PT - JOURNAL ARTICLE AU - Hugh Devlin AU - Tomos Williams AU - Jim Graham AU - Martin Ashley TI - A Comparative Study of Dentists’ Ability to Detect Enamel-only Proximal Caries in Bitewing Radiographs With and Without the use of AssistDent<sup>®</sup> Artificial Intelligence Software AID - 10.1101/2020.10.12.20211292 DP - 2020 Jan 01 TA - medRxiv PG - 2020.10.12.20211292 4099 - http://medrxiv.org/content/early/2020/10/14/2020.10.12.20211292.short 4100 - http://medrxiv.org/content/early/2020/10/14/2020.10.12.20211292.full AB - Enamel-only proximal caries, if detected, can be reversed by non-invasive treatments. Dental bitewing radiograph analysis is central to diagnosis and treatment planning and, when used to detect enamel-only proximal caries, it is an important tool in minimum intervention and preventive dentistry. However, the subtle patterns of enamel-only proximal caries visible in bitewing radiographs are difficult to detect and often missed by dental practitioners. This study measures the ability of dentists to detect enamel-only proximal caries in bitewing radiographs with and without the use of AssistDent® Artificial Intelligence (AI) software.23 dentists were randomly divided into a control arm, in which no Artificial Intelligence assistance was provided, and an experimental arm in which Artificial Intelligence assistance provided on-screen prompts for potential locations of enamel-only proximal caries. All participants analysed a set of 24 bitewings, gathered from one dental hospital and 11 general dental practices, which had previously been analysed independently by a panel of 5 dento-maxillofacial radiologists and 1 professor of restorative dentistry who, between them, identified a total of 65 enamel-only carious lesions and 241 healthy proximal surfaces.Results demonstrate that dentists using the assistive software found 75.8% of the enamel-only proximal caries compared to a 44.3% detection rate in the control group. This represents an absolute increase of 31.5% (relative increase in sensitivity of 71%). Participants in the experimental group incorrectly identified 14.6% of the healthy surfaces as having enamel-only proximal caries compared to 3.7% in the control group, an absolute increase of 10.9% (relative decrease in specificity of 11%).T-test analysis demonstrated a statistically significant difference (p&lt;0.01) between the two arms in sensitivity (true positive caries detection rate) and specificity (false positive rate).We conclude that AssistDent® Artificial Intelligence software significantly improves dentists’ ability to detect enamel-only proximal caries, with only a slight increase in false positives, and could be considered as a tool to support minimum intervention and preventive dentistry in general practice.Competing Interest StatementHD, JG and TW are employees, of Manchester Imaging Ltd. The Division of Dentistry, University of Manchester, purchased a software licence for AssistDent@[reg] from Manchester Imaging Ltd. MA is not an employee of Manchester Imaging Ltd and declares no conflict of interest.Clinical TrialManchester University Research Ethics Committee (Ref: 2020-9892-15955)Funding StatementNo external funding was receivedAuthor DeclarationsI confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained.YesThe details of the IRB/oversight body that provided approval or exemption for the research described are given below:Manchester University Research Ethics CommitteeAll necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived.YesI understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance).YesI have followed all appropriate research reporting guidelines and uploaded the relevant EQUATOR Network research reporting checklist(s) and other pertinent material as supplementary files, if applicable.YesAll data is provided in the main paper