1 Introduction
Colorectal cancer (CRC) is the third most commonly diagnosed cancer in male and the second most cancer in females worldwide, and the fifth leading cause of cancer-related deaths in China.[1,2] Colonoscopy is considered the gold standard for screening CRC via detection and removal of adenomatous polyps.[3–5] Adenoma detection rate (ADR) was recommended as an important quality indicator for screening colonoscopy.[6] High-quality clinical evidence shows that interval CRC (iCRC) rate was inversely correlated with provider ADR.[7] In addition, ADR itself would be affected by not only age, gender, withdrawal time, quality of bowel preparation, but also the time of colonoscopy and endoscopists fatigue.[8–11] Recent studies have shown that a higher ADR could be attained during colonoscopy performed in the morning, which was due to longer withdrawal time, better bowel preparation and full attention of endoscopists.[11,12]
Over the years, computer-aided detection (CADe), particularly based on deep learning algorithm, in colonoscopy draws increasing attention. CADe allows automated detection of a given target, mostly colorectal polyps, during real-time colonoscopy.[13,14] Such a CADe system could assist endoscopists in real time to avoid miss diagnosis of polyps by means of providing visual alerts on polyps that endoscopist might have missed due to inexperience or distraction caused by fatigue.[15] The CADe system showed high sensitivity and accuracy even when been compared to expert human endoscopists.[16] We have developed a polyp detection CADe system and demonstrated its positive impact in increasing ADR during colonoscopy through randomized clinical trials.[17–19] However, whether CADe system could help to eliminate ADR gap in afternoon colonoscopy due to endoscopists fatigue has not been analyzed. Hence, the aim of this study was to evaluate whether ADR differs by the timing (morning vs afternoon) of colonoscopy with CADe system, and to investigate the relevant factors for ADR in CADe colonoscopy.
2 Method
2.1 Study design
We retrospectively collected data of consecutive patients who underwent CADe colonoscopy from September 2017 to February 2019 in the Endoscopy Center of Sichuan Provincial People's Hospital, China and accessed the data at January 2020. All patients were prepared with 2 to 4 L of polyethylene glycol with split-dose. We excluded patients with a history of inflammatory bowel disease (IBD), CRC, colorectal surgery, and patients with a contraindication for biopsy, procedures that cecum were not reached were also excluded. Cecum intubation was defined by visualization of the ileocecal valve and the appendiceal orifice. Colonoscopies were performed with high definition colonoscopies (Fujifilm EC-L590, EC-580, EC-590) (Fujifilm, Tokyo, Japan) and high-definition monitors.
The CADe system (EndoScreener, Shanghai Wision AI Co., Ltd. China) is a real-time automatic polyp detection system developed on a deep learning architecture. In a preliminary study, the system was validated to have a per-image sensitivity of 94.38%, per-image specificity of 95.92% and an area under the receiver operating characteristic curve of 0.984 to detect colon polyps in colonoscopy report images. In addition, the system was also validated to have a per-polyp sensitivity of 100.00% (per-image sensitivity of 91.64%) and a per-image specificity of 95.40% in real-world colonoscopy videos. The system processes >25 frames per second with a latency lesser than 80 ms, an imperceptible latency for most human endoscopists. The CADe system captures and analyzes the video stream from the endoscopy processor and displays alert boxes on an adjacent monitor.[18]
The study was approved by the Institutional Review Board of Sichuan Academy of Medical Sciences & Sichuan Provincial People's Hospital. All data in this study was de-identified.
2.2 Schedule
Morning (AM) colonoscopy was defined as colonoscopic procedures starting between 8:00 am and 12:00 noon. Afternoon (PM) colonoscopy was defined as procedures from 2:00 pm and thereafter. Data on patient age, sex, body mass index (BMI), insertion time, withdrawal time, biopsy time, indication for procedure, quality of bowel preparation, type of sedation, as well as histology of polyps were obtained from database of our center for this retrospective analysis. The quality of bowel preparation rated by the endoscopists was based on the Boston Bowel Preparation Score (BBPS). Bowel preparation was defined adequate when total score was greater than 6.
2.3 Statistical analysis
Continuous variables were reported as mean with standard deviation, while categorical factors were reported as percentages. As continuous factors, age, BMI, and BBPS were further dichotomized as categories for the further analysis. Continuous variables were compared using the 2-sample t test and categorical variables were compared using the chi-square test. ADR was estimated by calculating the percentage of procedures in which at least one adenoma was found, while polyp detection rate (PDR) was the percentage of colonoscopies with any type of polyp found. Univariate analysis was done to identify association between ADR and various other covariates. Multivariate analysis was also performed for ADR, using logistic regression while adjusting for possible extraneous variables. A 2-tailed P value of <.05 was considered statistically significant. All statistical analyses were performed in the Statistical Package for the Social Science (SPSS Inc., Chicago, IL, USA) software version 23.0.
3 Result
3.1 Baseline characteristics (Table 1)
A total of 509 consecutive patients were enrolled, among which, 25 patients were excluded during colonoscopy because of meeting exclusion criteria. Finally, 484 were analyzed in the study, 53.93% (261/484) were done in the am and 46.07% (223/484) were done in the pm. There were averagely 4 cases in am and 3 cases in pm.
Table 1 -
Baseline information of patients between
am and
pm groups.
| Characteristics |
|
All patients (n = 484) |
am (n = 261) |
pm (n = 223) |
P value∗
|
| Age (y) |
Mean (SD) |
49.15 (13.19) |
49.67 (13.32) |
48.56 (13.05) |
.358 |
| Gender |
Female, n (%) |
243 (50.21) |
136 (52.11) |
107 (47.98) |
.412 |
|
Male, n (%) |
241 (49.79) |
125 (47.89) |
116 (52.02) |
|
| BMI |
Mean (SD) |
23.02 (3.20) |
22.90 (3.30) |
23.16 (3.07) |
.379 |
|
<25, n (%) |
346 (71.49) |
191 (73.18) |
155 (69.51) |
.642 |
|
25≤ BMI≤ 30, n (%) |
129 (26.65) |
65 (24.90) |
64 (28.70) |
|
|
30>, n (%) |
9 (1.86) |
5 (1.92) |
4 (1.79) |
|
| Indication |
Symptomatic, n (%) |
82 (16.94) |
210 (80.46) |
192 (86.10) |
.114 |
|
Screening, n (%) |
402 (83.06) |
51 (19.54) |
31 (13.90) |
|
| Anesthesia |
Yes, n (%) |
478 (98.76) |
258 (98.85) |
220 (98.65) |
.846 |
|
No, n (%) |
6 (1.24) |
3 (1.15) |
3 (1.35) |
|
∗P value correspond to t-test or Pearson's χ2.BMI = body mass index.
The baseline characteristics are presented in Table 1. Mean age was 49.15 ± 13.19 years, and 50.21% of the patients were female. The overall mean BMI was 23.02 ± 3.20. 82 patients were for screening colonoscopies (16.94%). There were no significant differences in terms of demographic data and indications for colonoscopy between the am and pm group (P > .05).
3.2 Endoscopic characteristics (Table 2)
A total of 4 physicians participated in the study, including 1 senior endoscopist (>10,000 cases experience), 2 mid-level endoscopists and 1 junior endoscopist (<500 cases experience). There were no statistically significant difference between the 2 groups in terms of endoscopic procedures. There were no complications reported. BBPS was 6.89 ± 1.38 in the am and 6.74 ± 1.44 in the pm, there was no significant difference between the 2 groups in terms of BBPS (6.89 vs 6.74, P = .232). Supervised anesthesia (midazolam or propofol) was performed in 478 (98.76%) procedures. The mean insertion time in the am and pm were 5.93 ± 4.24 minutes and 5.18 ± 3.57 minutes (P = .037), while the withdrawal time (7.53 vs 7.39, P = .42) and withdrawal time excluding biopsy (7.00 vs 6.85, P = .335) were similar between the 2 groups.
Table 2 -
Endoscopic characteristics between
am and
pm groups.
| Characteristics |
|
All patients (n = 484) |
am (n = 261) |
pm (n = 223) |
P value∗
|
| Anesthesia |
Yes, n (%) |
478 (98.76) |
258 (98.85) |
220 (98.65) |
.846 |
|
No, n (%) |
6 (1.24) |
3 (1.15) |
3 (1.35) |
|
| BBPS |
Mean (SD) |
6.82 (1.41) |
6.89 (1.38) |
6.74 (1.44) |
.232 |
|
Inadequate (sum < 6.0 or anyone < 2.0), n (%) |
69 (14.26) |
31 (11.88) |
38 (17.04) |
.118 |
|
Adequate (sum ≥ 6.0 and everyone ≥ 2.0), n (%) |
415 (85.74) |
230 (88.12) |
185 (82.96) |
|
| Endoscopist experience |
Senior, n (%) |
94 (19.42) |
47 (18.01) |
47 (21.08) |
.696 |
|
Midlevel, n (%) |
290 (51.24) |
159 (60.92) |
131 (58.74) |
|
|
Junior, n (%) |
142 (29.34) |
55 (21.07) |
45 (20.18) |
|
| Total time |
Mean (SD) |
13.05 (4.61) |
13.46 (4.83) |
12.56 (4.31) |
.032 |
| Insertion time |
Mean (SD) |
5.58 (3.96) |
5.93 (4.24) |
5.18 (3.57) |
.037 |
| Withdrawal time |
Mean (SD) |
7.47 (2.02) |
7.53 (1.98) |
7.39 (2.07) |
.420 |
| Withdrawal time excluding biopsy |
Mean (SD) |
6.93 (1.70) |
7.00 (1.71) |
6.85 (1.70) |
.335 |
∗P value correspond to t-test or Pearson's χ2.BBPS = Boston Bowel Preparation Score.
3.3 ADR
A total of 501 polyps were detected, of which 59.68% (299/501) were adenomas. 54.52% adenomas (163/299) were found in the am and 45.48% (136/299) in the pm. The mean number of adenomas detected per colonoscopy (0.98 am vs 1.09 pm, P = .882) and ADR (0.36 am vs 0.35 pm, P = .849) were equilibrated in the am and pm group (P > .05) (Table 3).
Table 3 -
Polyp and adenoma detection.
|
am (n = 261) |
pm (n = 223) |
P value∗
|
| PDR |
0.51 |
0.53 |
.648 |
| ADR |
0.36 |
0.35 |
.849 |
| Mean Number of Detected Polyp |
0.98 |
1.09 |
.420 |
| Mean Number of Detected Adenoma |
0.62 |
0.61 |
.882 |
∗P value correspond to t-test or Pearson's χ2.ADR = adenoma detection rate, PDR = polyp detection rate.
In univariate analysis (Table 4), the age (P < .001), gender (P < .001), BBPS (P = .004), BMI (P = .041), withdrawal time (P < .001) showed statistically significant determinants of ADR. However, the logistic regression analysis showed that only age (P < .001), gender (P = .003), withdrawal time (P < .001) and withdrawal time excluding biopsy (P < .001) have significant influence on ADR (Table 5).
Table 4 -
Factors associated with adenoma detection: the univariate analysis.
| Characteristics |
|
Adenoma (n = 172) |
Non-adenoma (n = 312) |
P value∗
|
| Age |
|
|
|
P < .001 |
|
<50y, n (%) |
67 (38.95) |
179 (57.37) |
|
|
>50y, n (%) |
105 (61.05) |
133 (42.63) |
|
| Gender |
|
|
|
P < .001 |
|
Female, n (%) |
65 (37.79) |
178 (57.05) |
|
|
male, n (%) |
107 (62.21) |
134 (42.95) |
|
| BBPS |
Mean ± SD |
6.57 ± 1.29 |
6.96 ± 1.45 |
P = .004 |
|
Inadequate (sum < 6.0), n (%) |
88 (51.16) |
116 (37.18) |
P = .004 |
|
Adequate (sum ≥ 6.0), n (%) |
84 (48.84) |
196 (62.82) |
|
| Doctor |
Senior, n (%) |
40 (23.26) |
54 (17.31) |
P = .143 |
|
Midlevel, n (%) |
103 (59.88) |
187 (59.94) |
|
|
Junior, n (%) |
29 (16.86) |
71 (22.75) |
|
| BMI |
< 25, n (%) |
111 (64.53) |
235 (75.32) |
P = .041 |
|
25≤ BMI≤ 30, n (%) |
58 (33.72) |
71 (22.76) |
|
|
30>, n (%) |
3 (1.75) |
6 (1.92) |
|
| Time |
AM, n (%) |
94 (54.65) |
167 (53.53) |
P = .849 |
|
PM, n (%) |
78 (45.35) |
145 (46.47) |
|
| Insert time |
Mean ± SD |
5.43 ± 3.88 |
5.67 ± 4.01 |
P = .520 |
| Withdrawal time |
Mean ± SD |
8.73 ± 2.39 |
6.77 ± 1.36 |
P < .001 |
| Withdrawal time excluding biopsy |
Mean ± SD |
7.55 ± 2.17 |
6.59 ± 1.26 |
P < .001 |
∗P value correspond to t test or Pearson's χ2.BBPS = Boston Bowel Preparation Score, BMI = body mass index.
Table 5 -
Factors associated with adenoma detection: a multivariable logistic regression analysis.
|
OR |
95% Confidence interval |
P value∗
|
| Age |
1.03 |
1.02–1.05 |
.000 |
| Male |
2.08 |
1.28–3.41 |
.003 |
| BMI |
1.02 |
0.95–1.10 |
.640 |
| BBPS |
0.97 |
0.81–1.15 |
.695 |
| Withdrawal time |
4.12 |
2.79–6.08 |
.000 |
| Withdrawal time excluding biopsy |
0.36 |
0.24–0.53 |
.000 |
∗P value correspond to t-test or Pearson's χ2.BBPS = Boston Bowel Preparation Score, BMI = body mass index.
4 Discussion
CRC is a leading cause of cancer related death worldwide. Thanks to screening colonoscopy and the removal of adenomatous polyps, the mortality and incidence of CRC in adults have significantly decreased by 51% and 32%, respectively in the United States over the past half-century.[20] Conversely, high quality researches have demonstrated failure to detect and remove adenomatous polyps during colonoscopy may increase the risk of interval cancer. Michal et al performed a study of 42 interval cancers identified during a period of 188,788 person-years and showed an inverse correlation between the adenoma detection rate and the risk of interval cancer.[21] Similarly, Douglas AC and members also showed the inversely relationship between the ADR and subsequent risk of colorectal cancer, that is every 1% increase in ADR predicts a 3% reduction in CRC incidence and 5% reduction in mortality.[22] ADR is commonly acknowledged as a major indicator to assess colonoscopy quality.
However, ADR would be affected by a plenty of factors, such as age, gender, BMI which are patient-related and quality of bowel preparation, withdrawal time, endoscopist level which belongs to operator-related factors.[9,23,24] Consistent with previous study,[25] we found ADR was higher in men (OR 2.08, 95%CI 1.28–3.41, P = .003) and increased with age (OR 1.03, 95% CI 1.02–1.05, P < .001). Moreover, longer pure withdrawal time, withdrawal time excluding biopsy, was also correlated with higher ADR (OR 0.36, 95% CI 0.24–0.53, P < .001). However, we found no linear correlation between ADR and the quality of bowel preparation as well as BMI during CADe colonoscopy (P = .695 and P = .0.64, respectively) based on the enrolled population, which may due to limited sample size.
The timing of colonoscopy was considered as a proxy to endoscopists fatigue. Prior studies reported the timing of colonoscopy may affect polyp detection.[10–12,26] A Korean study showed that compared to non-fatigued endoscopists, the ADR was significantly decreased in fatigued endoscopists (42.6% vs 25.0%, P = .008), and the results suggested that endoscopist fatigue may decrease the effectiveness of colonoscopy.[27] Until now, whether colonoscopy was performed in the morning or in the afternoon would affect the ADR and colonoscopy quality remains controversial.[28–30] Teng et al[11] showed afternoon colonoscopies led to a significantly reduced ADR, which might due to shorter withdrawal time and insufficient endoscopists attentiveness. Gurudu et al[31] reported the afternoon ADR was significantly lower than the morning ADR only when the endoscopists worked the full-day. The study also showed that afternoon ADR could maintain at high level if colonoscopies were performed in half-day blocks by different endoscopists, it was believed attributable to the decline of physician fatigue. Christopher et al have showed the ADR was higher in the early morning and the number of polyps gradually decreased as the day progressed.[12] The results indicated the decreased ADR and decline of colonoscopy quality in the later of day may due to the endoscopist fatigue, inattention caused by repetitive colonoscopies.[32] Thanks to the breakthrough of artificial intelligence technology in computer vision, computer-aided detection (CADe) of colorectal polyps is now attracting increased attention. Several studies have indicated that CADe could overcome the limitation of human endoscopist regarding diagnostic capability and consistent attentiveness, high-performance CADe has been also demonstrated in clinical trials to increase ADR by indicating the presence and location of polyps during colonoscopy in real time, in order to draw the endoscopists attention to some easy-to-miss polyps that are visible in the screen.[17–19,33–35]
Notably, with the assistance of CADe system, we found no significant difference in ADR between afternoon and morning colonoscopies, the ADR was 36% in the morning and 35% in the afternoon (P = .849). Furthermore, the mean number of detected adenoma was also stabled between morning and afternoon colonoscopies (0.62 vs 0.61, respectively, P = .882). We believe the result is generalizable because the endoscopists worked full-time during the CADe trials with a representative workload, averagely 12 colonoscopy procedures, including 3 to 4 procedures of CADe colonoscopy, were performed per half day.
Moreover, we found the endoscopist experience did not influence the ADR during CADe colonoscopy (P = .143). Worth mentioning, there was no decline of withdrawal time in the afternoon, the mean withdrawal time in the am and pm were 7.53 ± 1.98 minutes and 7.39 ± 2.07 minutes (P = .420). This indicates the CADe system may have double advantages, first to overcome the inexperience of junior endoscopist, second to spur all endoscopist to do better, just like the supervision of senior endoscopist,[36] without increasing fatigue level.
Although this study was based on a Chinese population whose average age was younger than 50, the ADR was 36.6% which was beyond the suggestion of guidelines of American Society for Gastrointestinal Endoscopy (ASGE)/American College of Gastroenterology (ACG),[37,38] therefore, the result could be deemed generalizable to other regions with qualified ADR in colonoscopy.
There are some limitations in this study. First is the small sample size and single center data, the result base on a 484 population from 1 endoscopy center may lack external validity. Larger and multi-source samples are needed to further confirm this result. The second is the retrospective design in nature which might have selection bias and introduce imbalance in baseline characteristics in 2 groups, but logistic regression was used to control known potential confounders to minimize this limitation. Last, the timing of colonoscopy (am vs pm) was used as an indirect marker to represent physician fatigue, because it is very difficult to directly measure physician fatigue. However, this is a commonly acknowledged methodology.
In conclusion, we found that the time of day (am vs pm) does not have a significant impact on ADR in CADe colonoscopies, which indicated that CADe colonoscopy could be used as an effective quality assurance that helps to overcome the afternoon fatigue and help endoscoposits to maintain high ADR. Larger-scaled controlled studies as well as cost effective studies on CADe colonoscopy are needed to confirm the effectiveness and cost effectiveness of the CADe system. In addition, we confirmed the ADR of CADe colonocopy was associated with the age, gender, and withdrawal time.
Author contributions
Conceptualization: Shan Lei, Mengtian Tu, Xun Xiao, Liangping Li, Xiaogang Liu, Pu Wang.
Data curation: Zhilan Wang.
Formal analysis: Shan Lei.
Investigation: Mengtian Tu, Peixi Liu, Lei Lei, Xun Xiao, Guanyu Zhou.
Methodology: Shan Lei, Guanyu Zhou.
Resources: Liangping Li, Xiaogang Liu.
Supervision: Liangping Li, Xiaogang Liu.
Writing – original draft: Shan Lei, Pu Wang.
Writing – review & editing: Pu Wang.
References
[1]. Chen W, Sun K, Zheng R, et al. Cancer incidence and mortality in China, 2014. Chin J Cancer Res 2018;30:1.
[2]. Liu S, Zheng R, Zhang M, et al. Incidence and mortality of colorectal cancer in China, 2011. Chin J Cancer Res 2015;27:22.
[3]. Citarda F, Tomaselli G, Capocaccia R, et al. Efficacy in standard clinical practice of colonoscopic polypectomy in reducing colorectal cancer incidence. Gut 2001;48:812–5.
[4]. Doubeni CA, Corley DA, Quinn VP, et al. Effectiveness of screening
colonoscopy in reducing the risk of death from right and left colon cancer: a large community-based study. Gut 2018;67:291–8.
[5]. Winawer SJ, Zauber AG, Ho MN, et al. Prevention of colorectal cancer by colonoscopic polypectomy. The national polyp study workgroup. N Engl J Med 1993;329:1977–81.
[6]. Rex DK, Petrini JL, Baron TH, et al. Quality indicators for
colonoscopy. Gastrointest Endosc 2006;63:S16–28.
[7]. Lam AY, Li Y, Gregory DL, et al. Association between improved
adenoma detection rate and interval colorectal cancer rates after a quality improvement program. Gastrointest Endosc 2020;pii: S0016-5107(20)30143-7.
[8]. Kang HW, Kim D, Kim HJ, et al. Visceral obesity and insulin resistance are risk factors for colorectal adenoma: a cross-sectional, case-control study. Am J Gastroenterol 2010;105:178–87.
[9]. Wang H, Wang P, Liu X, et al. Factors predicting the colorectal
adenoma detection rate in colonoscopic screening of a Chinese population: a prospective study. Medicine (Baltimore) 2019;98:e15103.
[10]. Sanaka MR, Deepinder F, Thota PN, et al. Adenomas are detected more often in
morning than in
afternoon colonoscopy. Am J Gastroenterol 2009;104:1659–64.
[11]. Teng TY, Khor SN, Kailasam M, et al.
Morning colonoscopies are associated with improved adenoma detection rates. Surg Endosc 2016;30:1796–803.
[12]. Chan MY, Cohen H, Spiegel BM. Fewer polyps detected by
colonoscopy as the day progresses at a Veteran's Administration teaching hospital. Clin Gastroenterol Hepatol 2009;7:1217–23.
[13]. Mahmud N, Cohen J, Tsourides K, et al. Computer vision and augmented reality in gastrointestinal endoscopy. Gastroenterol Rep 2015;3:179–84.
[14]. Berzin TM, Topol EJ. Adding artificial intellegence to gastrointestinal endoscopy. Lancet 2020;395:485.
[15]. Krause M, Pillitteri D, Scholz T, et al. Computer-aided diagnosis for
colonoscopy. Endoscopy 2017;49:813–9.
[16]. Alagappan M, Brown JRG, Mori Y, et al. Artificial intelligence in gastrointestinal endoscopy: the future is almost here. World J Gastrointest Endosc 2018;10:239–49.
[17]. Wang P, Berzin TM, Glissen Brown JR, et al. Real-time automatic detection system increases colonoscopic polyp and adenoma detection rates: a prospective randomised controlled study. Gut 2019;68:1813–9.
[18]. Wang P, Xiao X, Glissen Brown JR, et al. Development and validation of a deep-learning algorithm for the detection of polyps during
colonoscopy. Nat Biomed Eng 2018;2:741–8.
[19]. Wang P, Liu X, Berzin TM, et al. Effect of a deep-learning
computer-aided detection system on adenoma detection during
colonoscopy (CADe-DB trial): a double-blind randomised study. Lancet Gastroenterol Hepatol 2020;5:343–51.
[20]. Burke C, Kaul V, Pohl H. Polyp resection and removal procedures: insights from the 2017 digestive disease week. Gastroenterol Hepatol 2017;13:1–24.
[21]. Kaminski MF, Regula J, Kraszewska E, et al. Quality indicators for
colonoscopy and the risk of interval cancer. N Engl J Med 2010;362:1795–803.
[22]. Corley DA, Jensen CD, Marks AR, et al.
Adenoma detection rate and risk of colorectal cancer and death. N Engl J Med 2014;370:1298–306.
[23]. Ju Seok Kim, Sun Hyung Kang, Hee Seok Moon, et al. Impact of bowel preparation quality on adenoma identification during
colonoscopy and optimal timing of surveillance. Digest Dis Sci 2015;60:3092–9.
[24]. Jover R, Zapater P, Bujanda L, et al. Endoscopist characteristics that influence the quality of
colonoscopy. Endoscopy 2016;48:241–7.
[25]. Adler A, Wegscheider K, Lieberman D, et al. Factors determining the quality of screening
colonoscopy: a prospective study on adenoma detection rates, from 12 134 examinations (Berlin
colonoscopy project 3, BECOP-3)[J]. Gut 2013;62:236–41.
[26]. Munson GW, Harewood GC, Francis DL. Time of day variation in polyp detection rate for colonoscopies performed on a 3-hour shift schedule. Gastrointest Endosc 2011;73:467–75.
[27]. Lee CK, Cha JM. Kim WJ.Endoscopist fatigue may contribute to a decline in the effectiveness of screening
colonoscopy. J Clin Gastroenterol 2015;49:e51–6.
[28]. Lurix E, Hernandez AV, Thoma M, et al.
Adenoma detection rate is not influenced by full-day blocks, time, or modified queue position. Gastrointest Endosc 2012;75:827–34.
[29]. Junqi W, Sheng-Bing Z, Shu-Ling W, et al. Comparison of efficacy of
colonoscopy between the
morning and
afternoon: a systematic review and meta-analysis. Digest Liver Dis 2018;50:661–7.
[30]. Freedman JS, Harari DY, Bamji ND, et al. The detection of premalignant colon polyps during
colonoscopy is stable throughout the workday. Gastrointest Endosc 2011;73:1197–206.
[31]. Suryakanth R, Shiva K, Jonathan A, et al.
Adenoma detection rate is not inflfluenced by the timing of
colonoscopy when performed in half-day blocks. Am J Gastroenterol 2011;106:1466–71.
[32]. Almario CV, Spiegel BMR. Does endoscopist fatigue impact
adenoma detection rate? A review of the evidence to date. Gastrointest Endosc 2017;85:611–3.
[33]. Shin-ei Kudo, Masashi Misawa, Yuichi Mori, et al. Artificial intelligence-assisted system improves endoscopic identification of colorectal neoplasms. Clin Gastroenterol Hepatol 2019;pii: S1542-3565(19)30997-8.
[34]. Misawa M, Kudo S, Mori Y, et al. Artificial intelligence-assisted polyp detection for
colonoscopy: initial experience. Gastroenterology 2018;154:2027–9.
[35]. Ahmad OF, Soares AS, Mazomenos E, et al. Artificial intelligence and computer-aided diagnosis in
colonoscopy: current evidence and future directions. Lancet Gastroenterol Hepatol 2019;4:71–80.
[36]. Desai A, Tyberg A, Zerbo S, et al. Monitoring
colonoscopy withdrawal time significantly improves the
adenoma detection rate and the performance of endoscopists. Endoscopy 2016;48:256–62.
[37]. Rex DK, Schoenfeld PS, Cohen J, et al. Quality indicators for
colonoscopy. Am J Gastroenterol 2015;110:72–90.
[38]. Rex DK, Schoenfeld PS, Cohen J, et al. Quality indicators for
colonoscopy. Gastrointest Endosc 2015;81:31–53.
