The science and engineering behind TMDream
Evidence first: internationally accepted clinical guidelines, peer-reviewed research and specialized expertise. Then the engineering that brings it into clinical practice safely and traceably.
Built upon the best available scientific evidence
TMDream integrates internationally accepted guidelines, peer-reviewed research, specialized clinical expertise, thousands of verified clinical cases, and advances in artificial intelligence to support evidence-based clinical assessment while preserving independent clinical judgment.
1. Clinical guidelines
Internationally accepted diagnostic criteria and clinical recommendations.
DC/TMD · ICOP · ICSD-3 · AASM
2. Scientific evidence
Hundreds of peer-reviewed publications across multiple disciplines continuously inform the knowledge base.
3. Specialized clinical expertise
Knowledge contributed by experienced clinicians and researchers is translated into structured clinical reasoning.
4. Verified clinical cases
Thousands of curated and clinically verified real-world cases refine, evaluate and continuously improve the platform.
5,000+ curated clinical cases
5. Artificial intelligence
Advanced artificial intelligence integrates and applies complex clinical knowledge at the point of care.
How evidence becomes a clinical system
The five pillars are not an abstract principle: they shape the architecture. Every engine output is grounded in retrievable sources, and deployment is designed so clinical data never leaves the institution’s control.
Artificial intelligence in medicine demands the highest level of accuracy. The TMDream engine was developed and tested using real clinical datasets across orofacial pain and related disorders, built on rigorous academic foundations.
TMDream Engine Architecture
Proprietary three-layer stack: clinical data digitization, ensemble modeling, and validated medical RAG. Runs on-premises or in a private cloud with full data sovereignty.
OCR for unstructured notes and images, XGBoost for complex clinical patterns, and a RAG layer grounded in proprietary clinical JSON and medical literature, with no hallucinations.
Natural Language Processing (NLP)
Interprets free-text pain descriptions, extracting pain descriptors (such as 'pressure', 'burning', or 'throbbing') with high specificity based on evidence.
Multi-layered Neural Networks
Classifies complex profiles of orofacial pain and related disorders against a curated database of over 5,000 real-world clinical cases.
Validated Medical RAG
Ensures all triage suggestions are grounded in rigorous medical references, eliminating any risk of algorithmic hallucination.
From patient narrative to the right referral
Guided NLP triage, analysis in seconds, clinical prioritization, and direct routing to the right specialist (otolaryngology, general medicine, neurology, orthopedics, sleep medicine, and more), cutting unnecessary loops and care costs.
Four steps: semantic triage, AI engine in seconds, clinical prioritization, and referral to the appropriate specialist.
Sovereign Cloud Deployment
- Designed for LGPD-, GDPR- and HIPAA-compliant deployments
- Data Encryption at Rest and in Transit (TLS 1.3/AES-256)
- Anonymization of Personal Identifiers in Triage
Health Data Privacy & Compliance
Our architecture prioritizes corporate cybersecurity. The AI engine is hosted on client infrastructure, complying with the strictest requirements for medical data privacy.
Frictionless Interoperability
TMDream is designed to integrate into existing workflows in hospitals and clinics through modern REST APIs and connectors compatible with international standards such as HL7 and FHIR.