Systems and methods for generating biomarkers based on multivariate MRI and multimodality classifiers for disorder diagnosis
In some embodiments, the systems and methods of the disclosure can efficiently and accurately classify neurodegenerative disorder(s) and/or movement disorder(s) of a subject (e.g., a patient) using at least quantitative features associated with one or more regions of interest determined from one or more sets of image data of the subject's brain. The method may include processing one or more sets of MRI image data of the subject's brain to extract one or more quantitative features for one or more regions. The one or more quantitative features may include a first quantitative and a second quantitative feature. The method may further include classifying at least the one or more quantitative features into one or more classes associated with neurodegenerative dementia disorder, neurodegenerative movement disorder, non-neurodegenerative movement disorder and/or heathy control. The method may include generating a report including a classification of at least the one or more quantitative features.
1. A computer-implemented method for classifying at least one of neurodegenerative disorder and non-neurodegenerative movement disorder of a subject, the method comprising:
receiving subject data of a subject, the subject data including a plurality of sets of MRI image data of a brain of the subject, the plurality of sets of MRI image data including neuromelanin-sensitive MRI data and iron-sensitive MRI data;
processing each set of MRI image data to extract a plurality of quantitative features for one or more regions of the brain, the plurality of quantitative features for the one or more regions includes a first quantitative feature and a second quantitative feature;
the first quantitative feature determined using the neuromelanin-sensitive MRI data; and
the second quantitative feature determined using the iron-sensitive MRI data;
classifying the subject data using the plurality of quantitative features for the one or more regions into one or more classes from a plurality of classes, the plurality of classes including neurodegenerative dementia disorder, neurodegenerative movement disorder, non-neurodegenerative movement disorder, and/or heathy control; and
generating a report including the one or more classes.
2. The method according to claim 1 , wherein the neurodegenerative dementia disorder includes a parkinsonian class and a non-parkinsonian class.
3. The method according to claim 2 , wherein:
the parkinsonian class for the neurodegenerative dementia disorder includes one or more parkinsonian neurodegenerative dementia subclasses; and
the one or more parkinsonian neurodegenerative dementia subclasses includes at least one of Parkinson's disease dementia (PDD), dementia with Lewy bodies (DLB), and one or more other atypical parkinsonism dementia disorder subclasses.
4. The method according to claim 3 , wherein the one or more other atypical parkinsonism dementia disorder subclasses includes at least one of multiple system atrophy (MSA), progressive supranuclear palsy (PSP), and corticobasal degeneration (CBD).
5. The method according to claim 1 , wherein:
the non-neurodegenerative movement disorder includes one or more non-neurodegenerative movement disorder subclasses; and
the one or more non-neurodegenerative movement disorder subclasses includes psychogenic, essential tremor, and drug-induced.
6. The method according to claim 1 , wherein:
the neurodegenerative movement disorder includes a parkinsonian class;
the parkinsonian class for the neurogenerative movement disorder includes one or more parkinsonian movement disorder subclasses; and
the one or more parkinsonian movement disorder subclasses includes Parkinson's Disease (PD) and one or more other atypical parkinsonism movement disorder subclasses.
7. The method according to claim 6 , wherein the one or more other atypical parkinsonism movement disorder subclasses includes at least one of MSA, PSP, and CBD.
8. The method according to claim 1 , wherein:
the plurality of sets of MRI image data is acquired by one or more stored protocols, the one or more protocols including one or more protocols for iron-sensitive MRI sequences and one or more protocols for neuromelanin-sensitive MRI sequences; and
the plurality of quantitative features include (i) one or more of NM-MRI features and (ii) one or more of R2* features or one or more of qualitative susceptibility mapping (QSM) features.
9. The method according to claim 8 , wherein the plurality of quantitative features also include one or more of diffusion MRI features.
10. The method according to claim 8 , wherein the one or more regions includes substantia nigra pars compacta (SNc) and locus coeruleus (LC).
11. The method according to claim 10 , wherein the one or more regions also includes at least one of red nucleus and globus pallidus.
12. A system for classifying at least one of neurodegenerative disorder and non-neurodegenerative movement disorder of a subject, the system comprising:
at least one processor; and
a memory having stored thereon computer-executable instructions which are executable by the at least one processor to cause the system to perform at least the following:
processing a plurality of sets of MRI image data of a brain of the subject to extract a plurality of quantitative features for one or more regions of the brain, the plurality sets of MRI image data including neuromelanin-sensitive MRI data and iron-sensitive MRI data, the plurality of quantitative features for the one or more regions includes a first quantitative feature and a second quantitative feature for the one or more regions;
the first quantitative feature determined using the neuromelanin-sensitive MRI data; and
the second quantitative feature determined using the iron-sensitive MRI data;
classifying the subject data using the plurality quantitative features for the one or more regions into one or more classes from a plurality of classes, the plurality of classes including neurodegenerative dementia disorder, neurodegenerative movement disorder, non-neurogenerative movement disorder, and/or heathy control; and
generating a report including the one or more classes.
13. The system according to claim 12 , wherein:
the plurality of sets of MRI image data is acquired by one or more stored protocols, the one or more protocols including one or more protocols for iron-sensitive MRI sequences and one or more protocols for neuromelanin-sensitive MRI sequences; and
the plurality of quantitative features include (i) one or more of NM-MRI features and (ii) one or more of R2* features or qualitative susceptibility mapping (QSM) features.
14. The system according to claim 13 , wherein the plurality of quantitative features also include one or more diffusion MRI features.
15. The system according to claim 14 , wherein the neurodegenerative dementia disorder includes a parkinsonian class and a non-parkinsonian class.
16. The system according to claim 13 , wherein:
the neurodegenerative movement disorder include a parkinsonian class;
the parkinsonian class for the neurodegenerative dementia disorder includes one or more parkinsonian neurodegenerative dementia subclasses; and
the one or more parkinsonian neurodegenerative dementia subclasses includes at least one of Parkinson's disease dementia (PDD), dementia with Lewy bodies (DLB), and one or more other atypical parkinsonism dementia disorder subclasses.
17. The system according to claim 16 , wherein the one or more other atypical parkinsonism dementia disorder subclasses includes at least one of multiple system atrophy (MSA), progressive supranuclear palsy (PSP), and corticobasal degeneration (CBD).
18. The system according to claim 13 , wherein the one or more regions includes substantia nigra pars compacta (SNc) and locus coeruleus (LC).
19. The system according to claim 18 , wherein the one or more regions also includes at least one of red nucleus and globus pallidus.
20. The method according to claim 1 , wherein:
the subject data further includes at least one of clinical data, physiological data, demographic data, and genetic data;
the processing further includes processing the subject data to determine one or more clinical features; and
the classifying further includes using the one or more clinical features to classify the subject data into the one or more classes.
21. The system according to claim 12 , wherein:
the subject data further includes at least one of clinical data, physiological data, demographic data, and genetic data;
the processing further includes processing the subject data to determine one or more clinical features; and
the classifying further includes using the one or more clinical features to classify the subject data into the one or more classes.