Overview
The Master Orchestrator is the BRAIN of the Multi-Agent System. It coordinates all agents, manages workflows, handles state, and ensures proper sequencing and error handling across three integrated subsystems.
Key Responsibilities:
System Architecture
Three Integrated Subsystems
1. CardioAI Platform
ECG/Echo/Lab Analysis with 7 specialized AI agents
2. IoMT Platform
Device Management & Real-time Monitoring
3. PACS/DICOM Platform
Medical Imaging & Analysis
Subsystem 1: CardioAI Platform
Advanced cardiovascular AI analysis with seven specialized agents working in concert.
Capabilities:
ECGAgent
STEMI/NSTEMI detection, arrhythmia analysis
EchoAgent
HFrEF/HFpEF analysis, EF calculation
LabAgent
Troponin, BNP, lipids interpretation
RiskAgent
ASCVD risk stratification
DiagnosticAgent
Multi-modal diagnostic synthesis
TreatmentAgent
Evidence-based treatment recommendations
PrognosisAgent
5-year MACE prediction and mortality risk assessment
Task Types:
ECG_ANALYSIS- Real-time ECG interpretationECHO_ANALYSIS- Echocardiogram processingLAB_INTERPRETATION- Biomarker analysisCARDIAC_DIAGNOSIS- Complete diagnostic workflow
Subsystem 2: IoMT Platform
Internet of Medical Things platform for device management and real-time patient monitoring.
Device Management
- • Device registration & provisioning
- • Health monitoring & heartbeat
- • Credential generation
Protocol Support
- • MQTT (primary)
- • HTTP/REST
- • CoAP
- • WebSocket
Task Types:
DEVICE_REGISTRATION- Register new IoMT devicesTELEMETRY_PROCESSING- Process incoming vitalsALERT_GENERATION- Generate threshold alertsDEVICE_MONITORING- Monitor device health
Subsystem 3: PACS/DICOM Platform
Picture Archiving and Communication System for medical imaging analysis.
Ingestion
- • DICOM C-STORE
- • Study management
- • Orthanc integration
Analysis
- • Image segmentation
- • Quantification
- • Specialized analysis
Reporting
- • DICOM SR generation
- • PDF reports
- • Structured findings
Task Types:
DICOM_INGESTION- Ingest DICOM studiesIMAGE_SEGMENTATION- Segment cardiac structuresQUANTIFICATION- Calculate measurements (EF, volumes)IMAGING_DIAGNOSIS- AI-based diagnosisREPORT_GENERATION- Generate clinical reports
Automation Tier System
Four-tier system for managing clinical decision automation based on AI confidence and clinical context.
| Tier | Confidence | Auto Action | Ack Required | Timeout |
|---|---|---|---|---|
| Tier 1 | 0-60% | No | Optional | 600s (10m) |
| Tier 2 | 60-85% | No | Yes | 300s (5m) |
| Tier 3 | 85-95% | Yes* | Yes | 120s (2m) |
| Tier 4 | 95-100% |