We read with interest an article titled, "Short-Term Outcomes of Laparoscopic Total Gastrectomy Performed by a Single Surgeon Experienced in Open Gastrectomy: Review of Initial Experience" in your journal, published in September 2015 (doi: http://dx.doi.org/10.5230/jgc.2015.15.3.159), by Song et al.1 First, we congratulate the authors for the excellent clinicale srults obtained in the initial series of laparoscopic total gastrectomy (LTG). The short-term results included only 13.5% of patients with complications, which comprised no grade ≥3 complications according to the Clavien-Dindo classification or anastomosis leaks, and a small amount of bleeding (80.8±60.4 ml). On the basis of the moving average analysis of the operative time, the authors also reported that the learning curve of LTG is about 50 cases.

However, there is the possibility of misinterpreting that such excellent results can be achieved with 50 cases of LTG by any surgeon, if the surgeon's experience and clinical situation in Korea are not considered.

As described in the discussion, the surgeon had already performed more than 100 cases of open total gastrectomy (OTG) and 15 cases of laparoscopic distal gastrectomy (LDG) before performing LTG. This number of cases required to attain competence is too high for western countries and even for Japan.2 Although surgeons who are experienced with open gastrectomy need to understand laparoscopic surgical anatomy, visual-motor coordination, and changes in haptic sensation, understanding the surgical anatomy for many types of open gastrectomy procedures can be beneficial for facilitating the learning of laparoscopic gastrectomy, which was designed to simulate open procedures in many respects. The learning process involves repeating the same procedures; however, learning the surgical procedures involves more complex factors associated with various individual anatomic variations and unexpected surgical situations. Direct experience with various anatomies and clinical situations in open surgery and indirect experience from being involved in the Korean surgical society, where discussions and education about laparoscopic gastrectomy are provided, are not negligible factors in the learning curve, although the effect of indirect experience on the learning process in surgery has not been established.345 Therefore, the proposed learning curve is not the learning curve of the principle procedures of total gastrectomy; instead, it is a "learning curve of laparoscopic skills" by which the surgeon performs well-known, essential steps of total gastrectomy.

Furthermore, the same surgeon performed LDG in four more cases than LTG during the learning period of LTG. Thus, the surgeon already had experience with more than 50 cases of laparoscopic gastrectomy when he/she performed 8 cases of LTG (15 cases of LDG before LTG+8 of LTG+32 of LDG). As LTG and LDG have similar procedures, experience with LDG affects the surgeon's learning curve with LTG. This may be one of the reasons why the number of cases for overcoming the learning curve was smaller than that in a previous report that suggested 100 cases of LTG.6 This number of cases is also similar or less than that reported for the learning curve of LDG.78 Therefore, we would like to emphasize the authors' conclusion that more cases may be needed to achieve excellent results if a surgeon starts performing LTG without sufficient experience with OTG or LDG.

In addition, it is possible to misinterpret the plateau of the learning curves as achieving the surgical standard. A learning curve itself does not consider whether the surgical procedure meets acceptable quality standards. In Fig. 1, the learning curve of C reached the plateau faster than that of curve B. However, this does not guarantee that the quality of surgery C meets the surgical standard. Although curve E has a very long learning process (paradoxically called a steep learning curve), it reached the surgical standard, unlike curves C and D. In terms of the operative time, the learning curve of the surgeon in Song et al.'s study1 can be interpreted as curve A. In Fig. 3A of Song et al.'s article,1 the operation time was consistently less than 150 minutes after only a few initial cases. The gradual slope of the curve in Fig. 3B seems to be caused by the moving average values affected by the long operation time in the first couple of cases. If Fig. 3B was based on a smaller number of cases instead of the moving average of 15 cases, the plateau could have been reached with fewer than 54 cases.

A cumulative sum method can be an alternative tool for evaluating the learning process, which is based on whether the surgeon's procedure meets the surgical standard.9 One research study suggested 23 and 35 cases for learning lymph node dissection in LTG; the failure of lymph node dissection was defined as ≤25 retrieved lymph nodes, and the target success rate was set to 92.5%.10 Another research study assessed the occurrence of complications with a target success rate of 20%, and the authors suggested 45 cases as the plateau of complications for LTG.11 These different results indicate that there can be different learning curves for each outcome parameter, and the learning curve of the surgery cannot be evaluated by a single parameter such as the operative time.

Song et al.1 did not provide learning curves for parameters other than the operative time. However, the excellent average values of the 74 cases in the series, including the number of retrieved lymph nodes (39.9±11.8), blood loss (80.8±60.4 ml), an d rate of complications (13.5% without grade ≥3 complications), are comparable or even superior to those reported in previous studies conducted by experienced laparoscopic gastric surgeons of an earlier generation.12 Therefore, it was estimated that the learned skill set of LTG surpasses the surgical standard before 54 cases of LTG are performed. This implies that the learning process of laparoscopic surgery has been facilitated for surgeons of a new generation owing to the development and standardization of surgical procedures and the accumulated expertise of surgeons of an earlier generation.

In conclusion, we think that the excellent results of the initial series of LTG should be interpreted as a result from exceptional baseline experiences with OTG and LDG of a surgeon at an extremely high-volume center in Korea, and 54 as the number of cases required for a stabilized learning curve cannot be used by most surgeons worldwide. However, surgeons can achieve acceptable skill sets and outcomes of LTG with highly concentrated training and experience even before they reach the plateau of their personal learning curve.