Abstract
Background and study aims Esophagogastroduodenoscopy (EGD) is an effective and important diagnostic tool to detect gastric cancer (GC). Although previous studies show that examiner, patient, and instrumental factors influence the detection of GC, we analyzed whether assigning a different examiner to surveillance EGD would improve the detection of GC compared to assigning the same examiner as in the previous endoscopy.
Patients and methods We retrospectively reviewed patients who underwent two or more consecutive surveillance EGDs at a single center between 2017 and 2019. We identified factors associated with GC detection using multivariable regression analysis and propensity-score matching.
Results Among 7794 patients, 99 GC lesions in 93 patients were detected by surveillance EGD (detection rate; 1.2 %), with a mean surveillance interval of 11.2 months. Among the detected 99 lesions, 87 (87.9 %) were curatively treated with endoscopy. There were no differences in the clinicopathologic characteristics of GC detected by the same or different endoscopists. GC detection in the group examined by different endoscopists was more statistically significant than in the group examined by the same endoscopist, even after propensity-score matching (1.6 % and 0.7 %; P < 0.05). Endoscopic experience and other factors were not statistically significant between the two groups.
Conclusions In surveillance EGD, having a different endoscopist for each exam may improve GC detection rates, regardless of the endoscopist’s experience.
Introduction
Gastric cancer (GC) is the fifth most common cancer and the third leading cause of cancer death worldwide 1 . Screening with esophagogastroduodenoscopy (EGD) leads to GC mortality reduction in a healthy population 2 3 4 . Previous studies have shown the usefulness of EGD in GC detection, and early GC treatment also reduces GC mortality (hazard ratio [HR] 0.51) 5 6 . In addition, surveillance EGD after early GC treatment is important. High metachronous GC rates with a cumulative 3-year incidence of 5.6 % have been reported 7 . Although a Japanese cohort study of annual surveillance EGD reported that 96.2 % of metachronous GCs were curable with endoscopic resection 7 , they can lead to death 8 . Therefore, long-term surveillance is important and an efficient endoscopic surveillance strategy is required.
To detect early GC with EGD and reduce GC mortality, it is necessary to improve the accuracy of EGD. Factors related to EGD accuracy include examiner, patient, and endoscopic instrument factors. There are many challenges in improving the quality of screening EGD. The rate of missed GC within 3 years of surveillance is high: 4.7 % as reported in a multicentric cohort study 9 , and 9.4 % to 11.3 % in a systematic review and meta-analysis 10 11 . An endoscopist’s experience can influence missed GC diagnoses. The false-negative rate of GC detection by EGD is approximately 11.3 % to 25.8 %, depending on the number of years of endoscopic experience: 32.4 % for examiners with < 10 years of experience and 19.5 % for examiners with ≥ 10 years 12 . To prevent false negatives in GC detection, a systematic screening protocol for the stomach (SSS) 13 , increasing the examination time 14 , and increasing examiner experience 15 16 are effective strategies.
The possibility of missing a diagnosis can be expected to an extent if the diagnosis is made by a lone physician. The collective intelligence of multiple physicians can improve diagnostic accuracy 17 . In colonoscopy, it was reported that adenoma detection rates are improved with nursing observation, or a “second set of eyes” 18 . Similarly, observation by multiple endoscopists may improve detection of GC with EGD, but the application methods and usefulness in surveillance EGD are not clear. We focused on the use of a “second set of eyes” per patient by assigning a different endoscopist for each follow-up examination. In this study, we analyzed examiner factors involved in GC detection with surveillance EGD and how the duration of experience and performance of the test by a same or different endoscopist each time affected the GC detection rate.
Patients and methods
Study cohort
We performed a retrospective cohort study at a single center (Sendai Kousei Hospital, Japan). Patients were included if they underwent two or more consecutive surveillance EGDs between March 2017 and October 2019. Patients with only one screening EGD, EGD for a closer examination of an already diagnosed cancer, or treatment EGD were excluded. To identify suitable cases, we reviewed endoscopy reports and electronic medical health records.
This study was conducted in accordance with STROBE guidelines and was approved by the Institutional Review Board of the Sendai Kousei Hospital (authorization number: 1–99). Informed consent was obtained comprehensively in an opt-out format.
Primary outcome
The primary outcome was GC detection, classified as detected or non-detected.
Variable measurements
We collected patient (age, sex, surveillance objective, Helicobacter pylori status) and endoscopy (sedative agent used, scope type, endoscopist, and time interval between EGD) characteristics. Mucosal atrophy was assessed endoscopically using the Kimura-Takemoto classification 19 . Information about mucosal atrophy was obtained from the initial endoscopy. The U portion included the cardia, fundus, and upper body; the M portion included the mid-body and lower body; and the L portion included the angle and antrum 20 . The macroscopic GC types were classified according to the Paris classification as type 0-I (protruded), type 0-IIa (superficially elevated), type 0-IIb (flat), type 0-IIc (superficially depressed), or type 0-IIa + IIc (elevated with a central depression) 21 .
Surveillance EGD setting
The Endoscopic Department involved in the study site specializes in endoscopic diagnosis and treatment of patients with cancer and mainly receives referrals from local physicians. Most surveillance EGDs performed here are for cancer screening purposes and were targeted for those performed at approximately annual intervals. Endoscopists are assigned patients randomly on their assigned examination days. This accounts for the variability of the endoscopist for each surveillance EGD of the same patient. Twenty-nine endoscopists were engaged during the study period. The endoscopists checked images from a patient’s record of previous EGDs before performing a new EGD. In principle, the endoscopists including trainees performed the EGD alone. We focused on the latest EGD. The two last EGDs were analyzed for patients with non-detected GC. The EGD immediately before and at the time of cancer detection was analyzed for patients with detected GC. We defined the “same endoscopist” group as those who had the same endoscopist for the last two EGDs and the “different endoscopist” group as those who had different endoscopists for the last EGDs, regardless of whether GC was detected or not.
Endoscope type
We used high-definition and non-high-definition endoscopes for surveillance EGD. The high-definition endoscopes were GIF-H290, GIF-H260Z, GIF-H290Z (Olympus, Tokyo, Japan), and EG-L600ZW7 (FUJIFILM, Tokyo, Japan). The non-high-definition endoscopes were GIF-XP260N, GIF-XP290N GIF-PQ260 (Olympus), and EG-L580NW7 (FUJIFILM). Evis Lucera Elite (Olympus) and Lasareo 7000 (FUJIFILM) were used as processors. Patients were sedated (using midazolam or propofol) during EGD based on patient request or the endoscopist’s decision for the patient’s safety and pain relief. Safety during sedation was ensured over time by biomonitoring. Endoscopists with <10 years or ≥ 10 years of endoscopic experience were defined as non-experts (n = 12) or experts (n = 17), respectively 12 . All expert endoscopists were certified by the Japan Gastroenterological Endoscopy Society, and non-expert endoscopists were not certified. The observation procedure was left to the endoscopist.
Pathological evaluation
Post-treatment specimens were evaluated according to the 14 th Japanese Classification of Gastric Carcinoma (3 rd English edition) 20 , and curability with endoscopic treatment was evaluated according to the Gastric Cancer Treatment Guidelines version 5, published in January 2018 22 . Tubular adenocarcinoma and papillary adenocarcinoma were considered differentiated cancers, while poorly differentiated adenocarcinoma and signet-ring cell carcinoma were considered undifferentiated cancers. Curability was categorized as curative [eCuraA (curative section) and eCuraB (expanded indication) with a low metastasis risk] or non-curative [eCuraC (non-curative resection) and surgical cases] 22 .
Statistical analysis
All statistical analyses were performed using JMP 13 (SAS Institute, Cary, North Carolina, United States). The studentʼs t -test was used to analyze continuous variables and the χ2 test or Fisherʼs exact test was used to analyze categorical variables. The factors affecting GC detection by surveillance EGD were assessed using univariate and multivariable analyses. In multivariable logistic regression analysis, variables considered to be associated with the detection of GC (age, sex, co-existence of atrophic gastritis, previous GC history, sedation usage, scope type, same or different endoscopist on EGD, and endoscopic experience) were included. Odds ratios (ORs) and 95 % confidence intervals (CIs) were determined for significant variables on both analyses. Propensity-score matching (PSM) was performed with logistic regression analysis for age, sex, surveillance objective (atrophic gastritis and GC history), sedation, scope, and the experience of the same and different endoscopists. The c-statistic (area under the curve) for evaluating the discrimination was calculated. PSM was performed using nearest-neighbor matching algorithm without replacement and 1:1 matching. A caliper width of 0.05 of the standard deviation of the logit of the PSM was used. P < 0.05 was considered statistically significant.
Results
Characteristics of patients on surveillance EGD
Of the 17,829 patients who received EGD during this study period, 7794 were included in this study ( Fig. 1 ). Table 1 lists patient characteristics (total patients, patients examined by the same endoscopist and by different endoscopist). Of the included patients, 4605 (59.1 %) were male, and 4839 (62.1 %) were aged ≥ 65 years. For surveillance purposes, 5040 patients (64.7 %) had atrophic gastritis and 547 (7.0 %) had a GC history. During the latest EGD, 5911 patients (75.8 %) were sedated, high-definition scopes were used in 5421 patients (69.6 %), and experts screened 3578 patients (45.9 %). In 6555 patients (79.1 %), an endoscopist different from the previous one performed the EGD. The mean surveillance interval was 12.1 months.
Fig. 1.
Flowchart of patient enrollment. ESD/EMR, endoscopic submucosal dissection/endoscopic mucosal resection; PEG, percutaneous endoscopic gastrostomy.
Table 1. Characteristics of patients on surveillance esophagogastroduodenoscopy.
| Characteristics | Total (n = 7794) | Same endoscopist (n = 1239) | Different endoscopist (n = 6555) |
| Mean age, years | 66.5 ± 12.0 | 71.0 ± 10.7 | 67.0 ± 12.0 |
| Age, n (%) | |||
|
2955 (37.9) | 292 (23.6) | 2663 (40.6) |
|
4839 (62.1) | 947 (76.4) | 3892 (59.4) |
| Sex, n (%) | |||
|
4605 (59.1) | 712 (57.5) | 3893 (59.4) |
|
3189 (40.9) | 527 (42.5) | 2662 (40.6) |
| Objective for surveillance, n (%) | |||
|
2196 (28.2) | 304 (24.5) | 1892 (28.9) |
|
179 (2.3) | 61 (4.9) | 118 (1.8) |
|
5040 (64.7) | 792 (63.9) | 4248 (64.8) |
|
1563 (20.1) | 342 (27.6) | 1221 (18.6) |
|
547 (7.0) | 137 (11.1) | 410 (6.3) |
|
492 (6.3) | 87 (7.0) | 405 (6.2) |
|
32 (0.4) | 11 (0.9) | 21 (0.3) |
|
686 (8.8) | 107 (8.6) | 579 (8.8) |
| Sedation, n (%) | |||
|
5897 (75.7) | 951 (76.8) | 4946 (75.5) |
|
1897 (24.3) | 288 (23.2) | 1609 (24.5) |
| Scope, n (%) | |||
|
5339 (68.5) | 1139 (91.9) | 4200 (64.1) |
|
2455 (31.5) | 100 (8.1) | 2355 (35.9) |
| Endoscopist, n (%) | |||
|
1239 (15.9) | ||
|
6555 (79.1) | – | – |
| Endoscopic experience, n (%) | |||
|
4711 (60.4) | 828 (66.8) | 3883 (59.2) |
|
3083 (39.6) | 411 (33.2) | 2672 (40.8) |
| Patients detected GC, n (%) | 93 (1.2) | 9 (0.7) | 84 (1.2) |
| Mean surveillance interval, months | 12.1 ± 3.9 | 12.0 ± 4.5 | 12.0 ± 3.7 |
GC, gastric cancer.
GC clinicopathologic characteristics detected by surveillance EGD
GC was detected in 93 patients (1.2 %) with 99 lesions. The GC characteristics are summarized in Table 2 . The mean age of those with detected GC was 72.9 ± 7.9 years, and the mean surveillance interval was 11.2 ± 2.3 months. Of the 93 detected patients, 74 (79.6 %) were male, 84 (90.4 %) were Helicobacter pylori -infected (current and previously infected combined), and 61 (65.6 %) had severe atrophic mucosa of the stomach. Of the 99 lesions, 64 (64.6 %) were the depressed macroscopic type (0-IIc), 94 (94.9 %) were differentiated, and the median tumor size was 10 mm (range, 3–45 mm). The tumor locations were relatively evenly distributed. Eighty-six (86.9 %) lesions were intramucosal carcinomas, and eighty-seven (87.9 %) were curatively resected by endoscopy, although one patient had an advanced cancer which invaded the subserosa.
Table 2. Clinicopathologic characteristics of gastric cancer detected by surveillance esophagogastroduodenoscopy.
| Characteristics | Total 99 lesions (n = 93) | Same endoscopist 9 lesions (n = 9) | Different endoscopist 90 lesions (n = 84) | P value |
| Mean age, years | 72.9 ± 7.9 | 75 ± 8.6 | 73 ± 7.9 | 0.50 |
| Males, n (%) | 74 (79.6) | 7 (77.8) | 67 (79.8) | 1 |
| Helicobacter pylori status 1 , n (%) | ||||
|
14 (15.1) | 2 (22.2) | 12 (14.3) | 0.59 |
|
70 (75.3) | 6 (66.7) | 64 (76.2) | |
|
9 (9.7) | 1 (11.1) | 8 (9.5) | |
| Atrophy gastritis, n (%) | ||||
|
3 (3.2) | 0 (0.0) | 3 (3.6) | 0.64 |
|
28 (30.1) | 4 (44.4) | 24 (28.6) | |
|
61 (65.6) | 5 (55.6) | 56 (66.7) | |
|
1 (1.1) | 0 (0.0) | 1 (1.2) | |
| Tumor location, n (%) | ||||
|
18 (18.2) | 1 (11.1) | 17 (18.9) | 0.81 |
|
35 (35.3) | 4 (44.4) | 31 (34.4) | |
|
46 (46.5) | 4 (44.4) | 42 (46.7) | |
| Tumor circumference, n (%) | ||||
|
13 (13.2) | 0 (0.0) | 13 (14.4) | 0.50 |
|
33 (33.3) | 2 (22.2) | 31 (34.4) | |
|
33 (33.3) | 4 (44.4) | 29 (32.2) | |
|
20 (20.2) | 3 (33.3) | 17 (18.9) | |
| Macroscopic type, n (%) | ||||
|
1 (1.0) | 0 (0.0) | 1 (1.1) | < 0.05 |
|
22 (22.2) | 2 (22.2) | 20 (22.2) | |
|
7 (7.1) | 2 (42.2) | 5 (5.6) | |
|
64 (64.6) | 3 (33.3) | 61 (67.8) | |
|
5 (5.1) | 2 (22.2) | 3 (3.3) | |
| Histologic type, n (%) | ||||
|
94 (94.9) | 9 (100.0) | 85 (94.4) | 1 |
|
5 (5.1) | 0 (0.0) | 5 (5.6) | |
| Median tumor size, mm (range) | 10 (3–45) | 10 (7–28) | 9 (3–45) | 0.15 |
| Depth, n (%) | ||||
|
86 (86.9) | 8 (88.9) | 78 (86.7) | 1 |
|
2 (2.0) | 0 (0.0) | 2 (2.2) | |
|
10 (10.1) | 1 (11.1) | 9 (10.0) | |
|
0 (0.0) | 0 (0.0) | 0 (0.0) | |
|
1 (1.0) | 0 (0.0) | 1 (1.1) | |
| Curability of treatment, n (%) | ||||
|
87 (87.9) | 8 (88.9) | 79 (87.8) | 1 |
|
12 (12.1) | 1 (11.1) | 11 (12.2) | |
| Mean surveillance interval, months | 11.2 ± 2.3 | 12.0 ± 3.9 | 12.0 ± 2.4 | 0.36 |
M, mucosa; SM, submucosa; MP, muscularis propria; SS, subserosa.
SM1 is an invasion depth of less than 500 μm into the submucosa and SM2 is 500 μm or more.
Current: currently infected with Helicobacter pylori ; Post: After H. pylori eradication
Comparison clinicopathologic characteristics of detected GC between the same endoscopist and different endoscopist
The characteristics of GCs detected from EGDs performed by the same endoscopist compared to those performed by a different endoscopist are shown in Table 2 . Of the total 99 lesions, the same endoscopists detected nine GCs (9.1 %) and different endoscopists detected 90 GCs (90.9 %). Characteristics of GC detected were compared between the two groups. There were no statistically significant differences in age, sex, H. pylori status, atrophic gastritis, tumor location, tumor circumference, histologic type, tumor size, depth, curability of treatment, and surveillance interval. Only the macroscopic type of tumor showed significant differences ( P < 0.05).
Factors associated with GC detection in surveillance EGD
The logistic analysis of factors associated with GC detection in surveillance EGD is shown in Table 3 . In univariate logistic analysis, age ≥ 65 years (OR, 4.61; 95 % CI, 2.45–8.67), male sex (OR, 2.72; 95 % Cl, 1.64–4.52), atrophic gastritis (OR, 4.12; 95 % Cl, 2.19–7.75), GC history (OR, 2.18; 95 % Cl, 1.21–3.94), and high-definition scope (OR, 5.45; 95 % Cl, 2.52–11.8) were significant factors of GC detection by surveillance EGD. Conversely, there were no significant differences in sedation usage, fixed/different endoscopist, or endoscopic experience. In multivariable logistic analysis, it was suggested that a different endoscopist (OR, 2.47; 95 % Cl, 1.21–4.93), age ≥ 65 years (OR, 2.84; 95 % Cl, 1.47–5.47), male sex (OR, 2.83; 95 % Cl, 1.69–4.74), atrophic gastritis (OR, 2.79; 95 % Cl, 1.47–5.30), and high-definition scope (OR, 3.82; 95 % Cl, 1.69–8.60) were significant factors associated with GC detection by surveillance EGD. Sedation usage and endoscopic experience were not significantly different in the multivariable analysis.
Table 3. Factors associated with gastric cancer detection in surveillance esophagogastroduodenoscopy.
| Univariate Analysis OR (95 % CI) | P value | Multivariable analysis OR (95 % CI) | P value | |
| Age | ||||
|
Reference | < 0.05 | Reference | < 0.05 |
|
4.61 (2.45–8.67) | 2.84 (1.47–5.47) | ||
| Sex | ||||
|
2.72 (1.64–4.52) | < 0.05 | 2.83 (1.69–4.74) | < 0.05 |
|
Reference | Reference | ||
| Coexistence of atrophic gastritis | ||||
|
4.12 (2.19–7.75) | < 0.05 | 2.79 (1.47–5.30) | < 0.05 |
|
Reference | Reference | ||
| Previous history of GC | ||||
|
2.18 (1.21–3.94) | < 0.05 | 1.31 (0.72–2.41) | 0.37 |
|
Reference | Reference | ||
| Sedation | ||||
|
1.16 (0.71–1.92) | 0.63 | 1.18 (0.71–1.99) | 0.52 |
|
Reference | Reference | ||
| Scope | ||||
|
5.45 (2.52–11.8) | < 0.05 | 3.82 (1.69–8.60) | < 0.05 |
|
Reference | Reference | ||
| Endoscopist | ||||
|
Reference | 0.10 | Reference | < 0.05 |
|
1.77 (0.89–3.54) | 2.47 (1.21–4.93) | ||
| Endoscopic experience | ||||
|
1.06 (0.70–1.59) | 0.83 | 1.22 (0.80–1.85) | 0.35 |
|
Reference | Reference | ||
OR, odds ratio; CI, confidence interval; GC, gastric cancer.
Analysis after PSM
The comparison of patient characteristics between the different endoscopist and same endoscopist groups after PSM analysis is shown in Table 4 . We compared 1235 patients whose surveillance EGDs were performed by a different endoscopist with control patients whose EGDs were performed by a same endoscopist. They were adjusted using PSM based on covariates of age, sex, atrophic gastritis, GC history, sedation usage, scope type, and endoscopic experience (c statics = 0.765). There were 9 (0.7 %) GCs detected by same endoscopists and 20 (1.6 %) detected by different endoscopists ( P = 0.04, χ2 test). All matched factors after PSM were well-balanced in this study. Age, atrophic gastritis, sedation usage, type of scope, and endoscopic experience were not significantly different in the univariate analysis after PSM.
Table 4. Characteristics of patients after propensity-score matching analysis to compare the use of a same endoscopist and different endoscopists.
| Characteristics | Same endoscopist (n = 1235) | Different endoscopist (n = 1235) | P value |
| Age, n (%) | |||
|
292 (23.6) | 291 (23.6) | 0.96 |
|
943 (76.4) | 944 (76.4) | |
| Sex, n (%) | |||
|
709 (57.4) | 706 (57.2) | 0.90 |
|
526 (42.6) | 529 (42.8) | |
| Coexistence of atrophic gastritis, n (%) | |||
|
792 (64.1) | 795 (64.4) | 0.90 |
|
443 (35.9) | 440 (35.6) | |
| Previous history of GC, n (%) | |||
|
133 (10.8) | 132 (10.7) | 0.95 |
|
1102 (89.2) | 1103 (89.3) | |
| Detected GC, n (%) | 9 (0.7) | 20 (1.6) | < 0.05 |
| Sedation, n (%) | |||
|
973 (78.8) | 975 (79.0) | 0.92 |
|
262 (21.2) | 260 (21.0) | |
| Scope, n (%) | |||
|
1147 (92.9) | 1148 (93.0) | 0.94 |
|
88 (7.1) | 87 (7.0) | |
| Endoscopic experience, n (%) | |||
|
812 (65.7) | 811 (65.7) | 0.97 |
|
423 (34.3) | 424 (34.3) | |
GC, gastric cancer.
Discussion
In this study, we identified the key factors for GC detection in view of surveillance EGD by analyzing results of more than 7,000 EGD screenings. The frequency of GC detection was higher when different endoscopists performed consecutive EGDs than when the same endoscopist performed them on a given patient. The present study suggests that assigning a different examiner each year might be an effective and simple strategy for improving GC detection through surveillance EGDs without additional cost.
GC detected by surveillance EGD
The protocol for surveillance EGD for GC recommends annual follow-up 7 , and GC detected within the follow-up interval is assumed to have been missed. The GC detection rate in screening EGD is reported to be 0.91 % 14 . In a meta-analysis, it was reported that the median occurrence of metachronous cancer was 2.9 % in H. pylori -positive individuals and 1.2 % after eradication therapy 23 . The rate of metachronous cancer after endoscopic treatment was reported as 2.0 % per year 8 . We had a GC detection rate of 1.2 %, which was similar to that of a previous study supporting surveillance EGD.
GC detection is often difficult due to the inflammation caused by H. pylori in the background mucosa and epithelium with low-grade atypia on the surface of GC after H. pylori eradication 24 . Although the average size of missed cancers has not been mentioned in many studies 9 10 11 25 26 , Jin et al. reported that 87.4 % of GCs < 2 cm were detected by EGD within 2 years 27 , and Abe et al. reported that a missed cancer is usually small (< 20 mm) and differentiated intramucosal cancer 28 . Similarly, heterotopic multiple gastric carcinomas had a reported median size of 10 mm (range, 1–50 mm) on an annual surveillance EGD 8 . In a Japanese cohort study, 95 % of GCs detected by annual surveillance EGD were endoscopically resectable 8 . In our study, surveillance EGD at a mean interval of 12 months detected a median size of 10 mm (range, 3–45 mm) as small, and 87.9 % of the patients were endoscopically curable. As previously reported, endoscopic accuracy was maintained. We recommend annual surveillance in high-risk populations for early detection and treatment. As indicated in Table 2 , the only significantly different characteristic of GC detected in the same endoscopist and different endoscopist was the macroscopic type. This suggests that the factors contributing to GC detection may be related to factors in the endoscopic setting, rather than the GC itself. These 99 GCs had small tumors with a median size of 10 mm (range 3–45 mm). We retrospectively evaluated previous EGD images, but many of them could not be detected or were not in the EGD images. However, some of the large lesions and those that were deeply invasive GCs could be identified in the previous EGD images. This suggested that whether or not the presence of GCs was detected might be largely due to factors on the endoscopistʼs side, such as the observation procedure.
Improving the GC detection rate
The quality of EGD influences GC detection. There is limited evidence concerning EGD quality improvement methods. In a previous study, SSS 13 , training 15 , endoscopist experience 12 16 , longer examination time 14 29 , and the use of high-definition scopes 10 were reported as ways to increase the GC detection rate. Generally, the collective intelligence of multiple physicians can improve the diagnostic accuracy 17 . In colonoscopy, it was reported that adenoma detection rates are improved with multiple observations 18 . To the best of our knowledge, there have been few studies on whether same or different endoscopists contribute to the GC detection rate in surveillance EGD. In the present study, we found that changing the endoscopist remained a contributing factor toward GC detection. After PSM, the use of a different endoscopist had a significantly higher GC detection rate (1.6 %) than with a same endoscopist (0.7 %), indicating that changing the endoscopist can be a simple and effective way of increasing the GC detection rate in EGD screening. Occasionally, other lesions are detected during a repeat endoscopy of a GC lesion referred from another hospital. This may be due to the differences in endoscopic observation procedures, time, and the fact that some people are better or worse at observing certain parts, in addition to differences in modality. In this study, there were no significant differences in the characteristics of GCs detected among the two groups. We suspect that the different endoscopist group had a better detection rate for GC than the same endoscopist group because having a different endoscopist may correct for individual differences in endoscopic observation techniques, including procedures, time, and skill. However, depending on the number of endoscopists, the size of the institution, healthcare resources, and patient backgrounds, there may be no choice but to use the same endoscopist. Therefore, changing the endoscopist may not be applicable to all examinations and facilities. In situations where EGD with a change in endoscopist is possible, it may be recommended to improve the detection rate without additional costs.
In surveillance EGD, significant factors related to GC detection in multivariable analysis were age (≥ 65 years), male sex, background mucosal atrophy, use of a high-definition scope, and the use of different endoscopists. Older age (≥ 65 years), male sex, and background atrophic mucosa are known risk factors for GC 8 30 . The use of high-definition scopes has been reported to contribute to GC detection 26 , and this study supports those results. When a background-matched PSM analysis was performed to eliminate the influence of these factors, the different endoscopist group had a significantly higher GC detection rate. We consider this to be a simple and useful method that contributes to improved GC detection.
Park et al. reported that highly experienced endoscopists performed the exam for GC detection 26 , but in a multicenter cohort study, the endoscopistsʼ experience did not make a difference in GC detection 9 . In this study, the examiner’s expertise was not a factor that contributed GC detection. The lack of a significant difference in GC detection rates between the expert and non-expert examiners even after PSM may have been due to the inappropriate separation of the two groups using 10 years of experience as a cut-off. However, it is possible for non-experts to detect GCs, and we considered the possibility that individual differences in observation procedures, time, and individual peculiarities may have caused the difference, more than the cut-off problem. In a national Delphi survey of endoscopy education, 10 competencies have been identified that are essential for endoscopy education 31 . One of the essential competencies is that the experts should be able to control the procedure when trainees are unable to progress. This competency is related to the fact that collective intelligence can improve the diagnostic accuracy 17 . Regardless of endoscopic experience, our results suggest that trainees may be able to perform an examination without reducing the detection rate by changing the endoscopist for each exam with appropriate education. This simple method may also contribute to endoscopy education.
Limitations
This study has several limitations. First, this was a single-center, retrospective study. Second, although there was no significant difference, there were more EGDs performed by non-experts because of endoscopic training. Third, the ability to detect GC was not quantifiable. Based on a previous study 12 , we divided all examiners into two categories: non-experts with < 10 years of experience and experts with ≥ 10 years of experience. There is room for reconsideration of how to divide the two categories, as both trainees and experts with more than 20 years of experience were included. A particular problem was that non-experts included a small number of trainees who needed supervisors for endoscopy, as well as endoscopists with skills close to those of experts. Fourth, the background of patients for surveillance EGD included other cases with factors other than atrophic gastritis, intestinal epithelialization 32 , and history of GC 7 , which are commonly referred to as high-risk factors of GC. In addition, smoking 33 , genetic background and family history 34 , and some H. pylori statuses were not investigated. If the focus is only on the occurrence of GC, the number of cases should be further narrowed (e. g., excluding cases with mild atrophy). Fifth, although atrophic gastritis was determined endoscopically using the Kimura-Takemoto classification 19 , pathological evaluation should have been performed using the Sydney classification 35 to accurately assess the presence or absence of severe atrophy or intestinal epithelialization, which is a high-risk factor for GC. Sixth, all endoscopic images could not be reevaluated for mucosal atrophy by a single endoscopist; therefore, the study could not be tailored to the GC risk of mucosal atrophy grade.
Conclusions
In conclusion, on performing surveillance EGD every 12 months, GC was detected in 1.2 % of patients. Regardless of endoscopic experience, changing the endoscopist for each exam may contribute to improving the GC detection rate by correcting individual differences.
Acknowledgments
The authors thank the endoscopy staffs for their assistance in our daily endoscopic examinations. They also thank Editage (www.editage.com) for English language editing.
Footnotes
Competing interests The authors declare that they have no conflict of interest.
References
- 1.Sung H, Ferlay J, Siegel R L et al. Global Cancer Statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71:209–249. doi: 10.3322/caac.21660. [DOI] [PubMed] [Google Scholar]
- 2.Jun J K, Choi K S, Lee H Y et al. Effectiveness of the Korean National Cancer Screening Program in reducing gastric cancer mortality. Gastroenterology. 2017;152:1319–1.328E10. doi: 10.1053/j.gastro.2017.01.029. [DOI] [PubMed] [Google Scholar]
- 3.Hamashima C, Ogoshi K, Narisawa R et al. Impact of endoscopic screening on mortality reduction from gastric cancer. World J Gastroenterol. 2015;21:2460–2466. doi: 10.3748/wjg.v21.i8.2460. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Matsumoto S, Yoshida Y. Efficacy of endoscopic screening in an isolated island: A case-control study. Indian J Gastroenterol. 2014;33:46–49. doi: 10.1007/s12664-013-0378-2. [DOI] [PubMed] [Google Scholar]
- 5.Tsukuma H, Oshima A, Narahara H et al. Natural history of early gastric cancer: a non-concurrent, long term, follow up study. Gut. 2000;47:618–621. doi: 10.1136/gut.47.5.618. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Ezoe Y, Muto M, Uedo N et al. Magnifying narrowband imaging is more accurate than conventional white-light imaging in diagnosis of gastric mucosal cancer. Gastroenterology. 2011;141:2017–2025. doi: 10.1053/j.gastro.2011.08.007. [DOI] [PubMed] [Google Scholar]
- 7.Nakajima T, Oda I, Gotoda T et al. Metachronus gastric cancers after endoscopic resection: How effective is annual endoscopic surveillance? Gastric Cancer. 2006;9:93–98. doi: 10.1007/s10120-006-0372-9. [DOI] [PubMed] [Google Scholar]
- 8.Abe S, Oda I, Suzuki H et al. Long-term surveillance and treatment outcomes of metachronous gastric cancer occurring after curative endoscopic submucosal dissection. Endoscopy. 2015;47:1113–1118. doi: 10.1055/s-0034-1392484. [DOI] [PubMed] [Google Scholar]
- 9.Hernanz N, Rodríguez de Santiago E, Marcos Prieto H M et al. Characteristics and consequences of missed gastric cancer: A multicentric cohort study. Dig Liver Dis. 2019;51:894–900. doi: 10.1016/j.dld.2019.02.006. [DOI] [PubMed] [Google Scholar]
- 10.Pimenta-Melo A R, Monteiro-Soares M, Libânio D et al. Missing rate for gastric cancer during upper gastrointestinal endoscopy: A systematic review and meta-Analysis. Eur J Gastroenterol Hepatol. 2016;28:1041–1049. doi: 10.1097/MEG.0000000000000657. [DOI] [PubMed] [Google Scholar]
- 11.Menon S, Trudgill N. How commonly is upper gastrointestinal cancer missed at endoscopy? A meta-analysis. Endosc Int Open. 2014;02:E46–E50. doi: 10.1055/s-0034-1365524. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Hosokawa O, Hattori M DK et al. Difference in accuracy between gastroscopy and colonoscopy for detection of cancer. Hepatogastroenterology. 2007;54:442–444. [PubMed] [Google Scholar]
- 13.Yao K. The endoscopic diagnosis of early gastric cancer. Ann Gastroenterol. 2013;26:11–22. [PMC free article] [PubMed] [Google Scholar]
- 14.Kawamura T, Wada H, Sakiyama N et al. Examination time as a quality indicator of screening upper gastrointestinal endoscopy for asymptomatic examinees. Dig Endosc. 2017;29:569–575. doi: 10.1111/den.12804. [DOI] [PubMed] [Google Scholar]
- 15.Zhang Q, Chen Z Y, Chen di C et al. Training in early gastric cancer diagnosis improves the detection rate of early gastric cancer: An observational study in China. Medicine (United States) 2015;94:e384. doi: 10.1097/MD.0000000000000384. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Yamazato T, Oyama T, Yoshida T et al. Two years’ intensive training in endoscopic diagnosis facilitates detection of early gastric cancer. Intern Med. 2012;51:1461–1465. doi: 10.2169/internalmedicine.51.7414. [DOI] [PubMed] [Google Scholar]
- 17.Barnett M L, Boddupalli D, Nundy S et al. Comparative accuracy of diagnosis by collective intelligence of multiple physicians vs individual physicians. JAMA Netw Open. 2019;2:e190096. doi: 10.1001/jamanetworkopen.2019.0096. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Aslanian H R, Shieh F K, Chan F W et al. Nurse observation during colonoscopy increases polyp detection: A randomized prospective study. Am J Gastroenterol. 2013;108:166–172. doi: 10.1038/ajg.2012.237. [DOI] [PubMed] [Google Scholar]
- 19.Kimura K, Takemoto T. An endoscopic recognition of the atrophic border and its significance in chronic gastritis. Endoscopy. 1969;1:87–97. [Google Scholar]
- 20.Sano T, Kodera Y. Japanese classification of gastric carcinoma: 3rd English edition. Gastric Cancer. 2011;14:101–112. doi: 10.1007/s10120-011-0041-5. [DOI] [PubMed] [Google Scholar]
- 21.Lambert R LC. The Paris Endoscopic Classification of Superficial neoplastic lesions. Gastrointest Endosc. 2003;58:3–43. doi: 10.1016/s0016-5107(03)02159-x. [DOI] [PubMed] [Google Scholar]
- 22.Komatsu S, Otsuji E. Essential updates 2017/2018: Recent topics in the treatment and research of gastric cancer in Japan. Ann Gastroenterol Surg. 2019;3:581–591. doi: 10.1002/ags3.12284. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Sugimoto M, Murata M, Yamaoka Y. Chemoprevention of gastric cancer development after Helicobacter pylori eradication therapy in an East Asian population, Meta-analysis. World J Gastroenterol. 2020;21:1820–1840. doi: 10.3748/wjg.v26.i15.1820. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Kitamura Y, Ito M, Matsuo T et al. Characteristic epithelium with low-grade atypia appears on the surface of gastric cancer after successful Helicobacter pylori eradication therapy. Helicobacter. 2014;19:289–295. doi: 10.1111/hel.12132. [DOI] [PubMed] [Google Scholar]
- 25.Delgado G PG, Morales A VJ, Jimeno R M et al. Gastric cancer missed at esophagogastroduodenoscopy in a well-defined Spanish population. Dig Liver Dis. 2019;51:1123–1129. doi: 10.1016/j.dld.2019.03.005. [DOI] [PubMed] [Google Scholar]
- 26.Park M S, Yoon J Y, Chung H S et al. Clinicopathologic characteristics of interval gastric cancer in Korea. Gut Liver. 2015;9:167–173. doi: 10.5009/gnl13425. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Jin S, Jeon S W, Kwon Y. Optimal endoscopic screening interval for early detection of gastric cancer: a single-center study. J Korean Med Sci. 2018;33:1–1.0E167. doi: 10.3346/jkms.2018.33.e166. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Abe S, Oda I, Minagawa T et al. Metachronous gastric cancer following curative endoscopic resection of early gastric cancer. Clin Endosc. 2018;51:253–259. doi: 10.5946/ce.2017.104. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Park J M, Huo S M, Lee H H et al. Longer observation time increases proportion of neoplasms detected by esophagogastroduodenoscopy. Gastroenterology. 2017;153:460–4690. doi: 10.1053/j.gastro.2017.05.009. [DOI] [PubMed] [Google Scholar]
- 30.Asaka M, Kato M, Takahashi S I et al. Guidelines for the management of helicobacter pylori infection in Japan: 2009 revised edition. Helicobacter. 2010;15:1–20. doi: 10.1111/j.1523-5378.2009.00738.x. [DOI] [PubMed] [Google Scholar]
- 31.Kumar N L, Smith B N, Lee L S et al. Best practices in teaching endoscopy based on a delphi survey of gastroenterology program directors and experts in endoscopy education. Clin Gastroenterol Hepatol. 2020;18:574–5790. doi: 10.1016/j.cgh.2019.05.023. [DOI] [PubMed] [Google Scholar]
- 32.Kamada T, Haruma K, Ito M et al. Time trends in Helicobacter pylori infection and atrophic gastritis over 40 years in Japan. Helicobacter. 2015;20:192–198. doi: 10.1111/hel.12193. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Ladeiras-Lopes R, Pereira A K, Nogueira A et al. Smoking and gastric cancer: Systematic review and meta-analysis of cohort studies. Cancer Causes Control. 2008;19:689–701. doi: 10.1007/s10552-008-9132-y. [DOI] [PubMed] [Google Scholar]
- 34.Yaghoobi M, Bijarchi R, Narod S A. Family history and the risk of gastric cancer. Br J Cancer. 2010;102:237–242. doi: 10.1038/sj.bjc.6605380. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Dixon M F, Genta R M, Yardley J H CP. Classification and grading of gastritis. The updated Sydney System. International Workshop on the Histopathology of Gastritis, Houston 1994. Am J Surg Pathol. 1996;20:1161–1181. doi: 10.1097/00000478-199610000-00001. [DOI] [PubMed] [Google Scholar]