Takahiro Kinoshita1, Ichiro Uyama2, Masanori Terashima3, Hirokazu Noshiro4, Eishi Nagai5, Kazutaka Obama6, Yutaka Tamamori7, Toshinaga Nabae8, Michitaka Honda9, Takayuki Abe10. 1. Gastric Surgery Division, National Cancer Center Hospital East, Kashiwa, Japan. 2. Department of Surgery, Fujita Health University, Toyoake, Japan. 3. Division of Gastric Surgery, Shizuoka Cancer Center, Shizuoka, Japan. 4. Department of Surgery, Saga University Faculty of Medicine, Saga, Japan. 5. Department of Surgery and Oncology, Graduate School of Medical Sciences, Kyushu University, Fukuoka, Japan. 6. Department of Surgery, Graduate School of Medicine, Kyoto University, Kyoto, Japan. 7. Department of Gastroenterological Surgery, Osaka City General Hospital, Osaka, Japan. 8. Department of Surgery, JCHO Kyushu Hospital, Kitakyushu, Japan. 9. Department of Minimally Invasive Surgical and Medical Oncology, Fukushima Medical University, Fukushima, Japan. 10. Department of Preventive Medicine and Public Health, Biostatistics at Center for Clinical Research, Keio University School of Medicine, Tokyo, Japan.
Abstract
OBJECTIVE: A large-scale multicenter historical cohort study was conducted to investigate the efficacy of laparoscopic gastrectomy (LG) in comparison to open gastrectomy (OG) for locally advanced gastric cancer. BACKGROUND: LG is now practiced widely, but its applicability for advanced gastric cancer is still controversial. As oncologic outcomes of randomized trials are still pending, there is an urgent need for information that would be relevant to current practice. METHODS: Through a consensus meeting involving surgeons and biostatisticians, 30 preoperative variables possibly influencing the choice of surgical approach and associated with outcome were identified to enable rigorous estimation of propensity scores. A total of 1948 consecutive patients who underwent gastrectomy for clinical stage II/III gastric adenocarcinoma between 2008 and 2014 were identified, and their clinical data were collected from 8 participating hospitals. After propensity score matching, 610 cases (OG = 305, LG = 305) were finally selected for comparison of long-term outcomes. RESULTS: In the propensity-matched OG and LG populations, the mean observation period was 3.5 and 3.4 years, and the 5-year overall survival was 53.0% and 54.2%, respectively. The hazard ratio (LG/OG) for overall survival was 1.01 (95% confidence interval, 0.80-1.29), and noninferiority of LG was demonstrated statistically as the upper 95% confidence limit was less than the prespecified margin (1.33). The recurrence rate was 30.8% and 29.8% for OG and LG, respectively, and the hazard ratio for recurrence was 0.98 (95% confidence interval, 0.74-1.31). The patterns of recurrence in the 2 groups were similar. CONCLUSIONS: This observational study strictly adjusted for confounding factors has provided evidence to suggest that LG is oncologically comparable to OG for locally advanced gastric cancer. The validity of this result will be examined in ongoing randomized trials.
OBJECTIVE: A large-scale multicenter historical cohort study was conducted to investigate the efficacy of laparoscopic gastrectomy (LG) in comparison to open gastrectomy (OG) for locally advanced gastric cancer. BACKGROUND: LG is now practiced widely, but its applicability for advanced gastric cancer is still controversial. As oncologic outcomes of randomized trials are still pending, there is an urgent need for information that would be relevant to current practice. METHODS: Through a consensus meeting involving surgeons and biostatisticians, 30 preoperative variables possibly influencing the choice of surgical approach and associated with outcome were identified to enable rigorous estimation of propensity scores. A total of 1948 consecutive patients who underwent gastrectomy for clinical stage II/III gastric adenocarcinoma between 2008 and 2014 were identified, and their clinical data were collected from 8 participating hospitals. After propensity score matching, 610 cases (OG = 305, LG = 305) were finally selected for comparison of long-term outcomes. RESULTS: In the propensity-matched OG and LG populations, the mean observation period was 3.5 and 3.4 years, and the 5-year overall survival was 53.0% and 54.2%, respectively. The hazard ratio (LG/OG) for overall survival was 1.01 (95% confidence interval, 0.80-1.29), and noninferiority of LG was demonstrated statistically as the upper 95% confidence limit was less than the prespecified margin (1.33). The recurrence rate was 30.8% and 29.8% for OG and LG, respectively, and the hazard ratio for recurrence was 0.98 (95% confidence interval, 0.74-1.31). The patterns of recurrence in the 2 groups were similar. CONCLUSIONS: This observational study strictly adjusted for confounding factors has provided evidence to suggest that LG is oncologically comparable to OG for locally advanced gastric cancer. The validity of this result will be examined in ongoing randomized trials.