Cardiovascular early detection · Clinical AI

75% of cases are asymptomatic until rupture.

We detect lethal cardiovascular disease before it kills — with AI, medical imaging, and real clinical evidence.

See how it works Read the research
★ Harvard HSIL Hackathon 2026 · Winner ◆ JTCVS + STS · under review

Emergency surgery

50%

mortality

Silent cases

75%+

asymptomatic until rupture

Clinical speed

< 2 min

per case on GPU

Scroll

The problem

Our systems react to symptoms.
The deadliest diseases are silent.

Late diagnosis

Rupture happens before diagnosis. By the time symptoms appear, it's usually too late.

Limited infrastructure

40% of cardiovascular events go unregistered in rural Colombia.

Catastrophic cost

Emergency surgery costs 5–10× more than scheduled elective surgery.

Real case

"Patient with no symptoms. Two months later: rupture."

Mortality without surgery: 80–90%.

Our research

Two original studies. A decade of Colombian data.

Study I · Mortality

0

deaths registered

×0

annual deaths (2013→2023)

0 → 0

2013 vs 2023

JTCVS + STS

Published — under review

First national aortic-mortality registry in Colombia
Study II · Disease burden

+0

increase in DALYs

0 → 0

DALYs / 100k inhabitants

0

rupture below 55mm threshold

75%+

asymptomatic until rupture

Disease burden — DALYs, GBD 2019 + local registries

How it works

From medical image to quantified risk in under 2 minutes.

M1

Image ingestion

DICOM · CTA · Echo-TT

M2

Segmentation

Thoracic aorta · nnU-Net

M3

Morphometry

Diameters · lengths · curvature

M4

Risk score

Rupture · dissection · follow-up

Precision

3.19 mm

mean error vs. expert radiologist on aortic diameters.

Speed

< 2 min

per case on GPU — viable in real clinical workflow.

Dataset

43 CTA · 30 pts.

real Colombian studies, manually annotated.

Technology Readiness Level

TRL 3 current · TRL 7 target

TRL 13579

Impact

It's not just a diagnosis.
It's an intervention in the healthcare system.

Patient

  • 95%+ survival (elective repair)
  • 50% mortality (emergency surgery)
  • Pre-rupture detection
  • Individualized risk score
  • Longitudinal follow-up

Healthcare system

  • 5–10× lower cost: elective vs emergency
  • +136% DALYs: we can reverse it
  • Less ICU time
  • National standardization
  • Evidence-based triage

Rural Colombia

  • 40% unregistered CVD events
  • 1.4M DALYs lost (Colombia 2019)
  • Portable ECG + Echo-TT
  • No complex infrastructure
  • Active surveillance

Team

Doctors and engineers. Built from clinical practice.

CP
Surgeon · Epidemiologist

Carlos José Pérez

CEO · Galatea AI

NP
AI Engineer · Medical Imaging

Nicolás Pérez

CEO · Galatea AI

MV
Physician · Researcher

Maria Camila De Vivero

Galatea AI

DA
Intern · Researcher

Daniela Arbelaez

Galatea AI

Recognitions & credentials

★

Harvard HSIL Hackathon 2026

Winner

◆

Universidad El Bosque

Incubation

§

JTCVS + STS

Papers under review

◎

LATAM #1

Unique dataset

◈

Google DeepMind Scholar

CTO

Call to action

Let's turn a silent killer
into a preventable disease.

The question is no longer whether we can treat these diseases. The question is: why do we keep diagnosing too late?

Request Demo Contact the Team

nicolas@galateaai.net