High attrition risk
Cardiovascular toxicity is one of the main reasons candidates are discontinued, directly affecting whether a programme can move forward.
AI-driven vascular toxicology
Make vascular toxicity
at the candidate stage,
visible
For drug development and chemical safety assessment, VasoPT delivers an AI-driven, visual phenomics workflow for vascular toxicity — from live vascular imaging and automated interpretation to risk-graded reporting, producing quantifiable, comparable and decision-ready evidence in one standardized process.
In vivo, whole organism
96-well throughput
AI interpretation
Risk grading
Risk score
Intermediate-high
0
/ 100
0
compounds in a small-molecule library screen
0
chemicals of environmental health concern assessed
0
steps from scope to report
0
vascular injury phenotype classes
Why early vascular safety
Cardiovascular safety is
one of the hardest risks to remediate in drug development
Cardiovascular toxicity is a major reason candidates fail. The later it is found, the higher the cost — and it usually appears as complex, multi-organ effects.
The later the finding, the higher the cost
Safety issues found late in development sharply increase R&D cost and leave far less room to manoeuvre.
Complex risk patterns
Effects span the heart, vasculature, nervous system and development; no single endpoint captures the real risk.
Early identification matters
Early safety signals improve decision quality and concentrate resources on the candidates most likely to succeed.
The gap
Every existing method leaves a gap
In vitro models, isolated organs, animal studies and computational prediction each have strengths — yet between "whole vascular network" and "high-throughput screening" there has long been no early-stage method that delivers both.
01
In vitro cell models
hERG, iPSC-CMs, endothelial cells
StrengthMany model types, controllable conditions
LimitationPoor representation of the whole cardiovascular network
02
Ex vivo / organ models
Aortic rings, isolated hearts, tissue models
StrengthHigh human relevance
LimitationLimited coverage of vessel types
03
Animal models
Rodents, dogs, zebrafish
StrengthWhole-system readout
LimitationHigh cost, limited throughput
04
Computational prediction
QSAR, PBPK, machine learning
StrengthFast and inexpensive
LimitationHeavily dependent on data quality
No early, rapid screening method that covers both the whole vascular network and high-throughput detection VasoPT was designed to fill exactly this gap: reading the whole-organism vascular phenotype in vivo and using AI to turn morphological change into comparable numbers.
The platform
A visual assessment platform
built for early vascular toxicity screening
Built for early screening in drug development and chemical safety, VasoPT centres on morphological phenomics: it converts changes in the living vascular network into structured metrics that support rapid identification, quantitative assessment and prioritization of candidates.
-
In vivo, whole organismPreserves the spatial structure and circulatory environment of the whole network, not a single-cell readout.
-
High-throughput imaging96-well plates and high-content imaging acquire batches of samples under one standard.
-
AI interpretationAutomated region segmentation and multi-dimensional feature extraction reduce reader variability.
-
Risk gradingComposite score plus dose–response analysis, feeding directly into prioritization and decisions.
High-content imaging · 96-well
Core capabilities
From seeing vessels to delivering evidence
Three steps form one loop: imaging is the entry point, interpretation the core, grading the output.
Live vascular imaging
Vascular-fluorescent zebrafish embryos and larvae are imaged on a high-content system to capture the complete vascular network, preserving spatial structure and regional differences.
AI recognition and quantification
Automatically identifies vessel regions and key structural boundaries and extracts branching, density, length, diameter and regional injury parameters — turning morphology into computable objects.
Composite score
78.5
Intermediate-high
Multi-dimensional phenotype · dose response · risk grading
Composite scoring and grading
Multi-dimensional metrics are combined into a toxicity score; dose–response trends yield low / intermediate / high risk grades and a watch list, delivered as a readable report.
Comparison
Compared with common methods, the difference is more than extra metrics
The difference is the level of assessment: from single-endpoint readouts to whole vascular phenotype; from scattered results to a comparable metric matrix with unified grading.
| Dimension | Common methods | VasoPT |
|---|---|---|
| Level of assessment | Mostly single endpoints or local readouts | Whole vascular phenotype assessment |
| Result presentation | Scattered results, limited intuitiveness | Image visualization + regional injury localization |
| Data output | Limited metrics, weak cross-comparison | Multi-dimensional metric matrix and composite score |
| Analysis | Heavily dependent on manual reading | AI recognition and standardized analysis |
| Decision support | Hard to reach a consistent grading conclusion | Supports dose–response analysis and risk grading |
Applications
Three scenarios, already proven
From small-molecule libraries to nanomaterials to environmental chemicals — one method, different safety questions.
Case 01
Small-molecule library screening
Cardiovascular toxicity assessment and risk stratification of 500 small molecules: identify potential risk signals and output a low / intermediate / high risk list.
Low risk 52.0%Intermediate 32.4%High risk 15.6%
Case 02
Nanomaterial safety assessment
Black phosphorus nanomaterials, quantum dots and related nano drug-delivery systems: focus on vascular developmental abnormalities, local structural injury and dose-related risk.
Heart rate ↓12%Microvessel density ↓15%Regional injury ↑18%
Case 03
Vascular toxicity of environmental chemicals
150 chemicals of environmental health concern: clustering into six vascular injury phenotypes, with a composite risk ranking.
Six phenotype classesRegional injury patternsRisk ranking
Move cardiovascular risk to the decision table
Tell us your compound type, sample scale and assessment goal, and we will propose the matching test design and deliverables.