.
20. Gong JK, Cao YK, Wang YH, et al. Learning curve
for hand-assisted laparoscopic D2 radical gastrectomy. World J Gastroenterol. 2015;21:1606–1613.
21. Vinuela EF, Gonen M, Brennan MF, et al. Laparoscopic versus open distal gastrectomy for gastric cancer
: a meta-analysis of randomized controlled trials and high-quality nonrandomized studies. Ann Surg. 2012;255:446–456.
22. Luna A, Rebasa P, Montmany S, et al. Learning curve
for d2 lymphadenectomy in gastric cancer
. ISRN Surg. 2013;2013:508719.
23. Ajani JA, Bentrem DJ, Besh S, et al. Gastric cancer
, version 2.2013: featured updates to the NCCN Guidelines. J Natl Compr Canc Netw. 2013;11:531–546.
24. Martinez-Ramos D, Calero A, Escrig-Sos J, et al. Prognosis for gastric carcinomas with an insufficient number of examined negative lymph nodes. Eur J Surg Oncol. 2014;40:358–365.
25. Chi KC, Park JM. Laparoscopic gastrectomy
performed by an expert in open gastrectomy. J Gastric Cancer
26. Inokuchi N, Otsuki S, Fujimori Y, et al. Systematic review of anastomotic complications of esophagojejunostomy after laparoscopic total gastrectomy. World J Gastroenterol. 2015;21:9656–9665.
27. Song JH, Choi YY, An JY, et al. Short-term outcomes
of laparoscopic total gastrectomy performed by a single surgeon experienced in open gastrectomy: review of initial experience. J Gastric Cancer
28. Kim JH, Jung YS, Kim BS, et al. Learning curve
of a laparoscopy assisted distal gastrectomy for a surgeon expert in performing a conventional open gastrectomy. J Korean Gastric Cancer
29. Choi YY, Song HS, An JY. Reply: factors favorable to reducing the learning curve
of laparoscopic gastrectomy
for gastric cancer
. J Gastric Cancer
Keywords:Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.
learning curve; outcomes; laparoscopic gastrectomy; gastric cancer