Artificial Intelligence False Positives in Colonoscopy Are Not Time Consuming
Douglas K. Rex, MD, MASGE, reviewing Spadaccini M, et al. Gastrointest Endosc 2022 Jan 4.
Computer-aided detection (CADe), or artificial intelligence (AI), programs are now commercially available and have been shown to increase the adenoma detection rate. These AI programs also produce some false positives, which could potentially consume time during withdrawal as colonoscopists interrogate areas that are marked by the AI device. In reality, most false-positive hits are brief and instantly recognized as normal folds or contaminants.
In a study of 2 different video libraries with 2 CADe systems, the number of false-positive activations was almost identical between the software brands (25.5 vs 25.7 false positives per colonoscopy). In video review, the time required to assess each false activation was 0.2 ± 0.9 seconds, with only 1.6 to 1.8 false positives per colonoscopy requiring additional inspection. In both groups, less than 1% of mean withdrawal time was required for endoscopic inspection of false positives.
COMMENTThese data indicate that the overwhelming majority of false positives are instantly recognized as such. Only the occasional false positive requires a few seconds of interrogation to establish an accurate diagnosis.
Note to readers: At the time we reviewed this paper, its publisher noted that it was not in final form and that subsequent changes might be made.
CITATION(S)
Spadaccini M, Hassan C, Alfarone L, et al. Comparing Number and relevance Of false activations between two Artificial Intelligence CADe SystEms: the NOISE study. Gastrointest Endosc 2022 Jan 4. (Epub ahead of print) (https://doi.org/10.1016/j.gie.2021.12.031)
