Editorial 1, Issue 13.1

Will AI Finally Allow Dentists to Participate in Sleep Testing?

http://dx.doi.org/10.15331/jdsm.7414

Jean-François Masse, DMD, MSc, FACD, Diplomate, ABDSM
 
Editor-in-Chief Journal of Dental Sleep Medicine
Universite Laval, Quebec City, Quebec, Canada
 

Artificial intelligence is advancing at a pace that is reshaping diagnostic medicine. In barely three years, large language models have progressed from experimental novelties to highly accurate diagnostic tools supported by unprecedented computational resources and rapidly expanding medical benchmarks. These developments are not theoretical. They will force sleep medicine to reconsider who can, and who must, participate in the diagnostic workflow. Dentists who are treating OSA should be part of that reconsideration.

For years, skepticism about AI revolved around reliability. That position is becoming increasingly difficult to defend. In 2025, Microsoft released the AI Diagnostic Orchestrator (MAI-DxO), evaluated on 304 complex case challenges from the New England Journal of Medicine. When paired with an advanced reasoning model, MAI-DxO reached approximately 85% diagnostic accuracy, compared with roughly 20% for experienced physicians under the same conditions, while reducing unnecessary testing by 28%1-3. Additional independent research reinforces this trend. A randomized study from the University of Virginia showed that clinicians using ChatGPT Plus improved diagnostic accuracy modestly (from 73.7% to 76.3 %) , whereas ChatGPT Plus alone exceeded 90% accuracy on the same vignettes, suggesting ChatGPT Plus does better in certain situations without physicians than with4. Other evaluations demonstrate that GPT-4 performs at or above physician level for differential diagnosis and case-resolution tasks5–7. AI is rapidly evolving from an optional resource to a credible second opinion that patients, insurers, and regulators will soon expect clinicians to use.

The urgency of integrating AI is nowhere greater than in sleep medicine. Workforce limitations are longstanding. Collen estimated a ratio of roughly 43,000 patients per board-certified sleep physician in the United States, a gap that shows no sign of closing8. Meanwhile, the burden of obstructive sleep apnea (OSA) is expanding. Recent modeling projects that by 2050 nearly 77 million U.S. adults will have OSA, a 35% increase from 2020 and impacting approximately 46% of adults aged 30-699–11. These projections align with epidemiologic studies documenting a rise in moderate to severe sleep-disordered breathing over the past two decades12,13.
 
The medical workforce is also tightening. The Association of American Medical Colleges forecasts a national physician shortfall of up to 86,000 clinicians by 2036 including significant deficits in primary care—the very clinicians to whom simpler OSA cases are increasingly considered for delegation14–16. If prevalence rises, specialists remain scarce, and primary care contracts, the current diagnostic model becomes untenable. AI alone cannot compensate unless the system is redesigned around it.

Compounding this pressure is the growing recognition that single-night testing underestimates disease burden17. Punjabi et al. demonstrated that three-night portable monitoring frequently reclassifies OSA severity18. Lechat and colleagues confirmed striking night-to-night variability in AHI in large community samples, showing that diagnostic category can shift between nights and that this variability carries cardiometabolic consequences19,20. If multi-night testing is ever required for proper diagnosis, the current interpretive capacity of sleep medicine is insufficient. Telemedicine may improve access but cannot create additional clinician hours for reviewing multi-night data.

Dentists are uniquely positioned to help address this gap—if the profession is willing to consider a modernized, AI-supported diagnostic workflow. Millions of patients are examined weekly in dental practices, where clinicians routinely assess craniofacial structure, tongue posture, mandibular position, and other anatomic indicators associated with sleep-disordered breathing21. Oral appliance therapy is already recognized as a first-line treatment for many patients with mild to moderate OSA or CPAP intolerance22. Dentists regularly participate in collaborative models of care and increasingly play a role in sleep-related assessment and therapy.

As the sleep world is changing and some sleep physicians openly talk about leaving diagnosis of simple cases  to general practitioners in medicine, it is time for dentists who are treating OSA to think about what we can offer and step in.

If the sleep community does not lead in defining this new workflow, external forces will. Insurers, regulatory bodies, and technology companies are already developing frameworks for AI-supported diagnostics23-25. Without proactive guidance from clinicians, the system may be redesigned based on cost rather than care. As the AASM just published their new position statement on the use of AI in sleep medicine26, a joint initiative between the American Academy of Sleep Medicine and the American Academy of Dental Sleep Medicine should be considered to establish training standards, outline ethical and practical applications of AI in diagnostic workflows, and launch pilot programs evaluating AI-supported, dentist-assisted models.

Artificial intelligence has already changed what is possible. The prevalence of obstructive sleep apnea continues to rise, the diagnostic burden is increasing, and multi-night testing is moving toward necessity rather than luxury. With appropriate training and oversight, dentists—supported by AI—can help deliver timely, accurate evaluation to millions of patients who will otherwise face delays or remain undiagnosed. Sleep medicine is entering a period of transformation. Dentists must have a place at that table. This is not about taking the sleep doctor's place, but rather about being more involved as part of their team. 

CITATION

Masse, JF. Will AI Finally Allow Dentists to Participate in Sleep Testing? J Dent Sleep Med. 2026;13(1).

REFERENCES

  1. Microsoft Research. The path to medical superintelligence: evaluating MAI-DxO on NEJM case reports. Published June 30, 2025. https://microsoft.ai/news/the-path-to-medical-superintelligence/
  2. Kaku A, Yu T, Srivastava A, et al. Sequential Diagnosis Benchmark (SDBench): Evaluating Clinical Reasoning in Large Language Models. arXiv [Preprint]. 2025.
  3. Limb M. Microsoft claims AI diagnostic tool can outperform doctors. BMJ. 2025;390:r1385. doi:10.1136/bmj.r1385.
  4. Goh E, Gallo R, Hom J, et al. Large Language Model Influence on Diagnostic Reasoning: A Randomized Clinical Trial. JAMA Netw Open. 2024;7(10):e2440969. doi:10.1001/jamanetworkopen.2024.40969
  5. Katz U, Cohen E, Shachar E, et al. GPT versus resident physicians — a benchmark based on official board scores. NEJMâ¯AI. 2024;1(5):5. doi:10.1056/aidbp2300192.
  6. Hirosawa T, Harada Y, Tokumasu K, Ito T, Suzuki T, Shimizu T. Evaluating ChatGPTâ4’s diagnostic accuracy: impact of visual data integration. JMIR Med Inform. 2024;12:e55627. doi:10.2196/55627.
  7. Rao A, Pang M, Kim J, Kamineni M, Lie W, Prasad AK, Landman A, Dreyer KJ, Succi MD. Assessing the utility of ChatGPT throughout the entire clinical workflow. medRxiv [Preprint]. Published February 26, 2023. doi:10.1101/2023.02.21.23285886
  8. Collen JF, Wickwire EM, Capaldi V, Lettieri C. Losing sleep! Are we missing the future of sleep medicine? J Clin Sleep Med. 2020;16(4):473–474.
  9. Boers E, Barrett MA, Benjafield AV, Barnet JH, Ravelo LA, Kaye L, et al. Projecting the 30âyear burden of obstructive sleep apnoea in the USA: a prospective modelling study. Lancet Respir Med. 2025;13(12):1078–1086.
  10. ResMed Inc. Obstructive sleep apnea expected to affect nearly 77 million U.S. adults by 2050, new ResMed study finds [Internet]. Published August 27, 2025. Available from: https://investor.resmed.com/news-events/press-releases/detail/408/obstructive-sleep-apnea-expected-to-affect-nearly-77-million-u-s-adults-by-2050-new-resmed-study-finds
  11. Zou D, et al. Editorial: Novel technologies in the diagnosis and management of sleep-disordered breathing, volume II. Front. Sleep. 2025;4:1696478.
  12. Peppard PE, Young T, Barnet JH, Palta M, Hagen EW, Hla KM. Increased prevalence of sleepâdisordered breathing in adults. Am J Epidemiol. 2013;177(9):1006–1014. doi:10.1093/aje/kws342.
  13. Heinzer R, et al. HypnoLaus Study. Lancet Respir Med. 2015;3(4):310–318.
  14. Association of American Medical Colleges. The complexities of physician supply and demand: projections from 2021 to 2036 [Internet]. Washington (DC): Association of American Medical Colleges; 2024 Mar. Available from: https://www.aamc.org/media/75236/download
  15. Bazemore AW, Petterson SM, McCulloch KK. US primary care workforce growth: a decade of limited progress, and projected needs through 2040. J Gen Intern Med. 2025;40(2):339-346. doi:10.1007/s11606-024-09121-x
  16. Pendharkar SR, Blades K, Kelly JE, et al. Perspectives on primary care management of obstructive sleep apnea. J Clin Sleep Med. 2021;17(1):89–98.
  17. Tsai CW, Palomo JM, Link B, et al. Using a Wearable’s Multi-Night Capability to Mitigate Night-to-Night Variability in a Dental Clinic Cohort. J Dent Sleep Med. 2026;13(1)
  18. Punjabi NM, et al. Variability of OSA Severity. Chest. 2020;158:365–373.
  19. Lechat B, Naik G, Reynolds A, Aishah A, Scott H, Loffler KA, et al. Multinight prevalence, variability, and diagnostic misclassification of obstructive sleep apnea. Am J Respir Crit Care Med. 2022;205(5):563–569. doi:10.1164/rccm.202107â1761OC.
  20. Lechat B, Loffler KA, Reynolds AC, Naik G, Vakulin A, Jennings G, et al. High nightâtoânight variability in sleep apnea severity is associated with uncontrolled hypertension. npj Digit Med. 2023;6:57. doi:10.1038/s41746â023â00801â2.
  21. Kale SS, Kakodkar P, Shetiya SH. Assessment of oral findings of dental patients who screen high and no risk for obstructive sleep apnea (OSA) reporting to a dental college – a cross sectional study. Sleep Sci. 2018;11(2):112–117.
  22. Ramar K, Dort LC, Katz SG, Lettieri CJ, Harrod CG, Thomas SM, Chervin RD. Clinical practice guideline for the treatment of obstructive sleep apnea and snoring with oral appliance therapy: an update for 2015. J Clin Sleep Med. 2015;11(7):773–827. doi:10.5664/jcsm.4858.
  23. American Academy of Sleep Medicine. EnsoData receives FDA clearance for AIâpowered sleep diagnosis using pulse oximeters [Internet]. Published February 21, 2024.. Available from: https://aasm.org/ensodata-receives-fda-clearance-for-ai-powered-sleep-diagnosis-using-pulse-oximeters/
  24. American Academy of Sleep Medicine. Honeynaps receives FDA approval for AI sleep disorder diagnosis software [Internet]. Published September 26, 2023. Available from: https://aasm.org/honeynapsâreceivesâfdaâapprovalâforâaiâsleepâdisorderâdiagnosisâsoftware/
  25. PranaQ. PranaQ receives FDA 510(k) clearance for TipTraQ home sleep apnea test [Internet]. Published February 18, 2025. Available from: https://pranaq.com/pranaqâreceivesâfdaâ510kâforâtiptraqâhomeâsleepâapneaâtest/
  26. Oks M, Sachdeva R, Davenport MA, Husain AM, Kalra M, Krishnan V, et al. Artificial intelligence in sleep medicine: an updated American Academy of Sleep Medicine position statement. J Clin Sleep Med. 2025;21(11):1953-1955. doi:10.5664/jcsm.11832

SUBMISSION AND CORRESPONDENCE INFORMATION

Submitted in final revised form December 11, 2025
 
Address correspondence to: Jean-François Masse, DDS, MSc, FACD, D.ABDSM, Professor, Universite Laval, 2780 Masson #200, Quebec City, QC, G1P 1J6, Canada; Tel: 418871-1447; Fax: 418-871-4983; Email: jean-francois.masse@fmd.ulaval.ca

PDF