Pearl's AI platform analyzed 26 million dental x-rays across 15 million patients and found that untreated decay affects Americans at a rate 4.5 times higher than federal estimates — a gap with direct implications for how Canadian dental practices understand disease prevalence, treatment planning, and the limitations of self-reported oral health data.
As of August 2026, the disconnect between what patients report and what radiographs reveal has never been clearer. The Pearl Oral Health Index, released in June 2026, represents the first large-scale radiographic census of oral health, and its findings challenge decades of assumptions built on survey-based data. For dental professionals in Toronto and the Greater Toronto Area (GTA), the data raises pointed questions about screening protocols, patient communication, and the role AI diagnostics should play in Canadian practice.
What the Pearl Oral Health Index Found
Pearl, the dental AI company behind the FDA-cleared Second Opinion platform, built the Index from two complementary datasets covering adults 18 and older across all 51 U.S. states (including D.C.) and the United Kingdom. The data spans April 2024 through March 2026, encompassing 26 million dental x-rays, 15 million unique patients, and 737 million individual teeth.
The headline finding: Pearl's AI detected an average of 6.07 decayed teeth per patient. The National Health and Nutrition Examination Survey (NHANES), the U.S. federal benchmark for oral health statistics, reports an average of 0.7 decayed teeth per adult. That represents a 4.5-fold discrepancy.
This is not a rounding error. It suggests that survey-based self-reported data systematically underestimates the burden of untreated dental disease — and that the actual clinical picture, visible on radiographs, is far worse than public health data has indicated.
Why Survey Data Misses So Much
The NHANES model relies on clinical examination and patient self-reporting. Patients underreport symptoms — partly because early-stage decay is often asymptomatic, partly because recall bias distorts the picture. Clinical exams without radiographic support miss interproximal caries, early lesions beneath intact enamel, and recurrent decay around existing restorations.
AI-powered radiograph analysis removes these limitations. Pearl's platform reads bitewings, periapicals, and panoramic images at scale, applying consistent diagnostic criteria across millions of films. The result is a dataset that reflects what clinicians actually see (or would see, with proper imaging) rather than what patients remember or report.
Pro Tip: If your practice relies on visual examination alone for caries detection during recalls, consider how many interproximal lesions you might be missing. The Pearl data suggests that radiographic screening catches disease that visual and probe-based examination alone cannot.
What This Means for Canadian Dental Practices
Canada does not have an exact equivalent to NHANES for dental health. The Canadian Health Measures Survey (CHMS) collects some oral health data, but with smaller sample sizes and less frequent reporting cycles. The Pearl findings, while based primarily on U.S. and U.K. data, carry direct relevance for Canadian practice.
First, the same self-reporting biases that skew U.S. federal data almost certainly affect Canadian surveys. Patients in Mississauga and Markham are no more likely than patients in Manhattan to accurately recall their last cavity. If anything, the gap between perceived and actual oral health may be wider in populations with limited access to regular dental care — a group that has expanded significantly since the Canadian Dental Care Plan (CDCP) began enrolling previously uninsured Canadians in 2024.
Second, the influx of CDCP patients into Ontario practices means many dentists are now seeing patients with years of deferred care. The Pearl data provides a useful benchmark: expect higher caries rates than patients report, and plan treatment accordingly.
Screening Protocol Implications
The Royal College of Dental Surgeons of Ontario (RCDSO) does not mandate specific radiographic intervals, deferring to clinical judgment. But the Pearl findings make a strong case for standardized bitewing protocols at recall appointments, particularly for new patients, CDCP enrollees, and anyone who has not had radiographs in the past 18 months.
Pro Tip: For practices onboarding CDCP patients, consider implementing a baseline radiographic protocol — a full mouth series or panoramic plus bitewings — at the first visit. The Pearl data suggests that clinical exam alone will miss the majority of existing decay in this population.
The Broader AI Diagnostic Trend
Pearl is not the only company applying AI to dental radiographs. Overjet, Dentistry.AI, and VideaHealth have all received regulatory clearances for AI-assisted diagnostic tools. Health Canada has been evaluating AI dental diagnostic devices through its Medical Device pathway, and several are expected to receive Canadian licensure in 2026 and 2027.
The Pearl Oral Health Index is significant not just for its clinical findings but for what it demonstrates about AI's ability to generate population-level health data from routine clinical imaging. A dataset of 26 million x-rays, analyzed consistently and at scale, produces epidemiological insights that no survey can match.
For Canadian dental practices evaluating AI diagnostic tools, the question is shifting from "Does AI work?" to "What do we do with what AI finds?" The Pearl data suggests the answer involves rethinking how practices communicate with patients about disease that is present but asymptomatic — and how practices plan for treatment volumes that exceed what patients expect.
Patient Communication Challenges
One practical consequence of AI-detected disease is the conversation that follows. When a patient arrives for a routine cleaning and AI flags four interproximal lesions on bitewings, the treatment plan escalates from prophylaxis to restorative care. That conversation requires trust, transparency, and — increasingly — the ability to show patients what the AI found.
Practices in the GTA that have adopted AI diagnostic tools report that showing patients the AI-annotated radiograph, with lesions highlighted and confidence scores displayed, improves case acceptance. The patient sees what the clinician sees, and the AI serves as an objective second opinion.
Pro Tip: If you present AI findings to patients, frame the technology as a diagnostic aid that helps your team catch problems early — not as a replacement for clinical judgment. The Canadian Dental Association (CDA) has emphasized that AI tools support, rather than replace, the dentist's diagnostic responsibility.
Data Limitations Worth Noting
The Pearl Oral Health Index is not without caveats. The dataset skews toward patients who visit the dentist and have radiographs taken — a population that is, by definition, already engaged with dental care. The true burden of untreated decay in the broader population, including those who never visit a dentist, is likely higher still.
Additionally, the Index currently covers the U.S. and U.K. A Canadian-specific analysis, using radiographic data from Canadian practices, would be valuable for calibrating public health planning. The CHMS could incorporate AI-assisted radiographic analysis in future survey cycles to close the data gap.
Frequently Asked Questions
Q: What is the Pearl Oral Health Index?
The Pearl Oral Health Index is the first large-scale radiographic census of oral health, built from AI analysis of 26 million dental x-rays across 15 million patients. It found that untreated decay rates are 4.5 times higher than traditional survey-based estimates, with an average of 6.07 decayed teeth per patient compared to 0.7 reported by NHANES.
Q: How does AI-powered caries detection affect dental practices in Canada?
AI diagnostic tools analyze dental radiographs with consistent criteria at scale, catching interproximal lesions and early decay that visual examination misses. For Canadian practices — especially those onboarding previously uninsured CDCP patients — AI-assisted screening can reveal treatment needs that exceed patient expectations, requiring adjusted treatment planning and patient communication strategies.
Q: Are AI dental diagnostic tools available in Canada?
Several AI dental diagnostic platforms are in various stages of Health Canada evaluation. While some tools have received FDA clearance in the United States, Canadian dental practices should confirm Health Canada licensure before adopting any AI diagnostic device. The regulatory pathway for AI-based medical devices in Canada follows Health Canada's Medical Device framework.
EBIKO Dental will continue monitoring developments in AI-assisted dental diagnostics and their implications for Canadian dental practice.