METHOD FOR DIAGNOSING ALZHEIMER'S DISEASE Russian patent published in 2025 - IPC G16H30/00 A61B6/03 

FIELD: medicine.

SUBSTANCE: invention relates to analysis of medical data and is intended to support medical decisions when diagnosing the presence or absence of Alzheimer's disease in a patient with high reliability. Proposed method for diagnosing Alzheimer's disease involves conducting a study of the patient's brain by magnetic resonance imaging with the introduction of a contrast agent, determining values of magnetic susceptibility of basal veins in left L_BV and right R_BV cerebral hemispheres, thalamus veins of left L_TV and right R_TV cerebral hemispheres, according to obtained tomographic images, determining ranges of obtained values of magnetic susceptibility in accordance with intervals of their categorical values and encoding the obtained values into categorical values of magnetic susceptibility of basal veins and thalamus veins L_BV^*, R_BV*, L_TV*, R_TV*, accordingly, further, using discrete functional dependencies of variable categorical values of magnetic susceptibility in matrix form, the value of "decision-making root" is determined. Additionally determining the magnetic susceptibility values of internal cerebral veins of the left L_ICV and right R_ICV cerebral hemispheres, determining the ranges of the obtained values of magnetic susceptibility in accordance with the intervals of their categorical values and encoding the obtained values into the categorical values of the magnetic susceptibility of said internal cerebral veins L_ICV*, R_ICV*, respectively, then constructing a pre-trained "decision tree", which is a branching algorithm of actions corresponding to a block diagram built on the basis of data using an information-theoretic criterion, in which the "decision tree" node is a condition, and the leaf is a forecast, then the said "decision tree" algorithm is loaded into the computer memory, by means of computer keyboard input data of variable categorical values of magnetic susceptibility are inserted, algorithm of the trained "decision tree" processes the obtained data and displays one of the values of the variable values CON or AD on the computer monitor.

EFFECT: invention provides more reliable diagnosis of Alzheimer's disease in a patient.

1 cl, 3 dwg