Regulatory Status

Cavitech AI is currently investigational in the United States. A 510(k) submission covering radiograph, periodontal, and derivative AI features is in preparation, with a separate de novo pathway for the soft-tissue oral-lesion classifier. A Pre-Submission (Q-Sub) meeting with the FDA review team is scheduled. Commercial distribution is permitted during this period under 21 CFR §812 Investigational Device Exemption provisions and standard investigational-use labelling.

U.S. law permits commercial distribution of investigational medical-device software under three distinct but mutually reinforcing legal frameworks. (a) The FDA Clinical Decision Support (CDS) Final Guidance (September 2022) establishes a four-factor test under §3060 of the 21st Century Cures Act defining non-device CDS. Several Cavitech features — the ambient scribe, administrative intake summaries, scheduling, billing, and manual treatment-plan workflows — fall outside medical-device scope under this guidance. (b) For features that ARE medical devices (soft-tissue classifier, radiograph analysis), 21 CFR §812 establishes the Investigational Device Exemption framework allowing commercial distribution under informed-consent and IRB oversight pending clearance. (c) All commercial marketing complies with FDA misbranding prohibitions under 21 CFR §801 — we never use the terms "FDA approved" or "FDA cleared" while clearance is pending.

Cavitech is a modular platform. Some features are medical-device AI that require clearance; others are administrative or documentation-only features that sit outside medical-device scope. Both lists below are covered by a single Indications for Use statement filed with the regulator.

Our current evidence base for the investigational-use claim is the peer-reviewed published literature linked below — in particular Gomes 2023 (CNN classifier at 95.09% accuracy on 5,069 clinical photographs), the Rokhshad 2024 systematic review establishing AI-performance benchmarks of 74–100% across oral-mucosa devices, and the Finkelstein decision pathway catalogued in Al-Shehri 2025. Our methodology mirrors this peer-reviewed precedent: multi-reader ground-truth adjudication (three board-certified oral pathologists per soft-tissue case, histopathology where biopsied; two oral radiologists plus adjudicator for radiograph studies). We are in active discussions with clinical leads at academic dental centres in the United States, the United Kingdom, and South Africa to host our pre-submission reader studies under Good Clinical Practice and IRB/ethics oversight. No clinical-investigation site is confirmed at the time of this disclosure. As each site is confirmed and receives ethics approval, the protocol pre-registration and subsequent results will be published on this page and in peer-reviewed venues before being cited in any regulatory submission. No marketing claim of clinical validation is made while this programme is in planning.

Our submission cites the following already-cleared dental AI devices as the substantial-equivalence / reference basis for our classifiers. Each link below opens the public regulator record.

Commercial distribution during clinical evaluation is governed by the investigational-device provisions below. Every paying customer receives investigational-use labelling and signs a clinical-evaluation acknowledgement at onboarding.

Patient health information is handled in accordance with HIPAA (45 CFR Parts 160, 162, 164). We execute Business Associate Agreements with every infrastructure partner handling PHI. Our cleared AI inference runs on self-hosted GPU infrastructure — no PHI is transmitted to third-party APIs for device features. Non-device features using third-party AI (administrative summaries, scribe transcription) operate under enterprise zero-data-retention agreements with written BAAs.

Our commitment for every market:

• Never use the terms “FDA approved”, “FDA cleared”, “UKCA marked”, “CE marked”, “SAHPRA registered”, or “clinically validated” in marketing material or in-product labelling while clearance is pending.
• Never market AI clinical features directly to patients.
• Never output AI clinical content directly to patients without a licensed dentist in the loop.
• Every AI-feature screen carries a visible investigational-use disclosure.
• Every generated document (referral letters, reports, treatment plans) includes the dentist’s sign-off and investigational notice.

Every URL cited on this tab, consolidated. 21 primary sources · all public.

• Clinical Decision Support Software — Guidance for Industry and FDA StaffU.S. Food and Drug Administration · 2022Final guidance, September 2022 — defines non-device CDS four-factor testhttps://www.fda.gov/regulatory-information/search-fda-guidance-documents/clinical-decision-support-software
• 21st Century Cures Act — Section 3060 (Clarifying Medical Software Regulation)U.S. Congress — Public Law 114–255 · 2016Statutory basis for non-device CDS exclusionhttps://www.congress.gov/bill/114th-congress/house-bill/34/text
• 21 CFR Part 812 — Investigational Device ExemptionsU.S. Code of Federal RegulationsLegal framework for commercial distribution of investigational deviceshttps://www.ecfr.gov/current/title-21/chapter-I/subchapter-H/part-812
• 21 CFR Part 801 — Labelling of Medical DevicesU.S. Code of Federal RegulationsLabelling rules we comply with (investigational-use notice, misbranding prohibition)https://www.ecfr.gov/current/title-21/chapter-I/subchapter-H/part-801
• The 510(k) Program: Evaluating Substantial Equivalence in Premarket NotificationsU.S. Food and Drug Administrationhttps://www.fda.gov/regulatory-information/search-fda-guidance-documents/510k-program-evaluating-substantial-equivalence-premarket-notifications-510k
• Requests for Feedback and Meetings for Medical Device Submissions: The Q-Submission ProgramU.S. Food and Drug Administrationhttps://www.fda.gov/regulatory-information/search-fda-guidance-documents/requests-feedback-and-meetings-medical-device-submissions-q-submission-program
• Marketing Submission Recommendations for a Predetermined Change Control Plan for AI/ML-Enabled Device SoftwareU.S. Food and Drug Administration · 2024https://www.fda.gov/regulatory-information/search-fda-guidance-documents/predetermined-change-control-plans-medical-devices
• Use of Artificial Intelligence in the Classification of Elementary Oral Lesions from Clinical ImagesIJERPH 20:3894 (Gomes et al., 2023) · 2023CNN baseline, 95.09% accuracy on 5,069 clinical photographshttps://doi.org/10.3390/ijerph20053894
• Artificial intelligence for classification and detection of oral mucosa lesions on photographs: a systematic review and meta-analysisClinical Oral Investigations 28:88 (Rokhshad et al., 2024) · 2024Systematic review establishing AI-accuracy benchmarks (74–100%)https://doi.org/10.1007/s00784-023-05475-4
• The Current State of Clinical Diagnostic Algorithms for Mucosal Oral Lesions: A Scoping ReviewOral Diseases (Al-Shehri et al., 2025) · 2025Peer-reviewed scoping review catalogueing the Finkelstein pathwayhttps://doi.org/10.1111/odi.15388
• Diagnostic accuracy of artificial intelligence-assisted clinical imaging in the detection of oral potentially malignant disorders and oral cancerInternational Journal of Surgery 110(8):5034–5046 (Li et al., 2024) · 2024https://doi.org/10.1097/JS9.0000000000001469
• A Guide to Clinical Differential Diagnosis of Oral Mucosal LesionsFinkelstein MW, Lanzel E, Hellstein JW — Dentalcare CE Course CE110, University of Iowa College of Dentistryhttps://www.dentalcare.com/en-us/ce-courses/ce110
• HIPAA Privacy Rule — 45 CFR Parts 160 and 164 Subparts A and EU.S. Department of Health & Human Serviceshttps://www.hhs.gov/hipaa/for-professionals/privacy/index.html
• HIPAA Security Rule — 45 CFR Parts 160, 162, 164 Subpart CU.S. Department of Health & Human Serviceshttps://www.hhs.gov/hipaa/for-professionals/security/index.html
• Pearl Second Opinion — 510(k) K203292U.S. FDAPrimary predicate for radiograph CADe/CADxhttps://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpmn/pmn.cfm?ID=K203292
• Overjet Dental AI — 510(k) K210379U.S. FDAPredicate for caries and periapical lesion detectionhttps://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpmn/pmn.cfm?ID=K210379
• VideaHealth Dental AI — 510(k) K222392U.S. FDAPredicate for caries + periodontal radiographic analysishttps://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpmn/pmn.cfm?ID=K222392
• Diagnocat — 510(k) K231140U.S. FDAPredicate for periodontal bone-loss measurementhttps://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfpmn/pmn.cfm?ID=K231140
• Investigational Device Exemption (IDE) — FDA OverviewU.S. FDAhttps://www.fda.gov/medical-devices/premarket-submissions-selecting-and-preparing-correct-submission/investigational-device-exemption-ide
• Design Considerations for Pivotal Clinical Investigations for Medical DevicesU.S. FDAhttps://www.fda.gov/regulatory-information/search-fda-guidance-documents/design-considerations-pivotal-clinical-investigations-medical-devices
• Software as a Medical Device: Possible Framework for Risk Categorization (N12)International Medical Device Regulators Forum (IMDRF) · 2014https://www.imdrf.org/documents/software-medical-device-possible-framework-risk-categorization-and-corresponding-considerations

Review the same information for another market: