DentalClinicsAIHub

Dental CDSS

One radiograph, several readings, shown side by side. We attribute every finding to the platform that produced it, we point out where they agree and where they differ, and we stop there.

What it does

  1. One image goes to several platforms

    A bitewing, a periapical or a panoramic radiograph is sent to the platforms you selected, in the same exchange.

  2. Each reading comes back attributed

    Every finding carries the name of the platform that produced it and the version of the model that ran. Nothing is merged, rewritten or reworded.

  3. Agreement and disagreement are stated

    Findings reported by more than one platform are marked as such. Findings reported by only one are marked as such. Both are shown in full.

  4. The clinician reads and decides

    There is no verdict at the end of this page. The diagnosis is the practitioner's, and it always was.

FDI odontogram of the synthetic case. Tooth 26 is reported by all three mock platforms; tooth 16 by platforms A and C; tooth 15 by platform C only; tooth 36 by platforms A and C on two different surfaces; tooth 37 shows bone loss reported by platform B only; tooth 46 is charted as missing yet reported by platform B.
The synthetic case below, drawn on a chart. Three mock readers, not real platforms.

What this deliberately does not do. No confidence score. No priority level. No ranking of platforms against each other. No consensus finding, no arbitration, no combined summary.

Why. Each platform holds its own regulatory position for its own stated indication. A synthesis on top of them would be a new clinical output that none of those authorisations covers — a separate device, with its own regulatory path. We have chosen not to produce one.

See it on a synthetic case

Demo Synthetic case · three mock readers built to disagree · no image analysed, no vendor connected.

What a disagreement means

Different AI systems reading the same radiograph may report different findings. A tool that hid that behind a single averaged answer would be throwing away information the clinician should see. Showing the disagreement is the product.

Illustration: a dentist reviewing a radiograph on a monitor, the image softly out of focus.

Retention

The same as everywhere else here: the image is held in memory for the duration of the request and released when it closes. Comparing several readings does not mean storing them. Nothing is written to disk, including the comparison itself.

This is a decision-support display, not a diagnostic device. It produces no clinical finding of its own and holds no regulatory clearance. Each platform's status belongs to its own vendor.