Categories
Change Password!
Reset Password!
Computer-aided diagnosis (CADx) predicts polyp histology during colonoscopy, potentially plummeting unnecessary removal of nonneoplastic polyps, though its benefits and risks are unclear. This study was performed to assess the advantages and disadvantages of using CADx for the optical diagnosis of small (≤5mm) rectosigmoid polyps.
While computer-aided diagnosis can help manage small rectosigmoid polyps, it may not considerably improve outcomes compared to traditional approaches, guiding decisions on its routine use.
Computer-aided diagnosis (CADx) predicts polyp histology during colonoscopy, potentially plummeting unnecessary removal of nonneoplastic polyps, though its benefits and risks are unclear. This study was performed to assess the advantages and disadvantages of using CADx for the optical diagnosis of small (≤5mm) rectosigmoid polyps.
The article search began with exploring Medline, Scopus and Embase. Histologically verified diagnostic accuracy studies evaluating the real-time performance of physicians in predicting neoplastic changes in small rectosigmoid polyps, with or without CADx assistance during colonoscopy were included.
The clinical benefit and harm were assessed keeping in view the accuracy values of the endoscopist before and after the CADx assistance. GRADE was used to assess the evidence certainty.
The benefit was assessed by the proportion of nonneoplastic polyps spared removal with CADx, while harm was measured by the proportion of neoplastic polyps left in situ due to inaccurate CADx diagnosis. Histology was used as the benchmark for both the outcomes.
A total of ten studies involving 3,620 patients and 4,103 small rectosigmoid polyps were analyzed. CADx displayed a sensitivity of 87.3% (95% confidence interval [CI], 79.2% to 92.5%) and specificity of 88.9% (95% CI, 81.7% to 93.5%) in envisaging neoplastic variations in the studies evaluating CADx performance alone (9 studies; 3,237 polyps).
Among the studies comparing histology prediction before and after CADx assistance (4 studies; 2,503 polyps), no significant difference was found in the proportion of nonneoplastic polyps predicted to avoid removal or in the proportion of neoplastic polyps erroneously left in situ (moderate-certainty evidence), with the limitation that the simulated optical diagnosis may have influenced the operator's decision-making.
CADx did not offer any additional benefit or harm in managing small rectosigmoid growths during the colonoscopic examination.
Annals of Internal Medicine
Computer-Aided Diagnosis for Leaving Colorectal Polyps In Situ: A Systematic Review and Meta-analysis
Cesare Hassan et al.
Comments (0)