Ministry of Health Protocol

IRB Submission Package

System Detected Cancer Cells (SDCC): An Integrated Generative AI Framework for Early-Stage Oncology & Hepatology

📋

Submission Type

Initial Review (New Protocol)

🔬

Study Design

Retrospective / Observational Diagnostic Validation

⚖️

Risk Level

Minimal Risk (Retrospective Data Only)

Research Question

Can a multimodal Generative AI framework (Gemma-based) identify pre-cancerous "cold" tumor phenotypes and pre-cirrhotic liver stages (F2/F3) more accurately than current standard-of-care histology?

🎯 Primary Objective

To validate the sensitivity and specificity of the MedGemma model in staging liver fibrosis (F0-F4) using retrospective patient data from Abu Dhabi healthcare facilities.

💡 Innovation

A proprietary AI stack (C2S-Scale, MedGemma) that integrates single-cell transcriptomics with histopathology to detect molecular transition states before clinical visibility.

UAE National Health Alignment

Why This Research Matters for Abu Dhabi

📊

Disease Burden

The UAE has a projected rapid increase in liver disease (NAFLD/NASH) driven by high rates of Type 2 Diabetes and obesity. Early intervention is critical.

🏥

The Gap

Urgent need for non-invasive digital biomarkers to replace liver biopsies, which carry bleeding risks, high costs, and sampling errors.

✅

Population Benefit

Early detection of pre-cirrhotic (F2/F3) stages allows for lifestyle and therapeutic interventions that can reverse disease progression.

Study Design & Methodology

Study Population

1

Data Source

De-identified historical patient records from partner hospitals in Abu Dhabi (pathology slides + EHRs)

2

Sample Size

500 cases (250 Liver Fibrosis, 250 Oncology controls)

3

Gold Standard

AI output compared against pathologist-confirmed diagnosis

Inclusion & Exclusion

✓ Inclusion Criteria

  • • Patients aged 18+ years
  • • Confirmed histological reports for liver fibrosis or solid tumors
  • • Complete medical records available

✗ Exclusion Criteria

  • • Incomplete patient records
  • • Low-quality digital slide scans
  • • Missing key diagnostic information

UAE Regulatory Compliance

🔒

Data Sovereignty

Strict adherence to Federal Law No. 2 of 2019 (Health Data Law)

✓ Local Processing: All patient data remains within UAE borders
✓ gemma.cpp: On-premise inference (no cloud offloading)
✓ De-Identification: All PII removed, unique Study IDs assigned
✓ Encryption: At rest and in transit (ADHICS standards)
🤖

Responsible AI

Transparent, explainable AI that meets DOH ethical standards

✓ Explainability: SigLIP encoders highlight specific ROI on slides
✓ Human-in-the-Loop: AI provides "Second Opinion" only
✓ No Training: Validation only - no UAE data used for training
✓ Final Authority: Certified pathologist makes all diagnoses

Safety Monitoring & Validation

Performance Targets

  • • AUROC > 0.90 for Fibrosis Staging (matching literature benchmarks)
  • • Sensitivity & Specificity: Comprehensive reporting vs. gold standard
  • • Fail-Safe Protocol: Study paused if AI discordance exceeds 15%

Ethical Safeguards

  • • Minimal Risk: Retrospective data, no patient contact
  • • Waiver of Consent: Requested for archival data analysis
  • • Restricted Access: PI and authorized Co-Is only

Research Team

👨‍⚕️

Dr. Mostafa Abdelkader Elmourabea

Principal Investigator & System Architect

Creator of the Elmourabea System and lead researcher in predictive oncology and thermodynamic biology. Expert in integrating AI with molecular medicine.

📧 m.elmourabea@gmail.com
📧 ceo@eldoctooor.ae
📱 +971 50 7971431
📱 +971 52 270 9300

Core Team & Qualifications

AI

Technical Lead

Expert in Bioinformatics & Generative AI (Gemma Architecture)

MD

Pathology Consultant

Board-certified pathologist for AI validation and clinical interpretation

🔒

Data Privacy Officer

Ensures ADHICS compliance and UAE data protection standards

Note: All team members have completed DOH-mandated Research Ethics & GCP training

Future Therapeutic Potential

Beyond diagnosis - paving the way for programmable medicines

Phase II: ENSURE Platform

While this protocol focuses on diagnosis, success enables investigation of programmable tRNA therapies for patients with nonsense mutations (e.g., Arg-TGA).

  • → Target "orphan" genetic diseases in consanguineous populations
  • → Key area for UAE genetic research priorities
  • → Addresses 30 million patients with Stop Codon Diseases globally

Clinical Translation

Validated AI findings create pathways for therapeutic synergy licensing and precision medicine implementation.

  • → Silmitasertib + IFN-gamma synergy for cold tumors
  • → Non-invasive monitoring for chronic liver disease
  • → Population health screening programs

Request for Approval

Summary

✓ High Impact

Addresses critical liver disease & oncology burden in UAE

✓ Low Risk

Retrospective, observational, non-interventional study

✓ Compliant

Fully localized data processing (UAE Health Data Law)

For inquiries or additional information:

Dr. Mostafa Abdelkader Elmourabea

m.elmourabea@gmail.com | ceo@eldoctooor.ae

+971 50 7971431 | +971 52 270 9300