Practical Deep Learning Tool For Scoring Of Ulcerative Colitis Disease Activity In Central Reading

Aims To create software with a graphic user interface(GUI)to reduce the time to review&score a UC video & improve the accuracy of scoring.

Methods We built a web-based interface that read&write multiple databases&data stores displaying videos to be scored by a central reader,as well as the associated metadata required to improve the process.Our GUI shows a timeline under the video,with markers indicated for colon segments&sections that are blurry,poorly prepped,or otherwise highlighted in different colours.While we could also highlight video sections with the precise score assigned to it by AI,this would unnecessarily bias the central reader’s opinion.We hide the score generated by our AI models & instead display 3 colours for low,medium or high disease activity.When a video is loaded to be read,the playback marker is set to the first high disease activity section,consisting of a few seconds of video,that video is played back continuously in a loop until the reader selects the appropriate UCEIS or Mayo score.When the reader saves the section,the software moves the video cursor to the next highest scored section of the video.This way,the central reader can review only relevant video portions to confirm the score from each segment.If the reader’s scores do not align well with the AI scores the software continues to show more sections of the video,including sections it labelled as unscorable,that may be scorable.The interface allows changes in playback speed.

Results Reviews by 3 key opinion leaders(KOLs),user experience was overwhelmingly positive.Not only does the system allow the reader’s attention to be more efficiently used,but the interface allows both AI&central reader scores to be saved,allowing for the latter to be used iteratively to re-train&improve the underlying AI model(s).

Conclusions We have developed a practical AI tool that hopefully can improve the efficiency&accuracy of the central reading process in clinical trials for UC.Further improvements are ongoing.

Citation Byrne MF, East JE, Iacucci M et al. eP482 PRACTICAL DEEP LEARNING TOOL FOR SCORING OF ULCERATIVE COLITIS DISEASE ACTIVITY IN CENTRAL READING. Endoscopy 2021; 53: S256.

Publication History

Article published online:
19 March 2021