Systems and methods for processing medical images to define brain regions and to determine associated information

Abstract

The devices, systems, and methods can provide an automated, scalable NM-MRI approach to define regions, such as SNc and LC, and determine associated quantitative information, such as volume measurements. The methods may include generating an individualized region mask of a region of interest (ROI) of a brain region of an individual by converting a population region mask from standard space into individual space using the population region mask and MRI image data of the individual. In some examples, the method may further include generating an individualized reference ROI using the MRI image data. The method may further include segmenting the ROI from within the individualized region mask using information derived from the individualized reference ROI.

Description

SYSTEMS AND METHODS FOR PROCESSING MEDICAL IMAGES TO DEFINE BRAIN REGIONS AND TO DETERMINE ASSOCIATED INFORMATION CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. Provisional Application No. 68 / 687,210 filed August 26, 2024. The entirety of this application is hereby incorporated by reference for all purposes. BACKGROUND

[0002] Currently available imaging methods, such as neuromelanin-sensitive magnetic resonance imaging (NM-MRI) approaches, of brain regions, such as brainstem, basal ganglia, and cerebellar structures, continue to rely upon manual image processing methods that can introduce operator dependent error and that cannot be scaled for applications in large multi-site studies or clinical use. Furthermore, the brainstem reference regions in current NM-MRI methods are generally small, may include outlier voxels, and often have a non-Gaussian signal intensity distribution, which makes interpretation of results difficult. SUMMARY

[0003] Thus, there is a need for an automated, scalable NM-MRI approach to define regions and determine associated information, such as volume measurements.

[0004] The systems and methods of the disclosure can automatically and efficiently process medical images, such as magnetic resonance imaging (MRI) data to accurately define one or more regions within the brain (e.g., substantia nigra pars compacta (SNc) region, SNc subregions, locus coeruleus (LC), other regions and / or subregions, etc.) to enable accurate determination of quantitative information.

[0005] In some examples, the methods may include a computer-implemented that includes generating an individualized region mask of a region of interest (ROI) of a brain region of an individual by converting a population region mask from standard space into individual space using the population region mask and magnetic resonance imaging (MRI) image data of the individual. The method may further include generating an individualized reference ROI for the ROI using theMRI image data. The method may include segmenting the ROI from within the individualized region mask using information derived from the individualized reference ROI.

[0006] In some examples, the ROI may include the locus coeruleus (LC), substantia nigra pars compacta (SNc), other neuromelanin contrast regions, iron contrast regions, neuromelanin / iron contrast region overlap ROI, among others, or any combination thereof. The ROI may include the LC and / or the SNc.

[0007] In some examples, the segmenting may include segmenting the ROI from the quantitative contrast image data within the location of the individualized region mask to generate the segmented ROI and a segment-out region. The segment-out region may include voxels within the individualized region mask that are excluded from the segmented ROI based on signal intensity.

[0008] In some examples, the ROI may be the SNc. The segmenting the SNc may include positioning the individualized region mask for SNc with respect to the quantitative contrast image data based on signal intensity of edge mask voxels within the individualized region mask.

[0009] In some examples, the positioning the individualized mask for the SNc may include at least one iteration of: (i) positioning the individualized region mask within the quantitative contrast image data at a first location; (ii) generating a first edge mask at the first location by selecting an outermost layer of voxels within the individualized region mask; (ii) determining a first edge mask measure for the first location based on a signal intensity of the first edge mask; (iii) generating a second edge mask by moving the first edge mask to a second location by shifting a position of at least one voxel within the quantitative contrast image data; (iv) determining a second edge mask measure for the second location based on a signal intensity of the second edge mask; and (v) comparing the second edge mask measure to the first edge measure. In some examples, the location of the individualized region mask may be determined based on the comparing in at least one of the one or more iterations.

[0010] In some examples, the location of the individualized region mask may correspond to the first location when the second edge mask measure is greater than the first edge mask measure. In some examples, more than one iteration is performed until iteration in which a second edge mask measure is greater than the first edge mask measure. The more than one iteration may include a first iteration and a second iteration. In the second iteration, the second edge location and secondedge mask measure of the first iteration may correspond to the first edge location and first edge mask measure of the second iteration.

[0011] In some examples, the MRI image data may include T1 image data and quantitative contrast image data.

[0012] In some examples, the quantitative contrast image data may include neuromelanin (NM)-MRI image data and / or iron-sensitive MRI image data. The population region mask may be probabilistic and may be larger than a representative representation of the ROI.

[0013] In some examples, the method may further include registering the T1 image data and the quantitative contrast image data using partial and / or entire MRI image data in one or more registration iterations. The method may also include determining one or more transformation matrices based on the one or more registration iterations. The individualized region mask may be generated using the one or more transformation matrices.

[0014] In some examples the registering includes at least one iteration of: (i) a first registering that includes registering the entire T1 image data to the entire and / or partial quantitative contrast image data in the individual space; and / or (ii) a second registering that includes registering the partial T1 image data to the partial quantitative contrast image data. The one or more transformation matrices may be generated based on the at least one iteration of the registering. In some examples, the one or more transformation matrices may be generated based on more than one iteration of the registering.

[0015] In some examples, the generating an individualized reference ROI using the MRI image data may include positioning an initial individualized reference ROI within the MRI image data based on a location of the individualized region mask so that the initial individualized reference ROI is disposed adjacent to the individualized region mask. The initial individualized reference ROI may include one or more predefined geometric shapes having a predefined size. The generating may include removing outlier voxels from the initial individualized reference ROI based on signal intensity distribution of the initial individualized reference ROI resulting in the individualized reference ROI used to segment the ROI.

[0016] In some examples, the initial individualized reference ROI may be defined by an overlap between the two or more of the more than one predefined geometric shape. In some examples, the predefined size of the initial individualized reference ROI may depend on a size of the individualized region mask.

[0017] In some examples, the ROI may be LC. The initial individualized reference ROI for LC may be generated in a reference slice. For example, the reference slice may be based on a location of the individualized region mask for SNc and / or the individualized region mask for LC.

[0018] In some examples, the MRI image data may include one or more slices. In some examples, the generating the individualized reference ROI for LC may include determining a reference slice of the MRI image data based on a location of an SNc region mask and / or an individualized LC region mask. The generating the individualized reference ROI for LC may further include determining a center of mass of the individual region mask for LC defined by region masks disposed at opposing sides. The generating the individualized reference ROI for LC may also include generating the individualized reference ROI for LC in at least the reference slice based on the center of mass.

[0019] In some examples, the individualized reference ROI may further be generated in each slice of the MRI image data in which the individualized region mask for LC is present.

[0020] In some examples, if the individualized region mask for the SNc is present in a slice, the reference slice may be inferior to the slice that includes the individualized SNc region mask. If the one or more slices only includes the individualized LC region mask, the reference slice may be superior to a slice that includes the individualized LC region mask.

[0021] In some examples, the population mask for each ROI may be generated by receiving representative NM-MRI image data for a plurality of representative individuals. In some examples, the generation of the population mask for each ROI may further include defining an individualized representative region mask for the ROI in the representative NM-MRI image data for each representative individual so that the individualized representative region mask covers an area bigger than the ROI. The generation of the population mask for each ROI may also include converting each individualized representative region mask into the standard space to generate a standard representative region mask. The generation of the population mask for each ROI may further include averaging the standard representative region masks for the plurality of different representative individuals. The generation of the population mask for each ROI may further include determining the population region mask based on the averaging.

[0022] In some examples, the method may include generating a population reference ROI for the ROI using the population region mask.

[0023] In some examples, the ROI may include the LC. Each individualized representative region mask for the LC may be disposed at each center of mass of the LC of each respective representative NM-MRI Image data.

[0024] In some examples, the disclosure may include a computer-program product comprising instructions configured to cause one or more data processors to perform the method as described herein. In some examples, the disclosure may include a non-transitory machine-readable storage medium tangibly embodying a computer program product as described herein.

[0025] In some examples, the disclosure may include a system. The system may include one or more data processors. The system may further include one or more hardware storage devices having stored thereon computer-executable instructions which are executable by the one or more processors to cause the computing system to perform the method as described herein.

[0026] Additional advantages of the disclosure will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the disclosure. The advantages of the disclosure will be realized and attained by means of the elements and combinations particularly pointed out in the appended claims. It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure, as claimed. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The disclosure can be better understood with the reference to the following drawings and description. The components in the figures are not necessarily to scale, emphasis being placed upon illustrating the principles of the disclosure.

[0028] Figure 1 illustrates a flow diagram showing an example of a method of segmenting region(s) within MRI image data of a brain of an individual and determining associated quantitative information associated with each region according to some embodiments.

[0029] Figure 2A illustrates a flow diagram showing an example of a method of generating an individualized region mask, which can be used in the method of Figure 1, according to embodiments. Figure 2B illustrates a flow diagram showing an example of a method of registering the individual MRI data, which can be used in the method of Figure 2A, according to some embodiments.

[0030] Figure 3 illustrates a flow diagram showing an example of a method of segmenting SNc within image(s) of a brain of a subject and determining associated quantitative information according to some embodiments.

[0031] Figure 4 illustrates a flow diagram showing an example of a method of positioning an individualized SNc region mask within MRI image data, which can be used in the method of Figure 3, according to some embodiments.

[0032] Figure 5 illustrates an operational example of a method of segmenting SNc from individual MRI image data according to some embodiments.

[0033] Figure 6 illustrates an operational example of a method of generating a SNc population region mask according to some embodiments.

[0034] Figure 7 illustrates a flow diagram showing an example of a method of segmenting LC within image(s) of a brain of a subject and determining associated quantitative information according to some embodiments.

[0035] Figure 8A illustrates a flow diagram showing an example of a method of generating LC reference ROI, which can be used in the method of Figure 7, according to some embodiments. Figure 8B shows an illustrative example of selecting the reference slice used in Figure 8A according to embodiments.

[0036] Figure 9 illustrates an operational example of a method of segmenting LC from individual MRI image data according to some embodiments.

[0037] Figure 10 illustrates an operational example of a method of generating an LC population region mask according to some embodiments.

[0038] Figure 11 shows a block diagram illustrating an example of a system according to some embodiments. DESCRIPTION OF THE EMBODIMENTS

[0039] In the following description and Appendix, numerous specific details are set forth such as examples of specific components, devices, methods, etc., in order to provide a thorough understanding of embodiments of the disclosure. It will be apparent, however, to one skilled in the art that these specific details need not be employed to practice embodiments of the disclosure. In other instances, well-known materials or methods have not been described in detail in order to avoid unnecessarily obscuring embodiments of the disclosure. While the disclosure is susceptibleto various modifications and alternative forms, specific embodiments thereof are shown by way of example in the drawings and will herein be described in detail. It should be understood, however, that there is no intent to limit the disclosure to the particular forms disclosed, but on the contrary, the disclosure is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the disclosure.

[0040] In some embodiments, the systems and methods of the disclosure can automatically process medical images, such as magnetic resonance imaging (MRI) data (e.g., neuromelanin - sensitive (NM)-MRI image data, multi-echo gradient echo MRI data, diffusion MRI data, chemical exchange saturation transfer (CEST) MRI data, other MRI-pulse sequence data, etc.), to accurately define one or more regions within the brain (e.g., substantia nigra pars compacta (SNc) region, SNc subregions, locus coeruleus (LC), other regions and / or subregions, etc.). In some examples, the systems and methods of the disclosure can determine quantitative information, such as quantitative measurements (e.g., segmented ROI volume, individualized region mask ratio, contrast, metabolite measures, metabolic measures, etc.). In some examples, the systems and methods of the disclosure can generate a report using the information.

[0041] The information may be used to assist diagnosis and / or differential diagnosis of neurological disorder / disease, such as Parkinson’s disease, other (non-neurodegenerative) movement disorder, neurogenerative dementia disorders (e.g., parkinsonian, non-parkinsonian neurodegenerative dementia disorders, etc.). For example, the information may be used for detecting and quantitative monitoring of disease progression. The classification results can also be used to predict disease state.

[0042] Figure 1 shows an example of a flow diagram 100 illustrating a method of defining one or more regions in the brain of an individual and determining quantitative information associated with each region according to embodiments.

[0043] Operations in flow diagram 100 may begin at block 110 when medical image data of one or more regions of a brain of an individual (also referred to as an “subject”) acquired by medical imaging system(s) is obtained, such as receiving the MRI Image data 110 from an MRI imaging system and / or memory. The medical image data 110 may include one or more sets of MRI image data of one or more regions / subregions of the brain acquired using one or more stored protocols.

[0044] For example, the image data 110 can include but is not limited to image data of the brain of the individual acquired by the system using one or more stored protocols. In some embodiments, the protocols may relate to protocols for one or more MR imaging systems to acquire one or more sets of (individual) MRI image data. For example, the MRI image data may include quantitative contrast image data 112, T1 image data 114, among others, or any combination thereof.

[0045] For example, the quantitative contrast image data 112 may be obtained using pulse sequences that generate image contrasts sensitive to tissue characteristics. In some examples, the protocols may include but are not limited to protocols for pulse sequences for neuromelanin- sensitive MRI(e.g., explicit or incidental magnetization transfer contrast-based; NM-MRI magnetization transfer contrast, Magnetization Transfer (MT), (MT-GRE – GRE), MT’, MT ratio (MTR), etc.); iron-sensitive MRI sequences (e.g., T2 weighted imaging, R2* imaging, susceptibility weighted imaging, quantitative susceptibility mapping (QSM), etc.); diffusion MRI and its associated measures (e.g., free water, orientation dispersion index, other microstructural measures); resting and task-based functional MRI; chemical exchange saturation transfer (CEST) MRI contrasts; magnetic resonance spectroscopy (MRS), chemical shift imaging (CSI), proton density imaging; intravenous contrast enhanced MRI (e.g., gadolinium, iron based contrast, other ferromagnetic contrasts, contrasts relating to non-H1based MRI such as phosphorus or fluorine, etc.); other MRI contrasts, among others; or any combination thereof.

[0046] In some examples, the one or more regions of interest (ROI) may include but are not limited to one or more subregions and / or the region of substantia nigra pars compacta (SNc), locus coeruleus (LC), subthalamic nucleus, red nucleus, globus pallidus (total, pars interna and / or pars externa), putamen (lateral, medial, and / or total), caudate nucleus, cerebellar dentate nucleus, substantia nigra pars reticulata, middle cerebellar peduncle, superior cerebellar peduncle, inferior cerebellar peduncle, hippocampus (individual subfields and / or total), entorhinal cortex, occipital cortex (primary visual cortext, visual association cortext, and / or total), parietal cortex, cingulate gyrus, parahippocampal gyrus, frontal cortex (M1, premotor, supplementary motor area, Broca’s area, prefrontal, orbitofrontal, inferolateral frontal, and / or total), among others, or any combination thereof.

[0047] It will be understood that SNc region and subregions and LC region described in the examples are non-limiting examples of the ROI(s) and one or more of these otherregions / subregions may alternatively or additionally be the ROI(s). For example, the methods described herein may be adapted to local anatomy of a given ROI. By way of example, the details shown in the placement process for SNc can be abstracted to apply to other areas, but local anatomy to each ROI determines some aspects of the method for a given ROI. In this example, the method to shift the individualized SNc region mask to an optimal location may be shifted posteriorly in an iterative way, but the details of how an initial ROI is shifted can depend on local anatomy. For example, the shift may be different for subthalamic nucleus or putamen regions. In those structures, a more complex scheme of movements in multiple directions may be used.

[0048] In some examples, the individual MRI image data 110 of a brain region (e.g., brain and surrounding structures / tissues of the head) of the individual that includes one or more ROIs may be acquired with an MRI imaging system (e.g., 3T scanner, other MRI scanners, etc.) using the one or more protocols. For example, the MRI image data 110 may include but is not limited to quantitative contrast image data 112, such as NM-MRI image data acquired by one or more neuromelanin-sensitive MRI (NM-MRI) data acquisition protocols (e.g., reduced flip-angle magnetization prepared gradient echo (MT-GRE) sequence; acquired as separate measurements to enable off-line registration to address inter-measurement motion, etc.), iron-sensitive MRI data acquired by one or more iron-sensitive MRI acquisition protocols, among others, other types of quantitative contrast MRI image data, or a combination thereof; T1 image data 114 acquired using T1 MPRAGE data acquisition protocols; among others; or any combination thereof.

[0049] In some examples, other types of MRI data 110 may be co-acquisition. For example, the co-acquisition data may include but is not limited to: co-acquisition of multi-echo gradient echo MRI data (e.g., iron measurement within ROI); co-acquisition of diffusion MRI data (e.g., measurement of microstructural characteristics, including free water (FW), within ROI); co- acquisition of chemical exchange saturation transfer (CEST) MRI data within ROI (e.g., measurement of tissue molecular characteristics within ROI); co-acquisition of datasets with other MRI pulse sequences sensitive to tissue characteristics; among others, or a combination thereof.

[0050] In some examples, the individual MRI image data 110 (quantitative contrast image data 112 and / or the TI image data 114) may include one or more slices of the one or more ROIs.

[0051] At blocks 120 and 122, (i) one or more population region masks for the ROI(s) 120 and (ii) T1 template 122 may be obtained, such as retrieving from memory. By way of example, the one or more population region masks for the ROI(s) 120 may include a population region maskfor SNc, a population region mask for LC, a population region mask for other ROIs, among others, or any combination thereof. In some examples, a (standard space) population reference ROI associated with the ROI(s) 120 may be additionally obtained, such as retrieving from memory.

[0052] Each population region mask 120 may include a data structure stored in the memory and describe a region and / or subregion (e.g., SNc, SNc subregion, LC, etc.) within a reference brain in a standard space, such as the Montreal Neurological Institute (MNI)-space (e.g., MNI-152 space). For example, the dimensions and position (e.g., coordinates) of each population region mask 120 for an ROI may be quantitatively defined to include the anatomic location of the ROI, e.g., SNc or LC, and tissue adjacent to the main corpus of that structure so as to be optimized for selection of all voxels exhibiting the tissue contrast of interest (e.g., NM-MRI contrast while excluding voxels with contrast that is not referable to that structure (e.g., contrast originating from nearby blood vessels or ventricles)).

[0053] Each population region mask 120 may include a data structure stored in the memory and describe a region and / or subregion (e.g., SNc, SNc subregion, LC, etc.) within a reference brain in a standard space, such as the Montreal Neurological Institute (MNI)-space (e.g., MNI-152 space). In some examples, the (standard space) population reference ROI may be a representative bounding region of the brain generated and positioned in a quantitatively defined location in the brainstem relative to the associated population region mask.

[0054] In some examples, each population region mask may be a probabilistic mask that is generated from a plurality of representative images of the brain. By way of example, the probabilistic population mask may describe the ROI with probabilities for a presence of that ROI at different locations relative to a bounding region of the brain.

[0055] For example, the population region mask may be generating using representative NM- MRI image data of a brain region for a plurality of representative individuals. For each representative image data, an individualized region mask (also referred to as “individualized representative region mask”) may be generated to cover a ROI (e.g., SNc or LC) on one or two sides for each respective representative brain image data. The individualized representative region mask may be larger than the ROI. In some examples, the individualized region mask for each representative brain image data may include further processing (e.g., as discussed in Figure 10 for LC).

[0056] Next, a representative region standard mask may be generated for each individualized representative region mask by converting each individualized representative region mask into the standard (e.g., MNI) space.

[0057] Next, a (probabilistic) population region mask for the ROI may be generated in standard space using the generated individualized representative region standard space masks for the ROI. For example, all individualized representative region masks may be averaged in standard space (e.g., MNI-space) to generate the (probabilistic) population region mask.

[0058] In some examples, the (probabilistic) population region mask may be binarized (e.g., using a threshold (e.g., of about 0.5). The binarization may occur either before and / or after conversion of the (probabilistic) population region mask to individual space with a transformation matrix. For example, if the conversion is performed in standard space, a binarized population mask may be generated. When carried out in individual space, an (initial) individualized region mask may be generated.

[0059] In some examples, the (standard space) population reference ROI for each population region mask may (optionally) be generated using representative NM-MRI image data of a brain region for a plurality of representative individuals. For example, the population reference ROI may be generated using the respective population region mask. In some examples, the population mask may be binarized with a threshold (e.g., of about 0.5) and the associated population reference ROI (using one or more predefined geometric shapes (e.g., circles)) may automatically be generated lateral to the binarized ROI population mask.

[0060] In some examples, the population region mask and associated population reference ROI for SNc and LC may be generated according to the method described in Figures 8 and 10, respectively. In other examples, population region masks and / or associated population reference ROI for the SNc and / or LC, as well as masks for other regions, may be generated using different methods.

[0061] In some examples, the T1 template 122 may be a T1 standard space template. For example, the T1 template 122 may be a template specific to the standard space.

[0062] In some examples, next at block 130, one or more individualized region masks may be generated using the MRI image data 110 and the population region mask 120 for each ROI. In some examples, each population region mask 120 may be converted from standard space to anindividualized region mask in individual MRI space (e.g., NM-MRI space) based on registrations of the MRI data 110.

[0063] In some examples, each population region mask 110 may be converted from standard space to the individualized region mask for the respective ROI in individual MRI space using one or more transformation matrices generated from the registrations. In some examples, one or more transformation matrices may be generated based on the registrations between 1) individual quantitative contrast image data 112 and individual T1 image data 114; and 2) individual T1 image data 114 in the individual space and T1 template 122. In some examples, the one or more transformation matrices may be combined into a single, transformation matrix to be used to convert the population region mask to the individualized region mask. In some examples, the individualized region mask may then be binarized by thresholding.

[0064] For example, each individualized region mask may be generated by converting the respective population region mask 120 into individual space defined by the individual MRI image data 110 using the operations described in Figures 2A and 2B. By way of another example, each individualized region mask may be generated by converting the respective population region mask 120 into individual space defined by the individual MRI image data 110 using a different process.

[0065] By way of example, the generated individualized region mask may include more than one mask. For example, the individualized region mask for LC and SNc may include two individualized region masks disposed on opposing sides relative to the midline. Each of these individualized region masks may have its own individualized dimensions and position.

[0066] Figures 2A and 2B show an example of flow diagram 200 illustrating a method of converting each population region mask 120 from standard space into individual space to generate (respective) individualized region mask for an ROI according to some embodiments. Operations described in diagram 200 may be performed by a computing system, such as a computing system described below with respect to Figure 11.

[0067] Operations in flow diagram 200 may begin at block 210 with registering the individual MRI image data 110. In some examples, the registration may include two registrations that can be performed in parallel: a first registration 220 to register the individual T1 image data 114 and the quantitative contrast images 112; and a second registration 230 to register the individual T1 image data 114 and the T1 template 122.

[0068] In some examples, the first registration 220 may include one or more registrations with rigid transformations performed using the individual T1 image data 114 and the quantitative contrast images 112 of the individual. In some examples, the T1 image data 114 and the quantitative contrast image data 112 may be registered using partial and / or entire MRI image data in one or more registration iterations. For example, the one or more registrations may be performed using partial and / or entire TI images to partial and / or entire quantitative contrast images (with or without skull stripping).

[0069] For example, as shown in Figure 2B, the first registration 220 may include registering, at block 222, the entire T1 image 114 to the entire and / or partial quantitative contrast image (with or without skull stripping) to generate a registered whole, brain image in individual space. Using the registered whole brain image, a partial T1 image may be generated at block 224. Next, at block 226, the partial, T1 image with the same dimensions as the quantitative contrast image field of view may then be registered to the partial, quantitative contrast image (with skull stripping, without skull stripping, or iteratively both with and without skull stripping). Then, in some examples, steps 224-226 may be repeated for one or more additional iterations using different partial T1 images until certain criteria is met, such as predefined number of iterations. The one or more additional partial, T1 images may differ in location within the registered whole brain image (block 222). In other examples, operations performed at blocks 224 and 226 may be omitted.

[0070] In other examples, the first registration 220 to register the quantitative contrast images 112 and the individual T1 image data 114 may be performed using a different process.

[0071] In some examples, the second registration 230 may include registering the individual T1 images 114 to the standard space T1 template 122, for example, using linear registration followed by non-linear registration.

[0072] Next, as shown in Figure 2A, at block 240, one or more transformation matrices may be generated. In some examples, a matrix may be generated from each registration 220 and 230. For example, one or more matrices may be generated from the first registration 220; and a matrix may be generated from the second registration 230. By way of example, for the first registration, a matrix may be generated from the registration performed at block 222 and a matrix may be performed at each registration performed at block 226 (if generated). In some examples, the one or more transformation matrices may be combined into a single, transformation matrix to be used to convert the population region mask 120 to the individualized region mask.

[0073] Next, at block 250, the population region mask may be converted to the individual space using the one or more matrices (block 240) to generate the individualized region mask for the respective ROI.

[0074] Next, at block 260, the individualized region mask may be binarized by thresholding. The resulting individualized region mask may then be used in operations performed at blocks 140 and / or 150.

[0075] In some examples, one or more quality control operations may be performed during and / or after step 210. For example,(i) the individual T1 and quantitative contrast images registration 220 and / or (ii) the individual T1 to (MNI) T1 template registration 230 may be reviewed. In some examples, an agreement score that quantifies the percentage of non-overlap between the registered images for each registration may be calculated and compared to a criteria (e.g., threshold) to determine whether the registrations are acceptable for the proceeding steps. In some examples, if the agreement score for each registration does not meet the criteria, quality control may be performed to determine the cause of error.

[0076] Although not shown, the method 220 performed in Figure 2A may include additional operations. For example, as discussed below with regards to the LC region, the individualized region mask may be further processed between the operations performed at blocks 250 and 260.

[0077] As shown in Figure 1, at block 140, the individualized reference ROI in the individual space may be generated using (i) the respective individualized region mask (block 130) and / or population region mask 120 and (ii) the quantitative contrast image data 112 for the respective ROI. In some examples, the individualized reference ROI for each ROI may be generated using the respective individualized region mask 130 and the quantitative contrast image data 112 in the individual space. In some examples, the individualized reference ROI may be generated using an intensity distribution to remove high-intensity and low-intensity outlier voxels. By way of example, voxels that fall within a top percentage (e.g., top 1%-10% (e.g., top 5%-10%)) and a bottom percentage (e.g., bottom 1%-10 (e.g., bottom 5%-10%)) may be removed. The processing performed to generate the individualized reference ROI may differ based on the ROI. Examples of processing SNc region can be found at Figures 3-6 and LC region can be found at Figures 7-10.

[0078] The individualized reference ROI may include one or more reference regions in one or more slices for each ROI. For example, the individualized reference ROI may include two reference regions, three reference regions, four reference regions, five reference regions, etc. Insome examples, the reference ROI may include two or more reference regions. The reference regions may be the same or different. In some examples, if the generated reference ROI includes more than one reference region, the reference ROI may be treated as a single signal intensity distribution.

[0079] In some examples, an area for each reference region of the individualized reference ROI may be defined and bound by one or more predefined geometric shapes within a brain region (e.g., the cerebral peduncle just lateral of the SNc region mask and / or central pons ventral to the LC region mask) in individual space in each slice on each side where the respective ROI contrast is present that is consistent with the known anatomic shape of ROI. In some examples, the one or more predefined geometric shapes may include but is not limited to circles, squares, other shapes, or any combination. In some examples, each predefined geometric shape may have a predefined size. The predefined size may depend on the size of the individualized region mask.

[0080] In some examples, a location of each reference region within the quantitative contrast image data 112 may be based on a location of the individualized region mask. For example, each reference region may be placed so as to be adjacent (without overlapping) the individualized region mask. This way, a location of each reference region may be confined by the coordinates of the individualized region mask in the individual space.

[0081] In some examples, the generating the individualized reference ROI may first include positioning an initial reference ROI within the quantitative contrast image data 112 based on a location of the individualized region mask so that the initial reference ROI is disposed adjacent to the individualized region mask. In this example, the initial reference ROI may be defined by the overlap(s) between two or more of the more than one predefined geometric shape having the predefined size. For example, the initial reference ROI may include all voxels contained within the overlap(s) between the two geometric shapes. In some examples, the size of the initial reference ROI may depend on the size of the individualized region mask.

[0082] In some examples, the initial reference ROI may be generated in standard space and converted to individual space using the transformation matrix. For example, the population reference ROI may be converted to individual space using the transformation matrix.

[0083] For example, the individualized reference ROI may be generated selecting voxels in standard space, for example, adjacent to the population region mask in the cerebral peduncle (for SNc) or in the central pons (for LC), in a quantitatively defined position relative to the populationregion mask binarized at a particular probability threshold. The spatial relation may then be defined quantitatively such that the individualized reference ROI lies within the template brain voxels of the target structure (e.g., cerebral peduncle or central pons) and does not contain voxels from other bordering structures, such as ventricles, blood vessels, SNc, or LC, etc. The population reference ROI may also be defined quantitatively to be of sufficient size to ensure that enough voxels are included such that a non-gaussian signal intensity distribution will be unlikely to occur due to a small number of voxels. The population reference region ROI may then be converted to individual space, e.g., individual NM-MRI space, using the same transformation matrix that was generated and used to convert the population region mask from standard space into the individualized ROI region mask in individual space. Once in individual space, e.g., individual NM-MRI space, the population region mask may then be binarized at a defined threshold. The signal intensity distribution of the voxels may then be determined, and outlier voxels (high intensity and / or low intensity) may then be excluded from the mask. This can result in the individualized reference ROI mask. An image containing this mask may be generated, for example to support rapid, operationalized visual quality control.

[0084] In some examples, the placement of initial reference ROI (e.g., the one or more predefined geometric shapes) may be determined relative to the coordinates of the center of mass of the respective individualized region mask 130 in the individual space in each slice on each side where ROI contrast is present that is consistent with the known anatomical shape of the ROI. Next, outlier voxels (e.g., high intensity and low-intensity voxels) from the initial reference ROI may be removed based on signal intensity distribution of the initial reference ROI resulting in the individualized reference ROI used to segment the ROI. By way of example, voxels that fall within a top percentage (e.g., top 1%-10% (e.g., top 5%-10%)) and a bottom percentage (e.g., bottom 1%- 10% (e.g., bottom 5%-10%)) may be removed.

[0085] In some examples, the mean and standard deviation of signal intensity within the reference ROI may be determined. In some examples, an intensity threshold may be determined based on these measurements. For example, the intensity threshold may correspond to a number of standard deviations greater than the mean of signal intensity. For example, the intensity threshold could be about 1.5-6 standard deviations greater than the mean of signal intensity. In some examples, the intensity threshold may also be used for segmentation at block 150.

[0086] In some examples, one or more quality control operations may be performed. In some examples, a histogram of signal intensities in the reference region(s) of the individualized reference ROI may be generated to confirm that the distribution meets certain criteria, for example, the distribution is Gaussian. This way, voxels with extreme intensity values that can be caused by noise and anatomically inappropriate voxels (e.g., from SNc, ventricles, etc.) may be excluded from the reference ROI.

[0087] In some examples, if the distribution does not meet the criteria, quality control may be performed to determine the cause of error. For example, quality control may be performed to confirm the normality of the signal intensity distribution in the ROI, thereby improving the performance of features extracted with this pipeline (less variability).

[0088] Next, at block 150, the ROI may be segmented from the quantitative contrast image data 112 (i) within the location of the individualized region mask (block 130) and (ii) using the signal intensity distribution of the reference ROI in individual space (block 140). In some examples, the location of the individualized region mask within the reference ROI may optionally be determined before segmenting 150. For example, please see Figure 3 for an example of a flow diagram 300 illustrating a method of positioning an SNc individualized region according to some embodiments.

[0089] Next, the ROI region (e.g., LC and / or SNc)(also referred to as “segmented ROI region”) may be segmented within the positioned, individualized region mask based on signal intensity. For example, voxels within the individualized region mask that have a signal intensity greater than the intensity threshold determined at block 140 may be selected to be included in the ROI region.

[0090] Because the segmentation is based on thresholding informed by the characteristics of the signal intensity distribution of the reference ROI, which can therefore ensure a normal distribution, performance of the extracted features can be improved (such as, less variability due to noise in the reference region and resulting departures from a normal signal intensity distribution in the reference ROI).

[0091] In some examples, a ROI segment-out region may also be determined. The ROI segment-out region may include the voxels within the individualized region mask that are not in the segmented ROI (e.g., SNc or LC). For example, the out voxels (i.e., the voxels included in theROI segment-out region) may be the voxels within the individualized region mask whose signal intensity is less than the intensity threshold.

[0092] In some examples, if the quantitative contrast images 112 includes two sets of image data, such as NM-MRI and iron-sensitive data, the method 100 may optionally include, at block 160, registering the segmented ROI and / or the segment-out region to the other quantitative contrast image data. By way of example, steps 130-150 can be performed using one set of image data, such as NM-MRI, to segment one or more ROI regions (e.g., SNc and / or LC). The segmented ROI and / or the segment-out region may then each be registered to the other contrast quantitative image data (e.g., iron-sensitive data).

[0093] Next, at block 170, quantitative information associated with each segmented ROI and / or segment-out region for each ROI region (e.g., determined in blocks 150 and 160) for each set of MRI image data quantitative contrast image data may be determined. For example, the quantitative information for the segmented ROI and / or the segment-out region may include but is not limited to: volume measures, magnetization transfer contrast (MTC) measures, iron-sensitive (R2*) measures, MRI characteristics (e.g., including R2*, susceptibility, diffusion MRI measures (e.g., free water, ODI, NDI, FA, MD, etc.)), MRI Chemical Exchange Saturation Transfer (CEST) measures (e.g., amide, MT, phosphocreatine, glucose, pH, etc.)), other MRI measures, or any combination thereof.

[0094] For example, the one or more measures may include but are not limited to: segmented ROI volume (e.g., quantitative estimate of ROI volume in NM-MRI image data); segmented ROI volume / individualized mask ratio (e.g., ROI volume / individualized region mask volume NM-MRI data); segmented ROI MTC (e.g., mean MTC in the segmented ROI in NM-MRI image data); volume-weighted ROI MTC ratio (e.g., segmented ROI MTC X (segmented ROI volume / individualized mask volume NM-MRI image data)); segmented ROI R2* measure (e.g., mean R2* in the segmented ROI in iron-sensitive MRI data); volume-weighted ROI R2* ratio (e.g., segmented ROI R2* X (segmented ROI volume / individualized mask volume) in iron- sensitive MRI data); ROI segment-out region volume (e.g., quantitative estimate of ROI segment- out region volume in NM-MRI image data); ROI segment-out region MTC (e.g., MTC in the segment-out region in NM-MRI image data); volume-weighted ROI segment-out region MTC ratio (e.g., ROI segment-out region MTC X (segmented ROI volume / individualized mask volume) in NM-MRI image data); ROI segmented R2* measure (mean R* in the ROI segment-out regionin iron-sensitive MRI data); volume weighted ROI R2* ratio (e.g., ROI segment-out region R2* X (segmented ROI volume / individualized mask volume in iron-sensitive MRI data)); among others; or any combination thereof.

[0095] In some examples, the measures may include one or more contrast subregions in the segmented region. For example, the one or more contrast subregions may also include one or more neuromelanin / iron overlap regions. The one or more overlap regions may include but are not limited to control SN iron / neuromelanin contrast overlap region, PD SN iron / neuromelanin contrast overlap region, among others, or a combination thereof. For example, please see Langley J, Huddleston DE, Chen X, Sedlacik J, Zachariah N, Hu X. A multicontrast approach for comprehensive imaging of substantia nigra. Neuroimage.2015 May 15; 112:7-13; and Huddleston DE, Langley J, Sedlacik J, Boelmans K, Factor SA, Hu XP. In vivo detection of lateral-ventral tier nigral degeneration in Parkinson's disease. Hum Brain Mapp. 2017 May;38(5):2627-2634, which are hereby incorporated in their entirety. In some examples, one or more quantitative measures may be determined from the one or more overlap regions.

[0096] Next, at block 180, one or more reports including the associated information from block 170 may be generated. In some examples, the reports may also include generated images using the segmented images (e.g., ROI) and quantitative information. For example, the generated images may include but is not limited to qualitative parametric maps for each segmented ROI. For example, the generated images for each ROI (e.g., SNc and / or LC) may include but is not limited to anatomic region for the region contrast detection that can be in a color, the segmented region overlap in a different color, among others, or any combination thereof.

[0097] In some examples, one or more quality control operations may be performed in addition and / or alternative to those discuss above during one or more of the operations performed. Examples may include but are not limited to generating a visual QC review after: the registration(s) performed at block 130, the generation of the individualized reference ROI at block 140; and the segmentation of the ROI region at block 150; among others; or any combination thereof. In some examples, the report may include the images and / or quantitative information related to the one more quality control operations performed.

[0098] Operations described in diagrams 100 and 200 may be performed by a computing system, such as a computing system described below with respect to Figure 11. Although the flow diagrams 100 and 200 may describe the operations as a sequential process, in variousembodiments, some of the operations may be performed in parallel or concurrently. In addition, the order of the operations may be rearranged. An operation may have additional steps not shown in the figure. In some embodiments, some operations may be optional. Embodiments of the method may be implemented by hardware, software, firmware, middleware, microcode, hardware description languages, or any combination thereof. When implemented in software, firmware, middleware, or microcode, the program code or code segments to perform the associated tasks may be stored in a computer-readable medium such as a storage medium.

[0099] EXAMPLES

[0100] Now, having described the embodiments of the disclosure, in general, the examples describe some additional embodiments. While embodiments of the present disclosure are described in connection with the examples and the corresponding text and figures, there is no intent to limit embodiments of the disclosure to these descriptions. On the contrary, the intent is to cover all alternatives, modifications, and equivalents included within the spirit and scope of embodiments of the present disclosure.

[0101] Figures 3-10 show non-limiting examples of the methods described with respect to Figures 1 and 2 applied to specific regions. Figures 3-6 show non-limiting examples of flow diagrams 300-600 of the methods applied to the SNc region and Figures 7-10 show examples of flow diagrams 700-1000 of the methods applied to the LC region. It will be understood that the methods described with respect to Figures 1 and 2 may be applied to other regions.

[0102] Operations described in diagrams 300-1000 may be performed by a computing system, such as a computing system described below with respect to Figure 11. Although the flow diagrams 300-1000 may describe the operations as a sequential process, in various embodiments, some of the operations may be performed in parallel or concurrently. In addition, the order of the operations may be rearranged. An operation may have additional steps not shown in the figure. In some embodiments, some operations may be optional. Embodiments of the method may be implemented by hardware, software, firmware, middleware, microcode, hardware description languages, or any combination thereof. When implemented in software, firmware, middleware, or microcode, the program code or code segments to perform the associated tasks may be stored in a computer- readable medium such as a storage medium.

[0103] SNc Region Example

[0104] Figure 3 shows an example of the flow diagram 300 illustrating of a flow diagram 300 illustrating a method of segmented SNc, using an individualized SNc region mask and reference ROI, which can be used to determine quantitative information, according to embodiments. Like the flow diagram 100, operations in the flow diagram 300 may begin when MRI image data 310 of the SNc region of an individual are acquired by medical imaging system(s) is obtained, such as receiving the data from an imaging system and / or memory. In this example, the MRI image data 310 may include SNc quantitative contrast images 312 (e.g., NM-MRI image data, iron-sensitive MRI data, etc.) and T1 images 314 of the brain region that includes the SNc region. In some examples, the SNc quantitative contrast images 312 may include NM-MRI image data and / or iron- sensitive MRI image data.

[0105] At blocks 320 and 322, SNc population region mask 320 and T1 template 322 may be obtained, such as retrieving from memory. The SNc population region mask 320 may include a data structure stored in the memory and describes a region and / or subregion within a reference brain in a standard space, such as the Montreal Neurological Institute (MNI)-space (e.g., MNI-152 space). For example, in some examples, the SNc population region mask may be a probabilistic mask that is generated from a plurality of representative images of the brain. In some examples, the SNc population region mask may be generated using the method illustrated by the flow diagram 400 in Figure 4. In other examples, the SNc mask may be generated using different methods.

[0106] Next, like the flow diagram 100, at block 330, the individualized SNc region mask may be generated using the SNc quantitative contrast images 312 and the SNc population region mask 320. In some examples, the SNc population region mask 320 may be converted from standard space to an individualized SNc region mask in individual MRI space (e.g., NM-MRI space) based of registrations of the MRI data 310. In some examples, the SNc population region mask 320 may be converted from standard space to the individualized SNc region mask in individual MRI space using one or more transformation matrices generated from the registrations. In some examples, the individualized SNc region mask may be binarized by thresholding.

[0107] For example, the individualized SNc region mask may be generated using the method described at block 130 of the flow diagram 100 and the method illustrated by the flow diagram200 in Figures 2A and 2B. In other examples, the individualized SNc population region mask 320 may be generated using a different method.

[0108] Next, at block 340, the individualized SNc reference ROI (also referred as “reference ROI”) for SNc in the individual space may be generated using (i) the individualized SNc region mask and / or the SNc population region mask 320 and (ii) the quantitative contrast image data 312 (e.g., NM-MRI image data).

[0109] In some examples, the individualized SNc reference ROI may be generated using the individualized SNc region mask 330 and the SNc quantitative contrast image data 310 in the individual space using the method described in block 140 of the flow diagram 100. In some examples, an initial individualized reference ROI for SNc based on the location of the individualized SNc region mask within the quantitative contrast image data 312 may be generated. The high-intensity and / or low-intensity outlier voxels of the quantitative contrast image data 312 within the initial individualized SNc reference ROI may be removed according to a threshold based on intensity distribution generated for the quantitative contrast image data 312 data in order to remove both high-intensity and low-intensity outlier voxels. For example, the threshold may be about 1%-10%. The individualized SNc reference ROI for SNc resulting from the thresholding may be considered to be the individualized SNc reference ROI used for segmentation of the SNc. In some examples, the individualized SNc reference ROI may be generated using different methods.

[0110] In other examples, the individualized SNc reference ROI may be generated using the population SNc region mask 320. For example, an initial SNc reference ROI may first be generated in a standard space, such as an MNI-152 space, based on the location of the SNc population region mask 320 within the standard space. The SNc reference ROI in the standard space may then be projected back to the individual space to generate the individualized SNc reference ROI using a transformation determined by the one or more transformation matrices generated at block 330.

[0111] In some examples, the individualized SNc reference ROI for SNc may include one or more reference regions in one or more slices. For example, the reference ROI may include two, three, four, five, etc. In some examples, the reference ROI may include three or more reference regions. If the generated reference ROI includes more than one reference region, the reference ROI may be treated as a single, signal intensity distribution.

[0112] In some examples, an area for each individualized SNc reference ROI may be defined and bound by one or more geometric shapes (e.g., circle) within the cerebral peduncle just lateral of the SNc region mask in individual space in each slice on each side where SNc contrast is present that is consistent with the known anatomic shape of SNc. In some examples, each defined area may be confined by the coordinates of the individualized SNc region mask in the individual space. In some examples, the placement of one or more geometric shapes (e.g., circle) may be determined relative to the coordinates of the center of mass of the individualized SNc region mask in the individual space in each slice on each side where SNc contrast is present that is consistent with the known anatomical shape of SNc. Each defined area may correspond to a reference ROI region. In some examples, outliers may be removed based on thresholding.

[0113] In some examples, the one or more slices in which the individualized SNc reference ROI is generated may depend on the number of voxels present in the SNc individualized region mask. For example, the individualized SNc reference ROI may be generated if the number of voxels in the SNc individualized region mask exceeds a predefined number. The predefined number of voxels may include but is not limited to 50-1000 voxels. For example, the individualized SNc reference ROI may be generated within the cerebral peduncle on each slice in which at least a total of three hundred voxels is present in SNc region mask.

[0114] In some examples, like block 140, the mean and standard deviation of signal intensity within the individualized SNc reference ROI may be determined. In some examples, an intensity threshold may be determined based on these measurements. For example, the intensity threshold could be a number of standard deviations above the mean of signal intensity. For example, the threshold could be about 1.5-6 standard deviations above the mean of signal intensity. In some examples, the intensity threshold may also be used for segmentation at block 350.

[0115] Next, at block 350, the SNc may be segmented. In some examples, the segmentation may optionally include positioning the individualized SNc region mask (block 340) with respect to the quantitative contrast image data 312 at block 352. In some examples, the individualized SNc region mask may be automatically positioned with respect to the quantitative contrast image data 312 until the individualized region SNc region mask includes hyperintense voxels of SNc but does not include hyperintense voxels from adjacent areas. In some examples, the individualized SNc region mask may be automatically positioned using the method illustrated in flow diagram 400 ofFigure 4. In other examples, the individualized SNc region mask may be positioned using other methods. In further examples, the operations performed at block 350 may be omitted.

[0116] Figure 4 shows an example of flow diagram 400 illustrating a method of positioning the individualized SNc region mask (block 340) with respect to the quantitative contrast image data 312 according to some embodiments. Operations described in diagram 400 may be performed by a computing system, such as a computing system described below with respect to Figure 11.

[0117] At block 410, the individualized SNc region mask may be first positioned within the quantitative contrast image data 312 (e.g., NM-MRI image data) (also referred to as “first location”). Next, at block 420, an edge mask may be generated by selecting the outermost layer of voxels within the individualized SNc region mask.

[0118] Next, at block 430, an edge mask measure (also referred to as “first edge mask measure”) may be determined using the first location. In some examples, the first edge mask measure may include but is not limited to a percentage of edge mask voxels with a signal intensity above the intensity threshold from block 340. Next, at block 440, the edge mask may shift in position by one voxel within the quantitative contrast image data 312 (also referred to as “second location”). Then, at block 450, an edge mask measure (also referred to as “second edge mask measure”) may be determined using the second location. In some examples, like the first edge mask measure, the first edge mask measure may include but is not limited to a percentage of edge mask voxels with a signal intensity above the intensity threshold from block 340. At block 460, the first edge mask measure and the second edge mask measure may be compared. If the second edge mask measure is greater than the first edge mask measure (YES at block 460), the second location of the edge mask may be used to determine the location of the individualized SNc region mask with respect to the SNc quantitative contrast image data 312 (e.g., NM-MRI image data) at block 470.

[0119] If the second edge mask measure is less than the first edge mask measure (NO at block 460), operations at blocks 440-460 (also referred to as an “iteration”) may be repeated until the second edge mask measure is greater than the first edge mask measure (YES) at block 460. When the second edge mask measure is greater than the first edge mask measure, a minimum percentage of suprathreshold voxels in the edge mask is observed (YES at block 460) at the second location of the edge mask.

[0120] In subsequent iterations (e.g., after the first iteration), the second edge location and second edge mask measure of the previous iteration (e.g., the first iteration) corresponds to the first edge location and first edge mask measure of the second iteration.

[0121] Next, at block 354, the SNc may be segmented from the SNc quantitative contrast image data 312 (i) within the location of the individualized SNc region mask (block 352) and (ii) using the signal intensity distribution of the individualized SNc reference ROI in individual space (determined at block 340).

[0122] Next, the SNc may be segmented within the positioned, individualized SNc region mask by selecting voxels with the individualized SNc region mask that have a signal intensity greater than the intensity threshold determined at block 340.

[0123] In some examples, SNc segment-out region may also be determined. The SNc segment- out region may include all voxels within the SNc region mask that are not in the segmented SNc (SNc region).

[0124] In some examples, like block 160, the segmented SNc may optionally be registered to one or more other SNc quantitative contrast images to segment the SNc from the other images. For example, if the operations at blocks 330-350 were performed using NM-MRI image data, the segmented SNc may be registered to iron-sensitive MRI image data to determine the segmented SNc within those images. In some examples, the operations at block 360 may be omitted.

[0125] Next, at block 370, quantitative information associated with the segmented SNc and / or SNc segment-out region may be determined. For example, the quantitative information may include one or more measures described in block 170.

[0126] Next, at block 380, a report including the associated information from block 370 may be generated. For example, the report may include the quantitative information, segmented images, etc., for example, as described in block 180.

[0127] In some examples, one or more quality control operations may be performed as described in Figures 1 and 2 with regard to the respective operations. SNc Segmentation Example

[0128] Figure 5 shows an illustrative example 500 of segmenting SNc in individual space from NM-MRI image data 312, for example, by performing the operations at 330-350 of the flow diagram 300 shown in Figure 3, according to some embodiments.

[0129] As shown in the image 510, an individualized SNc region mask 512 may be generated by converting the SNc population region mask 320, for example, as described in Figure 6, in standard space (MNI) into individual space using the individual’s MRI image data 312, for example, as described in block 330. In this example, the mask was binarized using a threshold of 0.5.

[0130] Next, as shown in the image 520, an individualized SNc reference ROI 522 may be generated, for example, relative to the coordinates of the center of mass of the respective individualized SNc region mask 512, for example, as described in block 340.

[0131] Next, as shown in the image 530, the individualized SNc region mask 512 may be positioned for segmentation, for example, as described in block 350 and Figure 4. As shown, the individualized SNc region mask 512 can be shifted posteriorly to a different position with respect to the individualized SNc reference ROI 522 until its edge mask contains the minimum number of high-intensity voxels. As shown in image 530, the individualized SNc region mask 512 is positioned at a different position that is posterior to the SNc region mask 512 in the images 510 and 520.

[0132] Next, as shown at 540, the SNc 552 and the SNC segment-out region 562, as shown in the images 550 and 560, respectively, may be segmented using the positioned mask 512 from the area defined by the individualized SNc reference ROI 522.

[0133] In some examples, quantitative information for the SNc based on the segmented SNc regions 542 and / or 544 may be determined. Population SNc mask Generation Example

[0134] Figure 6 shows an illustrative example 600 of generating the SNc population region mask 320 according to some embodiments. For example, the generated SNc population region mask can be used in Figure 5 to generate the individualized SNc region mask 512. In this example, the generated LC population region mask may be probabilistic.

[0135] In some examples, the SNc population region mask may be generated in the standard space (e.g., MNI-152 space) using a voxel-selection method. In some examples, the SNc population region mask may be probabilistic.

[0136] In some examples, as shown in the image 610, NM-MRI image data of a brain region for a plurality of representative individuals may be received.

[0137] Next, as shown in the image 620, for each individual, individualized representative SNc region masks 622 may be defined to cover the SNc region on both sides in the respective NM-MRI image data. In this example, the representative SNc region masks 622 may be larger than the SNc regions.

[0138] Next, as shown in the image 630, each individualized representative SNc mask 622 may be converted into a representative SNc region standard space mask 632 into the standard (e.g., MNI) space.

[0139] Next, as shown in the image 640, the SNc population region mask 642 may be generated from the representative SNc region standard space mask 632 generated for the plurality of representative individuals. For example, all representative SNc region standard space mask 632 may be averaged in standard space (e.g., MNI-space) to generate the SNc population region mask 642.

[0140] In some examples, as shown in the image 650, a standard SNc reference ROI 652 may be generated in the standard space. In some examples, the SNc population region mask 642 may be binarized (e.g., using a threshold (e.g., of about 0.5) to generate the binarized SNc population region mask 652. Next, the (standard space) SNc population reference ROI 654 may be automatically generated lateral to the binarized SNc population region mask 652. In some examples, the (standard space) SNc population reference ROI 654 may be used to generate the initial SNc reference ROI. In some examples, this step may be omitted.

[0141] In some examples, the SNc population region mask 642 and / or the (standard space) SNc population reference ROI 654 may be generated using a different method.

[0142] LC Region Example

[0143] Figure 7 shows an example of the flow diagram 700 illustrating a method of segmented LC, using an individualized LC region mask and individualized LC reference ROI, which can be used to determine quantitative information, according to embodiments. Like the flow diagram 100, operations in the flow diagram 700 may begin when MRI image data 710 of the LC region of an individual are acquired by medical imaging system(s) is obtained, such as receiving the data from an imaging system and / or memory. In this example, the MRI image data 710 may include (individual) LC quantitative contrast image data 712 and T1 image data 714 of the brain region that includes the LC region of the brain. In some examples, the LC quantitative contrast imagedata 712 may include NM-MRI image data and / or iron-sensitive MRI image data. In some examples, the operations in the flow diagram 700 may be performed after the operations are performed in the flow diagram 400 for the SNc region. In some examples, one or more sets of the same MRI image data 710 may be used.

[0144] At blocks 720 and 722, LC population region mask 720 and T1 template 722 may be obtained, such as retrieving from memory. The LC population region mask 720 may include a data structure stored in the memory and describe a region and / or subregion within a reference brain in a standard space, such as the Montreal Neurological Institute (MNI)-space (e.g., MNI-152 space). For example, in some examples, the LC population region mask may be a probabilistic mask that is generated from a plurality of representative images of the brain. In some examples, the LC population region mask may be generated using the method illustrated by the flow diagram 1000 in Figure 10. In other examples, the LC mask may be generated using different methods.

[0145] Next, like the flow diagram 100, at block 730, the individualized LC region mask may be generated using the LC quantitative contrast images 712 and the LC population region mask 720.

[0146] In some examples, the LC population region mask 720 may be converted from standard space to an individualized LC region mask in individual MRI space (e.g., NM-MRI space) based of registrations of the MRI data 710. In some examples, the LC population region mask 720 may be converted from standard space to the individualized LC region mask in individual MRI space using one or more transformation matrices generated from the registrations. In some examples, the individualized LC region mask may be binarized by thresholding.

[0147] In some examples, the registrations performed for the SNc (Figure 3) may be used to generate the LC population region mask 720 if the LC quantitative contrast image data 710 included both the LC and SNc regions. In other examples, the registrations may be performed using the methods as described in Figures 1-2B.

[0148] For example, the individualized LC region mask may be generated using the method described in block 130 of the flow diagram 100 and the method illustrated by the flow diagram 200 in Figures 2A and 2B.

[0149] In some examples, the generated individualized LC region mask may be further processed. For example, the center of mass (CM) of the first (initial) generated individualized LC region mask on each side may be calculated and stored. Next, a geometric shape having apredefined shape having predefined dimensions may be placed in the center of the pons at coordinates in relation to the CM to define an area within the first generated individualized LC region mask. For example, the geometric shape may include but is not limited to a square (e.g., dimensions of about 2 mm x 2 mm (e.g., 5 X 5 voxels)), circle, among other shapes; or any combination thereof. By way of another example, the predefined dimensions may include 2 mm X 2mm and may be smaller or greater than 2 mm (e.g., 1- 7 mm). Voxels within the first (initial) generated individualized LC region mask may then be thresholded to identify and remove high- intensity and low-intensity voxels. For example, voxels that fall within a top percentage (e.g., top 1%-10%) and a bottom percentage (e.g., bottom 1%-10%) may be removed. The resulting mask may correspond to the individualized LC region mask for segmentation at block 750.

[0150] In other examples, the individualized LC population region mask may be generated using different methods.

[0151] Next, at block 740, the individualized LC reference ROI in the individualized space may be generated. In some examples, the individualized LC reference ROI may be generated using the method described in block 140 of the flow diagram 100.

[0152] In other examples, a different method may be used. For example, if there are multiple slices in which the individualized LC region mask is present, the individualized LC reference ROI may be generated using a reference slice. By way of example, Figure 8A shows an example of flow diagram 800 illustrating a generating an individualized LC reference ROI using a reference slice according to some embodiments.

[0153] At block 810, a reference slice may be determined. In some examples, the reference slice may be the most superior slice on which the reference ROI will be generated. For example, Figure 8B shows an example 850 of a subset of slices of the individual MRI contrast Image data 710 (e.g., NM-MRI Image data). In this example, the six slices 852-862 shown may be a subset of the slices received. For example, about 1-20 number of slices may be received.

[0154] For example, if there are multiple slices in which the individualized LC region mask is present, the slice on which the individualized LC reference ROI may be generated is located. In some examples, if the individualized SNc region mask has been generated, the individualized LC reference ROI region may be generated using the individualized SNc and LC masks in the MRI contrast Image data 710. In this example, the bottom slice of the MRI contrast image data that includes the individualized SNc region mask (block 330) may be determined. Using the bottomslice, the reference slice in which the individualized LC reference ROI may be generated may be determined. In this example, the reference slice may be considered the slice that is a number of slices (e.g., two slices) below (inferior) the bottom slice in which the individualized SNc region mask is disposed.

[0155] In the representative subset shown in Figure 8B, the slices 852-862 are shown from superior (top) to inferior (bottom). For example, the slice 852 would be considered the most superior (top) slice and slice 862 may be considered the most inferior (bottom) slice. By way of example, if the bottom of the SNc individualized region mask is observed in the slice 852, the slice 856 would be the reference slice in which the individualized LC reference ROI may be generated.

[0156] In other examples, if the individualized SNc region mask was not present in any of the slices in which the individualized LC region mask is present and / or was not generated, then the second most superior slice (e.g., slice below the top slice) on which the individualized LC region mask for segmentation is present may be considered to be the reference slice on which the individualized LC reference ROI may be generated. By way of example, if slice 854 is the most superior slice in which the individualized LC region mask was observed, the slice 856 would be the reference slice in which the individualized LC reference ROI may be generated.

[0157] Next, at block 820, after the reference slice is determined, a center of mass of the individualized LC region masks (disposed on either side) in the reference slice may be determined. For example, the center of mass may be located near the midline of the brainstem.

[0158] Next, at block 830, the individualized LC reference ROI may be generated at a location based on the center of mass. For example, a location at a distance from the center of mass may be identified. By way of example, the distance may include but is not limited to about 3 mm - 20 mm (e.g., 10 mm). The individualized LC reference ROI may be generated using this location. For example, the individualized LC reference ROI may correspond to a predefined geometric shape having a predefined size centered at this location. In some examples, the geometric shape may be a circle having a radius of 5 mm. In other examples, other shapes and sizes may be used. For example, the geometric shape is not limited to a circle having a radius of 2-20 mm and that other shapes and sizes may be used. In this example, the voxels selected at the location of this geometric shape may correspond to the individualized LC reference ROI.

[0159] In some examples, the individualized LC reference ROI may be generated on (i) the reference slice and (ii) all slices of MRI contrast data on which the individualized LC region mask is disposed on at least one side for segmentation is present inferior (below) to the reference slice.

[0160] In some examples, high-intensity and / or low-intensity outlier voxels in the individualized LC reference ROI in each slice in which it is generated may be removed based on intensity distribution. For example, voxels that fall within a top percentage (e.g., top 1%-10%) and a bottom percentage (e.g., bottom 1%-10%) may be removed.

[0161] In some examples, like blocks 140 and the 340, the mean and standard deviation of all voxels in the individualized LC reference ROI may be determined. Voxels within the individualized LC reference ROI that have an intensity value greater than a threshold determined based on mean, standard deviation, and a multiplier variable to define the number of standard deviations greater than the mean intensity at which to set the threshold. The multiplier variable may be based on signal-to-noise ratio). For example, the threshold may correspond to Mean + SD x multiplier variable.

[0162] In some examples, one or more quality control operations may be performed. In some examples, a histogram of signal intensities in the reference region(s) of the individualized LC reference ROI may be generated to confirm that the distribution meets certain criteria, for example, the distribution is Gaussian. This way, voxels with extreme intensity values that can be caused by noise and anatomically inappropriate voxels (e.g., from LC, ventricles, etc.) may be excluded from the reference ROI.

[0163] In some examples, one or more quality control operations may be performed as described in Figures 1 and 2.

[0164] Next, at block 750, the LC may be segmented from the LC quantitative contrast image data 712 (i) within the location of the individualized LC region mask (block 730) and (ii) using the signal intensity distribution of the individualized LC reference ROI in individual space (determined at block 740).

[0165] For example, the LC may be segmented within the positioned, individualized LC region mask by selecting voxels within the individualized LC region mask that have a signal intensity greater than the intensity threshold determined at block 740.

[0166] In some examples, a LC segment-out region may also be determined. The LC segment- out region may include all voxels within the individualized LC region mask that are not in the segmented LC (LC region).

[0167] In some examples, like block 160, the segmented LC may optionally be registered to one or more other LC quantitative contrast images to segment the LC from the other images at block 760. For example, if the operations at blocks 730-750 were performed using NM-MRI image data, the segmented LC may be registered to iron-sensitive MRI image data to determine the segmented LC within those images. In some examples, the operations at block 760 may be omitted.

[0168] Next, at block 770, quantitative information associated with the segmented LC and / or the LC segment-out region may be determined. For example, the quantitative information for each segmented region may include one or more measures described in block 170.

[0169] Next, at block 780, a report including the associated information from block 770 may be generated. For example, the report may include the quantitative information, segmented images, etc., for example, as described in block 180.

[0170] In some examples, one or more quality control operations may be performed as described in Figures 1 and 2 with regard to the respective operations. LC Segmentation Example

[0171] Figure 9 shows an illustrative example 900 of segmenting LC in individual space from NM-MRI image data 712, for example, by performing the operations at 730-750 of the flow diagram 700 shown in Figure 7 according to some embodiments.

[0172] As shown in Figure 9, at 902, an individualized LC region mask may be generated by converting the LC population region mask 720, for example, as described in Figure 10, in standard space (MNI) into individual space using the individual’s MRI image data 712, for example, as described in block 730. By way of example, as shown in the image 910, the initial individualized LC region mask 912 may be generated as described in block 130 of Figure 1. In this example, the mask was binarized using a threshold of 0.5.

[0173] Next, as shown in the image 920, the initial individualized LC region mask 912 may be further processed, for example, by performing the operations in block 730, to generate the individualized LC region mask 922. For example, as described in block 730, the center of mass (CM) of the individualized LC region mask 912 may be determined for each side of the mask 912so as to identify the center voxel. Using the coordinates of the CM, low-intensity voxels may be removed with respect to an area defined by a pre-defined geometric shape having predefined dimensions and centered at the CM. Voxels within the defined area of the first (initial) generated individualized LC region mask may then be thresholded to identify and remove low-intensity voxels to generate the mask 922.

[0174] Next, an individualized LC reference ROI 930 may be generated in the reference slice as shown in the image 940, for example, relative to the coordinates of the center of mass (CM) 934 of the respective individualized LC region mask 922, for example, as described in block 740 and Figure 8. For example, the reference slice may be determined, for example, by performing the operations described in Figure 8, such as the second slice below the bottom slice of the SNc. Next, the CM 934 of individualized region mask 934 (based on the opposing sides) in the reference slice may be determined. Next, using the coordinates of the CM of the individualized LC region mask 922, low-intensity voxels may be removed with respect to an area defined by a pre-defined geometric shape having predefined dimensions and centered at the CM. Voxels within the defined area of the defined area may then be thresholded to identify and remove low-intensity voxels to generate the individualized LC reference ROI 932 for segmenting. The threshold may be based on mean and standard deviation of intensity of all voxels in the individualized LC reference ROI and Sfactor as described in block 750.

[0175] Next, as shown at 940, the LC 952 and the LC segment-out region 954, as shown in the enlarged, partial images 950 and 960, respectively may be segmented using the positioned individualized LC region mask 922 from the area defined by the individualized LC reference ROI 932.

[0176] In some examples, quantitative information for the LC based on the segmented LC regions 952 and / or 962 may be determined. Population LC mask Generation Example

[0177] Figure 10 shows an illustrative example 1000 of operations to generate the LC population region mask according to embodiments. For example, the generated LC population region mask can be used in Figure 9 to generate the individualized LC region mask 922. In this example, the generated LC population region mask may be probabilistic.

[0178] In some examples, as shown in the image 1010, representative NM-MRI image data of the brain region for a plurality of representative individuals may be received.

[0179] Next, as shown in the image 1020, for each individual, the center of the LC on each side 1022 may be defined in the respective NM-MRI image data. For example, for each side 1022, a frame having a predefined geometric shape having predefined dimensions may be applied to cover the LC. The geometric frame, for example, may have a square shape of 3x3 voxels. In other examples, the frame may have a different shape and / or different size. The coordinates of the center voxel within the frame may be determined and recorded. This process can be repeated for the other side so that the coordinates of the center voxel of each side may be recorded. By identifying the center voxel, the maximum intensity voxel of the LC on each side may be identified.

[0180] Next, as shown in the image 1030, an individualized representative LC mask 1032 may be generated so that its center is located at the center voxel of the respective side. As shown in the image 1030, the representative individualized LC region mask may have a predefined geometric shape having predefined dimensions. For example, the individualized LC region mask may have a square shape with 10X10 voxels.

[0181] Next, as shown in the image 1040, each individualized representative LC mask 1032 may be converted into a standard representative LC mask 1042 into the standard (e.g., MNI) space.

[0182] Next, as shown in the image 640, the LC population region mask 1052 may be generated from the masks 1042 generated from the plurality of individuals. For example, all individualized representative LC masks 1042 may be averaged in standard space (e.g., MNI- space) to generate the LC population region mask 1052.

[0183] In some examples, although not shown, a (standard space) population LC reference ROI may be generated in the standard space. In some examples, the LC population region mask 640 may be binarized (e.g., using a threshold (e.g., of about 0.5) to generate a binarized LC population region mask 652. Next, the (standard space) population LC reference ROI may be automatically generated lateral to the binarized LC population region mask 652.

[0184] In other examples, a different method may be used to generate the LC population region mask.

[0185] System Example

[0186] Figure 11 depicts a block diagram of an example computing system 1100 for implementing certain embodiments. For example, in some aspects, the computer system 1100 may include computing systems associated with device(s) performing one or more operations (e.g., Figures 1-10) disclosed herein.

[0187] The block diagram illustrates some electronic components or subsystems of the computing system. The computing system 1100 depicted in Figure 11 is merely an example and is not intended to unduly limit the scope of inventive embodiments recited in the claims. One of ordinary skill in the art would recognize many possible variations, alternatives, and modifications. For example, in some implementations, the computing system 1100 may have more or fewer subsystems than those shown in Figure 11, may combine two or more subsystems, or may have a different configuration or arrangement of subsystems.

[0188] Other systems may also be used. It is also to be understood that the system 1100 may omit any of the modules illustrated and / or may include additional modules not shown.

[0189] By way of example, the system 1100 may include a computer system 1102 and a medical scanner 1710 (e.g., MR scanner capable of MRI image). In another example, the computer system 1102 may be part of the medical scanner 1160. In a further example, the computer system 1102 may be a part of an archival and / or image processing system, such as associated with a medical records database workstation or server, separate from the medical scanner 1160. In other examples, the computer system 1102 may be a personal computer, such as a desktop or laptop, a workstation, or combinations thereof. The computer system 1102 may be provided without other components for acquiring data by scanning a patient.

[0190] In the example shown in Figure 11, the computing system 1102 may include one or more processors 1110 and storage 1120. The processor(s) 1110 may be configured to execute instructions for performing various operations. The processor(s) 1110 may include one or more processing units, which may be any known processor or a microprocessor. For example, the processor(s) 1110 may include any known central processing unit (CPU), imaging processing unit (e.g., capable of processing medical image), graphical processing unit (GPU) (e.g., capable of efficient arithmetic), among others, or any combination thereof. The processor(s) 1110 may be communicatively coupled with a plurality of components within the computing system 1102. For example, the processor(s) 1110 may communicate with other components across a bus. The busmay be any subsystem adapted to transfer data within the computing system 1102. The bus may include a plurality of computer buses and additional circuitry to transfer data.

[0191] In some embodiments, the processor(s) 1110 may be coupled with and / or can comprise memory (e.g., memory 1120) and can be configured to execute instructions stored in the memory 1120 or storage to enable various apparatus, applications, or operating systems to perform operations and / or methods discussed herein. In some embodiments, the memory 1120 may offer both short-term and long-term storage and may be divided into several units. The memory 1120 may be volatile, such as static random access memory (SRAM) and / or dynamic random access memory (DRAM), and / or non-volatile, such as read-only memory (ROM), flash memory, and the like. Furthermore, the memory 1120 may include removable storage devices, such as secure digital (SD) cards. The memory 1120 may provide storage of computer readable instructions, data structures, program modules, audio recordings, image files, video recordings, and other data for the computing system 1102.

[0192] In some examples, the memory 1120 may be configured to store one or more (standard) population region masks 1122, associated (standard) population reference ROI 1124, and other parameters 1126. For example, the one or more population region masks 1122 and associated population reference ROI 1124, for example, may be generated using the methods illustrated in Figures 8 and 10. The one or more population region masks 1112 and associated reference ROI 1124 may be for regions including but not limited to SNc region, SNc subregion, LC region, among others, or any combination thereof. In some examples, one or more parameters 1126 used in the methods illustrated in flow diagrams 100-1000 may also be stored. For example, the one or more parameters 1126 may include but is not limited to predefined geometric shapes and associated dimensions, other parameters, among others, or any combination thereof.

[0193] In some embodiments, the memory 1120 may be distributed into different hardware modules. A set of instructions and / or code might be stored on the memory 1120. The instructions might take the form of executable code that may be executable by the computing system 1102, and / or might take the form of source and / or installable code, which, upon compilation and / or installation on the computing system 1102 (e.g., using any of a variety of generally available compilers, installation programs, compression / decompression utilities, and the like), may take the form of executable code.

[0194] In some embodiments, the memory 1120 may store a plurality of application modules, which may include any number of applications, such as applications for controlling input / output (I / O) devices 1140 (e.g., a sensor, a switch, a camera, a microphone or audio recorder, a speaker, a media player, a display device, etc.). The memory 1120 may store particular instructions to be executed by the processor(s) 1110.

[0195] In some embodiments, another system may assume the data analysis, image processing, or other functions of the processor(s) 1100. In response to commands received from an input device, the programs or data stored in the memory 1120 may be archived in long-term storage or may be further processed by the processor and presented on a display.

[0196] The computer system 1102 may include a communication subsystem 1130 that can include, for example, an infrared communication device, a wireless communication device and / or chipset (such as a Bluetooth® device, an IEEE 802.11 (Wi-Fi) device, a WiMax device, cellular communication facilities, and the like), NFC, ZigBee, and / or similar communication interfaces.

[0197] Depending on desired functionality, the communication subsystem 1130 may include separate transceivers to communicate with base transceiver stations and other wireless devices and access points, which may include communicating with different data networks and / or network types, such as wireless wide-area networks (WWANs), WLANs, or wireless personal area networks (WPANs). A WWAN may be, for example, a WiMax (IEEE 802.9) network. A WLAN may be, for example, an IEEE 802.11x network. A WPAN may be, for example, a Bluetooth network, an IEEE 802.15x, or some other types of network. The techniques described herein may also be used for any combination of WWAN, WLAN, and / or WPAN. In some embodiments, the communications subsystem 1130 may include wired communication devices, such as Universal Serial Bus (USB) devices, Universal Asynchronous Receiver / Transmitter (UART) devices, Ethernet devices, and the like. The communications subsystem 1130 may permit data to be exchanged with a network, the medical scanner 1160, other computing systems, and / or any other devices described herein. The communication subsystem 1130 may include a means for transmitting or receiving data, such as identifiers of portable goal tracking devices, position data, a geographic map, a heat map, photos, or videos, using antennas and wireless links. The communication subsystem 1130, the processor(s) 1110, and the storage 1120 may together comprise at least a part of one or more of a means for performing some functions disclosed herein.

[0198] The computing system 1130 may include one or more I / O devices 1140, such as a switch, a camera, a microphone or audio recorder, a communication port, or the like. For example, the I / O devices 1140 may include one or more touch sensors or button sensors associated with the buttons. In some examples, the I / O devices 1140 may also include, for example, a speaker, a media player, a display device, a communication port, or the like. For example, the display device may include an LED or LCD display and the corresponding driver circuit. The I / O devices 1140 may also include a text, audio, or video player that may display a text message, play an audio clip, or display a video clip.

[0199] The computing system 1102 may include a power device 1150 for providing electrical power to other circuits on the computing system 1102. In some examples, the power device 1150 may include a rechargeable battery. The power device 1150 may also include some power management integrated circuits, power regulators, power converters, and the like.

[0200] The computing system 1102 may be implemented in many different ways. In some embodiments, the different components of the computing system 1102 described above may be integrated into a same printed circuit board. In some embodiments, the different components of the computing system 1102 described above may be placed in different physical locations and interconnected by, for example, electrical wires. The computing system 1102 may be implemented in various physical forms and may have various external appearances. The components of computing system 1102 may be positioned based on the specific physical form.

[0201] The methods, systems, and devices discussed above are examples. Various embodiments may omit, substitute, or add various procedures or components as appropriate. For instance, in alternative configurations, the methods described may be performed in an order different from that described, and / or various stages may be added, omitted, and / or combined. Also, features described with respect to certain embodiments may be combined in various other embodiments. Different aspects and elements of the embodiments may be combined in a similar manner. Also, technology evolves and, thus, many of the elements are examples that do not limit the scope of the disclosure to those specific examples.

[0202] The foregoing method descriptions and the process flow diagrams are provided merely as illustrative examples and are not intended to require or imply that the operations of various embodiments must be performed in the order presented. As will be appreciated by one of skill in the art the order of operations in the foregoing embodiments may be performed in any order. Wordssuch as “thereafter,” “then,” “next,” etc. are not intended to limit the order of the operations; these words are simply used to guide the reader through the description of the methods. Further, any reference to claim elements in the singular, for example, using the articles “a,” “an” or “the” is not to be construed as limiting the element to the singular.

[0203] While the terms “first” and “second” are used herein to describe data transmission associated with a subscription and data receiving associated with a different subscription, such identifiers are merely for convenience and are not meant to limit various embodiments to a particular order, sequence, type of network or carrier.

[0204] Various illustrative logical blocks, modules, circuits, and algorithm operations described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and operations have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application, but such embodiment decisions should not be interpreted as causing a departure from the scope of the claims.

[0205] The hardware used to implement various illustrative logics, logical blocks, modules, and circuits described in connection with the embodiments disclosed herein may be implemented or performed with a general purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. A general-purpose processor may be a microprocessor, but, in the alternative, the processor may be any conventional processor, controller, microcontroller, or state machine. A processor may also be implemented as a combination of computing systems, (e.g., a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration. Alternatively, some operations or methods may be performed by circuitry that is specific to a given function.

[0206] In one or more example embodiments, the functions described may be implemented in hardware, software, firmware, or any combination thereof. If implemented in software, thefunctions may be stored as one or more instructions or code on a non-transitory computer readable medium or non-transitory processor-readable medium. The operations of a method or algorithm disclosed herein may be embodied in a processor-executable software module, which may reside on a non-transitory computer-readable or processor-readable storage medium. Non-transitory computer-readable or processor-readable storage media may be any storage media that may be accessed by a computer or a processor. By way of example but not limitation, such non-transitory computer-readable or processor-readable media may include RAM, ROM, EEPROM, FLASH memory, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium that may be used to store desired program code in the form of instructions or data structures and that may be accessed by a computer. Disk and disc, as used herein, include compact disc (CD), laser disc, optical disc, digital versatile disc (DVD), floppy disk, and Blu-ray disc where disks usually reproduce data magnetically, while discs reproduce data optically with lasers. Combinations of the above are also included within the scope of non- transitory computer-readable and processor-readable media. Additionally, the operations of a method or algorithm may reside as one or any combination or set of codes and / or instructions on a non-transitory processor-readable medium and / or computer-readable medium, which may be incorporated into a computer program product.

[0207] Those skilled in the art will appreciate that information and signals used to communicate the messages described herein may be represented using any of a variety of different technologies and techniques. For example, data, instructions, commands, information, signals, bits, symbols, and chips that may be referenced throughout the above description may be represented by voltages, currents, electromagnetic waves, magnetic fields or particles, optical fields or particles, or any combination thereof.

[0208] Terms, “and” and “or” as used herein, may include a variety of meanings that also is expected to depend at least in part upon the context in which such terms are used. Typically, “or” if used to associate a list, such as A, B, or C, is intended to mean A, B, and C, here used in the inclusive sense, as well as A, B, or C, here used in the exclusive sense. In addition, the term “one or more” as used herein may be used to describe any feature, structure, or characteristic in the singular or may be used to describe some combination of features, structures, or characteristics. However, it should be noted that this is merely an illustrative example and claimed subject matter is not limited to this example. Furthermore, the term “at least one of” if used to associate a list,such as A, B, or C, can be interpreted to mean any combination of A, B, and / or C, such as A, AB, AC, BC, AA, ABC, AAB, AABBCCC, and the like.

[0209] Further, while certain embodiments have been described using a particular combination of hardware and software, it should be recognized that other combinations of hardware and software are also possible. Certain embodiments may be implemented only in hardware, or only in software, or using combinations thereof. In one example, software may be implemented with a computer program product containing computer program code or instructions executable by one or more processors for performing any or all of the steps, operations, or processes described in this disclosure, where the computer program may be stored on a non-transitory computer readable medium. The various processes described herein can be implemented on the same processor or different processors in any combination.

[0210] For example, it will be appreciated that the disclosed methods and / or flow diagrams may be implemented as computer-executable instructions, in some embodiments. Thus, in one example, a computer-readable storage device (e.g., a non-transitory computer-readable medium) may store computer executable instructions that if executed by a machine (e.g., computer, processor) cause the machine to perform the disclosed methods and / or flow diagrams. While executable instructions associated with the disclosed methods and / or flow diagrams are described as being stored on a computer-readable storage device, it is to be appreciated that executable instructions associated with other example disclosed methods and / or block diagrams described or claimed herein may also be stored on a computer-readable storage device.

[0211] Where devices, systems, components or modules are described as being configured to perform certain operations or functions, such configuration can be accomplished, for example, by designing electronic circuits to perform the operation, by programming programmable electronic circuits (such as microprocessors) to perform the operation such as by executing computer instructions or code, or processors or cores programmed to execute code or instructions stored on a non-transitory memory medium, or any combination thereof. Processes can communicate using a variety of techniques, including, but not limited to, conventional techniques for inter-process communications, and different pairs of processes may use different techniques, or the same pair of processes may use different techniques at different times.

[0212] The disclosures of each and every publication cited herein are hereby incorporated herein by reference in their entirety.

[0213] While the disclosure has been described in detail with reference to exemplary embodiments, those skilled in the art will appreciate that various modifications and substitutions may be made thereto without departing from the spirit and scope of the disclosure as set forth in the appended claims. For example, elements and / or features of different exemplary embodiments may be combined with each other and / or substituted for each other within the scope of this disclosure and appended claims. .

Claims

CLAIMS What is claimed: A computer-implemented method comprising: generating an individualized region mask of a region of interest (ROI) of a brain region of an individual by converting a population region mask from standard space into individual space using the population region mask and magnetic resonance imaging (MRI) image data of the individual; generating an individualized reference ROI for the ROI using the MRI image data; and segmenting the ROI from within the individualized region mask using information derived from the individualized reference ROI. The method according to claim 1, wherein the ROI includes the locus coeruleus (LC), substantia nigra pars compacta (SNc), other neuromelanin contrast regions, iron contrast regions, neuromelanin / iron contrast region overlap ROI, among others, or any combination thereof. The method according to claim 2, wherein the ROI includes the LC and / or the SNc.The method according to claim 1, wherein: the segmenting includes segmenting the ROI from the quantitative contrast image datawithin the location of the individualized region mask to generate the segmented ROI and a segment-out region; and the segment-out region includes voxels within the individualized region mask that are excluded from the segmented ROI based on signal intensity. The method according to claim 4, wherein the ROI includes the SNc. The method according to claim 5, wherein the segmenting the SNc includes positioning the individualized region mask for SNc with respect to the quantitative contrast image data based on signal intensity of edge mask voxels within the individualized region mask.The method according to claim 6, wherein: the positioning the individualized mask for the SNc includes at least one iteration of: positioning the individualized region mask within the quantitative contrast image data at a first location; generating a first edge mask at the first location by selecting an outermost layer of voxels within the individualized region mask; determining a first edge mask measure for the first location based on a signal intensity of the first edge mask; generating a second edge mask by moving the first edge mask to a second location by shifting a position of at least one voxel within the quantitative contrast image data; determining a second edge mask measure for the second location based on a signal intensity of the second edge mask; and comparing the second edge mask measure to the first edge measure; and determining the location of the individualized region mask based on the comparing in at least one of the one or more iterations. The method according to claim 7, wherein the location of the individualized region maskcorresponds to the first location when the second edge mask measure is greater than the first edgemask measure. The method according to claim 8, wherein: more than one iteration is performed until iteration in which a second edge mask measure is greater than the first edge mask measure, the more than one iteration includes a first iteration and a second iteration; and in the second iteration, the second edge location and second edge mask measure of the first iteration corresponds to the first edge location and first edge mask measure of the second iteration. The method according to claim 1, wherein the MRI image data includes T1 image data and quantitative contrast image data. The method according to claim 10, wherein:the quantitative contrast image data includes neuromelanin (NM)-MRI image data and / or iron-sensitive MRI image data; and the population region mask is probabilistic and is larger than a representative representation of the ROI. The method according to claim 11, further comprising: registering the T1 image data and the quantitative contrast image data using partial and / or entire MRI image data in one or more registration iterations; and determining one or more transformation matrices based on the one or more registration iterations; wherein the individualized region mask is generated using the one or more transformation matrices. The method according to claim 12, wherein: the registering includes at least one iteration of: a first registering that includes registering the entire T1 image data to the entire and / or partial quantitative contrast image data in the individual space; and / or a second registering that includes registering the partial T1 image data to the partial quantitative contrast image data; and the one or more transformation matrices is generated based on the at least one iteration of the registering. The method according to claim 13, wherein the one or more transformation matrices is generated based on more than one iteration of the registering. The method according to 1, wherein the generating an individualized reference ROI using the MRI image data includes: positioning an initial individualized reference ROI within the MRI image data based on a location of the individualized region mask so that the initial individualized reference ROI is disposed adjacent to the individualized region mask, the initial individualized reference ROI including one or more predefined geometric shapes having a predefined size; andremoving outlier voxels from the initial individualized reference ROI based on signal intensity distribution of the initial individualized reference ROI resulting in the individualized reference ROI used to segment the ROI. The method according to claim 15, wherein the initial individualized reference ROI is defined by an overlap between the two or more of the more than one predefined geometric shape. The method according to claim 15, wherein the predefined size of the initial individualized reference ROI depends on a size of the individualized region mask. The method according to claim 15, wherein: the ROI includes LC; the initial individualized reference ROI for LC is generated in a reference slice; and the reference slice is based on a location of the individualized region mask for SNc and / or the individualized region mask for LC. The method according to claim 18, wherein the MRI image data includes one or more slices and generating the individualized reference ROI for LC includes: determining a reference slice of the MRI image data based on a location of an SNc region mask and / or an individualized LC region mask; determining a center of mass of the individual region mask for LC defined by region masks disposed at opposing sides; and generating the individualized reference ROI for LC in at least the reference slice based on the center of mass. The method according to claim 19, wherein the individualized reference ROI is further generated in each slice of the MRI image data in which the individualized region mask for LC is present. The method according to any of claims 18-20, wherein:if the individualized region mask for the SNc is present in a slice, the reference slice is inferior to the slice that includes the individualized SNc region mask; and if the one or more slices only includes the individualized LC region mask, the reference slice is superior to a slice that includes the individualized LC region mask. The method according to claim 2, wherein the population mask for each ROI is generated by: receiving representative NM-MRI image data for a plurality of representative individuals; defining an individualized representative region mask for the ROI in the representative NM-MRI image data for each representative individual so that the individualized representative region mask covers an area bigger than the ROI; and converting each individualized representative region mask into the standard space to generate a standard representative region mask; averaging the standard representative region masks for the plurality of different representative individuals; and determining the population region mask based on the averaging. The method according to claim 22, further comprising: generating a population reference ROI for the ROI using the population region mask. The method according to claim 22, wherein: the ROI includes the LC; and each individualized representative region mask for the LC is disposed at each center of mass of the LC of each respective representative NM-MRI Image data. A system, comprising: at one or more processors; and one or more hardware storage devices having stored thereon computer-executable instructions which are executable by the one or more processors to cause the computing system to perform at least the following:generating an individualized region mask of a region of interest (ROI) of a brain region of an individual by converting a population region mask from standard space into individual space using the population region mask and MRI image data of the individual; generating an individualized reference ROI using the MRI image data; and segmenting the ROI from within the individualized region mask using information derived from the individualized reference ROI. The system according to claim 25, wherein the ROI includes the locus coeruleus (LC), substantia nigra pars compacta (SNc), other neuromelanin contrast regions, iron contrast regions, neuromelanin / iron contrast region overlap ROI, among others, or any combination thereof. The system according to claim 26, wherein the ROI includes the LC and / or the SNc. The system according to claim 25, wherein: the segmenting includes segmenting the ROI from the quantitative contrast image data within the location of the individualized region mask to generate the segmented ROI and a segment-out region; and the segment-out region includes voxels within the individualized region mask that are excluded from the segmented ROI based on signal intensity. The system according to claim 26, wherein the ROI includes the SNc. The system according to claim 27, wherein the segmenting the SNc includes positioning the individualized region mask for SNc with respect to the quantitative contrast image data based on signal intensity of edge mask voxels within the individualized region mask. The system according to claim 30, wherein: the positioning the individualized mask for the SNc includes at least one iteration: positioning the individualized region mask within the quantitative contrast image data at a first location;generating a first edge mask at the first location by selecting an outermost layer of voxels within the individualized region mask; determining a first edge mask measure for the first location based on a signal intensity of the first edge mask; generating a second edge mask by moving the first edge mask to a second location by shifting a position of at least one voxel within the quantitative contrast image data; determining a second edge mask measure for the second location based on a signal intensity of the second edge mask; and comparing the second edge mask measure to the first edge measure; and determining the location of the individualized region mask based on the comparing in at least one of the one or more iterations. The system according to claim 31, wherein the location of the individualized region mask corresponds to the first location when the second edge mask measure is greater than the first edge mask measure. The system according to claim 32, wherein: more than one iteration is performed until iteration in which a second edge mask measure is greater than the first edge mask measure, the more than one iteration includes a first iteration and a second iteration; and in the second iteration, the second edge location and second edge mask measure of the first iteration corresponds to the first edge location and first edge mask measure of the second iteration. The system according to claim 25, wherein the MRI image data includes T1 image data and quantitative contrast image data. The system according to claim 34, wherein: the quantitative contrast image data includes neuromelanin (NM)-MRI image data and / or iron-sensitive MRI image data; and the population region mask is probabilistic and is larger than a representative representation of the ROI.The system according to claim 35, wherein the one or more processors are further configured to cause the computing system to perform at least the following: registering the T1 image data and the quantitative contrast image data using partial and / or entire MRI image data in one or more registration iterations; and determining one or more transformation matrices based on the one or more registration iterations; wherein the individualized region mask is generated using the one or more transformation matrices. The system according to claim 36, wherein: the registering includes at least one iteration of: a first registering that includes registering the entire T1 image data to the entire and / or partial quantitative contrast image data in the individual space; and / or a second registering that includes registering the partial T1 image data to the partial quantitative contrast image data; and the one or more transformation matrices is generated based on the at least one iteration of the registering. The system according to claim 37, wherein the one or more transformation matrices is generated based on more than one iteration of the registering. The system according to 38, wherein the generating an individualized reference ROI using the MRI image data includes: positioning an initial individualized reference ROI within the MRI image data based on a location of the individualized region mask so that the initial individualized reference ROI is disposed adjacent to the individualized region mask, the initial individualized reference ROI including one or more predefined geometric shapes having a predefined size; and removing outlier voxels from the initial individualized reference ROI based on signal intensity distribution of the initial individualized reference ROI resulting in the individualized reference ROI used to segment the ROI.The system according to claim 25, wherein the initial individualized reference ROI is defined by an overlap between the two or more of the more than one predefined geometric shape. The system according to claim 40, wherein the predefined size of the initial individualized reference ROI depends on a size of the individualized region mask. The system according to claim 40, wherein:the ROI includes LC; the initial individualized reference ROI for LC is generated in a reference slice; and the reference slice is based on a location of the individualized region mask for SNc and / or the individualized region mask for LC. The system according to claim 40, wherein the MRI image data includes one or more slices and generating the individualized reference ROI for LC includes: determining a reference slice of the MRI image data based on a location of an SNc region mask and / or an individualized LC region mask; determining a center of mass of the individual region mask for LC defined by region masks disposed at opposing sides; and generating the individualized reference ROI for LC in at least the reference slice based on the center of mass. The system according to claim 41, wherein the individualized reference ROI is further generated in each slice of the MRI image data in which the individualized region mask for LC is present. The system according to any of claims 42-44, wherein: if the individualized region mask for the SNc is present in a slice, the reference slice is inferior to the slice that includes the individualized SNc region mask; and if the one or more slices only includes the individualized LC region mask, the reference slice is superior to a slice that includes the individualized LC region mask.The system according to claim 26, wherein the population mask for each ROI is generated by: receiving representative NM-MRI image data for a plurality of representative individuals; defining an individualized representative region mask for the ROI in the representative NM-MRI image data for each representative individual so that the individualized representative region mask covers an area bigger than the ROI; converting each individualized representative region mask into the standard space to generate a standard representative region mask; averaging the standard representative region masks for the plurality of different representative individuals; and determining the population region mask based on the averaging. The system according to claim 46, wherein the one or more processors are further configured to cause the computing system to perform at least the following: generating a population reference ROI for the ROI using the population region mask. The system according to claim 47, wherein: the ROI includes the LC; and each individualized representative region mask for the LC is disposed at each center of mass of the LC of each respective representative NM-MRI Image data. A computer-implemented method comprising: generating an individualized region mask of a region of interest (ROI) of a brain region of an individual by converting a population region mask from standard space into individual space using the population region mask and magnetic resonance imaging (MRI) image data of the individual; generating an individualized reference ROI for the ROI using the MRI image data; and segmenting the ROI from within the individualized region mask using information derived from the individualized reference ROI.The method according to claim 49, wherein the ROI includes the locus coeruleus (LC), substantia nigra pars compacta (SNc), other neuromelanin contrast regions, iron contrast regions, neuromelanin / iron contrast region overlap ROI, among others, or any combination thereof. The method according to claims 49 or 50, wherein the ROI includes the LC and / or the SNc. The method according to any of claims 49-51, wherein: the segmenting includes segmenting the ROI from the quantitative contrast image data within the location of the individualized region mask to generate the segmented ROI and a segment-out region; and the segment-out region includes voxels within the individualized region mask that are excluded from the segmented ROI based on signal intensity. The method according to any of claims 49-52, wherein the ROI includes the SNc. The method according to claim 53, wherein the segmenting the SNc includes positioning the individualized region mask for SNc with respect to the quantitative contrast image data basedon signal intensity of edge mask voxels within the individualized region mask.The method according to claim 54, wherein:the positioning the individualized mask for the SNc includes at least one iteration: positioning the individualized region mask within the quantitative contrast image data at a first location; generating a first edge mask at the first location by selecting an outermost layer of voxels within the individualized region mask; determining a first edge mask measure for the first location based on a signal intensity of the first edge mask; and generating a second edge mask by moving the first edge mask to a second location by shifting a position of at least one voxel within the quantitative contrast image data; determining a second edge mask measure for the second location based on a signal intensity of the second edge mask; andcomparing the second edge mask measure to the first edge measure; and determining the location of the individualized region mask based on the comparing in at least one of the one or more iterations. The method according to claim 55, wherein the location of the individualized region mask corresponds to the first location when the second edge mask measure is greater than the first edge mask measure. The method according to claims 55 or 56, wherein: more than one iteration is performed until iteration in which a second edge mask measure is greater than the first edge mask measure, the more than one iteration includes a first iteration and a second iteration; and in the second iteration, the second edge location and second edge mask measure of the first iteration corresponds to the first edge location and first edge mask measure of the second iteration. The method according to any of claims 49-57, wherein the MRI image data includes T1 image data and quantitative contrast image data. The method according to claim 58, wherein: the quantitative contrast image data includes neuromelanin (NM)-MRI image data and / or iron-sensitive MRI image data; and the population region mask is probabilistic and is larger than a representative representation of the ROI. The method according to claims 58 or 59, further comprising: registering the T1 image data and the quantitative contrast image data using partial and / or entire MRI image data in one or more registration iterations; and determining one or more transformation matrices based on the one or more registration iterations; wherein the individualized region mask is generated using the one or more transformation matrices.The method according to any of claims 58-60, wherein: the registering includes at least one iteration of: a first registering that includes registering the entire T1 image data to the entire and / or partial quantitative contrast image data in the individual space; and / or a second registering that includes registering the partial T1 image data to the partial quantitative contrast image data; and the one or more transformation matrices is generated based on the at least one iteration of the registering. The method according to claim 61, wherein the one or more transformation matrices is generated based on more than one iteration of the registering. The method according to any of claims 49-62, wherein the generating an individualized reference ROI using the MRI image data includes: positioning an initial individualized reference ROI within the MRI image data based on a location of the individualized region mask so that the initial individualized reference ROI is disposed adjacent to the individualized region mask, the initial individualized reference ROI including one or more predefined geometric shapes having a predefined size; and removing outlier voxels from the initial individualized reference ROI based on signal intensity distribution of the initial individualized reference ROI resulting in the individualized reference ROI used to segment the ROI. The method according to claim 63, wherein the initial individualized reference ROI is defined by an overlap between the two or more of the more than one predefined geometric shape. The method according to claims 63 or 64, wherein the predefined size of the initial individualized reference ROI depends on a size of the individualized region mask. The method according to any of claims 49-65, wherein: the ROI includes LC;the initial individualized reference ROI for LC is generated in a reference slice; and the reference slice is based on a location of the individualized region mask for SNc and / or the individualized region mask for LC. The method according to claim 66, wherein the MRI image data includes one or more slices and generating the individualized reference ROI for LC includes: determining a reference slice of the MRI image data based on a location of an SNc region mask and / or an individualized LC region mask; determining a center of mass of the individual region mask for LC defined by region masks disposed at opposing sides; and generating the individualized reference ROI for LC in at least the reference slice based on the center of mass. The method according to claims 66 or 67, wherein the individualized reference ROI is further generated in each slice of the MRI image data in which the individualized region mask for LC is present. The method according to any of claims 66-68, wherein: if the individualized region mask for the SNc is present in a slice, the reference slice is inferior to the slice that includes the individualized SNc region mask; and if the one or more slices only includes the individualized LC region mask, the reference slice is superior to a slice that includes the individualized LC region mask. The method according to any of claims 49-69, wherein the population mask for each ROI is generated by: receiving representative NM-MRI image data for a plurality of representative individuals; defining an individualized representative region mask for the ROI in the representative NM-MRI image data for each representative individual so that the individualized representative region mask covers an area bigger than the ROI; and converting each individualized representative region mask into the standard space to generate a standard representative region mask;averaging the standard representative region masks for the plurality of different representative individuals; and determining the population region mask based on the averaging. The method according to any of claims 49-70, further comprising: generating a population reference ROI for the ROI using the population region mask. The method according to claim 71, wherein: the ROI includes the LC; and each individualized representative region mask for the LC is disposed at each center of mass of the LC of each respective representative NM-MRI Image data. A computer-program product comprising instructions configured to cause one or more data processors to perform the method of any of claims 1-24 or 49-72. A non-transitory machine-readable storage medium tangibly embodying a computer program product according to claim 73. A system comprising: one or more data processors; and one or more hardware storage devices having stored thereon computer-executable instructions which are executable by the one or more processors to cause the computing system to perform the method of any of claims 1-24 or 49-72.

Images