Efficacy of a computer-aided detection system in a fecal immunochemical test-based organized colorectal cancer screening program: a randomized controlled trial (AIFIT study)

Abstract

Background and aim: Computer-Aided Detection (CADe) increases adenoma detection in primary screening colonoscopy. The potential benefit of CADe in a fecal immunochemical test (FIT)-based colorectal cancer (CRC) screening programs is lacking. This study assesses whether the use of CADe increases the Adenoma Detection Rate (ADR) in FIT-based CRC screening program. Patients and methods: In a multicenter, randomized trial, 50-74 years old FIT-positive subjects undergoing colonoscopy, were randomized (1:1) to receive high-definition white light (HDWL) colonoscopy, with or without a real-time deep-learning CADe (CADEYE Fujifilm Co., Tokyo, Japan) by endoscopists with baseline ADR>25%. Main outcomes were ADR (primary outcome), mean number (SD) of adenomas per colonoscopy (APC) and advanced adenoma detection rate (Advanced-ADR). Subgroup analysis according to baseline endoscopists’ ADR (group 1: <40%, group 2: 41-45%, group 3 >46%) was also performed. Results: Eight hundred subjects (median age: 61.0 years, IQR: 55-67; 409 men) were included: 405 underwent CADe-assisted and 395 HDWL colonoscopy, respectively. ADR and APC were significantly higher in the CADe than in the HDWL arm: ADR: 53.6%; 95%CI: [48.6-58.5%] vs. 45.3%; 95%CI [40.3-50.45] (RR: 1.180; 95%CI: [1.026-1.361]); APC: 1.13 (1.54) vs. 0.90 (1.32), p=0.028. No significant difference in advanced-ADR (18.5%; 95%CI [14.8-22.6%] vs. 15.9%; 95%CI [12.5-19.9%]) was found. An increase in ADR was observed in all endoscopists' groups regardless of baseline ADR. Conclusions: Incorporating CADe significantly increases ADR and APC in the framework of a FIT-based CRC screening program. The impact of CADe appears to be consistent regardless of the endoscopist baseline ADR.

Artificial Intelligence-Aided Endoscopy and Colorectal Cancer Screening

“…Recently, deep learning-based algorithms have been developed aiming to assist in adenoma detection. After showing promising results in ex vivo studies [ 26 , 27 , 28 ], the use of artificial intelligence algorithms has been claimed to improve detection of colorectal lesions in several randomized controlled trials [ 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 , 40 , 41 ] ( Table 1 ).…”

“…However, it may be argued that a further increase the performance in adenoma detection of expert endoscopists, such as the ones involved in the first trials [ 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 , 38 , 39 , 40 , 41 ], may provide only little additional benefit.…”

Artificial intelligence and the push for small adenomas: all we need?

“…Some computer-aided polyp detection (CADe) systems have entered the field of clinical application, however, and have been studied extensively in randomized trials. At the time of writing, 19 such trials were available as full publications [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16][17][18][19]. They all show a more or less significant increase in detection of small adenomas.…”

Commentary

“…

Artificial intelligence (AI) is assumed to improve the detection of colorectal lesions, thus helping to reduce the risk of interval cancer after screening colonoscopy [1][2][3]. However, endoscopists need to be aware that AI systems should be used to assist and not to replace them.

…”