Methodologies linking patterns from multi-modality datasets
A method is disclosed to acquire imaging and non-imaging datasets from like objects. A linkage is found using a partial least squares (PLS) technique between imaging and non-imaging datasets. The linkage is then reduced to an expression of a single numerical assessment. The single numerical assessment is then used as an objective, quantified assessment of the differences and similarities between the objects. The data each dataset can be aspects of performance, physical characteristics, or measurements of appearance.
1. A computer-implemented method for assessing multi-modality datasets in the evaluation of a treatment for a disease, the method comprising:
receiving at a first computing device a plurality of longitudinal datasets regarding each of a plurality of subjects, wherein a first longitudinal dataset D i of the plurality of longitudinal datasets is a dataset of an imaging modality captured by an imaging system and received from the imaging system over a communications network, and a second longitudinal dataset D j of the plurality of longitudinal datasets is a dataset of a non-imaging modality received from a second computing device over the network, the D i and D j datasets being independent datasets and forming respective square matrices; and
executing instructions stored in memory of the first computing device, wherein execution of the instructions by a processor of the first computing device:
partitions each of the D i and D j datasets into a plurality of submatrices,
computes a linkage between the D i and D j datasets, wherein computing the linkage includes reading the D i and D j submatrices into memory one by one and processing the read submatrices one by one as needed to iteratively compute a single numerical assessment, wherein the single numerical assessment represents a differential effect of a treatment for a disease between at least two subsets of the plurality of subjects, and
reports the single numerical assessment to a user of the first computing device at a display interface of the first computing device.
2. The computer-implemented method of claim 1 , wherein the imaging modality is selected from the group consisting of: ultrasound; positron emission tomography, single photon emission tomography radiotracer, or other nuclear medicine procedure; structural, functional, perfusion-weighted, or diffusion-weighted magnetic resonance imaging; x-ray computed tomography; magnetic resonance spectroscopy measurements of N-acetyl aspartic acid, myoinositol, or other chemical compounds; electroencephalography, quantitative electroencephalography, event-related potentials, or other electrophysiological procedures; magnetoencephalography; a medical imaging measurement procedure; and a non-medical imaging measurement procedure.
3. The computer-implemented method of claim 1 , wherein the non-imaging modality is selected from the group consisting of: a set of biochemical measurements; a set of molecular measurements; a set of genetic measurements; a set of transcriptomic measurements; a set of proteomic measurements; a set of cognitive measurements or clinical ratings; a set of behavioral measurements; and a set of measurements from a non-medical non-imaging modality.
4. The computer-implemented method of claim 1 , further comprising defining at least one of the D i and D j datasets based on a time interval.
5. The computer-implemented method of claim 1 , wherein the method further comprises administering a treatment for the disease to a set of the subjects over a period of time.
6. The computer-implemented method of claim 1 , further comprising generating, based on the single numerical assessment, a report on a severity level of a progression of the disease.
7. The computer-implemented method of claim 1 , wherein computing the linkage between the D i and D j datasets comprises using a partial least squares analysis.
8. A computer-implemented method for assessing multi-modality datasets in the evaluation of a treatment for a disease, the method comprising:
receiving at a computing device a plurality of longitudinal datasets regarding each of a plurality of subjects, wherein a first longitudinal dataset D i of the plurality of longitudinal datasets is a dataset of a first imaging modality captured by one or more imaging systems and received from the one or more imaging systems over a communications network, and a second longitudinal dataset D j of the plurality of longitudinal datasets is a dataset of a second imaging modality received from the one or more imaging systems over the network, the D i and D j datasets being independent datasets and forming respective square matrices; and
executing instructions stored in memory of the computing device, wherein execution of the instructions by a processor of the computing device:
partitions each of the D i and D j datasets into a plurality of submatrices, computes a linkage between the D i and D j datasets, wherein computing the linkage includes reading the D i and D j submatrices into memory one by one and processing the read submatrices one by one as needed to iteratively compute a single numerical assessment, wherein the single numerical assessment represents a differential effect of a treatment for a disease between at least two subsets of the plurality of subjects, and
reports the single numerical assessment to a user of the computing device at a display interface of the computing device.
9. The computer-implemented method of claim 8 , wherein the first imaging modality is selected from the group consisting of: ultrasound; positron emission tomography, single photon emission tomography radiotracer, or other nuclear medicine procedure; structural, functional, perfusion-weighted, or diffusion-weighted magnetic resonance imaging; x-ray computed tomography; magnetic resonance spectroscopy measurements of N-acetyl aspartic acid, myoinositol, or other chemical compounds; electroencephalography, quantitative electroencephalography, event-related potentials, or other electrophysiological procedures; magnetoencephalography; a medical imaging measurement procedure; and a non-medical imaging measurement procedure.
10. The computer-implemented method of claim 8 , wherein the second imaging modality is selected from the group consisting of: ultrasound; positron emission tomography, single photon emission tomography radiotracer, or other nuclear medicine procedure; structural, functional, perfusion-weighted, or diffusion-weighted magnetic resonance imaging; x-ray computed tomography; magnetic resonance spectroscopy measurements of N-acetyl aspartic acid, myoinositol, or other chemical compounds; electroencephalography, quantitative electroencephalography, event-related potentials, or other electrophysiological procedures; magnetoencephalography; a medical imaging measurement procedure; and a non-medical imaging measurement procedure.
11. The computer-implemented method of claim 8 , wherein the method further comprises administering a treatment for the disease to a set of the subjects over a period of time.
12. The computer-implemented method of claim 8 , further comprising generating, based on the single numerical assessment, a report on a severity level of a progression of the disease.
13. The computer-implemented method of claim 8 , wherein computing the linkage between the D i and D j datasets comprises using a partial least squares analysis.
14. A computer-implemented method for assessing a linkage between multi-modality datasets, the method comprising:
receiving at a first computing device a plurality of longitudinal datasets regarding each of a plurality of subjects, wherein a first longitudinal dataset D i of the plurality of longitudinal datasets is a dataset of an imaging modality captured by an imaging system and received from the imaging system over a communications network, and a second longitudinal dataset D j of the plurality of longitudinal datasets is a dataset of a non-imaging modality received from a second computing device over the network, the D i and D j datasets being independent datasets and forming respective square matrices; and
executing instructions stored in memory of the first computing device, wherein execution of the instructions by a processor of the first computing device:
partitions each of the D i and D j datasets into a plurality of submatrices,
computes a linkage between the D i and D j datasets, wherein computing the linkage includes reading the D i and D j submatrices into memory one by one and processing the read submatrices one by one as needed to iteratively compute a single numerical assessment, wherein the single numerical assessment represents a differential effect of a presence of a specified characteristic between at least two subsets of the plurality of subjects, and
reports the single numerical assessment to a user of the first computing device at a display interface of the first computing device.
15. The computer-implemented method of claim 14 , wherein the characteristic is one or more of age, a weight measurement, a triglyceride measurement, or a body fat measurement.
16. The computer-implemented method of claim 14 , wherein the characteristic is the possession of a specified gene.
17. The computer-implemented method of claim 16 , wherein the gene is an apolipoprotein (APOE) gene.
18. The computer-implemented method of claim 14 , wherein the characteristic is the possession of at least one ε4 allele of the APOE gene.
19. The computer-implemented method of claim 14 , wherein the characteristic is the possession of two ε4 alleles of the APOE gene.
20. The computer-implemented method of claim 14 , wherein computing the linkage between the D i and D j datasets comprises using a partial least squares analysis.