INCREMENTAL YIELD OF ARTIFICIAL INTELLIGENCE IN A FECAL OCCULT BLOOD TEST BASED ORGANIZED SCREENING POPULATION PROGRAM

Aims Fecal Occult Blood Test (FOBT) is the first line test in organized Colorectal Cancer (CRC) screening settings in several countries. Artificial Intelligence (AI) can potentially improve diagnostic performance of colonoscopy, reducing adenoma miss rate and interval CRC. The aim of the study was to compare the diagnostic yield of AI-assisted colonoscopy versus standard colonoscopy (SC).

Methods This is a single-center RCT evaluating consecutive patients undergoing colonoscopy in a FOBT based screening population program. Patients were randomly assigned to SC or AI. Subjects with<6 score at Boston Bowel Preparation Scale were excluded from analysis. Polyp Detection Rate (PDR), Adenoma Detection Rate (ADR), Serrated Detection Rate (SDR), patients with advanced adenomas (i.e. villous histology, high-grade dysplasia or low-grade dysplasia>1cm) and patients with≥3 precancerous lesions were compared between the groups using χ2-test.

Results Data about 464 patients were collected: 231 assigned to AI and 233 to SC arm. Due to inadequate bowel cleansing, 22 patients (10 AI and 12 SC group) were excluded. PDR and ADR were both significantly higher in the AI group compared to SC group (respectively 81.9% [181/221] vs 71.5% [158/221]; p=0.01 and 69.2% [153/221] vs 60.2% [133/221]; p=0.04). Patients with advanced adenomas and with≥3 precancerous lesions were increased in the AI group. SDR was slightly increased in the SC arm. See Table 1.

Table 1

Outcomes	AI	SC	p-value
PDR	81.9% [181/221]	71.5% [158/221]	0.01
ADR	69.2% [153/221]	60.2% [133/221]	0.04
SDR	14.9% [33/221]	16.2% [36/221]	0.69
ADVANCED ADENOMAS	25.3% [56/221]	24.4% [54/221]	0.82
≥3 ADENOMAS	30.8% [68/221]	22.6% [50/221]	0.05

Conclusions AI improves ADR and PDR in a screening FOBT based setting. The impact of a AI supported colonoscopy in surveillance intervals needs to be further evaluated.

Publication History

Article published online:
14 April 2022