Survival outcomes after laparoscopy-assisted distal gastrectomy versus open distal gastrectomy with nodal dissection for clinical stage IA or IB gastric cancer (JCOG0912): a multicentre, non-inferiority, phase 3 randomised controlled trial

Articles

Survival outcomes after laparoscopy-assisted distal gastrectomy versus open distal gastrectomy with nodal dissection for clinical stage IA or IB gastric cancer (JCOG0912): a multicentre, non-inferiority, phase 3 randomised controlled trial Hitoshi Katai, Junki Mizusawa, Hiroshi Katayama, Shinji Morita, Takanobu Yamada, Etsuro Bando, Seiji Ito, Masakazu Takagi, Akinori Takagane, Shin Teshima, Keisuke Koeda, Souya Nunobe, Takaki Yoshikawa, Masanori Terashima, Mitsuru Sasako

Summary

Background Laparoscopy-assisted distal gastrectomy (LADG) is increasingly being used as an alternative to open distal gastrectomy (ODG) for gastric cancer treatment. Retrospective studies have shown equivalent survival with the two procedures, but these studies are limited by selection bias because LADG is more technically difficult than ODG. We aimed to evaluate whether LADG was non-inferior to ODG in terms of long-term survival outcomes. Methods We did an open-label, multicentre, non-inferiority, phase 3 randomised controlled trial at 33 institutions in Japan. Patients aged 20–80 years with histologically confirmed gastric adenocarcinoma (T1N0, T1N1, or T2[MP]N0), clinical stage I, in the middle or lower third of the stomach, Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1, with a body-mass index of less than 30 kg/m², were randomly assigned (1:1) to receive ODG or LADG. Randomisation was done by telephone, fax, or with a web-based system in the Japan Clinical Oncology Group Data Center; a minimisation method with a random component was used to adjust for institution and clinical stage (IA or IB). Only study-accredited surgeons performed ODG and LADG. The primary endpoint was relapse-free survival and was analysed according to the intention-to-treat principle. The non-inferiority margin (LADG vs ODG) was set at a hazard ratio (HR) of 1⋅54. The trial was registered with the UMIN Clinical Trials Registry, UMIN000003319. Findings Between March 15, 2010, and Nov 29, 2013, 921 patients were enrolled and randomly assigned to receive ODG (n=459) or LADG (n=462). 912 (99%) participants had the assigned surgery. 5-year relapse-free survival was 94⋅0% (95% CI 91⋅4−95⋅9) in the ODG group and 95⋅1% (92⋅7−96⋅8) in the LADG group. LADG was non-inferior to ODG for relapse-free survival (HR 0⋅84 [90% CI 0⋅56−1⋅27]), p=0⋅0075). The most common grade 3 or 4 adverse event was bowel obstruction, occurring in 11 (2%) of 455 patients in the ODG group and five (1%) of 457 patients in the LADG group. There were no treatment-related deaths. Interpretation This trial supports the non-inferiority of LADG compared with ODG for clinical stage I gastric cancer relapse-free survival, suggesting that LADG should be considered a standard treatment option when performed by experienced surgeons. Funding Japan National Cancer Center, Ministry of Health, Labour and Welfare of Japan, Japan Agency for Medical Research and Development. Copyright © 2019 Elsevier Ltd. All rights reserved.

Introduction Gastric cancer is one of the top three causes of cancerrelated death globally. Laparoscopy-assisted distal gastrectomy (LADG) is increasingly used as an alternative to open distal gastrectomy (ODG) for early-stage gastric cancer, on the basis of evidence for reduced intraoperative blood loss, lower postoperative pain, and faster recovery, among other advantages. Retrospective studies have shown equivalent survival after LADG compared with ODG.1,2 However, LADG is more technically difficult than ODG, which might introduce selection bias in comparative studies. Therefore, randomised controlled

trials with an assured high surgical skill level and quality control are necessary to confirm the non-inferiority of LADG compared with ODG. LADG is now considered an option for clinical stage I gastric cancer in the 2017 version of the Gastric Cancer Treatment Guidelines in Japan3 on the basis of clinical trials (JCOG0703 and JCOG0912)4,5 confirming the safety of LADG when conducted by experienced surgeons. However, the long-term oncological efficacy of laparo­ scopic approaches is controversial because laparoscopic manipulation and pneumo­ peritoneum effects might increase the risk of cancer cell

www.thelancet.com/gastrohep Published online November 19, 2019 https://doi.org/10.1016/S2468-1253(19)30332-2

Lancet Gastroenterol Hepatol 2019 Published Online November 19, 2019 https://doi.org/10.1016/ S2468-1253(19)30332-2 See Online/Comment https://doi.org/10.1016/ S2468-1253(19)30361-9 Department of Gastric Surgery (Prof H Katai MD, S Morita MD, T Yoshikawa MD) and Japan Clinical Oncology Group Data Center/Operations Office (J Mizusawa ME, H Katayama MD), National Cancer Center Hospital, Tokyo, Japan; Department of Gastrointestinal Surgery, Kanagawa Cancer Center Hospital, Yokohama, Japan (T Yamada MD); Division of Gastric Surgery, Shizuoka Cancer Center, Shizuoka, Japan (E Bando MD, M Terashima MD); Department of Gastroenterological Surgery, Aichi Cancer Center Hospital, Nagoya, Japan (S Ito MD); Department of Gastroenterological Surgery, Shizuoka General Hospital, Shizuoka, Japan (M Takagi MD); Department of Surgery, Hakodate Goryoukaku Hospital, Hakodate, Japan (A Takagane MD); Department of Surgery, National Hospital Organization, Sendai Medical Center, Sendai, Japan (S Teshima MD); Department of Surgery, Iwate Medical University, Morioka, Japan (K Koeda MD); Department of Surgery, Cancer Institute Hospital, Japanese Foundation for Cancer Research, Tokyo, Japan (S Nunobe MD); and Division of Upper Gastrointestinal Surgery, Department of Surgery,

1

Articles

Hyogo College of Medicine, Nishinomiya, Japan (Prof M Sasako MD) Correspondence to: Prof Hitoshi Katai, Department of Gastric Surgery, National Cancer Center Hospital, Tokyo 104–0045, Japan [email protected]

Research in context Evidence before this study We searched PubMed without language restrictions for studies published before March 1, 2015, with the terms “gastric cancer” and “laparoscopic gastrectomy”. Laparoscopy-assisted distal gastrectomy (LADG) is growing in popularity for gastric cancer treatment. Although some small-scale randomised controlled trials have suggested that LADG shows high efficacy, no confirmatory randomised controlled trials had been done to evaluate the efficacy and safety of LADG in comparison with open distal gastrectomy (ODG). A multicentre phase 2 trial of LADG (JCOG0703) validated the safety of LADG. A 2019 randomised controlled trial in Korea (KLASS-01) confirmed the non-inferiority of LADG compared with ODG for clinical stage I gastric cancer. Currently, LADG is not a standard treatment option in the Japanese Gastric Cancer Treatment Guidelines. The guidelines specify that ODG is the standard procedure and LADG is an investigational treatment. Confirmation of the non-inferiority of LADG compared with ODG is important given the known advantages of LADG, such as faster recovery and less severe postoperative pain. Added value of this study Around 1% of participants in this study were ineligible or did not undergo the assigned surgery; the JCOG0912 trial ensured

dissemination. Therefore, these guidelines stipulate that informed consent should be obtained from all patients after providing sufficient information, inclu­ ding the scarcity of data on long-term conseq­uences. Pivotal phase 3 studies have been done in Japan (JCOG0912) and Korea (KLASS-01)6 to address the comparative efficacy of LADG versus ODG. We have already reported that LADG performed by accredited surgeons is as safe as ODG in terms of adverse event frequency and short-term clinical outcomes.5 Here, we report long-term survival data from the JCOG0912 trial.

Methods

Study design This open-label, multicentre, non-inferiority, phase 3 randomised controlled trial was conducted within the framework of the Japan Clinical Oncology Group (JCOG) Stomach Cancer Study Group. Only surgeons specialising in both procedures at 33 Japanese institutions partici­ pated in the study. The study protocol was approved by the Protocol Review Committee of JCOG and by the institutional review board of each participating hospital. The design of the trial was reported previously.5,7

Participants The inclusion criteria for enrolment were as follows: aged 20–80 years, Eastern Cooperative Oncology Group performance status 0 or 1, body-mass index (BMI) of less than 30 kg/m², adequate organ function, histologically confirmed gastric adenocarcinoma in the middle or lower 2

good participant adherence. Less than 10% of the study population were found to have higher pathological stages than stage I. All LADG and ODG procedures were performed by certified surgeons with strict quality control, and close follow-up monitoring was done for 5 years. Therefore, this study was rigorously designed to detect potential differences in relapse-free survival and overall survival. Nonetheless, LADG was found to be non-inferior to ODG. Implications of all the available evidence LADG is now a standard surgical option for clinical stage IA or IB gastric cancer in many parts of east Asia. The current finding of the non-inferiority of LADG compared with ODG validates LADG as a standard option for stage I gastric cancer treatment in Japan. The efficacy and safety of laparoscopyassisted total gastrectomy and laparoscopy-assisted partial gastrectomy for clinical stage I gastric cancer have also previously been confirmed. Therefore, the efficacy and safety of laparoscopic gastrectomy has been shown for three major procedures used to treat clinical stage I gastric cancer in Japan. For advanced gastric cancer, the results of other clinical trials are awaited.

third of the stomach and curative resection predicted by distal gastrectomy, clinical stage IA (T1N0) or IB (T1N1, T2[MP]N0) according to the Japanese Classification of Gastric Carcinoma (second English edition),8 no invasion to the duodenum, no history of upper abdominal surgery or intestinal resection, no previous chemotherapy or radiation therapy against any other malignancies, and provided written informed consent. Patients were eligible if they had no indications for endoscopic mucosal resection (EMR) or endoscopic submucosal dissection (ESD) by clinical diagnosis. Patients with preceding EMR or ESD were eligible if the following conditions were fulfilled: (1) pathological findings showed that additional gastrectomy was required, (2) gastrectomy could be performed within 91 days of EMR or ESD, (3) no perforation occurred during the preceding EMR or ESD procedure, and (4) the resection margins of EMR or ESD did not reach the upper third of the stomach. Exclusion criteria for enrolment were synchronous or metachronous (within 5 years) malignancies other than carcinoma in situ, infectious disease requiring systemic therapy, body temperature of 38°C or higher, pregnancy or breastfeeding, severe mental illness, continuous systemic steroid therapy, unstable angina pectoris or history of myocardial infarction within 6 months, uncontrollable hypertension, uncontrollable diabetes or administration of insulin, and severe respiratory disease requiring continuous oxygen therapy. Clinical staging was determined at the discretion of each institution. The T category was evaluated by upper gastrointestinal

www.thelancet.com/gastrohep Published online November 19, 2019 https://doi.org/10.1016/S2468-1253(19)30332-2

Articles

endoscopy. Multidetector row CT was used for diagnosis of the T, N, and M categories.

Randomisation Eligible patients were registered by telephone, fax, or on a web-based system in the JCOG Data Center, which conducted randomisation and all data handling. Patients were randomly assigned (1:1) to receive ODG or LADG. The minimisation method with random components was used to balance the groups with respect to institution and clinical stage (IA or IB). The allocated procedure was not masked from investigators or patients. The JCOG Data Center also did central monitoring every 6 months to ensure patient eligibility, full and accurate data submission, protocol compliance, safety, and on-schedule study progress. Monitoring reports were independently reviewed by the JCOG Data and Safety Monitoring Committee.

Procedures Perioperative procedures were identical in both groups except for the surgical approach. The extent of nodal dissection was decided according to the surgical T and N stages, which were based on the third version of the Japanese Gastric Cancer Treatment Guidelines.9 In principle, lymphadenectomy dissection level D1 or D1+ was done for clinical stage IA tumours, and D2 dissection was done for clinical stage IB tumours. If nodal involvement was suspected during surgery, D2 lympha­ denectomy was performed. For clinical T1 gastric cancer with margins 4 cm or further from the pylorus, pyloruspreserving distal gastrectomy was allowed. Pyloruspreserving gastrectomy is a modified surgery that omits suprapyloric lymph node dissection from the distal gastrectomy dissection area. Bursectomy was not allowed but preservation of the omentum, vagus nerve, or both was discretionary. The recon­struction method was not specified. The method of gastric cancer lesion marking was not specified in the protocol and was at the discretion of the institution. Surgeons at partici­pating institutions performed open gastrectomy as routine medical care during the registration period. In the LADG group, a mini-laparotomy incision shorter than 6 cm was required. Gastrectomy and dissection were performed laparoscopically, and a mini-laparotomy was done when needed for specimen removal and reconstruction. If the intraoperative findings showed a stage of 2 or greater, LADG was converted to open surgery. Postoperative analgesia, such as epidural anaesthesia, was not specified by the protocol but use of analgesics on postoperative days 5–10 was recorded. Adjuvant chemotherapy with tegafur–gimeracil–oteracil for 1 year was recommended for patients with curative resection and pathological stage II, IIIA, or IIIB tumour, excluding T1 and T3N0 disease. Surgical quality assurance was facilitated by permitting only surgeons accredited by the study chair to perform

ODG and LADG. In the ODG group, only surgeons with experience of doing 60 or more open gastrectomies were accredited. In the LADG group, only surgeons with experience of performing 30 or more ODG and LADG procedures, as well as certification by the Japan Society for Endoscopic Surgery (or equivalent), were accredited. For all participants in the LADG group, procedures were evaluated from intraoperative photographs by the JCOG committee for quality control and assessment of surgery. Meetings were held three times a year to discuss quality control and quality assurance; we planned to discuss problems with surgical procedures at these meetings if there was a case that did not meet the required standards. To assess compliance with lymphadenectomy guidelines, the number of dissected nodes at all stations was recorded on case report forms and the results were monitored by the JCOG Data Center. All enrolled patients were followed up at least every 6 months for the first 2 years after surgery and then every year until 5 years after the last patient enrolment. Tumour markers were measured and chest x-ray, upper gastro­ intestinal endoscopy, and enhanced abdominal CT were done at least once yearly for the duration of follow-up. Surgical methods and pathology results were recorded according to the 13th and 14th editions of the Japanese Classification of Gastric Carcinoma (which correspond to the second and third English editions, respectively).8,10 Surgery time, blood loss, and all postoperative morbidities during hos­pitalisation were included on the case report forms for prospective data collection.

Outcomes The primary endpoint was relapse-free survival, defined as days from randomisation to relapse or death from any cause. Relapse-free survival was censored on the last day the patient was alive without any evidence of relapse. Secondary endpoints were overall survival of all randomly assigned patients, proportion of LADG completion, proportion of conversion to open surgery, adverse events, short-term clinical outcomes, and postoperative quality of life. Overall survival was defined as days from random­ isation to death from any cause and was censored on the last day the patient was alive. The proportion of LADG completion was defined as the fraction of patients receiving LADG (without conversion to open surgery) out of all patients in the LADG group receiving surgery (LADG or ODG). The proportion of conversion to open surgery was defined as the fraction of patients switched from LADG to ODG out of all patients diagnosed before gastrectomy with clinical stage IA or IB gastric cancer. The short-term clinical outcomes considered were: (1) time from the end of surgery until the first episode of flatus (a metric of recovery time); (2) the proportion of patients requiring an analgesic on postoperative days 5–10; (3) highest body temperature during the first 3 days after surgery; and (4) highest body temperature during hospitalisation; these results have been presented

www.thelancet.com/gastrohep Published online November 19, 2019 https://doi.org/10.1016/S2468-1253(19)30332-2

3

Articles

ODG (n=459)

921 patients enrolled

Age, years

64 (57–70, 27–80)

LADG (n=462) 63 (58–69, 25–80)

Sex

921 randomly assigned

Men

275 (60%)

289 (63%)

Women

184 (40%)

173 (37%)

T1

411 (90%)

411 (89%)

T2

48 (10%)

50 (11%)

Clinical T stage* 459 assigned to ODG

462 assigned to LADG

Unknown 1 ineligible*

0

1 (<1%)

Clinical N stage*

4 discontinued treatment 2 withdrew consent 2 refused assigned surgery

455 received ODG and included in safety analysis

5 discontinued treatment 1 ineligible 2 refused assigned surgery 2 other reasons

457 received LADG and included in safety analysis

N0

452 (99%)

451 (98%)

N1

7 (2%)

10 (2%)

Unknown

0

1 (<1%)

Clinical stage* IA

404 (88%)

401 (87%)

IB

55 (12%)

60 (13%)

Unknown

0

1 (<1%)

ECOG performance status 459 included in ITT analysis

462 included in ITT analysis

0

458 (>99%)

1

1 (<1%)

BMI, kg/m² Figure 1: Trial profile ODG=open distal gastrectomy. LADG=laparoscopy-assisted distal gastrectomy. ITT=intention-to-treat. *One patient in the ODG group who was ineligible had the assigned surgery and was included in the safety analysis.

<20 ≥20 to <25 ≥25

previously.5 Postoperative quality of life was evaluated using European Organisation for Research and Treatment of Cancer QLQ-C30 and QLQ-STO22 questionnaires. These quality of life results will be reported separately. Surgical morbidity was classified as having either an early adverse event or a late adverse event. Early adverse events included intraoperative or postoperative morbidity between surgery and first discharge date. Early adverse events were reported elsewhere.5 Late adverse events included postoperative morbidity after first discharge date. Postoperative morbidity was described according to the Common Terminology Criteria for Adverse Events version 4.0 and the JCOG postoperative complication criteria based on the Clavien−Dindo classification.11

Statistical analysis This randomised trial was designed to test whether LADG is non-inferior to ODG for relapse-free survival of patients with gastric cancer. Additional endpoints were adopted to evaluate whether LADG is less invasive than ODG, but these were considered to be exploratory. The initial primary endpoint was overall survival, and the planned sample size was 920 patients (460 patients per group) according to the method of Schoenfeld and Richter.12 We anticipated 5 years of follow-up after 5 years of accrual and assumed a 5-year overall survival of 90% in each group, so these numbers would be sufficient to determine non-inferiority within a margin of 5% for 5-year overall survival (yielding a non-inferiority margin hazard ratio [HR] of 1⋅54) with at least 80% power at a one-sided α of 5%. This non-inferiority margin was 4

22·6 (20·6−24·4, 16·2−29·7)

459 (99%) 3 (1%) 22·3 (20·6–24·4, 15·2−28·5)

86 (19%)

94 (20%)

283 (62%)

284 (61%)

90 (20%)

84 (18%)

Data are median (IQR, range) or n (%). ODG=open distal gastrectomy. LADG=laparoscopy-assisted distal gastrectomy. ECOG=Eastern Cooperative Oncology Group. BMI=body-mass index. *According to the 13th edition of the Japanese Classification of Gastric Carcinoma, which corresponds to the second English edition.

Table 1: Baseline characteristics of the intention-to-treat population

determined on the basis of a consensus reached by researchers in the JCOG Stomach Cancer Study Group. An HR of 1⋅54 was deemed acceptable considering the potential clinical benefits of LADG, including smaller surgical wounds, less postoperative pain, and quicker postoperative recovery, compared with ODG. However, before the first interim analysis on April 1, 2015, the primary endpoint was amended from overall survival to relapse-free survival because the surrogacy of relapse-free survival was shown in a meta-analysis13 and the number of events for overall survival was smaller than expected. After amending the primary endpoint, 68% power was maintained with a one-sided α of 5% and a non-inferiority HR of 1⋅54, which corresponds to a 3·6% difference in 5-year relapse-free survival, assuming a 5-year relapsefree survival of 93% in each group. We also planned two interim analyses. Multiplicity was adjusted using the Lan–DeMets method with the O’Brien–Fleming-type α-spending function.14 The statistical significance level for the primary analysis was 4·96%, considering the adjustment for multiple testing. The JCOG Data and Safety Monitoring Committee independently reviewed the interim analysis reports.

www.thelancet.com/gastrohep Published online November 19, 2019 https://doi.org/10.1016/S2468-1253(19)30332-2

Articles

ODG (n=455)

LADG (n=457)

100

Surgery time, min

194 (159–235, 48–445)

278 (230–327, 120–577)

80

Blood loss, mL

115 (60–225, 0–890)

38 (17–90, 0–1920)

90 Relapse-free survival (%)

Surgical results

Procedure Distal gastrectomy

333 (73%)

342 (75%)

Pylorus-preserving gastrectomy

122 (27%)

114 (25%)

Total gastrectomy

0

1 (<1%)

D0

0

0

D1

5 (1%)

5 (1%)

50 40 30

ODG LADG

20

Extent of lymph node dissection

10

D1+

319 (70%)

347 (76%)

D2

126 (28%)

101 (22%)

D2+

5 (1%)

4 (1%)

Pathological results T stage* T1a

201 (44%)

217 (48%)

T1b

197 (43%)

173 (38%)

T2

33 (7%)

50 (11%)

T3

19 (4%)

12 (3%)

T4a

5 (1%)

4 (1%)

T4b

0

0

TX

0

1 (<1%)

N stage* N0

395 (87%)

395 (86%)

N1

42 (9%)

49 (11%)

N2

17 (4%)

13 (3%)

N3

70 60

1 (<1%)

0

Pathological stage* IA

364 (80%)

354 (77%)

IB

52 (11%)

67 (15%)

II

28 (6%)

30 (7%)

IIIA

10 (2%)

0

HR 0·84 (90% CI 0·56–1·27); p=0·0075 0

Number at risk (number censored) ODG 459 (0) LADG 462 (0)

1

451 (4) 457 (1)

2

3

4 5 6 Years after randomisation

outcome analysis was one-sided because of the preset research objective (non-inferiority), whereas for all other analyses, p values were two-sided. All statistical analyses were performed with SAS version 9.2 or later (SAS Institute, Cary, NC, USA). The data presented in this article were obtained as of Nov 29, 2018. Analysis of efficacy endpoints was done according to the intention-totreat principle. In addition, supplementary analysis of efficacy endpoints was done in the per-protocol popu­ lation, defined as patients who received the assigned surgery. Analysis of safety endpoints was also done in the per-protocol population. This trial is registered with the UMIN Clinical Trials Registry, UMIN000003319.

6 (1%)

Role of the funding source

0

IV

1 (<1%)

0

The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

Relapse-free survival and overall survival curves were estimated by the Kaplan−Meier method. HRs and associated CIs for relapse-free survival and overall survival were estimated using the Cox proportional hazards model. As a sensitivity analysis, stratified Cox regression analysis was done with clinical stage as a stratum. Proportionality was checked using a log of negative log plot. Prespecified subgroup analyses of relapse-free survival were done for sex, age, body-mass index, histology, and clinical stage, cT, and cN. Differences in proportions were evaluated using Fisher’s exact test and differences in continuous variables by Wilcoxon’s rank sum test. The p value for the primary

9

Figure 2: Relapse-free survival There were 35 events in the ODG group and 30 events in the LADG group. ODG=open distal gastrectomy. LADG=laparoscopy-assisted distal gastrectomy. HR=hazard ratio.

0

Table 2: Surgical and pathological results in the safety population

8

444 (0) 432 (4) 423 (4) 395 (25) 220 (173) 123 (93) 36 (85) 0 (36) 451 (1) 447 (1) 430 (11) 393 (33) 231 (160) 117 (110) 36 (79) 0 (36)

IIIB

Data are median (IQR, range) or n (%). ODG=open distal gastrectomy. LADG=laparoscopy-assisted distal gastrectomy. *According to the 13th edition of the Japanese Classification of Gastric Carcinoma, which corresponds to the second English edition.

7

Results Between March 15, 2010, and Nov 29, 2013, 921 eligible patients with clinical stage I gastric cancer were enrolled by 48 surgeons from 33 institutions in Japan (appendix). Of these, 459 patients were randomly assigned to the ODG group and 462 to the LADG group (figure 1). After randomisation, one patient (<1%) in the ODG group was found to be ineligible because of non-curative resection by ESD for another synchronous gastric cancer, and one patient (<1%) in the LADG group was found to be ineligible because gastric adenocarcinoma was not histologically confirmed. However, according to the intention-to-treat principle, all 459 patients in the ODG group and 462 patients in the LADG group were included in the efficacy analyses.

www.thelancet.com/gastrohep Published online November 19, 2019 https://doi.org/10.1016/S2468-1253(19)30332-2

See Online for appendix

5

Articles

100 90

Overall survival (%)

80 70 60 50 40 30

ODG LADG

20 10 0

HR 0·83 (95% CI 0·49–1·40) 0

1

2

Number at risk (number censored) ODG 459 (0) 456 (3) 452 (0) LADG 462 (0) 461 (0) 457 (0)

3

4 5 6 Years after randomisation

442 (2) 454 (0)

7

8

9

435 (0) 421 (11) 303 (115) 177 (122) 54 (121) 0 (54) 448 (3) 430 (15) 312 (114) 173 (132) 56 (116) 0 (56)

Figure 3: Overall survival There were 31 events in the ODG group and 26 events in the LADG group. ODG=open distal gastrectomy. LADG=laparoscopy-assisted distal gastrectomy. HR=hazard ratio.

ODG (n=455) Grade 1–2

LADG (n=457) Grade 4

Grade 1–2

Grade 3

Grade 4

NA

2 (<1%)

0

NA

3 (1%)

0

Anastomotic stenosis

0

1 (<1%)

0

0

2 (<1%)

0

Cholecystitis

3 (1%)

4 (1%)

0

3 (1%)

2 (<1%)

0

Intra-abdominal abscess

Grade 3

Dumping syndrome

58 (13%)

0

0

56 (12%)

0

0

Gastro-oesophageal regurgitation

80 (18%)

0

NA

80 (18%)

0

NA

Bowel obstruction

2 (<1%)

1 (<1%)

2 (<1%)

4 (1%)

1 (<1%)

Ileus

1 (<1%)

10 (2%) 1 (<1%)

0

0

0

0

Pneumonia

2 (<1%)

0

0

4 (1%)

0

0

Wound infection

3 (1%)

0

0

1 (<1%)

0

0

Incisional hernia

8 (2%)

1 (<1%)

0

6 (1%)

0

0

Data are n (%). ODG=open distal gastrectomy. LADG=laparoscopy-assisted distal gastrectomy. NA=not applicable (grade was not defined in the Common Terminology Criteria for Adverse Events version 4.0).

Table 3: Late adverse events in the safety population

Four (1%) of 459 patients in the ODG group and five (1%) of 462 patients in the LADG group did not undergo their assigned surgery, and these patients were excluded from safety analyses. In the ODG group, two patients withdrew their consent and were censored at that time, and two patients refused the assigned surgery. One patient in the ODG group who was ineligible had the assigned surgery and was included in the safety analysis. In the LADG group, five patients were excluded from the safety analysis: one patient who was ineligible, one diagnosed preoperatively with widespread cancer requiring total gastrectomy, one whose surgery was cancelled soon after starting because of ischaemic heart disease, and two who refused the assigned surgery. Therefore, 455 (99%) of 459 patients in the ODG group and 457 (99%) of 462 patients in the LADG group were included in the safety analysis. 6

Baseline characteristics were similar in the two groups (table 1). Surgical and pathological results are sum­marised in table 2. Surgery was significantly longer in the LADG group than in the ODG group (median of 278 min [IQR 230–327] vs 194 min [159–235], p<0⋅0001), whereas blood loss was lower in the LADG group (38 mL [17–90] vs 115 mL [60–225], p<0⋅0001). The distribution of surgical procedures was similar in both groups. All procedures were deemed to be sufficiently high quality by assessment of intraoperative photographs by the JCOG committee. We found no significant differences in pathological stage distribution between the two groups. 40 patients (4%) had unsuspected pathological T3 or T4 tumour. 17 patients (2%) had tumours higher than stage II. Median follow-up for all patients after randomisation was 5⋅9 years (IQR 5⋅1−7⋅0). At the time of last follow-up on Nov 29, 2018, 57 (6%) of 921 patients had died, including 31 (7%) of 459 in the ODG group and 26 (6%) of 462 in the LADG group. 28 patients (3%) had recurrence, including 14 (3%) in the ODG group and 14 (3%) in the LDG group. Of these patients, six (1%) in the ODG group and two (<1%) in the LDG group had liver metastases, two (<1%) in the ODG group and three (1%) in the LADG group had peritoneal metastases, two (<1%) in the ODG group and three (1%) in the LADG group had lymph node metastases, and four (1%) in the ODG group and one (<1%) in the LADG group had lung metastases, bone metastases, or both. 20 patients (2%) died from recur­rence, comprising ten (2%) in the ODG group and ten (2%) in the LADG group. 37 (4%) of 921 patients died from other causes, including 21 (5%) of 459 in the ODG group and 16 (3%) of 462 in the LADG group. The primary outcome did not differ between groups. 3-year relapse-free survival was 95⋅8% (95% CI 93⋅5−97⋅3) in the ODG group and 97⋅4% (95⋅5−98⋅5) in the LADG group. 5-year relapse-free survival was 94⋅0% (91⋅4−95⋅9) in the ODG group and 95⋅1% (92⋅7−96⋅8) in the LADG group (figure 2). The HR for relapse-free survival (ODG vs LADG) was 0·84 (90% CI 0·56−1·27), which was within the predefined non-inferiority margin of 1⋅54 (p=0⋅0075, figure 2). Therefore, we found that LADG was non-inferior to ODG. The HR by sensitivity analysis using stratified Cox regression analysis was 0·85 (95% CI 0·52−1·38). The HR for relapse-free survival (ODG vs LADG) was 0·79 (0·48−1·29) by supplementary analysis in the per-protocol population. 3-year overall survival was 97⋅4% (95% CI 95·4−98⋅5) in the ODG group and 98⋅3% (96⋅6−99⋅1) in the LADG group. 5-year overall survival was 95⋅2% (95% CI 92⋅7−96⋅8) in the ODG group and 97⋅0% (94·9−98⋅2) in the LADG group (figure 3). The HR for overall survival (ODG vs LADG) was 0⋅83 (0⋅49−1⋅40). Conversion to open surgery was required for 16 (4%) of 457 patients who were randomly assigned to receive LADG. The reasons for conversion were reported elsewhere.5 LADG was completed in 440 (96%) of 457 patients who were randomly assigned to receive LADG.

www.thelancet.com/gastrohep Published online November 19, 2019 https://doi.org/10.1016/S2468-1253(19)30332-2

Articles

Sex Men Women Age ≤64 years ≥65 years Age ≤69 years ≥70 years BMI, kg/m2 <20 ≥20 to <25 ≥25 Histology* Differentiated Undifferentiated Clinical stage* IA IB cT* T1a T1b T2 cN* N0 N1 Overall

ODG, events/N

LADG , events/N

HR (95% CI)

26/275 9/184

22/289 8/173

0·80 (0·45–1·41) 0·94 (0·36–2·43)

15/237 20/222

13/266 17/196

0·79 (0·37–1·65) 0·92 (0·48–1·75)

19/334 16/125

20/348 10/114

1·02 (0·54–1·91) 0·63 (0·29–1·39)

10/86 20/283 5/90

6/94 16/284 8/84

0·52 (0·19–1·44) 0·80 (0·42–1·55) 1·72 (0·56–5.25)

19/208 16/251

18/211 12/250

0·92 (0·48–1·76) 0·75 (0·35–1·58)

27/404 8/55

22/401 8/60

0·80 (0·46–1·41) 0·96 (0·36–2·57)

6/148 21/263 8/48

6/137 17/274 7/50

1·05 (0·34–3·24) 0·78 (0·41–1·48) 0·91 (0·33–2·51)

35/452 0/7 35/459

29/451 1/10 30/462

0·82 (0·50–1·34) Not estimable 0·84 (0·52–1·37) 0·25

0·5

1

Favours LADG

2

4

8

Favours ODG

Figure 4: Subgroup analysis of relapse-free survival ODG=open distal gastrectomy. LADG=laparoscopy-assisted distal gastrectomy. HR=hazard ratio. BMI=body-mass index. *One patient was missing because their diagnosis of gastric cancer at a previous hospital was not confirmed at the participating hospital.

Late adverse event (grade 3 or 4) frequency did not differ between groups. 19 (4%) of 455 patients in the ODG group and 11 (2%) of 457 patients in the LADG group experienced one or more grade 3 or 4 adverse events, and there were no apparent differences in the frequencies of most specific adverse events (table 3), although individual sample sizes were small. However, bowel obstruction was less frequent in the LADG group than the ODG group. No treatment-related deaths occurred in either group. In subgroup analyses for relapse-free survival (figure 4), we found no differences by sex, age, histology, clinical stage, cN, or cT. However, patients with a BMI of 25 or higher tended to show poorer survival in the LADG group than in the ODG group, but the difference was non-significant.

Discussion This multicentre randomised controlled trial showed that relapse-free survival after LADG is non-inferior to that after ODG. We previously reported that LADG safety is equivalent to that of ODG, as evidenced by similar adverse event frequencies and short-term clinical outcomes.5 We found that LADG results in less blood loss, promotes faster recovery (as indicated by a shorter time to pass first flatus), and requires fewer analgesic doses after 5 days after surgery than does ODG, although it requires longer surgery times.5 The strengths of this study include a small number of ineligible patients, a small number of patients who did not have the allocated

surgery due to rigorous monitoring, and high quality of surgery due to advanced surgical skills and strict quality control. Therefore, our results support LADG as a standard surgical option for clinical stage IA or IB gastric tumours. Our findings are in accord with the Korean KLASS-01 trial6 that reported the non-inferiority of LADG compared with ODG. In the KLASS-01 study, the proportion of ineligible patients was 4%, and 6% did not receive the assigned surgery, compared with <1% ineligible patients and 1% not receiving the assigned surgery in our study. There are several reasons for the high adherence to surgical assignment among JCOG0912 study partici­ pants. Firstly, the JCOG Stomach Cancer Study Group recognises the importance of maintaining clinical trial quality and has become proficient at conducting clinical trials since it was formed in 1995. In participating institutions, surgeons treated patients in accordance with the Japanese Gastric Cancer Treatment Guidelines. The guidelines specify that ODG is the standard procedure and LADG is an investigational treatment. Therefore, although LADG is becoming more common, during the JCOG0912 registration period, surgeons did not do LADG in routine medical care. Surgeons carefully explained to the patient that they performed LADG only for clinical trial participants. The JCOG0912 participants understood the concept of this trial and agreed to undergo the assigned surgery without any objection. Other potential reasons for high adherence are the

www.thelancet.com/gastrohep Published online November 19, 2019 https://doi.org/10.1016/S2468-1253(19)30332-2

7

Articles

rigorous monitoring discussions held every 6 months to ensure trial quality, and the regular group meetings that were held to share information on ineligible or switched patients. For the generalisability of the results of this study, diagnostic accuracy must be considered. The proportion of patients having unsuspected pathological T3 or T4 tumour was 4·4% in this study, which is concordant with a previous Japanese report.15 However, this high accuracy might not be reproducible in countries other than Japan. If oncological safety of laparoscopic surgery is not secured for advanced disease, high contamination of advanced tumour could worsen the survival even though laparoscopic surgery is applied to clinical stage I. The surgery time is usually longer for LADG compared with that for ODG, but the amount of bleeding is reduced because of the magnifying effect. Although absolute surgery times were longer and blood loss volumes lower in the JCOG0912 study than in the KLASS-01 study for both ODG and LADG, and the absolute and relative differences between ODG and LADG in terms of surgery times and blood loss were smaller in KLASS-01,16 the characteristics of the two types of surgery are reflected in the results of two clinical trials. The reported numbers of harvested lymph nodes in JCOG0912 and KLASS-01 are similar, so it can be assumed that the two studies maintained oncological accuracy. With two studies in different countries independently achieving the same results, LADG has been shown to be a valid treatment option for stage I gastric cancer with the highest level of evidence. In the LACC prospective randomised trial,17 patients who underwent minimally invasive radical hysterectomy for early-stage cervical cancer had lower rates of diseasefree survival and overall survival and a higher rate of locoregional recurrence than patients who underwent open abdominal radical hysterectomy. These results call into question the findings in the literature suggesting that minimally invasive radical hysterectomy is associated with no difference in oncological outcomes in com­ parison with the open approach. Minimally invasive surgery for cancer does not automatically yield results equal to open surgery. These results show the necessity of high quality randomised controlled trials, even for early-stage cancer, and emphasise the importance of the JCOG0912 and KLASS-01 trials. In our study, surgical quality assurance was facilitated by permitting only surgeons who were accredited by the study chair. Strict inclusion criteria for surgeons usually lead to difficulty in establishing generalisability of the procedure to less experienced surgeons. To address this problem, we allowed accredited surgeons to attend the surgery as supervisor assistants for less experienced surgeons, to pass on technique-related knowledge of this procedure. We did subgroup analyses to identify potential clinicodemographic factors affecting the outcomes of 8

LADG and ODG, but the statistical power was insufficient to detect any interaction due to the small number of events. Furthermore, an effect of excess weight on relapse-free survival after LADG could have been suppressed by exclusion of obese patients (BMI ≥30). Nonetheless, we identified a trend for poorer relapse-free survival after LADG than ODG in patients with a BMI of 25 or higher. The HR for death was also increased in the high BMI group of KLASS-01,6 although no significant difference was found between BMI subgroups in the ODG and LADG groups. The surgical complication rate increased with higher BMI in the LADG group of JCOG0912,5 and there are many reports that postoperative complications are associated with poorer survival.18–20 A previous meta-analysis found that overweight and obesity not only increase surgical difficulty, but they also reduce survival of patients with gastric cancer, although most of the included studies evaluated the effect of high BMI on outcomes after open gastrectomy.21 By contrast, another meta-analysis found that the feasibility and safety of LADG were similar to those of ODG.22 Based on the available data, LADG might not be appropriate for patients with obesity and gastric cancer, but further examination is required to clarify this issue. This study has several limitations. First, the HR for noninferiority (1⋅54) was slightly larger than is usual in chemotherapeutic studies. However, HR generally rises when the baseline survival rate is high, as in this trial. We assumed the 5-year overall survival of 5% as a clinically significant difference and specified it as the non-inferiority margin. However, after we amended the primary endpoint to relapse-free survival, we still used the HR of 1⋅54, as this corresponds to only a 3⋅6% difference in 5-year relapse-free survival, which was a more precise difference. Second, we only included patients with clinical stage I gastric cancer treated by distal gastrectomy. The efficacy and safety of laparoscopic gastrectomy for more advanced disease and other procedures, such as total gastrectomy, must be confirmed in future clinical trials. We have already confirmed the safety of laparoscopic total gastrectomy and laparoscopic proximal gastrectomy for clinical stage I gastric cancer in the JCOG1401 trial.23 Our evidence for the efficacy of LADG might be extrapolated to laparoscopic total or proximal gastrectomy if surgeons accept the fact that the difficulty of lymph node dissection in laparoscopic total or proximal gastrectomy and LADG is equal. The CLASS-01 randomised clinical trial24 of advanced gastric cancer in China found that LADG did not result in inferior disease-free survival at 3 years compared with ODG. However, these results should be interpreted with caution because the differences in 3-year disease-free survival between the laparoscopic and ODG groups increased with tumour stage (although statistical significance was not reached in any of the stage subgroups), suggesting an effect of surgical difficulty. A randomised controlled trial to evaluate laparoscopic

www.thelancet.com/gastrohep Published online November 19, 2019 https://doi.org/10.1016/S2468-1253(19)30332-2

Articles

versus open surgery for advanced gastric cancer is ongoing (JLSSG0901) and long-term results are awaited.25,26 Third, the whole study cohort includes approximately 25% of pylorus-preserving gastrectomy. This is a modification from distal gastrectomy in terms of luminal resection and lymph node dissection (it modifies infra­ pyloric node dissection and omits supra-pyloric lymph node dissection). The oncological accuracy of the pyloruspreserving gastrectomy has been shown in large Japanese and Korean retrospective studies.27,28 The KLASS-04 trial (NCT02595086) is comparing distal gastrectomy with pylorus preserving gastrectomy in terms of functional outcomes, but no reliable results are available so far. Pylorus-preserving gastrectomy is popular only in Japan, Korea, and China; western surgeons are mostly not familiar with the procedure and it is not performed in Europe or the USA, either because of a scarcity of eligible patients or the debatable benefits over distal gastrectomy. The JCOG0912 study has the potential to influence the routine surgical practice worldwide, but the inclusion of this very specific modification of distal gastrectomy might weaken its impact. Another limitation of this study is that it is unclear whether the findings can be used to predict the outcomes of LADG when performed by less experienced surgeons. Furthermore, the number of observed events was lower than expected because of excellent prognosis. As a result, the CI for the HR was wide, although the upper limit of the CI did not exceed the non-inferiority margin. However, it was not feasible to increase the sample size or extend the follow-up period to observe the expected number of events because the relapse-free survival curve in both groups was almost at a plateau. Finally, it is uncertain whether these findings are applicable to obese patients or patients of non-Japanese ethnicity. In conclusion, two pivotal clinical trials (JCOG0912 and KLASS-01) have shown the non-inferiority of LADG compared with ODG for safe and efficacious treatment of clinical stage IA or IB gastric cancer. Therefore, LADG performed by an experienced surgeon should be considered a standard surgical option for the treatment of clinical stage IA or IB gastric cancer. Contributors HKatai, TYo, and MS designed the study and wrote the protocol. HKatai was the principal investigator. All authors except JM and HKatay recruited patients into the study. JM and HKatay were responsible for data management, statistical analysis, and interpretation. All authors contributed to writing or reviewing the report and approved the final version. Declaration of interests HKatai, JM, HKatay, SI, MTa, TYo, and MTe received grants from the Japan Agency for Medical Research and Development, during the conduct of the study. JM and HKatay received grants from the Ministry of Health, Labour, and Welfare of Japan, during the conduct of the study. HKatai has received lecture fees from Taiho Pharmaceutical, Medtronic, Kyowa Kirin, and Ono Pharmaceutical, outside the submitted work. JM has received lecture fees from Chugai Pharmaceutical, outside the submitted work. HKatay has received lecture fees from Johnson & Johnson, outside the submitted work. TYo has received lecture fees from Taiho Pharmaceutical, Chugai Pharmaceutical, Ono Pharmaceutical, Bristol-Myers Squibb,

Merck Sharp & Dohme, Terumo, Medtronic, and Johnson & Johnson, outside the submitted work. MTe has received lecture fees from Taiho Pharmaceutical, Chugai Pharmaceutical, Ono Pharmaceutical, Bristol-Myers Squibb, Eli Lilly, Yakult Honsha, Takeda Pharmaceutical, Pfizer, and Daiichi Sankyo, outside the submitted work. MS has received lecture fees from Johnson & Johnson, Medtronic, and Olympus, outside the submitted work. All other authors declare no competing interests. Data sharing Anonymised participant data reported in this article will be shared with investigators after proposed use is approved by investigators from the Stomach Cancer Study Group of the Japan Clinical Oncology Group. Proposals should be directed to [email protected]. The date will be available for achieving aims in the approved proposal. Acknowledgments We thank Haruhiko Fukuda for overseeing the management of this study. The study was supported in part by National Cancer Center Research and Development Funds (26-A-4, 29-A-3) and the Japan Agency for Medical Research and Development (grant number JP18ck0106317). References 1 Honda M, Hiki N, Kinoshita T, et al. Long-term outcomes of laparoscopic versus open surgery for clinical stage I gastric cancer: the LOC-1 study. Ann Surg 2016; 264: 214–22. 2 Kim HH, Han SU, Kim MC, et al. Long-term results of laparoscopic gastrectomy for gastric cancer: a large-scale case-control and case-matched Korean multicenter study. J Clin Oncol 2014; 32: 627–33. 3 Japanese Gastric Cancer Association. Japanese Gastric Cancer Treatment Guidelines 2014 (ver. 4). Gastric Cancer 2017; 20: 1–19. 4 Katai H, Sasako M, Fukuda H, et al. Safety and feasibility of laparoscopy-assisted distal gastrectomy with suprapancreatic nodal dissection for clinical stage I gastric cancer: a multicenter phase II trial (JCOG 0703). Gastric Cancer 2010; 13: 238–44. 5 Katai H, Mizusawa J, Katayama H, et al. Short-term surgical outcomes from a phase III study of laparoscopy-assisted versus open distal gastrectomy with nodal dissection for clinical stage IA/IB gastric cancer: Japan Clinical Oncology Group Study JCOG0912. Gastric Cancer 2017; 20: 699–708. 6 Kim HH, Han SU, Kim MC, et al. Effect of laparoscopic distal gastrectomy vs open distal gastrectomy on long-term survival among patients with stage i gastric cancer: the KLASS-01 randomized clinical trial. JAMA Oncol 2019; 5: 506–13. 7 Nakamura K, Katai H, Mizusawa J, et al. A phase III study of laparoscopy-assisted versus open distal gastrectomy with nodal dissection for clinical stage IA/IB gastric cancer (JCOG0912). Jpn J Clin Oncol 2013; 43: 324–27. 8 Japanese Gastric Cancer Association. Japanese Classification of Gastric Carcinoma: 2nd English Edition. Gastric Cancer 1998; 1: 10–24. 9 Japanese Gastric Cancer Association. Japanese gastric cancer treatment guidelines 2010 (ver. 3). Gastric Cancer 2011; 14: 113–23. 10 Japanese Gastric Cancer Association. Japanese classification of gastric carcinoma: 3rd English edition. Gastric Cancer 2011; 14: 101–12. 11 Katayama H, Kurokawa Y, Nakamura K, et al. Extended ClavienDindo classification of surgical complications: Japan Clinical Oncology Group postoperative complications criteria. Surg Today 2016; 46: 668–85. 12 Schoenfeld DA, Richter JR. Nomograms for calculating the number of patients needed for a clinical trial with survival as an endpoint. Biometrics 1982; 38: 163–70. 13 Oba K, Paoletti X, Alberts S, et al. Disease-free survival as a surrogate for overall survival in adjuvant trials of gastric cancer: a meta-analysis. J Natl Cancer Inst 2013; 105: 1600–07. 14 Lan KK, DeMets DL. Discrete sequential boundaries for clinical trials. Biometrika 1983; 70: 659–63. 15 Bando E, Makuuchi R, Irino T, Tanizawa Y, Kawamura T, Terashima M. Validation of the prognostic impact of the new tumornode-metastasis clinical staging in patients with gastric cancer. Gastric Cancer 2019; 22: 123–29. 16 Kim W, Kim HH, Han SU, et al. Decreased morbidity of laparoscopic distal gastrectomy compared with open distal gastrectomy for stage i gastric cancer: short-term outcomes from a multicenter randomized controlled trial (KLASS-01). Ann Surg 2016; 263: 28–35.

www.thelancet.com/gastrohep Published online November 19, 2019 https://doi.org/10.1016/S2468-1253(19)30332-2

9

Articles

17 Ramirez PT, Frumovitz M, Pareja R, et al. Minimally invasive versus abdominal radical hysterectomy for cervical cancer. N Engl J Med 2018; 379: 1895–904. 18 Shimada S, Sawada N, Ishiyama Y, et al. Impact of obesity on shortand long-term outcomes of laparoscopy assisted distal gastrectomy for gastric cancer. Surg Endosc 2018; 32: 358–66. 19 Tokunaga M, Tanizawa Y, Bando E, Kawamura T, Terashima M. Poor survival rate in patients with postoperative intra-abdominal infectious complications following curative gastrectomy for gastric cancer. Ann Surg Oncol 2013; 20: 1575–83. 20 Kubota T, Hiki N, Sano T, et al. Prognostic significance of complications after curative surgery for gastric cancer. Ann Surg Oncol 2014; 21: 891–98. 21 Wu XS, Wu WG, Li ML, et al. Impact of being overweight on the surgical outcomes of patients with gastric cancer: a meta-analysis. World J Gastroenterol 2013; 19: 4596–606. 22 Sun L, Zhao B, Huang Y, Lu H, Luo R, Huang B. Feasibility of laparoscopy gastrectomy for gastric cancer in the patients with high body mass index: a systematic review and meta-analysis. Asian J Surg 2019; published online April 29. DOI:10.1016/j.asjsur.2019.03.017. 23 Katai H, Mizusawa J, Katayama H, et al. Single-arm confirmatory trial of laparoscopy-assisted total or proximal gastrectomy with nodal dissection for clinical stage I gastric cancer: Japan Clinical Oncology Group study JCOG1401. Gastric Cancer 2019; 22: 999–1008.

10

24 Yu J, Huang C, Sun Y, et al. Effect of laparoscopic vs open distal gastrectomy on 3-year disease-free survival in patients with locally advanced gastric cancer: the CLASS-01 randomized clinical trial. JAMA 2019; 321: 1983–92. 25 Inaki N, Etoh T, Ohyama T, et al. A multi-institutional, prospective, phase ii feasibility study of laparoscopy-assisted distal gastrectomy with d2 lymph node dissection for locally advanced gastric cancer (JLSSG0901). World J Surg 2015; 39: 2734–41. 26 Lee HJ, Hyung WJ, Yang HK, et al. Short-term outcomes of a multicenter randomized controlled trial comparing laparoscopic distal gastrectomy with D2 lymphadenectomy to open distal gastrectomy for locally advanced gastric cancer (KLASS-02-RCT). Ann Surg 2019; published online Feb 9. DOI:10.1097/SLA.0000000000003217. 27 Morita S, Katai H, Saka M, Fukagawa T, Sano T, Sasako M. Outcome of pylorus-preserving gastrectomy for early gastric cancer. Br J Surg 2008; 95: 1131–35. 28 Oh SY, Lee HJ, Yang HK. Pylorus-preserving gastrectomy for gastric cancer. Gastric Cancer 2016; 16: 63–71.

www.thelancet.com/gastrohep Published online November 19, 2019 https://doi.org/10.1016/S2468-1253(19)30332-2