01
Test requirement definition
Define the test article, target endpoints and dose range.
Home / Technology
From image data to a closed risk-decision loop
Turning "visible vascular change" into "quantifiable, comparable, deliverable toxicology evidence" — that is the whole of what VasoPT does.
Workflow
Eight-step standardized workflow
From test requirement to risk report, every step has defined inputs and deliverables, keeping results comparable across batches and projects.
02
Exposure design
Set the exposure regimen, controls and replicates.
03
Live vascular imaging
Acquire zebrafish cardiovascular and vascular network images.
04
Image QC standardization
Unified positioning and selection of high-quality images.
05
AI recognition and segmentation
Automatically identify vessel regions and key structural boundaries.
06
Multi-dimensional feature extraction
Extract area, branching, density and morphology metrics.
07
Composite scoring and dose response
Produce a toxicity score and resolve dose-related change.
08
Risk grading and reporting
Deliver the risk grade, conclusions and a visual report.
Data accumulation
Standardized storage and structured management build a high-quality vascular toxicity database.
Model iteration
Continuous training and optimization improve accuracy and generalization.
Scope expansion
Extend scenarios and test articles, supporting multi-endpoint and multi-scale assessment.
AI interpretation
AI image recognition and scoring
Converting complex vascular images into numerical results: five steps that turn looking into reading.
STEP 01Image recognition
Automatically locate vessel structures, region boundaries and the main vascular network.
STEP 02Structural segmentation
Convert complex images into computable regions, lines and network objects.
STEP 03Feature extraction
Extract morphology, branching, density, length, diameter and regional injury parameters.
STEP 04Scoring and analysis
Combine multi-dimensional change into a toxicity score, risk grade and priority order.
STEP 05Data accumulation
Retain image data, phenotype parameters, dose relationships and risk scores.
Vessel skeleton · nodes · branch measurement
Six phenotype metrics, one comparable score
A single metric is easily misled by noise; only a multi-dimensional matrix reflects the true state of the vascular network. All metrics are extracted under one standard, enabling comparison across samples and batches.
Branch density
Vessel length
Branch points
Diameter change
Network complexity
Regional injury
Example: composite score for one compound
0–100
0
/ 100 · intermediate-high
Low risk
Intermediate
High risk
Phenomics
Why vascular morphological phenomics
Phenomics is the sum of traits expressed by individuals of a given genotype under defined environmental conditions. Applied to vessels, it means looking at the whole network, not a single point.
Naturally rich in metrics
A vascular network carries branching, density, diameter, orientation and regional variation — a quantifiable phenotype space.
The hard part is quantification and throughput
Quantifying vascular morphology is difficult, and many phenomic metrics resist high-throughput capture — precisely what AI and high-content imaging address.
Advantages for toxicity assessment
Intuitive, efficient, accurate and effective — well suited to large-scale screening and ranking at an early stage.
Scope of application VasoPT is positioned as an early screening and prioritization tool that rapidly identifies vascular toxicity signals among candidates; it does not replace formal regulatory nonclinical safety assessment. Endpoints, dose design and report use are as agreed before project start.
See what your compound looks like in the vasculature
Share your sample type and assessment goal, and we will propose a complete path from imaging to report.