Method for Classification of Platelet Aggregation

Publication: EP3809115A1
Published: 2021-04-21
Family Size: 1
Granted: No

Simple SummaryContent extracted from patent full text and abstract with AI.

This patent discloses a method that uses a trained convolutional neural network (CNN) to automatically detect, count, and classify platelet (thrombozyte) aggregates in blood samples—particularly whole blood—based on images produced by imaging flow cytometry. The system can distinguish between single platelets and aggregates of platelets of different sizes, providing a weighted diagnostic value to better understand and monitor blood clotting (aggregation) behavior under physiological conditions. The method eliminates manual boundary setting and subjective interpretation, delivering faster and more precise diagnostics.

Use CasesContent extracted from patent full text and abstract with AI.

  • Hospital or clinical laboratories for precise blood coagulation diagnostics.
  • Monitoring patients on anti-platelet or anti-coagulation medications to assess drug efficacy.
  • Evaluating bleeding or clotting risks prior to surgeries or invasive procedures.
  • Research on platelet function and aggregation dynamics in both healthy and disease states.
  • Pharmaceutical development for drugs targeting platelet function or aggregation.
  • Automation in blood analysis devices for routine screening or personalized medicine approaches.

BenefitsContent extracted from patent full text and abstract with AI.

  • Provides rapid, accurate, and reproducible classification of platelet aggregates, including small aggregates difficult to distinguish manually.
  • Minimizes sample preparation and uses whole blood, preserving physiological relevance and reducing costs.
  • Reduces subjective errors from manual boundary setting found in current methodologies.
  • Enables detailed monitoring of coagulation status and platelet function for personalized patient care or research.
  • Automates and streamlines the diagnostic process, enhancing laboratory workflow and throughput.
  • Improves detection sensitivity for small platelet aggregates (2-3 platelets), important for early or subtle changes in clotting function.
  • Supports the calculation of clinically meaningful weighted diagnostic parameters, assisting in risk stratification and therapy monitoring.

Technical Classifications (CPCs)

Main Classifications

Physics & Measurement

Sub Classifications

Computing & Calculating

Measuring & Testing

CPC Codes

G01N15/1429G01N15/147G06N3/045G06N3/0464G06N3/08G06N3/09

Inventors & Applicants

Applicants

Leibniz Institut Fuer Plasmaforschung und Tech E V

Patent Abstract

Die Erfindung betrifft eine Methode zur Detektion eines Thrombozyten-Objektes und zur Ermittlung einer Anzahl an Thrombozyten in dem Thrombozyten-Objekt in einer Probe, wobei ein Thrombozyten-Objekt mindestens einen Thrombozyten aufweist. Insbesondere weist die Probe Vollblut auf. Dabei wird mindestens ein Bild aus der Probe aufgenommen, insbesondere wobei das Bild ein Thrombozyten-Objekt zeigt, wobei das mindestens eine Bild insbesondere mittels eines bildgebenden Durchflusszytometers erzeugt ist. Das mindestens eine Bild der Probe wird mithilfe eines trainierten Faltungsnetzwerkes, welches insbesondere auf einer Computervorrichtung ausgeführt wird, untersucht. Das trainierte Faltungsnetzwerk ist dazu eingerichtet, auf dem mindestens einen Bild das Thrombozyten-Objekt zu erkennen, wobei das trainierte Faltungsnetzwerk dazu ausgebildet ist, eine Anzahl von Thrombozyten in dem erkannten Thrombozyten-Objekt zu bestimmen und auszugeben.

Key Information

Publication No.

EP3809115A1

Family ID

68296132

Publication Date

2021-04-21

Application No.

EP19204113A

Application Date

2019-10-18

Priority Date

2019-10-18

Granted

No

Possible Cooperation

For further information please contact the transfer office.