Truace
Subscribe
70/100

Moderate evidence

Cite
More
Export CSVExport JSONEmailLinkedInX

Review of 16 studies finds jaw-joint AI spots disc displacement better on MRI than on X-ray scans

Study title: Artificial Intelligence in Temporomandibular Joint Imaging: Diagnostic Performance Across Input Modality-Target Pairings-A Systematic Review and Diagnostic Test Accuracy Meta-Analysis

Evidence type
Systematic review or meta-analysis
Sources
1 sourcePeer-reviewed
Published
2 min read
Truvace record
TRV-2026-2861cite or check it

Study title: Artificial Intelligence in Temporomandibular Joint Imaging: Diagnostic Performance Across Input Modality-Target Pairings-A Systematic Review and Diagnostic Test Accuracy Meta-Analysis

A systematic review and meta-analysis pooled 16 studies (17 estimates) on standalone AI diagnosis of temporomandibular joint disorders from imaging. For disc displacement, AI reading MRI reached pooled sensitivity of 0.924 and specificity of 0.812.

Read the study abstract

To estimate standalone artificial intelligence (AI) diagnostic performance in temporomandibular joint (TMJ) imaging across input modality-target pairings. Four databases were searched through 18 June 2026. Eligible studies assessed standalone imaging AI and supplied an exact test-set 2 × 2 table in the report or after author contact. Four pairings were analysed by bivariate meta-regression. Two reviewers applied QUADAS-3 at estimate level. Sixteen studies contributed 17 estimates. For disc displacement, magnetic…

Artificial Intelligence in Temporomandibular Joint Imaging: Diagnostic Performance Across Input Modality-Target Pairings-A Systematic Review and Diagnostic Test Accuracy Meta-Analysis
"Vantage Atlas MRI scanner from Toshiba" by Mos.ru, CC BY 4.0.

The gain

A systematic review of 16 studies found that AI reading MRI scans for temporomandibular joint disc displacement reached pooled sensitivity of 0.924 and specificity of 0.812.

In brief

With panoramic or cone-beam CT input, where the disc is not directly visible, pooled sensitivity was 0.765 and specificity 0.741. The sensitivity gap was 0.159, and it stayed between 0.152 and 0.198 in leave-one-out and replicate analyses.

The authors say performance should be read by input-and-target pairing, not as a fixed trait of the algorithm. The formal test across pairings was not significant, and none of the estimates had low overall risk of bias, so the MRI advantage needs prospective testing.

Main points

  1. Four databases were searched through 18 June 2026; studies had to assess standalone imaging AI and supply an exact test-set 2 × 2 table, either published or obtained from the authors.
  2. Sixteen studies contributed 17 estimates, analysed with bivariate meta-regression across four input-target pairings.
  3. For MRI input, pooled sensitivity was 0.924 (95% CI 0.859-0.960) and specificity 0.812 (0.709-0.885).
  4. For panoramic or cone-beam CT input, pooled sensitivity was 0.765 (0.618-0.867) and specificity 0.741 (0.592-0.850).
  5. The sensitivity difference was 0.159, and 0.152-0.198 in leave-one-out and replicate analyses.
  6. The omnibus pairing test was not significant (χ²(6) = 10.38, P = .110).
  7. No estimate had low overall risk of bias under QUADAS-3, and several otherwise eligible estimates lacked exact published 2 × 2 data.

What the review found when it combined the studies

These are the review’s own pooled results, copied from its abstract. Truvace did not recalculate them.

  • Disc displacement detection by artificial intelligence on panoramic or cone-beam computed tomography: specificity

    Specificity 0.741 (95% CI 0.592 to 0.850)

    search to 18 June 2026

    The review’s own words
    For disc displacement, magnetic resonance imaging input yielded pooled sensitivity 0.924 (95% CI, 0.859-0.960) and specificity 0.812 (0.709-0.885), compared with 0.765 (0.618-0.867) and 0.741 (0.592-0.850) for panoramic or cone-beam computed tomography input.

  • Disc displacement detection by artificial intelligence on panoramic or cone-beam computed tomography: sensitivity

    Sensitivity 0.765 (95% CI 0.618 to 0.867)

    search to 18 June 2026

    The review’s own words
    For disc displacement, magnetic resonance imaging input yielded pooled sensitivity 0.924 (95% CI, 0.859-0.960) and specificity 0.812 (0.709-0.885), compared with 0.765 (0.618-0.867) and 0.741 (0.592-0.850) for panoramic or cone-beam computed tomography input.

  • Disc displacement detection on temporomandibular joint magnetic resonance imaging by artificial intelligence: specificity

    Specificity 0.812 (95% CI 0.709 to 0.885)

    search to 18 June 2026

    The review’s own words
    For disc displacement, magnetic resonance imaging input yielded pooled sensitivity 0.924 (95% CI, 0.859-0.960) and specificity 0.812 (0.709-0.885), compared with 0.765 (0.618-0.867) and 0.741 (0.592-0.850) for panoramic or cone-beam computed tomography input.

  • Disc displacement detection on temporomandibular joint magnetic resonance imaging by artificial intelligence: sensitivity

    Sensitivity 0.924 (95% CI 0.859 to 0.960)

    search to 18 June 2026

    The review’s own words
    For disc displacement, magnetic resonance imaging input yielded pooled sensitivity 0.924 (95% CI, 0.859-0.960) and specificity 0.812 (0.709-0.885), compared with 0.765 (0.618-0.867) and 0.741 (0.592-0.850) for panoramic or cone-beam computed tomography input.

Code checks every number above against the review’s text before it is shown. The short labels are an AI model’s reading of that text, so the review’s own sentence sits under each one. A pooled figure is only as strong as the studies behind it. How Truvace uses pooled results

The full story

Two reviewers rated risk of bias with QUADAS-3 at the level of each estimate. Only studies reporting an exact test-set 2 × 2 table, in the paper or after contacting authors, were included, so accuracy could be verified.

The review separated disc displacement read directly on MRI from displacement inferred from panoramic or cone-beam CT images that do not depict the disc. Sensitivity was higher in the direct-depiction group.

The reference standard used and the input pairing split the disc-displacement evidence identically, so the authors say design factors were entangled with imaging type.

Sources

1 on file
  1. Dentomaxillofacial Radiologydoi:10.1093/dmfr/twag076
    Peer-reviewed2026-11-12

Talk it over

Four questions to start onWho gets this benefit, and who was left out of the study or story?

Start with a question

Review of 16 studies finds jaw-joint AI spots disc displacement better on MRI than on X-ray scanshttps://truvace.com/item/artificial-intelligence-in-temporomandibular-joint-imaging-diagnostic-performanc

  1. Who gets this benefit, and who was left out of the study or story?

    The gain A systematic review of 16 studies found that AI reading MRI scans for temporomandibular joint disc displacement reached pooled sensitivity of 0.924 and specificity of 0.812.

    Answer in the thread

  2. How do we know? What can this kind of evidence show, and what can’t it?

    The evidence This rests on a systematic review or meta-analysis, from Dentomaxillofacial Radiology.

    Answer in the thread

  3. What does it leave open? What else would you want to know before acting on it?

    The record’s own caveat The cross-domain group contained only three estimates, the omnibus pairing test was not significant, design factors were entangled, and no estimate had low overall risk of bias.

    Answer in the thread

  4. Have you seen AI used this way where you study, work or live? Did it go the same way?

    Answer in the thread

The questions are the same on every record. The details under them are taken from this record as written.

What readers say

More from Health