CADe System and Mucosal Exposure Device Improve ADR in Screening Colonoscopy

Computer-aided detection (CADe) and a mucosal exposure device significantly improve adenoma detection rate (ADR) and advanced adenoma detection rate (AADR) compared with standard colonoscopy, according to study results published in Gastrointestinal Endoscopy.

The 4-arm, parallel, randomized controlled trial compared patients screened for colorectal cancer with CADe (CAD EYE, Fujifilm) alone, Endocuff (Endocuff Vision ARV 110; Olympus Corporation)-assisted colonoscopy (EAC) alone, or a combination of CADe and EAC (CADe+EAC) vs standard white light colonoscopy. A high-definition colonoscope (the ELUXEO 7000 system, EC 760ZP-V/L, Fujifilm Co) was used for all groups.

Eligible participants were asymptomatic and had undergone a routine screening colonoscopy or had received a positive fecal immunochemical test. The primary outcome was ADR, which was the percentage of participants with an adenoma detected during colonoscopy.

After screening, 1245 participants (mean age, 62.2 years; men, 40.4%) were randomly assigned to CADe alone (n=312), EAC alone (n=315), CADe+EAC (n=308), or to a control group (n=310).

At least 1 adenoma was detected in 644 participants. The ADR in the control, CADe, EAC, and CADe+EAC groups was 41.9%, 52.2%, 54.0%, and 58.8%, respectively. The absolute difference vs the control group was 10.3% (95% CI, 2.5%-18.1%; P =.01) in the CADe group, 12.0% (95% CI, 4.3%-19.8%; P =.002) in the EAC group, and 16.8% (95% CI, 9.1%-24.6%; P <.001) in the CADe+EAC group.

Overall, 1358 adenomas were detected, with a mean of 1.09 adenomas per colonoscopy (APC). The APC in the control, CADe, EAC, and CADe+EAC groups was 0.76, 1.11, 1.18, and 1.31, respectively. Each intervention group had a significantly higher incidence rate ratio (IRR) of APC vs the control group. The rate increased by 46% (95% CI, 16%-84%) in the CADe group, 55% (95% CI, 23%-95%) in the EAC group, and 73% (95% CI, 37%-217%) in the CADe+EAC group compared with the control group (P for both latter comparisons <.001). Minimal changes in the absolute mean ADR difference and IRR were found after adjustment for patient- and procedure-related characteristics.

The proximal adenoma detection rate (pADR) for the control, CADe, EAC, and CADe+EAC groups was 25.2%, 33.3%, 34.9%, and 37.0%, respectively. Each intervention group had a significantly higher pADR vs the control group. The absolute difference compared with the control group was greatest in the CADe+EAC group (11.9%; 95% CI, 4.6%-19.1%; P =.001).

The number of proximal adenomas per colonoscopy (pAPC) in the control, CADe, EAC, and CADe+EAC groups was 0.42, 0.67, 070, and 0.78, respectively. The differences between pADR and IRR for pAPC were comparable after adjustment for patient and endoscopic characteristics.

The AADR for the control, CADe, EAC, and CADe+EAC groups was 7.7%, 8.3%, 8.3%, and 13.6%, respectively. Only the CADe+EAC group had a significant unadjusted absolute difference compared with the control group (5.9%; 95% CI, 1.04%-10.8%; P =.02). The IRR in the CADe+EAC group was still statistically significant vs the control group after adjustment for other characteristics in the sensitivity analysis.

Study limitations include not blinding the endoscopists and the single-center design. Also, trainees were involved, and the study was not powered to make pairwise comparisons of individual study groups.

“We demonstrated that using CADe or EAC alone was associated with increased adenoma and proximal adenoma detection,” the study authors noted. “Furthermore, incorporating CADe with EAC resulted in an additional benefit of detecting adenoma, proximal adenoma, and advanced adenoma.”