| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Article |
1 Center for Gastric Cancer, National Cancer Center, 809 Madul-dong, Ilsandong-gu, Goyang-si, Gyeonggi-do, 411–769, South Korea
2 Cancer Registration Branch, Research Institute for National Cancer Control and Evaluation, National Cancer Center, 809 Madul-dong, Ilsandong-gu, Goyang-si, Gyeonggi-do, 411–769, South Korea
3 Cancer Biostatistics Branch, Research Institute for National Cancer Control and Evaluation, National Cancer Center, 809 Madul-dong, Ilsandong-gu, Goyang-si, Gyeonggi-do, 411–769, South Korea
Correspondence: Address correspondence and reprint requests to: Keun Won Ryu, MD, PhD; E-mail: docryu{at}korea.com
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Methods: The authors prospectively reviewed the data of all patients who underwent total gastrectomy with D2 lymph node dissection during a 4-year period. Retrieved lymph node number was used as a surrogate marker of oncological outcome. The retrieved lymph node number cut-off value required for satisfactory D2 lymph node dissection was defined as >25. Cumulative sum analysis was used to examine the learning curves of individual surgeons at target accuracy rates of 85%, 90%, 92.5%, 95%, and 98%.
Results: Two junior staff surgeons performed 198 curative-intent total gastrectomies with D2 lymph node dissections during the study period; their success rates exceeded 90%. Operating time decreased with operative experience (Pearson correlation coefficient = –0.515, P < 0.001). The learning period for total gastrectomy with D2 lymph node dissection for these two junior members of staff was calculated as 23–35 cases, presuming a 92.5% success rate.
Conclusions: The current study suggests that the surgical learning period for D2 lymph node dissection extends to at least 23 cases or 8 months. In clinical trials containing gastric cancer surgery, the learning curve for qualified surgery from the standpoint of oncological outcome should be considered to minimize bias due to surgeon-associated factors.
Key Words: Gastric cancer • Total gastrectomy • D2 lymph node dissection • Learning curve • Cumulative sum analysis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Although there has been some debate on the issue, D2 lymph node dissection is regarded as essential for the treatment of gastric cancer.3–5 It not only increases the radicality of surgical therapy but also provides adequate lymph node staging. Two European randomized controlled trials that compared D1 and D2 gastrectomy revealed a high operative mortality exceeding 10% in the D2 group.6,7 Based on such reports, D2 gastrectomy is regarded as a procedure that should be restricted to specialized centers where adequate training and supervision can be provided.
Clinical trials concerning the adequacy of lymph node dissection extent in gastric cancer patients, including the above two European randomized controlled trials and a recent chemoradiation trial in gastric cancer patients, have been criticized because of lack of surgical quality control.6–8 In a Dutch trial, the proportion of noncompliance (patients who did not complete D2 lymph node dissection) was 51%.6
Quality control is the most essential factor in both surgical and clinical trials for patients who undergo surgery.9 A lack of surgical standardization or quality control can make it difficult to correctly interpret clinical trial results. However, no report has been issued concerning the minimal oncologic surgical experience required for surgical quality control in clinical trials.
Thus, the aim of the present study was to evaluate the learning curve for qualified D2 lymph node dissection in a specialized center in patients with gastric cancer.
|
PATIENTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
All 198 study subjects underwent D2 lymph node dissection, and lymph nodes along the upper border of the pancreas to the posterior gastric artery were dissected in the spleen-preserved group.10 Informed consent was obtained from all patients. Standard reconstruction was performed using Roux-en-Y esophagojejunostomy. The number of retrieved lymph nodes exceeded 15 in all 198 patients except one. Tumors were staged according to the International Union Against Cancer (UICC) staging system.11
All specimens were sent for pathological examination immediately after resection. A single pathologist (M. C. K.) retrieved all the lymph nodes using palpation under gross inspection. There was no size limitation of lymph nodes for harvest. All pathologic examinations were performed in a standard manner.
Surgeon Training
Both of the junior staff surgeons involved had completed a 2-year fellowship course on gastric cancer surgery in university hospitals. During this period, each surgeon initially performed gastric cancer surgeries as an assistant for more than 200 cases annually. They joined the National Cancer Center in 2001 (surgeon A) and 2002 (surgeon B) and, for institutional standardization purposes, each performed gastric cancer surgeries as a first assistant for more than 3 months.
Cumulative Sum (CUSUM) Analysis
CUSUM (Sn) of a series of observations is defined as Sn = 
Xi –Xo, where Xi = 0 for success, Xi = 1 for failure, and Xo = a predetermined acceptable failure rate such that if the target success rate is set at 95%, then the acceptable failure rate is 5%. CUSUMs are determined from consecutive observations. For example, if surgeon X performs five consecutive observations and the first, second, and fifth are successes but the third and fourth are failures, CUSUM values become –0.05, –0.1, 0.85, 1.80, and 1.75. These values are then plotted against the number of observations made to produce the curve. A positive slope indicates that the acceptable failure rate has been exceeded. Conversely, when the curve assumes a negative slope, the failure rate is less than the set acceptable failure rate.
For the purposes of CUSUM analysis, each observation is recorded as a success or failure. Success was defined as the retrieval of >25 lymph nodes and failure, as the retrieval of 
25. CUSUM functions were calculated for targeted success rates of 85%, 90%, 92.5%, 95%, and 98% for the institution and for individual surgeons. These functions were plotted against operation number to produce CUSUM graphs.
A typical learning curve produced by CUSUM analysis demonstrates an initial upward slope, corresponding to a period when the incidence of failure exceeds the acceptable failure rate. This upward slope is followed by either a plateau (indicating that the observed incidence of failure is equal to the acceptable failure rate) or a downward slope (indicating that the observed failure rate is lower than the acceptable failure rate).
Statistical Analysis
The point at which the curve either plateaus or acquires a consistent negative slope indicates the number of attempts required to learn to perform the process under study at the required standard.
All statistical analyses were carried out using SPSS 9.0 for Windows (SPSS, Chicago, IL). Clinicopathologic variables were analyzed using the two-tailed chi-squared test and Student’s t-test. P <0.05 was considered statistically significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
compares patient characteristics for surgeons A and B. The average number of retrieved lymph nodes was 40.5 (range 14–98). Surgeon A’s patients were younger than surgeon B’s patients (P = 0.032), and the depth of tumor invasion encountered by surgeon B was greater than that encountered by surgeon A (P = 0.019).
|
. Figure 2a and b shows the learning curves for D2 lymph node dissection for surgeons A and B, respectively. A visual inspection of the CUSUM plot in Figure 1 shows that failure was more frequent (i.e., slopes were more positive) at the beginning of the series and that these improved after the forty-ninth consecutive case for a 90% success rate. Following an initial training period of 50 cases, the success rate remained stable up to case 145 and then showed an improvement.
|
|
Postoperative Complications
Twenty-five of the 198 patients (12.6%) experienced postoperative complications (Table 2). The most common complications were atelectasis (5 cases, 2.5%), following by postoperative bleeding (4 cases, 2.0%) and wound infection (4 cases, 2.0%). Postoperative complication rates were similar for the first 99 cases and the second 99 cases (13.1% vs. 11.1%, P = 0.873).
|
Analysis of Cases of Failure
No difference was found between "failed" and "successful" operations in terms of patient or tumor factors. Detailed results are listed in Table 3. The median number of retrieved lymph nodes in failed cases was 21 (range 14–25), whereas in successful cases this was 39 (range 26–98).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The key step when evaluating the learning curve of any surgical procedure is the selection of an appropriate outcome measure that demonstrably measures the ability of a surgeon to perform the particular task.12 A previous publication used mortality and morbidity rates as outcome measures and suggested that the learning curve in gastrectomy with D2 lymph node dissection ranges 15–25 cases in a series of 38 patients.13 This previous study focused primarily on the feasibility of D2 lymph node dissection, used a restricted sample size, and presented limited information on oncologic outcome. However, in the case of cancer surgery, oncologic outcome should be considered so that clinical trial results are correctly interpreted.
We selected number of retrieved lymph nodes as a surrogate marker of oncologic quality for D2 lymph node dissection based on the results of a previous anatomic study.14 Wagner et al.14 reported that the mean number of retrieved lymph nodes after D2 lymph node dissection was 26 based on cadaver dissection. The number of retrieved lymph nodes reflects the performance of an institution, surgeons, and pathologists; thus, the pathologist learning curve should be considered together with that of surgeons.9 Our national cancer center was founded in 2001 and employs four gastric surgeons, two medical oncologists, three gastroenterologists, one pathologist, one radiologist, and one anesthesiologist. Thus, the single pathologist at our center examines all pathologic specimens from gastric cancer patients, and we were able to justifiably overlook pathologist influence during the present study.
The present study shows that fewer than the expected number of retrieved lymph nodes (<26) were retrieved from 8.6% of patients. In a recent multi-institutional randomized controlled trial of D1 vs. D2 gastric cancer surgery, the noncompliance rate (defined as percentage of patients who did not complete D2 lymph node dissection) was 51% despite quality control; participating surgeons received a videotape and booklet about the technique and were instructed in the operating room by an expert gastric cancer surgeon over the first 4 months of the intake period.6
Another prospective study of D1 vs. D2 gastric cancer surgery in Germany revealed that participating surgeons performed quality-assured surgeries.15 However, 38% of patients received less than D2 surgery. The reasons for the discrepancies between our success rate and those of previous studies are probably attributable to the incidence of gastric cancer in Korea and a nationwide policy for gastric cancer treatment. D2 lymph node dissection is considered a standard and safe procedure in Korea, where 20,000 cases of gastric cancer are diagnosed annually.1 General surgeons are taught the technique early during their surgical training, and it should be noted that the two junior staff surgeons of the present study had completed a 2-year subspecialty training program in gastric cancer before study commencement, which has been reported to influence gastrectomy outcome.16
In the present study, the learning period for D2 lymph node dissection was determined to be 23–35 cases, presuming a 92.5% success rate, and it took the two junior surgeons 8 and 13 months, respectively, to gain this level of experience. Analysis of the curves reveals that the success rate plateaued between 90% and 95%. Achievement of a success rate of 90% at the outset indicates that a 92.5% success rate was a reasonable choice in the present study. Performance between surgeons differs widely.17,18 Thus, although our estimate of the time involved may appear high even in a high-volume center like ours, this study suggests that our results should be used as a guide only for surgical oncology. Despite such considerations, it appears that substantial amounts of time and experience are required before required levels of surgical quality are achieved.
The lymph node metastasis rate of failed cases tended to be lower than that of successful cases. This may have been because surgeons understaged tumors during operations and reduced the extent of lymph node dissections intentionally. The effect of stage migration could provide another possible explanation as retrieved lymph node numbers were smaller in failed than in successful cases and nodal stages could be downstaged.19
All patients included in this study underwent total gastrectomy with D2 lymph node dissection. Moreover, because dissection extent depends on primary tumor location, it can differ among subtotal gastrectomy patients with D2 lymph node dissection. To reduce lymph node dissection extent variations caused by primary tumor locations, we included only patients who had undergone total gastrectomy with D2 lymph node dissection.
In the present study, complication rates were not reduced by experience despite a significant reduction in operating times, and this result was maintained on examining the results of the two surgeons independently. One possible explanation for this apparent paradox is that the two junior staff surgeons were not unfamiliar with D2 lymph node dissection because of their early training as house staff and fellows. Although the overall complication rate was higher for surgeon B, he treated a greater proportion of elderly patients and patients with a T4 lesion.
Retrieved lymph node number is probably not an absolute surrogate marker for D2 lymph node dissection. Excluding surgeon- and pathologist-associated factors, a number of other factors are known to influence retrieved node numbers, e.g., body mass index (BMI).20 However, we found no relationship between retrieved node numbers and BMI or any other tumor or patient factors in this series. Considering that a large difference exists between the BMIs of Eastern and Western patients, Western gastric surgeons may require more learning cases. Therefore, adjuvant chemoradiation for Western patients with advanced gastric cancer may be a reasonable treatment option.8 Recurrence and survival are being followed in our study cohort to further determine the effects of the learning period.
The present study suggests that the surgical learning period for D2 lymph node dissection extends to at least 23 cases or 8 months. Possible strategies to improve the quality of D2 lymph node dissection include subspecialty training courses in gastric cancer surgery, close intraoperative supervision by expert practitioners, and assistance from other well-trained staff. Moreover, in clinical trials involving cancer surgery, the learning period required to achieve qualified surgery from the viewpoint of oncological outcome should be considered to minimize bias due to surgeon-associated factors.
Received for publication January 26, 2006. Accepted for publication April 5, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONPATIENTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  K. W. Ryu, Y.-W. Kim, J. H. Lee, B.-H. Nam, M.-C. Kook, I. J. Choi, and J.-M. Bae Surgical Complications and the Risk Factors of Laparoscopy-Assisted Distal Gastrectomy in Early Gastric Cancer Ann. Surg. Oncol., June 1, 2008; 15(6): 1625 - 1631. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. H. Lee, Y.-W. Kim, K. W. Ryu, J. R. Lee, C. G. Kim, I. J. Choi, M. C. Kook, B.-H. Nam, and J.-M. Bae A Phase-II Clinical Trial of Laparoscopy-Assisted Distal Gastrectomy with D2 Lymph Node Dissection for Gastric Cancer Patients Ann. Surg. Oncol., November 1, 2007; 14(11): 3148 - 3153. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
