What was studied?
The reviewers analyzed 14 studies covering 50,080 radiographic images. Some systems made predictions for an entire patient, while others assessed individual teeth or images. Those are different diagnostic tasks, so their results must be interpreted separately.
Periodontitis may cause loss of supporting bone visible on an X-ray. However, an actual diagnosis also requires a clinical examination, including gum measurements and assessment of inflammation.
How accurate was AI?
At the patient level, pooled sensitivity was 93% and specificity was 88%. At the tooth or image level, the figures were 90% and 94%. These results suggest promising performance as an additional screening tool, not perfect detection. False alarms and missed cases remain possible.
The researchers called for prospective studies, stronger external validation and transparent reporting before routine clinical deployment. Evidence from lower- and middle-income settings was also limited.
Will software replace the dentist?
No. An algorithm may highlight suspicious bone changes, but it cannot replace examination of gum bleeding, pocket depth, tooth mobility or a complete medical and dental history. A dentist must interpret findings, confirm the diagnosis and plan treatment.
Bleeding during brushing, persistent bad breath or loose teeth deserve a professional assessment, regardless of whether a clinic uses AI software.
A practical question to ask
Ask what the radiograph shows, whether a periodontal examination is indicated and how gum health will be monitored. Technology is most valuable when it supports careful clinical judgment.
Primary research sources
- Artificial intelligence for detection of periodontitis on radiographic images: a systematic review and diagnostic test accuracy meta-analysis · Journal of Dentistry · 2026-09-17 · DOI: 10.1016/j.jdent.2026.107054
This article is for general information and does not replace an examination or advice from a dentist. Mentioning a technology does not mean the clinic provides it.
