Source/Disclosures
Disclosures: Spadaccini reports no relevant financial disclosures. Please see the full study for all other authors’ relevant financial disclosures.
June 30, 2020
1 min read
Save

Artificial intelligence helps detect colorectal neoplasia

Source/Disclosures
Disclosures: Spadaccini reports no relevant financial disclosures. Please see the full study for all other authors’ relevant financial disclosures.
You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact customerservice@slackinc.com.

Adding artificial intelligence systems during colonoscopies helped increase the detection of colorectal neoplasia, according to a meta-analysis published in Gastrointestinal Endoscopy.

Marco Spadaccini, MD, of Humanitas Research Hospital and University in Italy, and colleagues wrote that missed lesions due to recognition failure or technical issues can lead to interval colorectal cancer.

“By addressing these pitfalls, artificial intelligence is expected to reduce the risk of miss rate and consequently of interval CRC,” they wrote. “The aim of our systematic review and meta-analysis is to assess the relationship between the increased detection led by [computer aided polyp detection (CADe)] and the main features of the detected lesions.”

Investigators searched the literature for studies that reported the diagnostic accuracy of CADe systems in detection of colorectal neoplasia. The primary outcome was pooled adenoma detection rate. Secondary outcomes included adenomas per colonoscopy (APC) according to size, morphology and location, advanced APC (AAPC), as well as polyp detection rate (PDR), polyp per colonoscopy (PPC), and sessile serrated lesion per colonoscopy (SPC).

Researchers identified five trials comprising 4,354 patients that fit their criteria. The pooled ADR was high in the CADe group (36.6%) compared with the control group (25.2%; RR = 1.44; 95% CI, 1.27-1.62). The APC was also higher in the CADe group (0.58 vs. 0.36; RR = 1.7; 95% CI, 1.53-1.89), included higher rates among adenomas 5 mm or less in size, between 6 mm and 9 mm in size, and at least 10 mm in size. The APC was also higher for proximal and distal polyps, as well as in polyps that were flat or with polypoid morphology.

The CADe systems helped produce a higher SPC (RR = 1.52; 95% CI, 1.14-2.02) and a trend for AADR that was not significant.

“Lesion detection by AI is not impacted by factors such as size and morphology that are known to affect detection by human observers,” Spadaccini and colleagues wrote. “According to the current evidence, there is substantial and convergent evidence for the incorporation of artificial intelligence to increase detection of colorectal neoplasia during colonoscopy.”