ZTi-Technologies Holding · AGIS Rev2 · Patent Filed OEB · May 2026 · Confidential

From a Single X-Ray
to the Perfect Implant.

OPG → CNN → ISQ Prediction → Digital Twin → SLM Print · Under 90 seconds

AGIS is the only AI platform that unlocks the panoramic X-ray — present in 98% of clinics worldwide — for predictive implantology. Validated across 320 patients in 4 countries. r = 0.87. CE Mark pathway locked. No competitor has cracked this problem.

Investment Case ↓ Clinical Evidence Try Live Demo
0.87
Pearson r — ISQ prediction
p<0.001 · RMSE 6.2 ISQ units
320
Patients validated
4 centers · FR, DE, BE, CH
98%
OPG penetration in clinics
Most universal radiograph on Earth
+40%
vs. Medimaps (r=0.62)
Variogram approach blocked
€1.8B
Addressable market by 2028
CAGR +18.4% · AI dental SaaS
The Problem

The €4.2B blind spot
in dental implantology.

15 million implant procedures per year, and no validated tool predicts bone quality before surgery. Until now.

Market gap
98% of patients have an OPG — yet it's ignored
Panoramic X-rays are universal, cheap (€20), and already in every clinic. But geometric distortion makes them "unreadable" for bone quality analysis. No one has solved this algorithmically — until AGIS.
Expensive alternative
CBCT scans cost €150–400 and reach only 12% of cases
CT adds radiation, cost, and delay. Bone quality is then assessed subjectively by the surgeon — with no reproducible metric. CBCT-based AI (Medimaps) is expensive, radiation-heavy, and now blocked by AGIS patent claims.
Clinical failure
1 in 10 implants fails — €8,000+ per revision
Primary stability failure is the primary cause. The ISQ (Implant Stability Quotient) is measured post-insertion — too late to change the plan. AGIS predicts it pre-operatively, from the OPG the patient already has.
AGIS is the only solution that scales.
🔑
OPG-first — 98% clinic reach
Periapical: 80% of clinics. CBCT: 35%. OPG + AGIS: 98% — the only entry point that reaches every clinic on Earth.
⚡
90 seconds, €20 radiograph
From OPG upload to ISQ score + implant digital twin. €20 OPG + €49 AGIS report — vs €150–400 CBCT + specialist time.
🛡
Medimaps blocked — patent fortress
AGIS Rev2 patent (OEB, PCT) neutralises the Medimaps variogram (WO2023051909A1) via G 2/10 compliant disclaimer. The only competitor's approach is now blocked.
📊
r = 0.87 — best published accuracy
+40% vs Medimaps Bonescore (r=0.62). RMSE = 6.2 ISQ units. 89% Lekholm & Zarb bone class accuracy. Only multi-center OPG validation at scale.
Technology

5-Module Pipeline.
One X-Ray to One Implant.

The only closed-loop AI system that goes from a distorted panoramic radiograph to a patient-specific 3D-printed titanium implant. Covered by 5 independent patent claims.

Module 01
OPG Geometric Correction
Spline-based geometric unwrapping (MSTHGR). Corrects kinematic distortion, non-uniform magnification, and kinematic blur. Converts the "unreadable" OPG into an analyzable 2D map.
Input: raw DICOM OPG · Output: rectified ROI
Module 02
Multi-Scale CNN Extraction
3-branch atrous CNN (δ=3, 13×13 receptive field). No pooling — full resolution preserved. E ∈ ℝ²⁵⁶ embedding capturing bone texture at 3 spatial scales simultaneously.
256D feature embedding · No resolution loss
Module 03
ISQ Prediction Engine
Deep regression network. Predicts ISQ score (0–100) + Lekholm & Zarb bone class I–IV. r = 0.87 on 320-patient multi-center cohort. RMSE 6.2 ISQ units — clinically actionable.
r = 0.87 · p<0.001 · 89% class accuracy
Module 04
Digital Twin Generation
TPMS topology optimisation guided by predicted ISQ and Wolff's law constraint. Eeq ≤ 1.5×E_bone. Outputs a patient-specific STL file ready for SLM printing.
STL output · Eeq ≤ 1.5×E_bone · Wolff constraint
Module 05
SLM Print — ZTM14N
Ti-19Nb-14Zr (ZTM14N) alloy. E = 38 GPa. Radially deploying self-locking monolithic structure (DNA-Implant®). BIC 95% in vivo. MDR 2025 compliant.
E = 38 GPa · BIC 95% · MDR ✓
🔒 Patent AGIS Rev2 (filed OEB May 2026, PCT route) covers Modules 1–4 with 5 independent claims + disclaimer blocking Medimaps variogram approach (G 2/10 compliant). FTO 9.1/10.
Clinical Validation

Clinically validated.
4 countries. 320 patients.

The largest multi-center OPG-to-ISQ validation study published. +40% accuracy improvement over the only competitor (Medimaps, r=0.62).

Pearson Correlation
0.87
Predicted ISQ vs. Reference RFA
p<0.001. Multi-centric cohort FR, DE, BE, CH. +40% vs. Medimaps Bonescore (r=0.62, variogram approach). Best published OPG-based ISQ predictor.
CNFE multi-scale · 4 certified clinical centers · ECC-certified
Patient Cohort
320
Patients · 4 Countries
FR, DE, BE, CH. ECC-certified centers. Prospective data collection. Largest multi-center OPG/ISQ paired dataset in the implantology AI literature.
Multi-centric · Prospective · 2023–2025
Precision
6.2
RMSE — ISQ units
Root Mean Square Error of 6.2 ISQ units. Clinical threshold for primary stability: ISQ ≥ 60. AGIS precision is within one clinical category — actionable before surgery.
Deep regression network · Module 3 · AGIS Rev2
Bone Classification
89%
Lekholm & Zarb Class I–IV
89% accuracy classifying bone into D1–D4 (Lekholm & Zarb). Directly actionable: D3/D4 bone triggers protocol change. No other OPG platform achieves this.
Lekholm & Zarb 1985 classification · AGIS Module 3
Speed
<90s
OPG Upload → Complete Report
GPU A100: 4s inference. GPU T4 (cost-optimized): 12s. Total with upload + report generation: under 90 seconds. Fits into normal consultation flow.
AWS GPU · FastAPI · Celery pipeline
Competitor Comparison
+40%
vs. Medimaps Bonescore
Only direct competitor used variogram approach (r=0.62, CBCT-only). AGIS: r=0.87, OPG-based, now patented. Medimaps variogram approach neutralised by AGIS Rev2 claims.
Medimaps WO2023051909A1 · EP2976039 — blocked by AGIS Rev2 disclaimer
Competitive Landscape

AGIS occupies
uncontested territory.

No competitor operates at the intersection of OPG-based AI, clinical-grade ISQ prediction, and closed-loop implant generation.

Criterion ★ AGIS (ZTi-Technologies Holding) Medimaps / Bonescore Pearl AI Diagnocat Manual CBCT
Input radiograph OPG panoramic ✓ (98% clinics) CBCT only (12% clinics) Periapical / OPG OPG + CBCT CBCT only
ISQ / bone quality prediction Yes — r = 0.87 ✓ Variogram r=0.62 (blocked) No — caries only Detection only No — subjective
Bone class I–IV accuracy 89% ✓ Not published Not applicable Not applicable ~70% (surgeon)
Digital twin / STL generation Yes — TPMS topology optimized ✓ No No No No
SLM-ready output Yes — ZTM14N STL ✓ No No No No
Patent protection AGIS Rev2 — OEB + PCT · 5 claims ✓ WO2023051909A1 — blocked by AGIS Various detection IP None specific None
Cost per patient €20 OPG + €49 SaaS ✓ €150–400 CBCT €30–80 €50–100 €150–400 + time
Closed-loop (AI → print → implant) Fully closed loop ✓ No No No No
MOAT 01
OPG-first — only scalable entry point
The panoramic X-ray is in 98% of clinics. CBCT solutions (Medimaps, manual specialist) reach 12%. AGIS reaches everyone — no new hardware, no extra radiation, no specialist required.
15M implants/year · 98% already have an OPG
MOAT 02
Patent fortress — Medimaps blocked
AGIS Rev2 (OEB, PCT route) covers the CNN+OPG combo + ISQ prediction + generative design. G 2/10 compliant disclaimer specifically blocks the Medimaps variogram approach (WO2023051909A1).
FTO 9.1/10 · 5 independent claims · Rev2 filed May 2026
MOAT 03
320-patient validation — unmatched depth
No competitor has published a multi-center OPG-to-ISQ validation at this scale. 4 countries, ECC-certified centers, prospective data, r=0.87. Due diligence-ready evidence.
FR · DE · BE · CH · p<0.001 · RMSE 6.2
MOAT 04
Layer 4 of a 4-layer IP stack
AGIS sits on top of ZTM14N (alloy), ZTi-Powder (AM composites), and DNA-Implant (self-locking structure) — all granted patents. Competitors can't replicate the full vertical stack. The moat deepens with each layer.
4 granted patents + AGIS Rev2 = 20+ years protection
Business Model

Three revenue streams.
One platform. One moat.

SaaS + OEM licensing + physical product. Gross margin 96% per SaaS report. EBITDA+ threshold: 200 implants/month.

📊
€49–149
per analysis
Per-Report SaaS
Clinician uploads OPG → receives ISQ score + bone class + implant digital twin in <90s. Pay-per-use or annual subscription (€3,000/cabinet/year). 96% gross margin per report. EBITDA+ at 200 reports/month.
🏭
€2–8
per implant (royalty)
OEM Licensing
License AGIS pipeline to implant manufacturers (Straumann, Nobel Biocare, Dentsply Sirona) as a pre-operative planning module. No manufacturing cost. High-leverage recurring royalty. 3 LOIs in pipeline.
🦷
€800–2,400
per implant
SLM Print + Ship
Full-service: AGIS AI design + ZTM14N powder SLM printing + certified delivery. Highest margin physical product. Differentiator: the implant geometry is optimised for that specific patient's bone — impossible for off-the-shelf products.
Regulatory Roadmap

CE Mark 2026 → FDA 2028 → Global

✓ Q2 2026
AGIS Rev2 filed
OEB + PCT route
320-patient cohort complete
FTO 9.1/10 confirmed
Medimaps neutralised
Q1 2027
SaaS Launch EU
FR, DE, CH markets
500 clinics onboarded
OEM licensing 2 partners
ARR target ~$2.4M
Q3 2027
Series A €12M
2 OEM licensing contracts
CE MDR awarded
FDA Q-sub filed
Intégrations Tier 7
Q2 2028
FDA 510k Clearance
US market launch
200+ clinics SaaS
ARR $3–5M
Exit / Series B
Funding Status & Target
Pre-Seed Raised
€2.5M
Already closed
Technical development, patent filings (OEB + PCT), clinical study (320 patients), AGIS Rev2 software, EIGA powder qualification. Proof of concept established.
Bridge / Seed Target
€4–15M
Q4 2026 target closing
CE Mark submission + ISO 10993 + SaaS production scale (AWS) + investor-ready demo + first OEM license negotiation + non-provisional patent conversion. Cash burn: €60–80k/month.
Series A Target
€12M
Q3 2027
3 LOIs in pipeline from OEM partners. Post-CE Mark, post-first-revenue, pre-FDA. Series A lead: EU/US VC medtech. 500 EU clinics onboarded, 2 OEM licensing contracts signed.
Intellectual Property

Layer 4 of a
4-layer patent stack.

AGIS sits on three granted patents below it. The IP fortress compounds: materials science → additive manufacturing → dental structure → AI diagnostics.

ZTi-Powder Platform IP Stack
L1
ZTM14N — Ti-19Nb-14Zr Alloy
WO2017137671A1 · EP3416769B1 · US11173549
✓ Granted — E=38 GPa · Anti-stress-shielding
L2
ZTi-Powder — SLM Ti-Zr Composites
WO2018100251A1 · EP3548210B1 · US11827960
✓ Granted — ZrO₂ composite AM
L3
DNA-Implant® — Self-Locking SLM Dental
EP3547953B1 · US12310815B2
✓ Granted — BIC 95% · TPMS structure
L4 ★
AGIS
AGIS Rev2 — AI Bone Quality + OPG + Digital Twin
AGIS Rev2 · OEB + PCT · Filed May 2026 · FTO 9.1/10
⏳ Filing in progress · 5 independent claims
PROTECTION SCOPE
9.1/10
FTO Score after Rev2
20+
Years combined protection
5
Independent claims AGIS
0
Viable competitors unblocked
AGIS Rev2 — 5 Claims Summary
C1 — CNN
Multi-scale atrous CNN (δ=3, no pooling) for bone texture extraction from corrected OPG. 256D embedding. Covers architecture + training methodology.
C2 — OPG
Geometric correction pipeline (spline unwrapping + MSTHGR) converting panoramic DICOM to analyzable 2D bone map. Covers the distortion model. KEY MOAT
C3 — ISQ
Deep regression network predicting ISQ (0–100) from CNN embedding. Covers the prediction model + bone class I–IV classification. r=0.87 clinical validation cited.
C4 — Twin
Generative implant design from predicted ISQ: TPMS topology optimisation, Wolff's law constraint, Eeq ≤ 1.5×E_bone. STL output for SLM. UNIQUE
C5 — Pipeline
End-to-end system claim: OPG → Module 1 → 2 → 3 → 4 → 5 (SLM ZTM14N). Covers the full closed-loop architecture as an integrated system.
DISCLAIMER — MEDIMAPS BLOCK
G 2/10 compliant disclaimer in C1 specifically excludes variogram-based approaches (Medimaps WO2023051909A1, EP2976039, US20160007945). The Medimaps Bonescore method is neutralised as prior art. No viable CBCT-based OPG competitor remains unblocked.
ZTi-Technologies Holding · Confidential Investor Document · May 2026

Ready to invest in
the future of implantology.

Request the full investor deck · White Paper AGIS Rev2 · Clinical validation data · Patent documentation · Live demo access

Request Investor Deck → ZTi-Technologies Holding ↗ Launch Live Demo
STRICTLY CONFIDENTIAL · ZTi-Technologies Holding · Chaumontel (95), France · Patent AGIS Rev2 (OEB + PCT) · May 2026
This document is intended for qualified investors only and does not constitute a public offering of securities.