USING COMPUTER-AIDED POLYP DETECTION SYSTEM(CADE) TO MAINTAIN THE HIGH QUALITY IN ADENOMA RATE DURING COMMUNITY-BASED COLORECTAL CANCER SCREENING IN THAILAND: A RANDOMIZED TRIAL

Introduction: Due to limited endoscopy resources and personnel in Thailand, a high-volume colonoscopy (50-100 colonoscopies/day) performed within a few days by a group of community endoscopists gathering from different centers has been utilized to clear stranded patients with FIT positive especially after COVID-19 pandemic. Because of the rush in procedure time to achieve a high turnover rate (>15 colonoscopies/endoscopist/day), these endoscopists may develop fatigue. This in turn may contribute to a decline in effectiveness. Computer-aided polyp detection system (CADe) has been shown to improve ADR and could help to maintain the adenoma detection rate (ADR) and other indices in this situation. Therefore, we aimed to evaluate the benefit of CADe guidance colonoscopy compared to colonoscopy without CADe in high-volume-colonoscopy setting.
Methods: At two community-endoscopy centers, all FIT-positive individuals aged 50-75 during 3-10/2022 were recruited for a high-volume CRC screening. All subjects were randomly assigned to undergo colonoscopy performed with 8 days with or without CADe guidance in a 1:1 ratio. Twenty-eight endoscopists participated in the study. Every endoscopist was assigned to perform colonoscopy in both CADe and control groups. High definition colonoscopes without magnification were used in both groups. The image enhancement mode was not used in polyp detection. The CADe system with real-time bounding box notifications and voice alarms projected on the monitor was activated before colonoscope insertion until finish. The primary outcome was the ADR. Secondary outcomes were the proximal adenoma detection rate (pADR), advanced adenoma detection rate (AADR), and the number of adenomas/proximal adenomas/advanced adenomas per colonoscopy (APC, pAPC, and AAPC, respectively). The differences in measured outcomes were compared to controls using a linear regression model.
Results: A total of 467 participants (34.7% male) with a mean age of 59.9±6.4 were randomized to CADe (n=226) and control (n=241). Patient and procedural characteristics were comparable between the two groups (Table 1). The mean historical ADR of participated endoscopists was 37.9±7.54. The overall ADR in the control group dropped to 27.8% while the ADR in the CADe group was 40.3% (p=0.004). CADe also showed significantly higher mean APC when compared to that of from the controls (0.64 vs. 0.42; p=0.01). However, CADe showed no statistically significant differences in the pADR (11.9% vs. 7.9%; p=0.14), AADR (12.4% vs. 10%; p=0.41), pAPC (0.15 vs. 0.10; p=0.21), and AAPC (0.12 vs. 0.11; p=0.71).
Conclusion: During high-volume colonoscopy with the risk in dropping endoscopists' performance, CADe guidance colonoscopy can significantly help the endoscopists to maintain their high ADR level. In addition, CADe can be of help for the higher APC than conventional colonoscopy.