Method for determining the temporal progression of a biological phenomenon and associated methods and devices
Inventors
Durrleman, Stanley • Schiratti, Jean-Baptiste • Allassonniere, Stéphanie • Colliot, Olivier
Assignees
Centre National de la Recherche Scientifique CNRS • Assistance Publique Hopitaux de Paris APHP • Ecole Polytechnique • Institut National de la Sante et de la Recherche Medicale INSERM • Institut National de Recherche en Informatique et en Automatique INRIA • Sorbonne Universite • Institut du Cerveau et de La Moelle Epiniere ICM
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-10832089-B2
Publication Date
2020-11-10
Expiration Date
Abstract
Provided herein is a method for determining the temporal progression of a biological phenomenon which may affect a studied subject, the method including the steps of providing first data relative to biomarkers for the studied subject, the biomarkers being relative to the progression of the biological phenomenon, providing a numerical model, converting the first data into at least one point on the same Riemann manifold, and using a numerical model to determine a temporal progression for the biological phenomenon for the studied subject, the numerical model being a function in a Riemann manifold, the numerical model associating to values of biomarkers a temporal progression trajectory for the biological phenomenon and data relative to the dispersion of the progression trajectory for the biological phenomenon among a plurality of subjects, the numerical model being obtained by using a stochastic approximation in an expectation-maximization technique on data relative to biomarkers.
Core Innovation
The invention determines the temporal progression of a biological phenomenon that may affect a studied subject by using biomarker data. First data are provided as data relative to biomarkers for the studied subject, where the biomarkers are relative to the progression of the biological phenomenon, and a numerical model in a Riemann manifold associates values of biomarkers with a temporal progression trajectory and data relative to the dispersion of the progression trajectory among a plurality of subjects.
The numerical model is obtained by using stochastic approximation in an expectation-maximization technique on data relative to biomarkers taken at different time points for a plurality of subjects. The method converts the first data into at least one point on the same Riemann manifold and uses the numerical model to determine the temporal progression of the biological phenomenon for the studied subject.
The disclosed approach is described as a spatiotemporal mixed-effects model for longitudinal, manifold-valued biomarker data, mapping biomarker trajectories to a temporal progression trajectory and its dispersion across subjects, while operating on a numerical model defined on a Riemann manifold. Geometric mechanisms are used to uniquely decompose spatial versus temporal components and to personalize a studied subject’s progression by converting the biomarker data into points on the same manifold.
Claims Coverage
The independent claim is directed to a method for determining temporal progression using a numerical model defined on a Riemann manifold with stochastic approximation inside an expectation-maximization technique, followed by converting subject biomarker data into points on the same manifold. The inventive features below summarize the main elements of the independent claim, with additional inventive features stated in some dependent claims as refinements.
Riemann-manifold numerical model for temporal progression and dispersion
The numerical model is a function in a Riemann manifold, associating to values of biomarkers a temporal progression trajectory for the biological phenomenon and data relative to the dispersion of the progression trajectory among a plurality of subjects.
Stochastic approximation within expectation-maximization to obtain the numerical model
The numerical model is obtained by using a stochastic approximation in an expectation-maximization technique on data relative to biomarkers taken at different time points for a plurality of subjects.
Convert subject biomarker data into points on the same Riemann manifold
Converting the first data into at least one point on the same Riemann manifold.
Determine temporal progression for the studied subject using the numerical model
Using the numerical model to determine a temporal progression for the biological phenomenon for the studied subject.
Monte Carlo Markov Chain stochastic approximation expectation-maximization
The expectation-maximization technique uses a Monte Carlo Markov Chain stochastic approximation when providing the numerical model.
Subject-count threshold for constructing the numerical model
The number of subjects is greater than or equal to 100 when providing the numerical model.
Overall claim coverage centers on fitting and using a numerical model on a Riemann manifold that outputs temporal progression together with dispersion among subjects. The core workflow includes stochastic-approximation estimation within expectation-maximization, converting the studied subject’s biomarker data into points on the same manifold, and then using the model to determine the subject’s temporal progression.
Stated Advantages
Not explicitly described in patent.
Documented Applications
Risk prediction.
Diagnosis.
Therapeutic target or biomarker identification.
Compound screening.
An Alzheimer’s disease experimental implementation using ADAS-Cog 13 from ADNI cohorts, including symptom-onset timeline estimation and reported predictive error characteristics.
Interested in licensing this patent?