Transitional impact of short‐ and long‐term outcomes of a randomized controlled trial to evaluate laparoscopic versus open surgery for colorectal cancer from Japan Clinical Oncology Group Study JCOG0404

Abstract Background The JCOG0404 randomized controlled trial conducted to compare laparoscopic surgery (LAP) with open surgery (OP) for stage II/III colon cancer showed better short‐term outcomes and equal long‐term outcomes of LAP versus OP. Technical instrumentation of surgery and anticancer agents given during the registration period might have affected the outcomes. Aim To evaluate outcomes according to the registration periods. Methods The overall registration period was divided into three periods (first: 2004‐2005, second: 2006‐2007 and third: 2008‐2009). Short‐term and long‐term outcomes were compared between registration periods. Results In total, 1057 patients were registered. Numbers of patients undergoing each approach for each of the three periods (1st/2nd/3rd) were 528 for OP (106/244/178) and 529 for LAP (106/246/177). Operation time (minutes) did not change between the periods for OP (160/156/161) or LAP (205/211/219). Blood loss (mL) gradually decreased in the latter two periods: (119/80/75) for OP and (35/28/25) for LAP. Incidence of complications (%) decreased in the latter periods for OP (27.6/20.3/21.3), whereas that for LAP remained consistently low (14.3/14.8/13.6). There was no particular trend in 5‐year overall survival and recurrence‐free survival depending on the period regardless of treatment. D3 dissection rates were 95% or more for all periods in both groups. Conclusions Operation time and survival rates did not change over time, whereas blood loss in OP improved in the latter periods. Quality control applied in this trial might have been effective in producing such safe endpoints. (ClinicalTrials.gov, number NCT00147134, UMIN Clinical Trials Registry, number C000000105.)


| INTRODUC TI ON
At the beginning of this century, several randomized controlled trials for colorectal cancer from all over the world showed excellent short-term results and equal long-term results of laparoscopic surgery compared with open surgery. [1][2][3][4][5][6][7][8] Most of these trials showed that laparoscopic surgery for colon cancer has been accepted as a standard therapy. In Japan, the Japan Clinical Oncology Group (JCOG) conducted a randomized controlled trial to confirm the efficacy of laparoscopic surgery for stage II or III colon cancer (JCOG0404). [9][10][11] The result of JCOG0404 was interpreted to be that laparoscopic surgery could be an acceptable treatment option for stage II or III colon cancer.
In terms of the quality control of this trial, our clinical question was whether the duration of this study could affect the clinical out- Therefore, the purpose of the present study was to evaluate trial outcomes based on different periods within the overall registration period for JCOG0404. We analyzed the changes in short-term and long-term outcomes over time for each registration period and investigated whether quality of the surgical technique was maintained in this trial.

| Summary of JCOG0404
JCOG0404 was a multi-institutional trial to confirm the non-

| Measured outcomes
Adequateness of D3 dissection was evaluated by central review.
Cases with distant metastases or severe invasion to adjacent organs were excluded from the photographic analysis. Short-term outcomes of operative time, blood loss, length of postoperative hospital stay, and incidence of early complications were analyzed. Early complications were defined as occurrences within 30 days of surgery.
Terminology and grading of complications were described according to the Common Terminology Criteria for Adverse Events 3.0. 14 Incidence of early complications included all grades in this study.
Long-term outcomes of overall survival and relapse-free survival were compared between three periods within the overall registration period for each approach. Definitions of these outcomes are reported elsewhere. 11

| Statistical analysis
All patients were divided into one of three periods to assess the transitional impact of the trial. colorectal cancer, laparoscopic colectomy, randomized controlled trial, transitional impact overall survival and relapse-free survival, and the log-rank test was used to compare the three periods. A Cox proportional hazard model was used to estimate the hazard ratio (HR) of the latter two periods to the first period in overall survival and relapse-free survival.
Overall survival and relapse-free survival were analyzed by intention-to-treat. Short-term outcomes of operative time and incidence of early complications were analyzed in the patients who had surgery as assigned, and those patients who did not undergo the assigned approach were excluded. All P values were two-sided. All statistical analyses were carried out with SAS version 9.2.  (Table 1).  However, the incidence was consistently low in all periods in the laparoscopic surgery group (1st period: 14.3%, 2nd period: 14.8%, 3rd period:

| Short-term outcomes
13.6%, P = 0.95). There were no significant changes in the incidence of ≥grade 3 early complications over time in either group. In the analysis of the details of complications, there were no significant changes in any diseases in either group. Decreasing tendencies of incisional wound complications (1st period: 13.3%, 2nd period: 9.5%, 3rd period: 8.6%, P = 0.42) and anastomotic leakage (1st period: 6.7%, 2nd period: 1.7%, 3rd period: 4.0%, P = 0.06) were detected in the open surgery group, although they were not significant statistically. (Figure 3).   A matched-control study reported that operation time for laparoscopic surgery shortened in the latter period. 15 Another study also reported shortened operation time in latter period, regardless of surgical experience or patient factors, as a result of the evolution of surgical apparatus. 16 Therefore, we expected that the tendency for improvement would be especially remarkable in the short-term outcomes of laparoscopic surgery. However, the results of the present study in laparoscopic surgery slightly contradicted our expectation.     There was no particular trend in overall survival based on the periods in the two groups. Median survival time of patients with metastatic disease and/or recurrent colorectal cancer has gradually been extended by the development of chemotherapy in recent years. [21][22][23] Therefore, we expected the overall survival of the patients in the latter two periods to be improved. In fact, the chemotherapy regi-  27 Qualified surgeons were shown to improve the safety of laparoscopic surgery in urology and pediatric surgery. 28,29 Especially for colorectal surgery, the safety of laparoscopic surgery might be adequately guaranteed under the supervision of a surgeon qualified by JSES. 30 The endoscopic surgical skill qualification system of JSES originated in Japan, and certification is very difficult to obtain. Therefore, we surmised that the quality of laparoscopic surgery was maintained throughout the trial.

| D ISCUSS I ON
In the CLASICC study, the conversion rate improved from 38% in the first year of registration to 16% in the final year. 3 Difference in short-term outcomes according to registration period was reported in a case-matched study. 15 There was some anxiety over how the different registration periods would influence the results because laparoscopic surgery was a new procedure at the beginning of that study. However, there were no differences in short-term and longterm results between the periods in the laparoscopic surgery. The conversion rate of the present study was low at 5.4%, whereas it was reported to range between 10% and 20% in other randomized controlled studies. [1][2][3][4][5] As the reason, it seems that severe regulation of laparoscopic surgeons influenced the results. In particular, the introduction of JSES-qualified surgeons might have had a very large influence. A qualification system to evaluate technical skill appears to be important in clinical trials relating to surgical procedures.
There are some limitations in the present study. First, we do not have any data on the energy devices used in both arms of the trial.
Therefore, the influence of technological improvements on the outcomes can only be surmised. Second, we do not have data on chemotherapy given for recurrences in each period. In fact, we believe that neither the development of surgical devices nor the improvement of chemotherapy influenced the short-and long-term results during the study period. In