FDI Tooth Numbering for AI Scribes — Why It Matters in Canada

Three numbering systems, one mouth

Three tooth-numbering systems are in active use globally. FDI (also called ISO 3950) is the international two-digit system: the first digit is the quadrant (1 upper right, 2 upper left, 3 lower left, 4 lower right) and the second is the tooth position from the midline. Tooth 11 is the upper right central incisor; tooth 36 is the lower left first molar.

Universal is the American system: teeth are numbered 1 to 32 starting from the upper right third molar and finishing at the lower right third molar. Tooth 8 is the upper right central incisor; tooth 19 is the lower left first molar.

Palmer is a quadrant-grid shorthand still common in UK and some Canadian academic settings, but rare in modern Canadian operatory software.

Canada charts in FDI

Canadian dental schools teach FDI. The CDA references FDI. Provincial dental colleges expect FDI in clinical records. Almost every Canadian practice management system uses FDI as the primary tooth identifier. When a Canadian dentist says "tooth 36" in the operatory, the AI scribe needs to chart the lower left first molar — not look up tooth 36 in a Universal table (where it doesn't exist; Universal stops at 32).

US-built scribes default to Universal. When deployed in a Canadian practice, they either ask the practice to switch to Universal (which breaks every existing chart and every insurance submission), or they offer a "Canadian mode" that translates on the way in. Translation introduces a class of bugs that's particularly painful in dentistry: silent off-by-one errors. "Tooth 9" in Universal is the upper left central incisor; tooth 9 doesn't exist in FDI. "Tooth 19" in Universal is the lower left first molar; tooth 19 in FDI doesn't exist either. A misconfigured translation can place a restoration on the wrong tooth and the audit log will look clean.

Where AI scribes get FDI wrong

The failure modes are predictable:

• Defaulting to Universal in the transcript and translating to FDI only at chart write. Any speech-recognition error in the number pollutes the chart silently.
• Failing to handle primary (deciduous) teeth — FDI uses quadrants 5–8 for primaries (e.g. tooth 51 is the upper right primary central incisor). US-trained models rarely see this corpus.
• Mis-handling supernumerary teeth and tooth-fragment notations, which FDI codes differently from Universal.
• Losing the FDI/Universal mapping when audio is unclear and the model has to guess — the guess is biased toward Universal.

What "native FDI" actually means

A scribe that handles FDI natively has the FDI numbering as its first-class chart identifier — not a translation target. In practice that means:

• Speech-recognition lexicons trained on FDI pronunciations ("three-six" and "thirty-six" both map to tooth 36).
• Primary dentition handled directly (51–85).
• Quadrant inference from clinical context ("we'll restore the mesial of 36" pulls the correct quadrant automatically).
• Chart, treatment plan, and patient handout all displayed in FDI.
• Optional Universal display only as a courtesy view for clinicians who trained in the US — never as the source of truth.

Nurvivo Dental was built around FDI as the source of truth, with provincial code systems wired in on top. Any Universal display is derived; the chart, the audit log, and the billing handoff are FDI.