What can detect:

General:

• Automatism (Bradycardia, Tachycardia, СА Block (II, III), Asystole)
• Excitability (Rhythm: Sinus, Atrial, Nodal, Ventricular)
• Conductivity (AV Block (I, II, III), AV Dissociation, Intraventricular Block)
• Pre-excitation (Wolf-Parkinson-White syndrome, Lown-Ganong-Levine syndrome)

Ectopia: Supraventricular, Ventricular, Tachycardia, Fibrillation/Flatter

PQRST: PQ int/segment, QT (QTc), QRS, ST abnormalities

• Increased depth of analysis (80+ detection classes, ST, HRV, etc.)
• 24/7 online access to ECG data for doctors
• Ability to process cardiac data from any type of ECG device *
• Complete medical report on the patient’s ECG at a professional level
• The speed and accuracy of ECG annotation and interpretation is one of the best on the market
• Localization in Ukrainian/English languages

* Wellness devices e.g. AppleWatch, SmartWhatches, and ECG Cardiographs are not supported. Already validated: Bittium, Cortium, LifeSignals, Livetec, BTL, ScottCare Holter, Amedtec, Phillips.

• For Physicians and attending Clinicians all the above transforms into high-quality reports that frequently contain references to the strips that would be otherwise hard or even impossible to spot with conventional Holter processing pipelines.
• The software resides completely in the cloud, leading to another strong point – the ability of doctors to access the evidence online, from anywhere.
• The report becomes clickable for all the statistics and strips, leading to the signal in context.
• This allows for any depth of exploration, particularly for patients with multiple arrhythmia types. It opens the internals of the record to the attending clinician, helping to build trust between the doctor and a diagnostic center.

• AI’s classification ability is the quality bottom line, supporting diagnostic companies with high-quality output.
• The Cardio.AI already understands almost any possible arrhythmia or conduction issue, and the team is constantly working towards reaching full coverage, including even rare arrhythmias, essentially a superhuman state on the ECG interpretation.
• The process of improving the AI is conceptually very simple, the team just adds the strips on which AI performs poorly to the training pipeline, eventually converging the quality to be above 90% (compared to human’s average 74%) on any particular arrhythmia class, with some classes, like Atrial Fibrillation or Ventricular ectopic rhythms, to be above 98%.