Learning curve of intracorporeal anastomosis in laparoscopic colectomy for right side colon cancer: a cumulative sum analysis

Introduction

It has been over 30 years since the first laparoscopic colectomy [1] was reported, and it is now a procedure performed globally. Laparoscopic colectomy includes two methods of anastomosis: intracorporeal and extracorporeal. Initially, surgeons conducted extracorporeal anastomosis, where bowel anastomosis was conducted after specimen extraction from a minilaparotomy so that the anastomotic technique of open surgery could be applied [2]. In recent years, intracorporeal anastomosis has gained popularity due to evolution of the laparoscopic linear stapler device and improved techniques from laparoscopic surgeons. Advantages of intracorporeal anastomosis include the following: extensive detachment of the mesentery of the colon is not required, openings for specimen extraction can be set anywhere in the abdomen, and risk of mesenteric twisting is low [2,3]. Furthermore, intracorporeal anastomosis is very useful in obese patients, whose thick abdominal wall and relatively short mesentery make it difficult to pull the bowel outside the abdomen [4]. It has also been reported in two randomized clinical trials that intracorporeal anastomosis was superior than extracorporeal anastomosis due to the shorter wound length, quicker recovery of bowel function, and less post-operative analgesia in the former [5,6] and it is thought to further evolve in the future due to its potential as a minimally invasive procedure.

However, there are technical difficulties associated with intracorporeal anastomosis and anastomotic leakage has been reported, albeit infrequently [4,5]. A side-to-side iso-peristaltic anastomosis (overlap anastomosis) is often conducted for intracorporeal anastomosis, and the entry hole of the bowel with the stapler is closed by running sutures of the monofilament [5,6,7]. It has also been reported that, for entry hole closure methods, the double layer closure has a significantly lower incidence of anastomotic leakage and shorter post-operative hospital stay than the single layer closure [7]. Improving the technique of laparoscopic sewing, which is a major part of intracorporeal anastomosis, is important for laparoscopic colorectal surgeons. Therefore, it is important to examine the learning curve for intracorporeal anastomosis from the viewpoint of both education and safety management as the success of the procedure is susceptible to differences in the skill of the laparoscopic surgeon and takes time to achieve proficiency.

In the industry for quality management, the cumulative sum (CUSUM) method has been used in order to evaluate the learning curve of procedures that require skill acquisition [8]. More recently, it has also been used in surgical procedures [9,10]. To the best of our knowledge, there are no studies evaluating the learning curve of intracorporeal anastomosis in laparoscopic colectomy. This study aimed to clarify the number of cases that are required for laparoscopic surgeons to master the technique of intracorporeal anastomosis, and to that end, we conducted a cumulative sum analysis on learning curves of intracorporeal anastomosis in laparoscopic colectomy for right side colon cancer.

Materials And Methods

Patients and study design

Between July 2018 and March 2020, consecutive 62 patients underwent laparoscopic colectomy with intracorporeal overlap anastomosis at Tokyo Medical University Hospital for primary cancers that existed in the right side colon. In this retrospective single-center study, 51 consecutive patients were selected as subjects who underwent intracorporeal overlap anastomosis, after excluding six patients who underwent other intracorporeal anastomotic procedures, three patients whose procedures were conducted by another surgeon, and two patients for whom video of the surgery was not available (Fig. 1). No extracorporeal anastomosis was performed during this period. All surgeries were conducted by one laparoscopic surgeon who was a “qualified surgeon” according to the endoscopic surgical skill qualification system specified by the Japan Society for Endoscopic Surgery [11]. This surgeon was trained 3 times in intracorporeal anastomosis at dry labo and wet labo. One laparoscopic surgeon performed the first intracorporeal anastomosis by the surgical technique shown below since July 2018 and completed in March 2020. The reason for selecting cases from this period was that the surgeon was performing intracorporeal anastomosis for the first time and was immobilized with a fixed assistant surgeon to perform the surgical technique. Clinicopathological and perioperative data were obtained from our database. Colon cancer was classified according to the American Joint Committee on Cancer classification guidelines [12]. Anastomotic leakage was defined as CT guided drainage of the intra-abdominal abscess and contrast medium contrasting the intestinal tract. Those that did not were defined as surgical site infection (SSI). This study was approved by the Tokyo Medical University Ethics Committee (approval no. T20-0054). Written notification and oral informed consent were obtained from all patients prior to the entry of their data in the database.

Surgical technique

Mechanical and oral antibiotic bowel preparation was conducted the day before surgery. Under general anesthesia, a total of five ports were placed in the abdomen in a square position, after which laparoscopic colectomy was initiated. An appropriate lymph node dissection was conducted with the medial approach. The bowel on the oral side and anal side of the specimen was cut inside the abdominal cavity using a linear stapler. Intracorporeal overlap anastomosis was conducted as per the following procedure (Fig. 2).

a. A small incision was made as an entry hole for inserting a linear staple near the staple line of the oral side bowel

b. Similarly, an entry hole was created at 6 cm from the staple line of the anal side bowel

c. The 60 mm stapler jaw was inserted into both entry holes in coordination with my assistant surgeon

d. Intracorporeal overlap anastomosis was performed using a linear stapler

e. Double layer closure of the entry hole was conducted by laparoscopic running sewing that used a 3 − 0 PDS* II (polydioxanone)

The specimen was extracted typically through a Pfannenstiel incision. The wound for specimen extraction was closed using absorbable interrupted muscular sutures and fascial stitch (1–0 VICRYL Plus). For the measurement of surgical times, the time required for processes a-e was defined as the intracorporeal anastomosis time (IAT), that for a-d as stapling time (ST), and that for e as laparoscopic sewing time (LST); these were measured using the video of the surgery.

Statistical Analysis

The learning curves of IAT, ST, and LST were created using the CUSUM method [13], which is a type of time-weighted control chart method. The score of the first case was set as the difference between the IAT, ST, and LST required for the first case and average time for all cases. The score of the second case was set as the difference between the IAT, ST, and LST required for the first case and average time for all cases, to which the score of the first case is added. scores for the third case onwards were calculated using the following equation. CUSUM_N=CUSUM_N−1 + (Time_N – Time_mean), where N is the number of cases [7, 14]. The point at which the amount of change in CUSUM_N was maximum was set as the phase change. All statistical analyses were performed using software JMP Pro version 16.0.0 (SAS Institute, Inc., Cary, NC). Continuous variables are presented as median (range). Comparisons between the learning curves of IAT of phase 1 and phase 2 groups for clinicopathological, operative factors of all continuous parameters and postoperative complications of Clavien-Dindo Classification ≥ II were conducted using Student’s t test. The chi-square test was used for comparing categorical data of small samples. A two-tailed p value of < 0.05 was considered statistically significant.

Results

The Table 1 shows the consecutive 51 patient characteristics as follow; Age, 69 (47–87) year; gender, male 29 (56.9%); body mass index 22.9 (15.0-31.9) Kg/m²; pretreatment carcinoembryonic antigen 3.1 (1.2–97.1) ng/ml, tumor location, cecum 13 (25.5%), ascending colon 23 (45.1%), transverse colon 8 (15.7%), descending colon 5 (9.8%), sigmoid colon 2 (3.9%); comorbidities any, 18 (37.3%). Figure 3 shows the learning curve for intracorporeal anastomosis using the CUSUM of the IAT. The CUSUM score increased as the number of operative cases progressed, up to the 20th case (Phase 1), after which the CUSUM score started to decrease from the 21st case (Phase 2). Figure 4a shows the CUSUM of the ST. The CUSUM score increased until the 9th case (Phase 1), plateaued until the 17th case (Phase 2), and then started to decrease from the 18th case (Phase 3). Figure 4b shows the CUSUM of the LST. The CUSUM score increased until the 22nd case (Phase 1), after which it began to decrease from the 23rd case (Phase 2).

Table 1

Patient characteristics
	n = 51
Age, years	69 (47–87)
Gender, n (%)
Male	29 (56.9)
Female	22 (43.1)
BMI (Kg/m²)	22.9 (15.0-31.9)
Pretreatment CEA, ng/ml	3.1 (1.2–97.1)
Tumor location, n (%)
Cecum	16 (31.4)
Ascending colon	25 (49.0)
Transverse colon	10 (19.6)
Comorbidities any, n (%)	18 (37.3)
Cardiovascular	8 (15.7)
Diabetes	5 (9.8)
Respiratory	3 (5.9)
Renal	2 (3.9)
BMI body mass index, CEA carcinoembryonic antigen

The perioperative outcomes were compared between Phase 1 and Phase 2 for the CUSUM of the IAT, the IAT of Phase 2 was significantly faster than that of Phase 1 (p < 0.001) and the postoperative hospital stay of Phase 2 was significantly shorter than that of Phase 1 (p = 0.021). No significant differences were observed in terms of operative time, blood loss, number of lymph nodes harvested, site of skin incision, length of skin incision, pathological stage and time to flatus. For postoperative complications, the organ / space SSI of Phase 2 was significantly lesser compared to that of Phase 1 (p = 0.009). No significant differences were observed in terms of anastomotic leakage, anastomotic stenosis, superficial incisional/deep incisional SSI, ileus, deep vein thrombosis, pneumonia and mortality within 30 days (Table 2).

Table 2 Perioperative outcomes and postoperative complications


	Phase 1 (n = 20)	Phase 2 (n = 31)	p value
Operative time, min	176 (99-313)	164 (108-268)	0.459
Intracorporeal anastomotic time, min	23.5 (14.2-35.5)	16.2 (11.5-23.9)	< 0.001*
Blood loss, ml	21 (1-45)	16 (5-78)	0.398
Number of lymph nodes harvested, n (%)	20 (7-45)	19 (5-51)	0.842
Length of skin incision, mm	42 (31-87)	49 (30-79)	0.539
Pathorogical stage, n (%)			0.888
I	7 (35.0)	9 (29.0)
II	8 (40.0)	11 (35.5)
III	4 (20.0)	9 (29.0)
IV	1 (5.0)	2 (6.5)
Time to flatus, day	2 (1-3)	2 (1-3)	0.767
Postoperative hospital stay, day	8 (6-33)	7 (6-17)	0.021*
Anastomotic leakage, n (%)	0	0	-
Anastomotic stenosis, n (%)	1 (3.2)	0	0.417
Superficial incisional/deep incisional SSI, n(%)	1 (5.0)	0	0.209
Organ/space SSI, n (%)	4 (20.0)	0	0.009*
Ileus, n (%)	2 (10.0)	1 (3.2)	0.315
Deep vein thrombosis, n (%)	0	1 (3.2)	0.417
Pneumonia, n (%)	1 (5.0)	0	0.209
Mortality within 30 days	0	0	-
SSI surgical site infection

Discussion

The findings of this study using cumulative sum analysis denote that a laparoscopic surgeon required 20 cases to master the procedure of intracorporeal overlap anastomosis. The CUSUM score of IAT increased up to the 20th case in Phase 1 (initial learning). The surgical technique was immature in Phase 1, peaked at the 21st case, and the score gradually decreased during Phase 2 (mastery), and a mature state was reached [7]. According to other reports, the number of cases necessary for the learning curve of laparoscopic colectomy was 15–35 cases [15, 16, 17]. These studies stated the learning curves for the overall surgical procedure. Meanwhile, our study was unique in that it focused on intracorporeal anastomosis, which is a process that requires advanced laparoscopic skills and is prone to the smallest differences in technical skill. During our analysis, the CUSUM score of ST was divided between Phase 1 (initial learning), Phase 2 (consolidation), and Phase 3 (mastery). When guiding and inserting the stapler into the small entry hole that was created in the bowel under the motion restriction by the laparoscopic port, there is a need for coordination between the surgeon who controls the stapler and the assistant who holds the bowel. In this process, the learning curve comprises of multiple factors related to the surgeon and assistant; therefore, it is thought that the CUSUM score of ST plateaued (consolidation phase) before transitioning into the downward trend of Phase 3 (mastery). Reducing the learning curve of this process requires shortening of Phase 2 by increasing training for coordination between the surgeon and assistant through simulations such as wet lab training. Meanwhile, the CUSUM score of LST depends on the skill of a single surgeon and was separated into 2 phases, with a peak reached at 22 cases. In general, there are three surgical sewing techniques: hand sewing, laparoscopic sewing, and robotic sewing. Hand sewing, which is the oldest, is the basis of open surgery and familiar to most surgeons. Robotic sewing, which uses advanced technology, is useful for intracorporeal anastomosis due to joint function [18]. Although there is no doubt that robotic surgery will become widespread among minimally invasive surgeries, it will still take some time to be accepted and available worldwide due to its high cost [19], and laparoscopic surgery will account for a large proportion of minimally invasive surgeries until then. Laparoscopic sewing is difficult because it is needle controlled and sutured with straight forceps without a joint function, and it takes time to master this technique. Therefore, it is of great significance to analyze the learning curve of the laparoscopic sewing. The number of cases required to shift to a downward trend in the CUSUM score of LST is lower than that of ST. In other words, it may be suggested that improving the skill of laparoscopic sewing may lead to a shortened learning curve of intracorporeal anastomosis in laparoscopic colectomy.

We have previously reported that the incidence of organ/space SSI was high in a study where saline that underwent lavage during surgery around the anastomosis was sampled after intracorporeal anastomosis and an examination of bacterial culture was conducted [20]. In this study, the incidence of organ/space SSI was significantly lower in Phase 2 than in Phase 1 of IAT for postoperative complications. This is thought to be due to the shortened Phase 2 in IAT, which reduced the intestinal release time and exposure time of the intestinal bacteria to the intracorporeal space. Furthermore, intracorporeal anastomosis has been reported to increase post-operative inflammatory responses such as white blood cells and C-reactive protein [20, 21]. Prolonged post-operative inflammation increases the duration of antibiotic use, time taken for first diet intake, and recovery of intestinal function. In this study, all four cases with organ/space SSI were drained by computed tomography-guided aspiration. For these reasons, it is thought that the postoperative hospital stay was significantly extended in Phase 1 relative to that in Phase 2. It was suggested that shortening of the learning curve of intracorporeal anastomosis may reduce the time for intestinal release into the abdominal cavity. This may lead to reduced postoperative complications and promote postoperative recovery.

The learning curve of a surgical procedure is thought to be governed by organizational factors (facilities, equipment), surgical team-related factors (experience, cooperation), case complexity, and surgeon-related factors (previous experience, natural abilities, motivation) [22]. Additionally, the operation time is affected by the surgeon’s competency, disease severity, and patient characteristics. Therefore, conventional chronological analysis or logistic regression analysis for operation time is not suitable for learning curve analysis, unless only cases with the same disease and patient factors are included. Furthermore, these analyses, even with selective inclusion of perfectly homogeneous groups, cannot reflect the actual clinical situation [23]. Meanwhile, the CUSUM method is suitable for observing the slow learning curve process from quantitative skills to qualitative changes, and it is more practical and accurate due to the small sample size and lack of need for grouping [24, 25]. The CUSUM analysis applied in this study is a good statistical method for determining the likelihood of changes and adverse events [22]. However, CUSUM analysis is not well suited for comparing differences between surgeons, and therefore, only one surgeon was included in this study. When introducing a new surgical procedure, it is important to understand the number of cases required to achieve skill proficiency and safety from the perspective of both efficient surgical education and patient safety management. Anastomotic leakage, which is the most critical complication in this study, was not observed in Phase 1 (initial learning) or Phase 2 (mastery). It was suggested that a surgeon who is a “qualified surgeon”, according to the Endoscopic Surgical Skill Qualification System, can safely perform intracorporeal anastomosis and achieves expertise after working on 20 cases. To the best of our knowledge, there have been no reports of research on learning curves that are specific to intracorporeal anastomosis in laparoscopic colectomy, and it is thought that these results will serve as a reference when facilities that have been conducting extracorporeal anastomosis introduce intracorporeal anastomosis in the future.

This retrospective observational study has several limitations. The number of cases was small (51 cases), only one surgeon with extensive experience in laparoscopic procedures participated, and only the CUSUM method was used to evaluate the learning curve, with multiple other methods not adopted. In the future, a multicenter clinical trial that involves multiple surgeons with difference levels of surgical experience and uses multiple evaluation criteria is necessary in order to evaluate the learning curve of intracorporeal anastomosis.

In conclusion, in our study, 20 cases were required for a laparoscopic surgeon to achieve expertise when conducting intracorporeal anastomosis in laparoscopic colectomy for colon cancer. Compared to the initial learning phase, the master phase had a significantly faster IAT, significantly decreased incidence of organ/space SSI, and significantly decreased postoperative hospital stay.

References

Schlinkert RT (1991) Laparoscopic-assisted right hemicolectomy. Dis Colon Rectum 34:1030–1031
Tarta C, Bishawi M, Bergamaschi R (2013) Intracorporeal ileocolic anastomosis: a review. Tech Coloproctol 17:479–485
Stein SA, Bergamaschi R (2013) Extracorporeal versus intracorporeal ileocolic anastomosis. Tech Coloproctol 17:35–39
Reggio S, Sciuto A, Cuccurullo D et al (2015) Single-layer versus double-layer closure of the enterotomy in laparoscopic right hemicolectomy with intracorporeal anastomosis: a single-center study. Tech Coloproctol 19:745–750
Bollo J, Turrado V, Rabal A et al (2020) Randomized clinical trial of intracorporeal versus extracorporeal anastomosis in laparoscopic right colectomy (IEA trial). Br J Surg 107:364–372
Allaix ME, Degiuli M, Bonino MA et al (2019) Intracorporeal or Extracorporeal Ileocolic Anastomosis After Laparoscopic Right Colectomy: A Double-blinded Randomized Controlled Trial. Ann Surg 270:762–767
Sugishita T, Tsukamoto S, Imaizumi J et al (2022) Evaluation of the learning curve for robot-assisted rectal surgery using the cumulative sum method. Surg Endosc 36:5947–5955
Kestin IG (1995) A statistical approach to measuring the competence of anaesthetic trainees at practical procedures. Br J Anaesth 75:805–809
Jimenez-Rodriguez RM, Diaz-Pavon JM, de la Portilla de Juan F et al (2013) Learning curve for robotic-assisted laparoscopic rectal cancer surgery. Int J Colorectal Dis 28:815–821
Foo CC, Law WL (2016) The learning curve of robotic assisted low rectal resection of a novice rectal surgeon. World J Surg 40:456–462
Kazama K, Numata M, Aoyama T et al (2021) Does the Endoscopic Surgical Skill Qualification System improve patients' outcome following laparoscopic surgery for colon cancer? A multicentre, retrospective analysis with propensity score matching. World J Surg Oncol 53. doi: 10.1186/s12957-021-02155-z.
Amin MB, Edge SB (2017) AJCC cancer staging manual. New York, NY: Springer
Nasseri Y, Stettler I, Shen W et al (2021) Learning curve in robotic colorectal surgery. J Robot Surg 15:489–495.
Foo CC, Law WL (2016) The Learning Curve of Robotic-Assisted Low Rectal Resection of a Novice Rectal Surgeon. World J Surg 40:456–462.
Giani A, Veronesi V, Bertoglio CL et al (2022) Multidimensional evaluation of the learning curve for laparoscopic complete mesocolic excision for right colon cancer: a risk-adjusted cumulative summation analysis. Colorectal Dis 24:577–586.
Tsai KY, Kiu KT, Huang MT et al(2016) The learning curve for laparoscopic colectomy in colorectal cancer at a new regional hospital. Asian J Surg 39:34–40.
Akiyoshi T, Kuroyanagi H, Ueno M et al (2011) Learning curve for standardized laparoscopic surgery for colorectal cancer under supervision: a single-center experience. Surg Endosc 25:1409–1414.
Huscher CGS, Lazzarin G, Marchegiani F et al (2022) Robotic right colectomy with robotic-sewn anastomosis: a pilot case series. J Robot Surg 26.
Jayne D, Pigazzi A, Marshall H et al (2017) Effect of Robotic-Assisted vs Conventional Laparoscopic Surgery on Risk of Conversion to Open Laparotomy Among Patients Undergoing Resection for Rectal Cancer: The ROLARR Randomized Clinical Trial. JAMA 318:1569–1580.
Ishizaki T, Katsumata K, Enomoto M et al (2021) Comparison of Intra-Abdominal Infection Risk Between Intracorporeal and Extracorporeal Anastomosis in Laparoscopic Right Hemicolectomy for Colon Cancer: A Single-Center Retrospective Study. Am Surg 87:341–346.
Kayano H, Nomura E, Ueda Y et al (2019) Short- and Long-term Outcomes of 2-Step Stapled Intracorporeal Versus Extracorporeal Anastomosis in Laparoscopic Colectomy for Colon Cancer. Anticancer Res 39:6393–6401.
Haidenberg J, Kendrick ML, Meile T et al (2003) Totally extraperitoneal (TEP) approach for inguinal hernia: the favorable learning curve for trainees. Curr Surg 60:65–68.
Rhu J, Sung K, An CH, Cho J (2022) Learning curve analysis using the cumulative summation method for totally extraperitoneal repair of the inguinal hernia. Langenbecks Arch Surg 407:3101–3106.
Nasseri Y, Stettler I, Shen W et al (2021) Learning curve in robotic colorectal surgery. J Robot Surg 15:489–495.
Szymczak P, Grzybowska ME, Sawicki S et al (2021) Laparoscopic pectopexy-CUSUM learning curve and perioperative complications analysis. J Clin Med. 10:1052.