Systems and methods for full body circulation and drug concentration prediction

Inventors

VARSHNEY, Jyotika

Assignees

Verisim Life Inc

Abstract

A method for predicting drug concentration levels includes receiving at least one subject characteristic of the subject, and executing a full body circulation model by: determining a first concentration of the drug in a first blood flow entering a first organ determining a second concentration of the drug in the first organ, determining a third concentration of a drug in a third blood flow entering a second organ, the third blood flow downstream of the first organ, and determining, using the second organ model a fourth concentration of the drug in the second organ.

Core Innovation

The invention provides a system for predicting drug concentration levels as a function of time in one or more organs of a subject. It includes a subject database with a plurality of subject characteristics and a drug database with at least one drug characteristic of a drug to be administered. One or more processors train organ-specific drug concentration prediction models using subject profiles that include a drug identifier, known concentration-time pairs for a respective organ, and a characteristic of the training subject.

Each organ drug concentration prediction model is trained by iteratively providing inputs to output predicted concentration-time pairs, comparing predicted concentrations to known concentrations to generate a comparison result, and adjusting at least one model parameter responsive to the comparison result until the comparison result is minimized or is less than a threshold difference. After training, the system determines blood flow rates to organs and determines a concentration of the drug in blood entering an organ based on an initial drug dosage and the blood flow rate. The trained models are then modified using subsets of subject characteristics to determine drug concentrations in organs and to propagate downstream blood entering downstream organs.

The system provides an output indicating a recommendation for administering the drug to the subject responsive to comparing organ drug concentrations to organ-specific maximum concentration values and comparing organ exposure via first and second area under curve (AUC) values to first and second maximum AUC values over first and second time durations. The method applies the same structure, including training first and second organ drug concentration prediction models, modifying the trained models using different subsets of subject characteristics, determining concentrations in organs and downstream blood, and issuing the recommendation based on the concentration and AUC comparisons.

Claims Coverage

The document provides two independent claims, a system claim and a method claim. The inventive features focus on organ-specific machine-learning drug concentration prediction models trained with known concentration-time pairs, blood-flow-rate-based concentration propagation across organs using subject characteristics and initial drug dosage, and output recommendations based on organ maximum concentration and maximum AUC comparisons.

Across the independent claims, the core coverage is for training and applying organ-specific drug concentration prediction models using known concentration-time pairs, integrating subject-derived blood flow rates and initial drug dosage to propagate concentrations across organs, and generating a recommendation based on organ-specific maximum concentration and maximum AUC comparisons.

Stated Advantages

Documented Applications