Research

Digital Surface Labs

Canine-first tooth scaling is a different product, not a faster human one

Dogs skip FDA 510(k). They do not skip anesthesia, veterinary licensure, or a later human filing.
2026-09-13 tooth-cleaning-robotveterinaryregulationgo-to-marketdentistry

There is a real canine path. It is not a shortcut to the human scaling robot.

A veterinary-only scaler or clinic robot can be sold in the United States without a 510(k) or PMA. That federal gap is genuine. Everything else that makes Joe's current concept a product — an awake patient, a bite-referenced intraoral ring, a mall visit, and later a human Class II file — does not transfer. Professional canine scaling is a general-anesthesia medical procedure on a registered veterinary premises. The awake "dog teeth cleaning" market exists, and every boarded veterinary dental college treats it as cosmetic scraping.

This is a product-strategy memo, not legal, dental, or veterinary advice.

The question, split in three

"Canines to speed time to market" mixes three businesses:

Path What you sell FDA premarket Who may operate it Anesthesia Speeds a later human scaler?
A. Awake pet-spa / NAD service Cosmetic tartar scraping on a conscious dog Irrelevant; state law blocks the service In California, a scaler is a veterinary dental operation Professionally rejected No. It stains the human medical brand.
B. Clinic device under anesthesia Capital equipment that helps a COHAT None, if labeled animal-only DVM, or RVT under DVM supervision, on a registered premises Required for standard of care Almost none for clinical effectiveness
C. Research dogs / cadavers No sale; a still mouth for perception and force IDE-exempt if truly veterinary or lab-animal IACUC / VCSC Yes, in a lab or teaching hospital Useful as preclinical safety/handling, not as a predicate

Joe's current concept is a DryShield-style ring distal to human canines, for an awake mall patient, with licensed hygiene as the operator constraint. Path A is the analog of that visit and is the wrong analog. Path B is a real business with a different machine. Path C is ordinary medical-device animal work.

Current human position

The August 24, 2026 decision memo and the California legal notes already bound the human path:

  • Scaling is licensed practice. A robot does not become a hygienist. In California an RDH may scale under general supervision (BPC 1910); unlicensed "robot techs" may not.
  • A human ultrasonic scaler is Class II (21 CFR 872.4850). Dental robotic navigation (Yomi) is Class II clinician-controlled implant guidance, not autonomous cleaning.
  • QMSR applies to finished devices intended for human use. Human testing needs IRB; significant-risk work needs an IDE.
  • The fundable first human product in that memo is still a plaque map plus bounded supragingival soft-plaque removal, not autonomous subgingival scaling.

A canine pivot does not relax those human rules. It creates a second product.

What actually gets faster

FDA's Center for Veterinary Medicine is explicit:

The FDA does not require submission of a 510(k), PMA, or any pre-market approval for devices intended for animal use.

Manufacturers who exclusively make animal devices need not register, list, or file MDRs. Quality system regulation (21 CFR 820) governs finished devices intended for human use. The remaining federal duties are not nothing: the device must not be adulterated or misbranded, labeling must name the species and give adequate directions, and radiation-emitting hardware still goes through CDRH. CVM can still seize or warn.

That is the whole federal shortcut. It is large.

It disappears if intended use is dual. Objective intent is not the word "veterinary" on the box. Labeling, ads, design, and founder statements count (21 CFR 801.4). A camera-guided bite-referenced robot that is publicly described as a human hygiene machine, then sold "for dogs first," is the fact pattern that pulls CDRH in.

Animal field success is also not a human 510(k) predicate. CDRH's animal-study guidance is about safety of a device intended for human use, in GLP nonclinical studies, not client-owned clinic cases. Substantial equivalence requires the same intended use as a legally marketed human device. The Animal Rule (efficacy in animals instead of people) is for CBRN countermeasures, not dentistry.

USDA-APHIS licenses biologics (vaccines). It does not license a scaler. Anesthesia drugs are already approved animal drugs under CVM; the robot maker is not filing a new NADA unless it sells a novel drug.

What does not get faster

California still requires a veterinarian

BPC 4826(d): performing a dental operation on an animal is the practice of veterinary medicine. 16 CCR 2037 defines that operation as using any instrument, device, or scaler on tooth, gum, or related tissue, including removal of calculus. The toothbrush/gauze/floss exception is explicit. A robotic scaler is not a toothbrush.

The 2011/2012 California rule change was aimed at groomers using hand scalers. Unlicensed NAD is unlicensed veterinary practice. AB 516 (2026) lets RVTs do dental care, including extractions, under veterinary supervision. It does not legalize a spa tech.

Premises must be registered (BPC 4853). A mall booth is not a workaround unless it is a registered veterinary premises with a responsible licensee manager.

So the human constraint is "RDH in a dental office." The canine constraint is "DVM (and usually an RVT) in a veterinary hospital, under general anesthesia." The license did not go away. It changed professions.

Anesthesia is the procedure

Adult dogs have 42 teeth, a scissor bite, and carnassials (maxillary P4 over mandibular M1). They do not sit for subgingival curettage, six-point probing, or full-mouth films. AVDC, AAHA, and AVMA all put professional scaling under general anesthesia with a cuffed endotracheal tube.

AVDC's April 2024 position (they now prefer "anesthesia-free tooth scraping"):

  • It is not a replacement for anesthetized dentistry.
  • It is not effective prevention.
  • Subgingival access on every tooth is impossible in an unanesthetized veterinary patient.
  • Visible tartar removal is purely cosmetic.

AAHA 2019: anesthesia-free dentistry "has not been shown to be safer or comparable… and is therefore unacceptable." AVMA policy (2014): periodontal probing, intraoral radiography, dental scaling, and extraction should be performed under anesthesia.

A 2026 JAVMA prospective comparison (Niemiec et al.) found no medical benefit from anesthesia-free dentistry versus an anesthetized dental: post-AFD periodontal scores stayed high (~4.35) versus ~0.09 after the anesthetized procedure.

Healthy-dog anesthetic death is on the order of 1 in 1,800 within 48 hours (CEPSAF). That is why owners skip dentistry, and why NAD sells. It is not a reason the specialty colleges will accept a mall robot.

The ring does not fit

Joe's fixture is a bite-referenced ring with pads distal to human canines, like DryShield. That load path is patient bite → pads → extraoral frame.

In a dog:

  • The canine teeth are the tallest in the mouth, not a convenient landmark behind a human-shaped block.
  • An anesthetized dog does not bite the pad. The ground has to become a table-mounted speculum, packing, and ET-tube clearance.
  • Interincisal opening under anesthesia ranges from about 40 mm (tiny dogs) to 180 mm (giant breeds).
  • Brachycephalic muzzles do not accept a human isthmus. Crowding and BOAS make intubation the hard part of the day.
  • Dog enamel is thin: mean roughly 0.3–0.5 mm, near the CEJ a film. Human occlusal enamel is millimeters. A force envelope tuned for people is not conservative enough.

Same slogan (tool poses in fixture coordinates). Different part.

The bottleneck is not the scaler operator

A human Peninsula hygiene visit is expensive because the RDH is expensive. A canine COHAT is expensive because of anesthesia, monitoring, radiographs, and extractions. US GP clean-only commonly runs $300–$700; with extractions $700–$2,000+. NAD cosmetics run $100–$300.

Vet tech median pay is about half of RDH pay. Clinics already own piezo or magnetostrictive scalers from iM3, Midmark, Dentalaire, Acteon. A robot that only scrapes crowns faster, still under GA, still with a DVM in the building, has not removed the cost owners fear.

Board-certified small-animal dentists are on the order of ~280 people in the US (AVDC total ~309 including equine). The GP clinic count is ~34,000. That is a real capital-equipment channel. It is not a mall.

US pet/veterinary dental spend is roughly $2.6–3.1 billion (vendor research, noisy), services ~70%. Human cleaning-visit payments in the August 24 memo sit in a $46 billion envelope. The 80%-of-dogs-by-age-three line is a high-end anesthetized/necropsy figure. Primary-care records show diagnosed periodontal disease closer to 10–20%; anesthetized probing finds much more, especially in extra-small breeds.

Has this been done?

Honest pattern: tools flow from human dentistry into veterinary operatories. Surgical and dental robots do not use dogs as a commercial beachhead for later FDA clearance.

The rare "vet SKU, then human" stories

Example What it was Did vet sales become a human product?
AliveCor Veterinary Heart Monitor (Sep 2012) then Kardia (510(k) Nov 2012) Same phone-case ECG. Vet SKU launched while the human file was already in. Timing and cash, not a substitute for the human 510(k). Cleanest device example, and a poor analog for a contact robot.
Merck ivermectin (livestock 1981 → Heartgard 1987 → Mectizan 1987) Same small molecule, different programs. Yes, for an antiparasitic. Irrelevant to a scaler.
VetStem (horses 2004, dogs 2007) → Personalized Stem Cells human IND 2019 Vet clinical volume and GMP used to open a human IND. IND acceleration. No approved human product.
PetVivo Spryng (vet device 2021) + human subsidiary Somatrix Explicit "vet devices skip premarket" strategy. Human OA still pipeline. Vet revenue. Human still TBD.
Zoetis Cytopoint (caninized IL-31 mAb, 2016) vs Nemluvio (different human mAb, 2024) Parallel immunology, not the same antibody. Concept, not a port.
Zoetis Librela (canine anti-NGF, 2023 US) vs tanezumab (human anti-NGF, killed 2021) Same target class. Dogs commercial, humans failed on rapidly progressive osteoarthritis. Anti-pattern. Dog success did not rescue the human class.

Dual-use tools (the actual industry)

Cavitron, EMS Piezon, Acteon, Planmeca, Midmark, Butterfly iQ, Masimo pulse ox, LiteCure/Companion lasers, PulseVet shockwave: human first, then a vet SKU, preset, or bite-resistant housing. iM3 and Dentalaire got rich staying in veterinary dentistry. They did not become human dental companies.

Dental and surgical robots

  • Neocis Yomi: FDA-cleared human implant robot, 100,000+ osteotomies. No veterinary commercial line.
  • Perceptive: human crown-prep robot, first-in-human work, no dogs-as-customers story.
  • da Vinci in vet: Cornell used a used Si-e in 2023 for a client-owned gastropexy. Intuitive did not enter veterinary. New systems are ~$2 million.
  • No commercial robotic hygienist exists in dogs or humans. Academic robotic dentistry on dogs is mandibular osteotomy, implant, or cadaver work — not prophylaxis.

Anesthesia-free pet dental (the cautionary market)

Pet Dental Services (25+ years, hospital-based NAD, BIOLASE diode deal), HealthySmiles, Creature Comforts, and a long tail of mobile spas sell $100–$400 awake scrapings. The National Pet Dental Association's own history is 1980s groomer "awake dentals" until states closed the loophole. This is a real cash business and a permanent fight with AVDC/AAHA. Partnering with it brands a later human medical device against the people who set standard of care.

Transferability (0–10)

Layer Score Why
Perception stack 4 Cameras, ultrasound, endoscopy as modalities transfer. Trained tooth graphs, lighting, and awake motion do not. An anesthetized dog is an easier imaging plant.
Force-control stack 5 Same physics (sub-newton lateral, water on, side of tip). Numeric limits must be retuned for thin enamel. Tip shapes and carnassial access differ.
Fixture 2 Human bite ring is the wrong object.
Business model 3 Awake high-frequency hygiene vs low-frequency GA medical day.
Regulatory dossier 7 as its own vet SKU; 1 as a stepping-stone No 510(k) for animal-only. Almost no reuse toward a human file. Dual-use labeling can make the human path worse.
Brand 2 "We scaled dogs" does not help Hillsdale. NAD-spa branding is toxic with specialists.

Overall reuse of the same robot: about 2–3/10.

Weighted TTM (weights = 100)

This scores time to getting paid to scale teeth, canine-first versus staying on the human mall path.

Weight Factor Canine-first Human-first
18 FDA premarket Pro. No 510(k)/PMA if animal-only. Con. Class II or De Novo, QMSR, likely clinical. Years after a working system.
16 Fixture and anatomy Con. Ring is scrap. Size × skull-type SKU explosion. Pro. One species, a size ladder, architecture already chosen.
14 Legality of the visit Con for NAD. Neutral-to-con for COHAT: still a DVM, GA, rads, registered premises. Con in a CA mall unless it is a real dental office (already in the legal notes). Known act.
12 Still patient vs service density Pro for engineering (still mouth). Con for a 10-minute slot business (you inherited a surgical day). Con for engineering (motion, gag). Pro for slot density.
10 Demand vs bottleneck Con. Demand is real; bottleneck is owner GA fear and price, not scaler speed. Pro if Peninsula RDH labor is the constraint. That is the mall thesis.
8 Incumbents Con. Competing with a few-thousand-dollar piezo, not a $0 workflow. Mixed. Human chairs already have ultrasonics.
8 Clinical completeness Con. Standard of care is rads + probing + subgingival ± extraction. Scaling-only is NAD with better PR. Mixed. Human hygiene is scaling/polish ± exam. Closer to a complete visit.
6 Liability / delayed diagnosis Con for NAD-shaped service. Managed if a DVM owns a full COHAT. Managed with a dentist of record; ugly if an unlicensed kiosk claims "cleaning."
5 Data reuse toward the other species Near zero. Do not count this as acceleration. Near zero.
3 Brand optionality Con. Dog-first stains a human mall brand. Con the other way if you later sell vet. Pick.

Net: canine-first is faster to a legal animal device in a vet OR. It is not faster to a scaling service that matches standard of care, because that service is already COHAT and the missing piece is not a robot. It is slower to the mall hygiene service the ring is for.

Realistic clocks after a clinic-ready system, not after earplug transfer:

  • Vet-only capital device, honestly labeled: ~3–18 months to first legal revenue (sales cycle and clinic proof, not FDA).
  • Human Class II / De Novo robot-assisted scaler: ~2–5+ years of Q-Sub, testing, review, and clinical work. A vanilla scaler 510(k) is months; a camera-guided contact robot is not vanilla. Yomi spent years before the 2017 clearance, then many follow-on 510(k)s.
  • CA NAD spa: not a 2026 legal revenue path.

Recommendation

Do not pivot the commercial product to dogs in order to reach humans faster.

  1. If the goal is the human mall scaler, stay on the human path already in the August 24 memo: measurement, bounded supragingival contact, licensed dental office. Use typodonts, extracted teeth, and (if needed later) GLP/IACUC animal safety studies of the human device. Do not sell a vet SKU with human-pivot language on the site.

  2. If the goal is earlier cash, the only professionally defensible canine product is a B2B clinic robot that shortens anesthetized COHAT time (mapping, subgingival access, less fatigue) for GP or specialty veterinary dentistry. Judge it as its own TAM against iM3/Midmark, not as a stepping-stone. Kill it if it does not save anesthesia minutes.

  3. Do not build an awake dog-cleaning service. It is the NAD market: legally constrained in California, medically rejected by AVDC/AAHA/AVMA, and poisonous next to a later human 510(k) story.

  4. Anesthetized research dogs are a testbed, not a GTM. A still, intubated mouth is an easier plant than an awake mall patient for closing a perception-and-contact loop. That is Path C. It requires IACUC (dogs are AWA-covered), and if the study is really to support a human file it is not "solely veterinary use."

Concrete next steps (only if Joe wants a canine option)

None of these start hardware motion, spending, or outreach.

  • Keep human and veterinary intended-use language in separate documents. One public sentence about "dogs then people" can collapse the CVM exemption.
  • If a clinic SKU is worth a week of design: list what would have to change in the ring (table ground, ET-tube clearance, size classes, no cooperative bite) before any CAD.
  • If a research-dog protocol is worth considering: IACUC/VCSC path, not a service business. Treat it as force and imaging physics, not clinical claims.
  • A short CVM AskCVM letter on "veterinary use only" wording, plus California veterinary-board counsel on premises, is the cheap check before any canine commercial label. Do not send those letters unless Joe asks.

Open questions

  • Would a GP vet pay capex to save 10–20 minutes of a COHAT, or only if the machine also improves radiographs, probing, or extraction planning?
  • How much of Joe's force-control research is worth doing on extracted dog teeth versus human teeth versus phantoms? Extracted-tooth enamel studies already exist in the vet literature.
  • If the human first product stays soft plaque, does a canine calculus-heavy mouth even teach the right contact problem?
  • EU has no harmonized veterinary-device 510(k) analog either; a vet SKU is not a European human MDR shortcut.

Sources

FDA / CFR

California

Standard of care

Market, anatomy, precedents