A landmark narrative review published in PMC (PubMed Central) has mapped the entire landscape of FDA-cleared AI solutions in dental imaging, confirming that AI-assisted radiograph analysis has crossed from early-adopter novelty into clinical infrastructure. As of September 2026, two platforms — Pearl and Overjet — hold a combined 18 FDA 510(k) clearances, covering caries detection, bone loss measurement, periapical radiolucency identification, and 3D CBCT segmentation. For Canadian dentists evaluating whether to integrate AI into their diagnostic workflow, the evidence base now extends well beyond marketing claims.
As of September 2026, the dental AI imaging landscape looks materially different than it did even 18 months ago. What was once a pair of niche software add-ons is now a category with regulatory depth, clinical validation across multiple study designs, and integration paths for most major imaging systems. The question for Canadian practices is no longer whether AI imaging works — it is which platform fits their clinical workflow, patient volume, and imaging hardware.
What the PMC Narrative Review Actually Found
The review, published in a peer-reviewed open-access journal indexed by PubMed Central, systematically catalogued every FDA-cleared AI solution currently available for dental imaging. Unlike manufacturer white papers or press releases, narrative reviews synthesize published evidence across studies, identify gaps in the literature, and contextualize clinical claims against actual regulatory submissions.
The core finding: dental AI imaging has moved from proof-of-concept to regulated, clinically validated tooling. The platforms with FDA clearances have undergone the 510(k) pathway, demonstrating substantial equivalence to predicate devices — meaning each cleared module had to show that it performs at least as well as existing diagnostic methods for the specific indication.
This matters because an FDA clearance is not a marketing badge. Each 510(k) submission requires clinical data, bench testing, and a defined intended use. A platform with ten clearances has been evaluated ten separate times for ten distinct diagnostic tasks.
Pearl: Eight FDA Clearances and the Broadest Clinical Coverage
Pearl's Second Opinion platform holds eight 510(k) clearances as of early 2026. The foundational clearance (K210365, March 2022) established real-time pathology detection on two-dimensional dental radiographs as a viable AI-assisted diagnostic category.
What Pearl detects on a single radiograph is extensive: up to 18 distinct findings per image, including caries, periodontal bone loss, calculus, periapical radiolucencies, defective restoration margins, and existing restorations. The platform integrates with bitewing, periapical, and panoramic radiographs — the three formats that account for the vast majority of dental imaging in general practice.
According to Pearl's published validation data, Second Opinion achieves over 95% accuracy in identifying dental conditions. The training dataset is described as the world's largest dental radiograph collection, annotated by board-certified dentists and radiologists. While "largest" is a manufacturer claim that is difficult to independently verify, the breadth of the training data is reflected in the number of distinct conditions the platform can identify — a model trained on a small dataset would not achieve clearance across eight separate clinical indications.
Beyond two-dimensional imaging, Pearl has expanded into 3D and panoramic analysis, with separate clearances for CBCT segmentation, pediatric caries detection, and periapical lesion contouring. This is significant because it means a practice using Pearl can apply AI analysis across multiple imaging modalities without switching platforms.
Overjet: Ten FDA Clearances and a Focus on Workflow Integration
Overjet leads the field in total FDA-cleared modules with ten as of 2026. The clearances cover caries and calculus detection for both pediatric and adult patients, periapical radiolucency identification, automated dental charting, and image enhancement.
The distinction between Overjet and Pearl is not primarily about which detects more conditions — both cover the core diagnostic categories. The difference is in workflow philosophy. Overjet has emphasized integration with practice management systems and insurance workflows, positioning AI analysis as a tool that simultaneously improves clinical accuracy and documentation completeness.
Automated charting is a notable Overjet capability. Rather than simply flagging pathology on an image, the platform can generate chart entries from radiographic analysis, reducing the manual documentation burden on the clinical team. For a busy practice handling 30 or more patients per day, this saves measurable time per appointment.
Pro Tip: When evaluating AI imaging platforms, request a trial period of at least 30 days. Most practices need two to three weeks before the clinical team integrates AI analysis into their natural diagnostic rhythm. A one-week demo is rarely enough to assess real-world utility.
What This Means for Canadian Practices
Neither Pearl nor Overjet is currently subject to Health Canada device licensing requirements in the same way that a physical diagnostic device would be — software as a medical device (SaMD) regulation in Canada is evolving, and practices should confirm the current regulatory status before purchasing. Health Canada's August 2026 guidance update on AI medical devices (covered in a previous EBIKO News post) signals that formal Canadian regulatory requirements for dental AI are coming, but the timeline remains uncertain.
That said, FDA clearance is a meaningful quality signal. Canadian practices that adopt an FDA-cleared AI platform are using software that has been evaluated against defined clinical benchmarks, not a prototype or an unregulated consumer product.
The practical considerations for a Canadian practice evaluating AI imaging are:
- Imaging hardware compatibility: Both platforms integrate with major sensor manufacturers, but confirm compatibility with your specific hardware before committing. A platform that requires a sensor swap adds significant cost.
- Practice management integration: AI analysis is most valuable when it flows directly into charting and treatment planning. Ask whether the platform integrates with your existing PMS (Dentrix, Open Dental, Curve, etc.).
- Per-image pricing vs. subscription: Pricing models vary. Some charge per image analyzed, others use monthly subscription pricing. A high-volume practice (1,000+ images per month) will see dramatically different economics depending on the model.
- Training and change management: The clinical team needs to understand what the AI is flagging and how to interpret confidence scores. A 95% accuracy rate means that roughly 1 in 20 findings may be a false positive — clinicians must remain the final diagnostic authority.
Pro Tip: Before purchasing, run a blinded test: have the AI analyze 50 radiographs your team has already diagnosed. Compare the AI findings against your existing treatment plans. This gives you a concrete, practice-specific accuracy benchmark rather than relying on manufacturer-published figures.
The Emerging Competitive Landscape Beyond Pearl and Overjet
While Pearl and Overjet dominate the FDA-cleared category, several other platforms are building AI imaging capabilities for dental. Dental Intelligence announced new AI-powered features launching in Q2 2026, focusing on practice analytics and workflow automation rather than pure diagnostic imaging. ScanO, VideaHealth, and other entrants are pursuing their own clearance pathways.
The competitive pressure is healthy for the market. As more platforms achieve regulatory clearance, pricing will likely compress, integration options will expand, and clinical validation data will deepen. For Canadian practices, this means that waiting another 12 to 18 months before adopting may yield better pricing and broader compatibility — but it also means missing out on the diagnostic accuracy improvements that early adopters are already capturing.
Clinical Evidence: What the Studies Show
The PMC narrative review compiled evidence across multiple study designs. The consistent finding is that AI-assisted radiograph analysis detects pathology that clinicians miss, particularly for early-stage caries and subtle bone loss patterns. One widely cited statistic: combining cloud imaging with AI-powered analysis increased caries detection accuracy by 20% across participating dental practices.
This does not mean that dentists are poor diagnosticians. It means that radiographic interpretation is inherently subjective, influenced by viewing conditions, fatigue, patient volume, and the specific training of the clinician. AI analysis provides a consistent second read — unaffected by the 3 PM fatigue factor that degrades human diagnostic accuracy in every clinical setting.
For practices seeing 25 or more patients per day, the cumulative impact of catching even one additional early-stage lesion per week is substantial over a year. Early intervention means smaller restorations, lower treatment costs for patients, and better long-term outcomes.
Pro Tip: Track your AI-detected findings for 90 days and calculate your incremental detection rate — the number of clinically significant findings the AI caught that your team initially missed. This metric makes the ROI conversation concrete.
What Canadian Dental Regulators Are Watching
The Royal College of Dental Surgeons of Ontario (RCDSO) has not issued specific guidance on AI-assisted imaging, but the College's existing standards on diagnostic accuracy and record-keeping apply regardless of the tools used. A dentist remains professionally responsible for every diagnosis, whether AI-assisted or not.
The Canadian Dental Association (CDA) has acknowledged AI as a transformative technology for the profession and is monitoring international regulatory developments. Health Canada's recent guidance update on AI medical devices signals that a formal framework for dental AI is under development, though the specific requirements and timeline remain to be finalized.
For Ontario practices, the practical guidance is straightforward: document your diagnostic process, maintain professional liability coverage that accounts for AI-assisted decision-making, and never defer clinical judgment entirely to software. AI is a diagnostic aid, not a diagnostic replacement.
Frequently Asked Questions
Q: Are AI dental imaging platforms like Pearl and Overjet approved for use in Canada?
Both Pearl and Overjet hold multiple FDA 510(k) clearances in the United States. Health Canada's regulatory framework for software as a medical device is evolving, and practices should confirm the current Canadian regulatory status before purchasing. FDA clearance is a meaningful quality signal but is not equivalent to Health Canada licensing.
Q: How accurate is AI dental imaging compared to a dentist reading radiographs?
Published validation data indicates that leading AI platforms achieve over 95% accuracy in identifying dental conditions. Studies show that AI-assisted analysis increases caries detection accuracy by approximately 20% compared to clinician-only interpretation, primarily by catching early-stage lesions that are difficult to detect visually.
Q: What does AI dental imaging cost for a typical practice?
Pricing varies by platform and model. Some platforms charge per image analyzed (typically $0.50 to $2.00 CAD per radiograph), while others offer monthly subscriptions. A mid-volume general practice processing 500 to 1,000 radiographs per month should expect annual costs between $3,000 and $15,000 CAD, depending on the platform and feature tier selected.