Why it matters
Without this system, misreads or delayed recognition of fracture risk can lead to career-ending injuries, missed entries, or welfare compromises; timely actionable data is essential for prevention.
From design to safety
Design
AI-fracture risk imaging design
In Design, the yard defines data flows from standing CT, standing MRI, fetlock imaging, bone density, and radiologist reports to drive risk scoring and reporting templates. Templates specify thresholds for flag tiers and audit trails for compliance.
- Standing CT Scanner
- Standing MRI Unit
- Fetlock Imaging Protocol
Build
Build imaging data pipeline
Build integrates scanners, data storage, and AI modules with standardized imaging protocols, consent, and annotation schemes, ensuring traceability and regulatory readiness. It maps file schemas to the racing regulator's data dictionary and establishes encryption at rest.
- Standing CT Scanner
- Standing MRI Unit
- Fetlock Imaging Protocol
- Bone Density Scan
Operation
Operate imaging workflow
During Operation, the AI ingests daily scan data, runs fracture-risk models, flags anomalies, updates patient records, and notifies clinicians for review. It also cross-checks with clinical signs, tracks scan-to-symptom latency, and generates weekly risk dashboards.
- Bone Density Scan
- Radiologist Reports
- Standing CT Scanner
- Standing MRI Unit
Maintenance
Maintain imaging integrity
Maintenance monitors scanner calibration, image quality checks, data integrity, and backup routines, plus regular audits of imaging protocols and software rules. Procedures document error budgets, alert thresholds, and version-controlled rule updates with rollback.
- Radiologist Reports
- PACS/Imaging Archive
- Standing CT Scanner
- Standing MRI Unit
Safety
Safety and regulatory alignment
Safety ensures leg health surveillance, regulatory compliance for imaging, appropriate consent, and escalation to the yard veterinary surgeon for welfare decisions. Audits verify radiologist sign-offs, patient isolation procedures for infectious findings, and withdrawal time adherence.
- Radiologist Reports
- PACS/Imaging Archive
- Bone Density Scan
Part by part
- 1
Standing CT Scanner
- Watch for
- motion artefacts or patient movement compromising image clarity
- AI action
- detects subtle cortical changes; triggers radiologist review and repeat scan if needed
- 2
Standing MRI Unit
- Watch for
- scan gaps due to fixation or patient tolerance limits
- AI action
- assesses soft tissue integrity; flags high-risk fetlock structures for priority review
- 3
Fetlock Imaging Protocol
- Watch for
- fetlock joint images with poor resolution or misalignment
- AI action
- measures joint space and subchondral changes; flags risk elevation for clinical follow‑up
- 4
Bone Density Scan
- Watch for
- variable readings due to hydration or recent exercise
- AI action
- estimates bone mineral density; assigns fracture-risk level and recommended imaging cadence
- 5
Radiologist Reports
- Watch for
- inconsistencies between imaging findings and clinical signs
- AI action
- integrates imaging metrics with narrative reports; suggests additional tests or scans
- 6
PACS/Imaging Archive
- Watch for
- data integrity gaps or missing patient identifiers
- AI action
- routes scans to clinicians; auto-tags with case identifiers and annotations
Worked example
Imaging QA: Fetlock scan anomaly detection
Illustrative scenario with sample data.
A rider reports subtle fetlock discomfort; AI reviews CT and MRI to pre-empt fracture risk.
Signal in
standing CT: 2D slices; MRI: T1/T2; fetlock imaging: 3 views; bone density: BMD score; radiologist notes: last report date
System out
risk flag: moderate; recommended follow-up MRI in 6 weeks; radiologist to annotate suspected microfracture regions; update patient record with new imaging cadence
Human sign-off: Yard veterinary surgeon validates risk assessment; licensed trainer approves follow-up schedule and welfare adjustments
How to roll it out
Phase 1
Weeks 1 to 4
Baseline data capture, scanner calibration, and initial AI rule setup with radiologist collaboration.
Phase 2
Weeks 5 to 12
Pilot imaging cases; refine thresholds and alerting; establish report templates and clinician sign-off workflows.
Phase 3
Weeks 13 to 20
Full integration into daily welfare checks; formal training for staff; regulatory documentation alignment.
Where people stay in charge
- Only a licensed veterinarian or radiologist may interpret clinical risk flags and authorize imaging repeats.
- AI decisions must be reviewed against clinical signs, with human override available for welfare concerns.
- All data handling complies with medical record standards and racing regulator requirements.
Questions owners and crews ask
What data sources feed the AI for fracture risk assessment?
The AI ingests standing CT scans, standing MRI scans, fetlock imaging views, bone density measurements, and corresponding radiologist reports to produce an integrated risk assessment.
Who signs off on AI-generated risk flags?
The yard veterinary surgeon and the licensed trainer jointly review and sign off on AI-generated risk flags before any action is taken.
How is image quality ensured for AI processing?
Regular calibration, standardized imaging protocols, and automated quality checks ensure images meet minimum clarity before AI processing.
How will owners or syndicates be informed about fracture risk findings?
A formal imaging report, annotated by the radiologist and supported by AI metrics, is shared with owners or syndicate managers per governance protocols.