Methods and apparatuses for guiding treatment of back pain

Abstract

Methods and apparatuses for detecting and metrics described herein, e.g., disc budge (e.g., beaking), nuclear tail, nucleus boundary and nucleus disruption score may be highly predictive and may be used as part of a patient treatment. In particular, described herein are methods and apparatuses for identifying from one or more images of a patient's spine, one or more of: a disc budge score (e.g., beaking score), nuclear tail score, nucleus boundary score and nucleus disruption score, and using this score to propose, or modify a patient treatment.

Description

METHODS AND APPARATUSES FOR GUIDING TREATMENT OF BACK PAINCLAIM OF PRIORITY

[0001] This patent application claims priority to U.S. provisional patent application no. 63 / 651,947, filed May 24, 2024, titled “METHODS AND APPARATUSES FOR GUIDING TREATMENT OF BACK PAIN,” and herein incorporated by reference in its entirety.BACKGROUND

[0002] Back pain is a very common complaint for the middle-aged and the elderly. Structural back pain is often caused by degeneration of the intervertebral discs. Treatments for this condition start by examining the patient’s MRI. Physicians first have to identify the source of the pain before determining a course of treatment. This is complicated for the following reasons. First, the mere detection of degeneration in the MRI is not a good predictor for lower back pain since disc degeneration does not always cause pain. Second, patients may exhibit lower back pain along with a myriad of other symptoms which include areas like shoulder, limbs, buttocks, etc. These symptoms, while originating from the spine, are hard to localize at the disc level.

[0003] Currently, the physician examines the patients for symptoms in the context of the radiologist’s report and determines the disc / discs to be treated. The current standard for radiological analysis is the Modified Pfirrmann Score, which, in general terms, is a quantification of the loss of hydration in the discs. The basic idea is that, if a disc loses hydration, it loses its elasticity, ergo its ability to properly support the person’s weight and absorb the energy associated with movements. However, the Modified Pfirrmann Score has been shown to have a wide interobserver variance. This variability, combined with the difficulty in tracing pain symptoms directly to the disc leads to a large degree of subjectivity in the diagnosis and treatment of lower back pain.

[0004] What is needed are methods and apparatuses for identifying the likelihood of back pain developing or worsening, and in particular, methods and apparatuses for identifying a likely or actual cause of back pain and dysfunction.SUMMARY OF THE DISCLOSURE

[0005] Described herein are methods and apparatuses (e.g., devices, system, etc. including software) to compute a diagnostic metrics. These methods and apparatuses may optionally also include a large language model agent (e.g., an LLM-agent) driven pain questionnaire. In some examples the apparatuses described herein may analyze the input,patient’s MRI and (optionally) the patient’s responses questionnaire and may identify the source of actual or likely disc-related pain (e.g., the pain generator disc(s)) along with a confidence indicator of the prediction. Unlike the Modified Pfirrmann score, the computation of this measure does not have any subjectivity. While the user (e.g., surgeon, clinician, etc.) has the ability to tune the parameters for specific clinical purposes, the computation itself is not subjective.

[0006] In addition, the methods and apparatuses described herein may include a computational pipeline which combines computer vision and artificial intelligence algorithms and supports the ingestion of patient MRIs and corresponding outcome measures. The surgeon can use this system to create a standard of care by tuning the weights associated with metrics if they deem that the demographics of the population used to build the system are not compatible with the surgeon’s. In any of these methods and apparatuses, the methods and apparatuses may be configured to diagnose and / or treat a patient with for a spinal (e.g., disc) medical issue. In some cases these methods and apparatuses may be computational, rather than relying on machine learning.

[0007] The metrics, or the metrics and / or the pain questionnaire can be weighted to suit the surgeon’s population demographics and specific clinical problems. For instance, the system can be used to select target discs for non-surgical intervention. In some examples, the system can be trained using machine learning algorithms with appropriate training data. The training data could consist of patients’ MRI studies, pain description, and a label to indicate which, if any, disc was selected for the given non-surgical treatment. Alternatively, in some cases the apparatus (e.g., system) and method may instead use computational techniques (rather than machine learning techniques). The system may compute the one or more novel metrics for the discs and learn the decision making process. The system can be used to replicate the decision making process for unseen MRI. The training would need to be redone if the expert’s decision making process changes for any reason or if the characteristics of the population in the original training dataset differs from the characteristics of the current set.

[0008] In some examples, the system may permit the use of traditional measures of disc pathology; any combination of traditional and novel metrics could be used in the pipeline. In some cases the system may, however, enforce that the set of algorithms used in training the machine learning model be the same as those used in operational use.

[0009] These methods and apparatuses for performing them may include accessing or receiving one or more images of a patient’s spinal anatomy; estimating, in a processor one or more of: a beaking score, a nuclear tail score, a nucleus boundary score, or a nucleus disruption score; generating, using the one or more of: the beaking score, the nuclear tailscore, the nucleus boundary score, or the nucleus disruption score, a pathology output; and outputting the pathology output.

[0010] For example, described herein are methods and apparatuses (e.g., systems, devices, including software, hardware and / or firmware) for identifying a pain generator disc. For example, a method may include: receiving a scan volume data including one or more spinal discs; processing the scan volume data to generate a plurality of virtual slices through the scan volume data, comprising virtual axial slices and virtual sagittal slices; segmenting the plurality of virtual slices to identify the one or more spinal discs; detecting beaking in any of the one or more spinal discs by estimating a beaking score for each disc of the one or more spinal discs and generating a first subset of the plurality of virtual slices in which the beaking score corresponds to a likelihood of herniation of the disc greater than a beaking threshold; detecting a nuclear tail in any of the one or more spinal discs by estimating a nuclear tail score for each disc of the one or more spinal discs and generating a second subset of the plurality of virtual slices in which the nuclear tail score corresponds to a likelihood of nuclear degeneration of the disc greater than a nuclear tail threshold; detecting an internal disruption in any of the one or more spinal discs by estimating an internal disruption score for each disc of the one or more spinal discs and generating a third subset of the plurality of virtual slices in which the internal disruption score corresponds to a likelihood of degeneration of the disc greater than an internal disruption threshold; modifying the one or more virtual slices of the first subset of the plurality of virtual slices to indicate a beaking including a representation of the corresponding beaking score; modifying the one or more virtual slices of the second subset of the plurality of virtual slices to indicate a nuclear tail region including a representation of the corresponding a nuclear tail score; modifying the one or more virtual slices of the third subset of the plurality of virtual slices to indicate internal disruption in a region of a disc of the one or more virtual slices including a representation of the corresponding internal disruption score for each region that has an internal disruption greater than an internal disruption threshold; and outputting, on a user interface, the modified one or more virtual slices of the first subset, the modified one or more virtual slices of the second subset and the modified one or more virtual slices of the third subset.

[0011] In any of these methods and apparatuses, estimating a beaking score for each disc of the one or more spinal discs may comprise estimating a change in a slope around a circumference of each disc. For example, a change in slope around the circumference of the disc may be determined as a radial difference between two consecutive spans extending between a centroid of the disc and a perimeter of the disc, wherein consecutive spans are separated by between about 2 and 30 degrees. In some cases estimating a beaking score foreach disc of the one or more spinal discs comprises estimating a deviation of the slope from a smooth contour fitted over each disc from a segmented contour of the disc.

[0012] In any of these methods and apparatuses, estimating a nuclear tail score for each disc of the one or more spinal discs may comprise determining a width of a region of a nucleus of the disc having a height that is greater than a minimum threshold height of the nucleus when height is measured perpendicular to a long axis of the spinal disc. Any of these methods may include comprising normalizing the plurality of virtual slices to align and / or resize the virtual slices. Any of these methods may include segmenting the plurality of virtual slices to identify a nucleus in each of the spinal discs.

[0013] In any of the methods and apparatuses described herein estimating the internal disruption score for each disc of the one or more spinal discs may comprise further segmenting each virtual slice of the plurality of virtual slices to determine a boundary between an annulus and a nucleus, dividing each virtual slice in a plurality of regions comprising nucleus region and annulus regions and estimating disruption rations from an amount of nucleus outside of a nucleus boundary and / or an amount of annulus region within the nucleus boundary for each region of the plurality of regions. In any of these methods, dividing each virtual slice comprises dividing into 12 or more radial regions. For example, estimating the amount of nucleus outside of the nucleus boundary and / or the amount of annulus region within the nucleus boundary may comprise estimating a number of nucleus pixels outside of the boundary and the / or a number of annulus pixels inside the boundary. Determining the boundary may comprise receiving user input to refine the boundary.

[0014] In any of these methods and apparatuses, estimating the internal disruption score for each disc of the one or more spinal discs may comprise generating a digital twin for each disc of the one or more spinal discs and comparing each disc of the one or more spinal discs to its corresponding digital twin. Generating the digital twin may comprise algorithmically determining the digital twin based on an elliptical curve using an axial view of the disc. Any of these methods and apparatuses may include receiving one or more user inputs of the user interface to switch adjust one or more of: the beaking threshold, the nuclear tail threshold, and / or the internal disruption threshold.

[0015] In any of these methods and apparatuses, outputting may comprise outputting an anatomical identifier for each disc of the modified one or more virtual slices of the first subset, the modified one or more virtual slices of the second subset and the modified one or more virtual slices of the third subset. In any of these methods and apparatuses, processing the scan volume data to generate the plurality of virtual slices through the scan volume data may comprise identifying a representative slice for each disc of the one or more spinal discsand including only the representative slices in the plurality of virtual slices. Processing the scan volume data to generate the plurality of virtual slices may comprise one or more of: resizing the one or more virtual slices, histogram equalizing the one or more virtual slices, reorienting the one or more virtual slices, and / or cropping the one or more virtual slices.

[0016] In any of these methods and apparatuses segmenting may comprise generating a mask of the spinal disc for each of the one or more virtual slices. Any of these methods and apparatuses may include aligning the discs from the segmented virtual slices to a standard alignment. Any of these methods and apparatuses may include outputting one or more candidates for the pain generator disc based on the modified one or more virtual slices of the first subset, the modified one or more virtual slices of the second subset and the modified one or more virtual slices of the third subset.

[0017] Also described herein are apparatuses for performing any of these methods. For example, an apparatus may include: a processor; a memory coupled to the one or more processors, the memory storing computer-program instructions, that, when executed by the one or more processors, perform a computer-implemented method comprising: receiving a scan volume data including one or more spinal discs; processing the scan volume data to generate a plurality of virtual slices through the scan volume data, comprising virtual axial slices and virtual sagittal slices; segmenting the plurality of virtual slices to identify the one or more spinal discs; detecting beaking in any of the one or more spinal discs by estimating a beaking score for each disc of the one or more spinal discs and generating a first subset of the plurality of virtual slices in which the beaking score corresponds to a likelihood of herniation of the disc greater than a beaking threshold; detecting a nuclear tail in any of the one or more spinal discs by estimating a nuclear tail score for each disc of the one or more spinal discs and generating a second subset of the plurality of virtual slices in which the nuclear tail score corresponds to a likelihood of nuclear degeneration of the disc greater than a nuclear tail threshold; detecting an internal disruption in any of the one or more spinal discs by estimating an internal disruption score for each disc of the one or more spinal discs and generating a third subset of the plurality of virtual slices in which the internal disruption score corresponds to a likelihood of degeneration of the disc greater than an internal disruption threshold; modifying the one or more virtual slices of the first subset of the plurality of virtual slices to indicate a beaking including a representation of the corresponding beaking score; modifying the one or more virtual slices of the second subset of the plurality of virtual slices to indicate a nuclear tail region including a representation of the corresponding a nuclear tail score; modifying the one or more virtual slices of the third subset of the plurality of virtual slices to indicate internal disruption in a region of a disc of the one or more virtual slicesincluding a representation of the corresponding internal disruption score for each region that has an internal disruption greater than an internal disruption threshold; and outputting, on a user interface, the modified one or more virtual slices of the first subset, the modified one or more virtual slices of the second subset and the modified one or more virtual slices of the third subset.

[0018] The method and apparatuses described herein may be configured to detect, in any of the one or more spinal discs, one or more of: breaking, a nuclear tail, and / or an internal disruption. For example, a method of identifying a pain generator disc may include: receiving a scan volume data including one or more spinal discs; processing the scan volume data to generate a plurality of virtual slices through the scan volume data, comprising virtual axial slices and virtual sagittal slices; segmenting the plurality of virtual slices to identify the one or more spinal discs; detecting, in any of the one or more spinal discs, one or more of: breaking, a nuclear tail, and / or an internal disruption, wherein: detecting beaking comprises estimating a beaking score for each disc of the one or more spinal discs and generating a first subset of the plurality of virtual slices in which the beaking score corresponds to a likelihood of herniation of the disc greater than a beaking threshold, detecting a nuclear tail comprises estimating a nuclear tail score for each disc of the one or more spinal discs and generating a second subset of the plurality of virtual slices in which the nuclear tail score corresponds to a likelihood of nuclear degeneration of the disc greater than a nuclear tail threshold, and detecting an internal disruption comprises estimating an internal disruption score for each disc of the one or more spinal discs and generating a third subset of the plurality of virtual slices in which the internal disruption score corresponds to a likelihood of degeneration of the disc greater than an internal disruption threshold; modifying the one or more virtual slices of the first subset of the plurality of virtual slices to indicate a beaking including a representation of the corresponding beaking score; modifying the one or more virtual slices of the second subset of the plurality of virtual slices to indicate a nuclear tail region including a representation of the corresponding a nuclear tail score; modifying the one or more virtual slices of the third subset of the plurality of virtual slices to indicate internal disruption in a region of a disc of the one or more virtual slices including a representation of the corresponding internal disruption score for each region that has an internal disruption greater than an internal disruption threshold; and outputting, on a user interface, the modified one or more virtual slices of the first subset, the modified one or more virtual slices of the second subset, and / or the modified one or more virtual slices of the third subset.

[0019] Similarly, an apparatus may include: a processor; a memory coupled to the one or more processors, the memory storing computer-program instructions, that, when executed bythe one or more processors, perform a computer-implemented method comprising: receiving a scan volume data including one or more spinal discs; processing the scan volume data to generate a plurality of virtual slices through the scan volume data, comprising virtual axial slices and virtual sagittal slices; segmenting the plurality of virtual slices to identify the one or more spinal discs; detecting, in any of the one or more spinal discs, one or more of: breaking, a nuclear tail, and / or an internal disruption, wherein: detecting beaking comprises estimating a beaking score for each disc of the one or more spinal discs and generating a first subset of the plurality of virtual slices in which the beaking score corresponds to a likelihood of herniation of the disc greater than a beaking threshold, detecting a nuclear tail comprises estimating a nuclear tail score for each disc of the one or more spinal discs and generating a second subset of the plurality of virtual slices in which the nuclear tail score corresponds to a likelihood of nuclear degeneration of the disc greater than a nuclear tail threshold, and detecting an internal disruption comprises estimating an internal disruption score for each disc of the one or more spinal discs and generating a third subset of the plurality of virtual slices in which the internal disruption score corresponds to a likelihood of degeneration of the disc greater than an internal disruption threshold; modifying the one or more virtual slices of the first subset of the plurality of virtual slices to indicate a beaking including a representation of the corresponding beaking score; modifying the one or more virtual slices of the second subset of the plurality of virtual slices to indicate a nuclear tail region including a representation of the corresponding a nuclear tail score; modifying the one or more virtual slices of the third subset of the plurality of virtual slices to indicate internal disruption in a region of a disc of the one or more virtual slices including a representation of the corresponding internal disruption score for each region that has an internal disruption greater than an internal disruption threshold; and outputting, on a user interface, the modified one or more virtual slices of the first subset, the modified one or more virtual slices of the second subset, and / or the modified one or more virtual slices of the third subset.

[0020] A processor may include hardware that runs the computer program code. Specifically, the term ‘processor’ may include a controller and may encompass not only computers having different architectures such as single / multi-processor architectures and sequential (Von Neumann) / parallel architectures but also specialized circuits such as field- programmable gate arrays (FPGA), application specific circuits (ASIC), signal processing devices and other devices.

[0021] The difficulty in interpreting scan data, including MRI and CT scan data, to identify spinal discs that are causing pain to a patient and / or that are likely to lead to further, potentially more serious problems for the patient in the future is well known, and present aparticularly challenging technical problem. It is currently difficult, or impossible, to rapid and effectively (with greater than 90% effectiveness) identify which spinal disk is causing (or is likely to cause) pain, particularly as many of the spinal defects may be subtle, and difficult to detect. In addition, currently available techniques for reviewing such scan data are not able to accurately determine boundaries (regions) that may be causing the pain or that may lead to pain. These problems are technical in nature; the ability to analyze scan data is limited by the inability to understand how to analyze scan data and apply rigorous analysis (free of bias and interobserver bias). The methods and apparatuses described herein solve these technical problems by processing (e.g., pre-processing) scan data to systematically and rigorously estimate technical scores for at least beaking (providing a likelihood of herniation of the disc greater than a beaking threshold), nuclear tail (e.g., corresponds to a likelihood of nuclear degeneration of the disc greater than a nuclear tail threshold) and / or internal disruption in any of the one or more spinal discs (e.g., corresponding to a likelihood of degeneration of the disc greater than an internal disruption threshold).

[0022] All of the methods and apparatuses described herein, in any combination, are herein contemplated and can be used to achieve the benefits as described herein.BRIEF DESCRIPTION OF THE DRAWINGS

[0023] A better understanding of the features and advantages of the methods and apparatuses described herein will be obtained by reference to the following detailed description that sets forth illustrative embodiments, and the accompanying drawings of which:

[0024] FIGS. 1 A-1B illustrate an example of estimating a beaking score from a disc mask (FIG. 1 A) and the average slope (FIG. IB) at different orientations around the disc for a healthy disc.

[0025] FIGS. 2A-2B illustrate an example of estimating a beaking score from a disc mask (FIG. 2A) and the average slope (FIG. 2B) at differ orientations for a disc having acute beaking.

[0026] FIG. 3 schematically illustrates an example of a method of detecting and / or quantifying nuclear tail.

[0027] FIG. 4 shows an example in which a nuclear tail was detected as described herein.

[0028] FIGS. 5A-5B illustrate examples of detection of the nucleus boundary in a healthy (FIG. 5 A) and an unhealthy (FIG. 5B) disc.

[0029] FIG. 5C illustrate an example of estimating properties of a disc from a scan of disc by dividing the scan into scan / disc into multiple sections.

[0030] FIGS. 6A and 6B illustrate one example of a method of identifying and refining the contour of a nucleus in a disc.

[0031] FIG. 7 illustrates an image showing the projection of the active contour output onto a disc.

[0032] FIG. 8 illustrates an example of an estimate of a digital twin from one example of an image of a disc.

[0033] FIG. 9 illustrates one example of a method for accurately determining a digital twin of a disc.

[0034] FIGS. 10A-10B show an example of the change in the extreme points of the sagittal boundary in two slices before (FIG. 10) and after (FIG. 10B) axial correction.

[0035] FIGS. 11 A-l IB illustrate the impact of the adjusted end points on the nucleus boundary of an image (slice) through the disc.

[0036] FIGS. 12A-12B illustrate the impact of the adjusted end points on the nucleus boundary of an image (slice) through the disc.

[0037] FIG. 13 illustrates an example of an area used for computing a disruption score of a disc (e.g., of a slice though a disc).

[0038] FIG. 14 illustrates an example of an area used for computing a disruption score of a slice through a disc.

[0039] FIGS. 15A-15B illustrate one example of a sagittal MRI slice alone (FIG. 15 A) or with an output mask from 3 class Unet segmentation (FIG. 15B).

[0040] FIG. 16A shows an example of a region of interest (ROI) for extraction of a disc. FIG. 16B shows an example of the segmented disc of FIG. 16A.

[0041] FIG. 17A shows an example of an extracted, rotated and re-sized disc image. FIG. 17B shows internal segmentation of the disc of FIG. 17A.

[0042] FIG. 18 A illustrates an example of an original axial MRSI slice. FIG. 18B shows an output of an axial disc segmentation model as described herein.

[0043] FIG. 19A shows an example of an original axial MRI slice. FIG. 19B shows the output of an axial spinal cord segmentation model, as described.

[0044] FIG. 20A shows an original axial MRI slice. FIG. 20B shows the output of axial posterior element segmentation model.DETAILED DESCRIPTION

[0045] Any of the methods and apparatuses described herein may be configured to quantify stenosis, based at least in part on one or more images of the patient’s spine. The images may be taken as part of these methods, or may be received, accessed, and / orotherwise used by the methods and apparatuses described herein. Images may include any appropriate imaging modality, such as CT scan images, MRI images, ultrasound images, X- ray images, etc. Any of these methods and apparatuses may be included or incorporated as part of the imaging system and / or methods, or any of the methods and / or system may access or otherwise communicate with an imaging system, a repository of patient images, and / or a patient medical record.

[0046] A healthy intervertebral disc consists of a thick outer layer called the annulus and a softer, gel-like interior called the nucleus. When a disc loses its structural integrity, the annulus is compromised, in a non-uniform manner, and the nucleus is no longer properly contained. The disc also exhibits distortion, the weaker portions of the annulus distend and could push against nerves or in some cases, they push against the spinal canal. This pathology is called stenosis. Specifically, it is called foraminal stenosis when the disc distends and pushes against the foraminal nerves, i.e., the point where the nerves exit the spinal region and extend into the body. It is called central canal stenosis when it pushes against the spinal cord.

[0047] Severe cases of stenosis are easy to identify, and the resultant pain is typical of nerve compression. Most cases of stenosis are often mild and even experts will find it difficult to correlate it to pain symptoms since there is no easy way to grade or quantify milder cases of stenosis.

[0048] When the annulus is compromised in this manner, the weak portions continue to give way, progressively weakening over time since the weight of the body is constantly exerting pressure on the disc. Quantification of the pathology, especially when the stenosis is mild, could be beneficial in designing early interventions.

[0049] As used herein, the term ‘beak’ may refer to a mild case of herniation. A disc beak can be defined as an acute posterior midline intervertebral disc protrusion that may push the thecal sac in the midline region resulting in central canal stenosis. The method or apparatus (e.g., system) may look for the presence of disc beaking in the T2W axial slices. This is clinically important, while predicting the early stages of disc protrusion with minimal annulus fibrosus disruption. Measurement of disc beaking quantifies how acutely the annulus fibrosus has protruded into the dural sac with a relatively intact posterior longitudinal ligament.

[0050] When the annulus in the posterior portion of the disc is compromised, the distention of the disc causes pressure against the spinal canal. In extreme cases, this pathology is easy to discern. However, damage to the annulus could take the form of a vertical tear in the posterior portion of the disc or a horizontal tear where the nucleus material appears to be in the process of being extruded through a tear. The horizontal tear, whensubtle, is not discerned as stenosis since the annulus would not have yet distended to be clearly classified as a central canal stenosis.

[0051] Again, the detection and quantification of the damage to the posterior portion of the annulus, especially when central canal stenosis is not detected, could form the basis for early intervention.

[0052] As used herein, the term “nuclear tail” may refer to a narrow extension of the centrally located nucleus pulposus towards its posterior end. Unlike a healthy nucleus, the fragments of degenerated nucleus pulposus migrate through the pre-existing fissures in the posterior part of the annulus fibrosus to form a tail. The presence of the tail is an indication of the presence of fissures, ergo indicative of a weakening / injury to the posterior portion of the annulus. The presence of a nuclear tail clinically signifies the early stage of nuclear degeneration and its protrusion through the pre-existing posterior annular fissures

[0053] The methods and apparatuses described herein may detect and / or quantify disruption of the internal structure of the vertebral disc based on one or more images.

[0054] Sometimes, weakening of the intervertebral discs does not result in a stenosis-like appearance. Examination of the disc through an MRI study would show that the disc no longer has a well-defined nuclear and annular separation. A healthy disc should exhibit a bright and mostly smooth nucleus and a darker but smooth annulus. This is typically apparent in all but the most lateral of the slices in the sagittal series of the MRI. In a disc which is degraded, the clear separation is lost. The brighter nucleus material is seen to enter the annular regions and the nucleus material shows darker regions, reflecting localized loss of hydration.

[0055] Annular tears and loss of disc height appear in the early stages of disc degeneration and can cause the centrally lying nucleus pulposus to dissipate peripherally into the annulus fibrosus. The disruption can irritate the surrounding nerves resulting in pain.With further degeneration, microscopic calcifications can occur in the cartilaginous endplates causing damage. Therefore, the nucleus pulposus extending to the annulus fibrosus or vice versa in a disc indicates disruption and hence its degeneration.

[0056] In general, the methods and apparatuses described herein may include, in some cases, methods and apparatuses to quantify the presence and amount of disc bulge, methods and apparatuses to quantify the pathology in the posterior annular complex, and / or methods and apparatuses to quantify the extent of disruption of the internal structure of the intervertebral discs.

[0057] For example, a method or apparatuses for detecting beaking may include segmentation (e.g., semantic segmentation) algorithms that may be applied on the axial MRIslices. In some examples, one or more artificial intelligence algorithms can be used to outline the disc in axial MRI. These methods and apparatuses may use, e.g., a UNet-based neural network. The output segmentation mask may be used for computation, e.g., of the centroid of the disc. The centroid of the disc, e.g., the segmentation mask, may be computed. Since a typical axial series has more than one slice going through a given intervertebral disc (or level), the mask corresponding to the most medial of the slices may be selected for this computation. The methods and apparatuses may also estimate a distance, measured along a radial line from the centroid, and angle, measured against a horizontal, 3 o’clock position, of the boundary points of the mask from the centroid are computed. In any of these examples, the mask may be divided into sectors (e.g., of 5 degrees, 8 degrees, 10 degrees, 12 degrees, 15 degrees, 18 degrees, 20 degrees, 22 degrees, 25 degrees, etc. or more) and the average distance may be computed for each span of angle between the centroid and boundary points.

[0058] Any of these methods and apparatuses may estimate a slope of the average distances is computed either sequentially or using an overlapping of angles. For example: assuming that the sectors have a 15 degrees span, the sequential computation will sum the distances between 0 and 15 degrees for the first span, while the overlapping summation would span from 352.5 ( - 7.5) degrees to 7.5 degrees. Both methods are valid

[0059] As used herein, the slope may refer to the simple difference between the average radial distance of two consecutive spans. The methods and apparatuses described herein may plot the slopes (e.g. may compute and / or graph them). The plot of the slope over the perimeter of the disc may be used to quantify the beaking in the disc. For example, a healthy discs exhibits a smooth slope curve. A disc is seen to have a sharp beak in the foraminal and central canal region if there is a sign reversal of the slopes in the posterior side between 60- 120 degrees. Subtler beaking can be detected by detecting the deviation of the actual slopes from a smooth contour which could be fitted over the same set of points.

[0060] An example of the output for nearly healthy and beaked discs are respectively shown in FIGS. 1 A-1B. FIGS. 1 A-1B illustrate a disc mask (FIG. 1 A) and the average slope (FIG. IB) at different orientations around the disc make for a healthy (e.g., relatively healthy) disc. In contrast, FIGS. 2A-2B illustrate an example of a mask (FIG. 2A) and the average slope (FIG. 2B) at differ orientations for a disc having acute beaking. The slope plot of FIG. 2B shows a disc exhibiting pronounced beaking in the central canal region, as seen by the acute slope reversal.

[0061] Any of these methods and apparatuses may also estimate nuclear tail from the one or more images. For example a system may detect the presence of a nuclear tail in T2W sagittal sections. The method of detecting disc beaking may therefore include the applicationof successive semantic segmentation algorithms on the sagittal MRI slices. One or more of a variety of different artificial intelligence algorithms may be used to outline the disc in axial MRI image(s). In one example a UNet based neural network may be used. The first pass may isolate the disc from the main MRI. Computer vision algorithms may be applied to refine the outline of the disc and create a discernible edge where the intervertebral disc meets the vertebrae. The second pass of segmentation (e.g., semantic segmentation) may isolate the nucleus from the annulus in the mask of the disc isolated in the first pass. The output segmentation mask of the nucleus may be used for the computation of the centroid of the disc (e.g., the segmentation mask), and / or the nucleus region, which may be divided into a number of non-overlapping vertical patches of a few pixel widths starting from the posterior side till the centroid of the nucleus mask. The height of each patch, hpatch and the height of the center patch (the vertical patch going through the centroid), hcenterpatch may be computed. After that, these method and apparatuses may compute the ratio, Ri between the height hpatch of ithpatch from the posterior side with the height of the central patch hceterPatch, i.e. Ri = hpatch /

[0062] Any of these methods may calculate the number of patches, Pnumber, for which the height ratios, RiS are below a value, thr, which can be set by the user (e.g., doctor, technician, clinician, medical expert, etc.). For example, 1(.) may be an indicator function which returns 1 whenever its argument function is true, else it returns 0.

[0063] If the value of Pnumber is more than a threshold value, which may (again) be set by the user, a Nuclear Tail is considered to be present in the disc image. An example of this technique is shown in FIG. 3. FIG. 3 schematically illustrates sliding the nucleus as described above. An applied example is shown in FIG. 4, showing an example in which a nuclear tail was detected. In FIG. 4, the highlighted portion indicates the region where the threshold was exceeded.Internal Disruption / Nucleus-Annulus Ratio

[0064] Any of the methods and apparatuses described herein may also or alternatively detect and quantify internal disruption. For example, any of these methods may detect and compute the distortion of the nucleus pulposus and the annulus fibrosus for each disc slice in the T2W sagittal Images. This metric (internal disruption) may also be used to signify the health of the endplate. Successive segmentation (e.g., semantic segmentation) algorithms may be applied on the sagittal MRI slices. As mentioned above, in some cases any one of severalartificial intelligence algorithms can be used to outline the disc in axial MRI; for example, a UNet-based neural network may be used, as shown herein. In some cases the first pass may isolate the disc from the main MRI. Alternatively or additionally, computer vision algorithms may be applied to refine the outline of the disc and create a discernible edge where the intervertebral disc meets the vertebrae. In some examples a second pass may be used in which a segmentation algorithm may isolate the nucleus material from the annulus and end plates, e.g., using a UNet segmentation model. The output segmentation mask of the nucleus may be used, e.g., for computation. For example, the apparatus may estimate the shape and location of the nucleus boundary for each slice of the disc based on the internal nucleus material. In some examples, the nucleus boundary may be estimated.

[0065] Based on the nucleus boundary, two primary regions within the disc may be distinguished: the interior nucleus region and the exterior region close to the disc endplate. For example, the nucleus boundary may be estimated; a disc region outside of the boundary may be considered to be the ideal healthy annulus and endplate of the disc. A disruption score may be estimated that measures the amount of nucleus outside this boundary and annulus inside the boundary. For example, the disruption score may be estimated by computing: (1) the number of annulus pixels inside of the boundary divided by the total number of pixels; and / or (2) the number of nucleus pixels outside of the boundary divided by the total number of pixels. A perfectly healthy disc should have a disruption of zero, indicating that all nucleus and annulus materials are in their respective places and that the endplate is intact.

[0066] In some examples, the disc may be divided into a plurality of sections (e.g., four quadrants, five sections, six sections, 7 sections, 8 sections, 9 sections, 10 sections, 11 sections, 12 sections, 13 sections, 14 sections, 15 sections, 16 sections, four or more sections, five or more sections, six or more sections, 8 or more sections, 10 or more sections, 12 or more sections, etc.), thereby permitting the measuring of the disruptions in various regions (e.g., posterior inferior, posterior superior, anterior inferior, and anterior superior regions) of the disc. In some examples at least two ratios are computed for each of these quadrants: (1) number of annulus pixels inside the boundary in a quadrant divided by the total number of pixels in the quadrant; and (2) number of nucleus pixels outside the boundary in a quadrant divided by total number of pixels in the quadrant.

[0067] Any of these methods and apparatuses may also include an estimate of the nucleus boundary. For example, FIGS. 5A-5B illustrate examples of detection of the nucleus boundary in a healthy (FIG. 5A) and an unhealthy (FIG. 5B) disc. The nucleus boundary may be computed for each medial slice of a disc.

[0068] In some cases the disc may be divided into six or more sections. For example,FIG. 5C shows an example of an image of a disc that has been divided into 16 sections (Al- A12 and N1-N4). Dividing the disc into smaller regions may enhance the analysis and ensure that each region of the disc can reflect its unique characteristics. In general, any of the methods described herein may include separating the disc into regions as describe below. For example, for each rectangular disc region of interest (ROI), the centroid may be computed from the disc mask image moments, which are statistical properties that include spatial distribution. The centroid (ex, cy) may be estimated or computed. For example, the centroid may be computed as:

[0069] In some examples, the moments may be calculated from the pixel intensities of the image. For example, pixel intensities may be estimated by calculating or approximating:

[0070] Mpq= Y.x y xpyqI(x, y), / (x, y) is pixel intensity at position (x,y)

[0071] Where Moois the spatial moment (area), M01and M10are the first order moments.In FIG. 5C, the Nl, N2, N3, and N4 are the inside boundary regions of the four quadrants formed by x and y axes through this centroid as shown in the FIG. 6A. In this example, the area outside the nucleus boundary is divided into 12 additional regions by further separating each quadrant into three parts. This is achieved by rotating the x-axis at increments of angles 15, 30, 45, 45, 30, 15, 15, 30, 45, 45, 30, and 15 degrees; these increments of angles may be regular (e.g., in some cases, by dividing the number of regions, e.g., twelve, by 360 degrees), or irregular. In FIG. 5C, the regions outside the nucleus boundary are labeled Al to A12, as illustrated. In general, an image of a disk may be divided up in a plurality of outside regions (Al -An, where / ? = 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, etc.) and a plurality of inside regions (Nl-Nn, wherein n = 2, 3, 4, 5, 6, 7, 8, etc.).

[0072] Thus, in some cases the internal disruption ratios may be computed at (or for) each of these regions (in this example, regions A1-A12 and / or N1-N4), for example, for the inside regions by calculating the number of annulus pixels inside the boundary in a region divided by the total number of pixels in the region; and for the number of outside regions by determining the number of nucleus pixels outside the boundary divided by total number of pixels in the region. This division to smaller regions allows the user to vary the weighs given to disruption in different areas of the disc. For example, a user may give more importance on the posterior side of the disc for specific pain symptoms and assign a higher weightage to the disruption on the regions A6, A7, A8, A9, A10, and Al 1 for their interpretation.

[0073] A nucleus mask may be used through the application of semantic segmentation models described above. The boundary may be estimated using the active contour or snakestechnique which moves an initial shape defined by a set of contour points to the lowest energy point using an iterative method. The active contour algorithm is defined such that it minimizes three energies: continuity, curvature and gradient of the image. In some examples, the initial shape may be the corresponding disc contour shrunk by a small amount in all directions. The initial contour is shown in green in FIG. 6A. An initial estimate may be determined by the application of the active contour to move the initial shape close to the edges of the nucleus in the disc, shown in blue color in FIG. 6B.

[0074] The output of the active contour algorithm may provide a projection of the active contour onto the disc. For example, FIG. 7 illustrates an image showing the projection of the active contour output onto a disc. A user (e.g., surgeon) can validate the contour created by the algorithm at this point. In FIG. 7, the disc image shows the nucleus boundary.

[0075] An axial view may be used for refining the nucleus boundary. In some cases an active contour method may result in an overly aggressive expansion of the boundary for unhealthy disc with annular tear, thereby underestimating the pathology. The basic idea is that, since so much of the nucleus material has escaped into the annular space, the active contour algorithm may mistakenly create a large contour, consequently, underestimating the amount of nucleus outside of its expected location. In general the methods and apparatuses described herein may use a digital twin for every disc analyzed. This digital twin is assumed to have a healthy and intact nucleus but otherwise sharing the original’s size and position. Once the shape and size of the digital twin’s nucleus boundary is computed, it may be projected onto the disc with the pathology and used to compute the disruption score.

[0076] In FIG. 8, it is initially assumed that the twin’s nucleus boundary would be obtained from computation described in the preceding section. The active contour algorithm may yield the blue boundary as shown below. This is an incorrect estimation of the twin’s nucleus since a healthy nucleus would not extend so far into the posterior of the disc. To solve this problem, the axial view of the disc may be used. The basic idea is that an elliptical curve may first be constructed in the axial view of the disc. The user may adjust the size and position of the ellipse. Once accepted as final, the position of sagittal slices is mapped onto the axial view. Then, the extreme points, the most anterior and posterior points, of the nucleus boundary may be projected onto the axial view. The position of these extreme points may be evaluated against the elliptical curve. The following computation estimates where the digital twin’s nucleus would be. First, the extreme points of the nucleus boundary at the posterior and anterior sides may be projected to the best, i.e., the most medial, axial disc slice. These methods may assume that the extreme points of a healthy disc, i.e., the digital twin’s disc, would lie along an elliptical curve in the axial plane. The elliptical curve may becomputed by fitting an elliptical curve to the segmentation mask of the disc from the axial view. This elliptical curve may be shown in gray in the FIG. 9.

[0077] FIG. 9 shows the projection of the original extreme points from the initial estimate of the nucleus boundary in red. Each red point is adjusted to lay on the elliptical curve and the adjusted points are shown in green.Example of the adjustment of end points on an L5-S1 disc.

[0078] The modified points, e.g., the green points in FIG. 9, may then be projected back to the sagittal plane. FIGS. 10A-10B show the change in the extreme points of the sagittal boundary in the slices 9 and 10 after axial correction. FIG. 10 A shows images of slices 9 and 10 before axial correction, and FIG. 10B shows the same slices after axial correction as described above.

[0079] These methods may also include projection of the end points of the sagittal view, as shown in FIGS. 11A-1 IB for slice 9 (disc: L5-S1) and in FIGS. 12A-12B for slice 10 (disc: L51-S1). FIGS. 10A-10B show projections of the end points to the sagittal view.

[0080] FIGS. 11 A-l IB and 12A-12B illustrate the impact of the adjusted end points on the nucleus boundary. For example, FIG. 11 A shows an example of an original nucleus boundary (slice 9), and FIG. 1 IB shows the adjusted nucleus boundary (slice 9). FIG. 12A shows an example of an original nucleus boundary (slice 10), and FIG. 12B shows the adjusted nucleus boundary (slice 10). Once the nucleus boundaries are adjusted, the resultant contours may be used to compute the disruption score as described above.

[0081] For example, FIG. 13 illustrates an example of an area used for computing a disruption score of slice 9, FIG. 14 illustrates an example of an area used for computing a disruption score of slice 10.

[0082] The metrics described herein, e.g., disc budge (e.g., beaking score), nuclear tail, nucleus boundary and disruption score may be highly predictive, e.g., alone or in combination with each other for selecting target discs for surgical and non-surgical interventions. They have also been shown to be predictive of pain generators when compared to discograms.Examples

[0083] FIGS. 15A-15B illustrate one example of a method for a sagittal segmentation model as described herein. In this example, scan files (e.g. a DICOM medical image file) may be received by the system and segmented. In this example, the scan file, which may be from an MRI scan or other imaging modality, may be received. In some cases the apparatus (software, firmware, hardware, etc.) may include or be part of an imaging system, such as an imaging system configured to take MRI scan. Initially, the scan files in a particular seriesdescription may be converted to an array for mathematical processing, such as a numpy array, which typically includes all the sagittal slices in sequence from left to right using the left-to- right value, e.g., in metadata (e.g., in an ‘ImagePositionPatient’ metadata) from the scan file. The input slice may then be horizontally flipped, and resized, e.g., to (512, 512) size, and may be histogram equalized as a preprocessing step. Both medial and lateral UNet segmentation models may be applied on each slice. For example, for a 1.5 Tesla MRI slices, some portions on the left and right may be cropped out and then histogram equalized in order to maintain the brightness of the slice and improve segmentation results. In any of these examples, 3 or more class segmentation masks for each of disc, vertebra and spinal cord may be the output with size, e.g., (1728, 1728), and original size.

[0084] Also described herein are methods and apparatuses for performing sagittal disc segmentation. A refined disc mask may be obtained by training a model (e.g., a TransUnet model) for segmentation of disc in a regio of interest (ROI) around each disc. As shown FIG. 16A, the ROI region contains the disc, a region from vertebrae above and below the disc, and the spinal cord at the posterior side of the disc. The ROI region may be selected around each disc obtained from the disc mask of the 3 class segmentation model shown in FIGS. 15A- 15B.

[0085] FIG. 16A shows an example of a region of interest (ROI) for extraction of a disc. FIG. 16B shows an example of the segmented disc of FIG. 16A.

[0086] Also described herein are sagittal internal material segmentation. In this example, each sagittal disc image may be obtained from a disc segmentation model (such as that shown in FIG. 16B) and may be rotated to align horizontally. The disc images may then be resized to (512, 256) before internal material segmentation. In any of these examples, a three-class segmentation model may be applied on each rotated and resized disc to obtain the nucleus pulposus and annulus fibrosus in the disc.

[0087] FIG. 17A shows an example of an extracted, rotated and re-sized disc image. FIG. 17B shows internal segmentation of the disc of FIG. 17A.

[0088] Also described herein are methods and apparatuses for axial disc segmentation. For example, in some cases the scan files (e.g., DICOM files) are that are initially in a particular series description may be converted to an array (e.g., a numpy array) which includes all the axial slices in sequence from bottom to top using the height value in metadata (e.g., in TmagePositionPatient’ metadata). In this example a UNet segmentation model is used to detect disc / vertebrae in all the axial series descriptions. The input axial slices are resized, e.g., to (512, 512, 3) before it is used to predict. The prediction threshold is kept at 0.5. The output mask is of shape (512, 512) and (384, 384). This is illustrated in FIGS. 18A-18B, showing the mask for a disc / vertebra 1805.

[0089] FIG. 18 A illustrates an example of an original axial MRSI slice. FIG. 18B shows an output of an axial disc segmentation model as described herein.

[0090] Also described herein are methods and apparatuses for axial spinal cord segmentation. Initially, the scan files (e.g., DICOM files) in a particular series description may be converted to an array (e.g., a numpy array) which may include all the axial slices in sequence from bottom to top using the height value, e.g., in metadata (e.g., ‘ImagePositionPatient’ metadata). A segmentation technique (e.g., a UNet segmentation model) may be used to detect the spinal cord in all the axial series descriptions. The input axial slices may be resized, e.g., to (512, 512, 3), before they are used to predict as described above. This is illustrated in FIGS. 19A-19B. The prediction threshold may be kept at, e.g., 0.5. In this example, the output mask is of shape (512, 512) and (384, 384). FIG. 19A shows an example of an original axial MRI slice. FIG. 19B shows the output of an axial spinal cord segmentation model, as described.

[0091] Also described herein are axial posterior element segmentation. For example, a scan file (e.g., a DICOM file) may include a particular series description that may be initially converted to an array (e.g., a numpy array) which may include all the axial slices in sequence, e.g., from bottom to top using the height value in metadata (e.g., in TmagePositionPatient’ metadata). A segmentation technique (e.g., a UNet segmentation model) may be used to detect facet joints in all the axial series descriptions. The input axial slices may be resized to, e.g., (512, 512, 3) before used to predict. In some examples, the prediction threshold may be kept at about 0.5. The output mask(s) may be, e.g., a shape of (512, 512) and (384, 384). FIGS. 20A-20B show one example of this, showing an original scan image and the scan image (FIG. 20A) and the scan image showing the output of the posterior element segmentation model (FIG. 20B).

[0092] The methods and apparatuses described herein may include the determination of one or more scores or indexes to indicate the heath or status of a spinal disc. These scores may be estimated using the image processing techniques descried herein. In some cases described herein are methods and apparatuses, including user interfaces for receiving scan data including one or more medical (e.g., spinal) scans, such as but not limited to MRI, CT, etc. scans, and modifying the one or more images from the scan, e.g., by masking and / or segmenting, and estimating a score based on the modified scan, as described herein. This estimating step may be performed algorithmically and / or using a trained machine learning agent. In some cases the apparatus may be configured to receive and process the scan data, including displaying the one or more images (scans) of the spinal disc to illustrate thefeatures used to estimate the one or more scores, for review and analysis by a user (e.g., doctor, surgeon, nurse, etc. or other medical practitioner). The apparatus may be configured to output the one or more scores and / or to generate a prediction based on the one or more scores.

[0093] For example, an apparatus as described herein may be configured to estimate and display one or more of the scores described herein, such as (but not limited to) an internal disruption score and / or a nucleus-annulus ratio, a beaking score, a nuclear tail score, and / or a nuclear boundary score. In some cases the apparatus may display an original and / or modified image of a spinal disc (e.g., showing segmentation and / or indicator of measured features) as well as outputting (in some cases onto the image(s)) the one or more scores. In some cases the method or apparatus may allow the user to select between different scores. In some cases the user may modify the scores. In some cases the apparatus or method may include a confidence level for the score. The confidence level indicator may initially be based on one or more factors related to the variability in the estimate. In any of these methods and apparatuses the confidence level indicator may be modified by the user.

[0094] The methods and apparatuses described herein may apply the estimated score(s). For example, the methods and apparatuses may determine if a patient is in pain, and where the patient is in pain, which may in term be used to guide treatment, including identifying what region or regions of the spinal disc should be modified based on the analysis including the scoring. The methods and apparatuses described herein may also be used to predict one or more outcomes. For example, these methods may be used to predict how a patient will respond to a treatment. In practice, the methods and apparatuses described herein have been found to predict with greater than 90% accuracy the outcome of treatment using the scoring techniques described herein. Further these methods and apparatuses for estimating a score are significantly more accurate than existing techniques (e.g., may predict with >90%) and are less susceptible to bias and inaccuracies.

[0095] Any of these methods and apparatuses may also be used to monitor patient progress / recovery over time. As mentioned, these methods and apparatuses may be applied with greater impartiality and consistency, allowing more effective tracking of treatments, including surgical interventions.

[0096] All publications and patent applications mentioned in this specification are herein incorporated by reference in their entirety to the same extent as if each individual publication or patent application was specifically and individually indicated to be incorporated by reference. Furthermore, it should be appreciated that all combinations of the foregoing concepts and additional concepts discussed in greater detail below (provided such conceptsare not mutually inconsistent) are contemplated as being part of the inventive subject matter disclosed herein and may be used to achieve the benefits described herein.

[0097] Any of the methods (including user interfaces) described herein may be implemented as software, hardware or firmware, and may be described as a non-transitory computer-readable storage medium storing a set of instructions capable of being executed by a processor (e.g., computer, tablet, smartphone, etc.), that when executed by the processor causes the processor to control perform any of the steps, including but not limited to: displaying, communicating with the user, analyzing, modifying parameters (including timing, frequency, intensity, etc.), determining, alerting, or the like. For example, any of the methods described herein may be performed, at least in part, by an apparatus including one or more processors having a memory storing a non-transitory computer-readable storage medium storing a set of instructions for the processes(s) of the method.

[0098] While various embodiments have been described and / or illustrated herein in the context of fully functional computing systems, one or more of these example embodiments may be distributed as a program product in a variety of forms, regardless of the particular type of computer-readable media used to actually carry out the distribution. The embodiments disclosed herein may also be implemented using software modules that perform certain tasks. These software modules may include script, batch, or other executable files that may be stored on a computer-readable storage medium or in a computing system. In some embodiments, these software modules may configure a computing system to perform one or more of the example embodiments disclosed herein.

[0099] As described herein, the computing devices and systems described and / or illustrated herein broadly represent any type or form of computing device or system capable of executing computer-readable instructions, such as those contained within the modules described herein. In their most basic configuration, these computing device(s) may each comprise at least one memory device and at least one physical processor.

[0100] The term “memory” or “memory device,” as used herein, generally represents any type or form of volatile or non-volatile storage device or medium capable of storing data and / or computer-readable instructions. In one example, a memory device may store, load, and / or maintain one or more of the modules described herein. Examples of memory devices comprise, without limitation, Random Access Memory (RAM), Read Only Memory (ROM), flash memory, Hard Disk Drives (HDDs), Solid-State Drives (SSDs), optical disk drives, caches, variations or combinations of one or more of the same, or any other suitable storage memory.

[0101] In addition, the term “processor” or “physical processor,” as used herein, generally refers to any type or form of hardware-implemented processing unit capable of interpreting and / or executing computer-readable instructions. In one example, a physical processor may access and / or modify one or more modules stored in the above-described memory device. Examples of physical processors comprise, without limitation, microprocessors, microcontrollers, Central Processing Units (CPUs), Field-Programmable Gate Arrays (FPGAs) that implement softcore processors, Application-Specific Integrated Circuits (ASICs), portions of one or more of the same, variations or combinations of one or more of the same, or any other suitable physical processor.

[0102] Although illustrated as separate elements, the method steps described and / or illustrated herein may represent portions of a single application. In addition, in some embodiments one or more of these steps may represent or correspond to one or more software applications or programs that, when executed by a computing device, may cause the computing device to perform one or more tasks, such as the method step.

[0103] In addition, one or more of the devices described herein may transform data, physical devices, and / or representations of physical devices from one form to another. Additionally or alternatively, one or more of the modules recited herein may transform a processor, volatile memory, non-volatile memory, and / or any other portion of a physical computing device from one form of computing device to another form of computing device by executing on the computing device, storing data on the computing device, and / or otherwise interacting with the computing device.

[0104] The term “computer-readable medium,” as used herein, generally refers to any form of device, carrier, or medium capable of storing or carrying computer-readable instructions. Examples of computer-readable media comprise, without limitation, transmission-type media, such as carrier waves, and non-transitory-type media, such as magnetic-storage media (e.g., hard disk drives, tape drives, and floppy disks), optical-storage media (e.g., Compact Disks (CDs), Digital Video Disks (DVDs), and BLU-RAY disks), electronic-storage media (e.g., solid-state drives and flash media), and other distribution systems.

[0105] A person of ordinary skill in the art will recognize that any process or method disclosed herein can be modified in many ways. The process parameters and sequence of the steps described and / or illustrated herein are given by way of example only and can be varied as desired. For example, while the steps illustrated and / or described herein may be shown or discussed in a particular order, these steps do not necessarily need to be performed in the order illustrated or discussed.

[0106] The various exemplary methods described and / or illustrated herein may also omit one or more of the steps described or illustrated herein or comprise additional steps in addition to those disclosed. Further, a step of any method as disclosed herein can be combined with any one or more steps of any other method as disclosed herein.

[0107] The processor as described herein can be configured to perform one or more steps of any method disclosed herein. Alternatively or in combination, the processor can be configured to combine one or more steps of one or more methods as disclosed herein.

[0108] When a feature or element is herein referred to as being "on" another feature or element, it can be directly on the other feature or element or intervening features and / or elements may also be present. In contrast, when a feature or element is referred to as being "directly on" another feature or element, there are no intervening features or elements present. It will also be understood that, when a feature or element is referred to as being "connected", "attached" or "coupled" to another feature or element, it can be directly connected, attached or coupled to the other feature or element or intervening features or elements may be present. In contrast, when a feature or element is referred to as being "directly connected", "directly attached" or "directly coupled" to another feature or element, there are no intervening features or elements present. Although described or shown with respect to one embodiment, the features and elements so described or shown can apply to other embodiments. It will also be appreciated by those of skill in the art that references to a structure or feature that is disposed "adjacent" another feature may have portions that overlap or underlie the adjacent feature.

[0109] Terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. For example, as used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and / or "comprising," when used in this specification, specify the presence of stated features, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, steps, operations, elements, components, and / or groups thereof. As used herein, the term "and / or" includes any and all combinations of one or more of the associated listed items and may be abbreviated as " / ".

[0110] Spatially relative terms, such as "under", "below", "lower", "over", "upper" and the like, may be used herein for ease of description to describe one element or feature's relationship to another element(s) or feature(s) as illustrated in the figures. It will be understood that the spatially relative terms are intended to encompass different orientations of the device in use or operation in addition to the orientation depicted in the figures. For example, if a device in the figures is inverted, elements described as "under”, or "beneath"other elements or features would then be oriented "over" the other elements or features. Thus, the exemplary term "under" can encompass both an orientation of over and under. The device may be otherwise oriented (rotated 90 degrees or at other orientations) and the spatially relative descriptors used herein interpreted accordingly. Similarly, the terms "upwardly", "downwardly", "vertical", "horizontal" and the like are used herein for the purpose of explanation only unless specifically indicated otherwise.

[0111] Although the terms “first” and “second” may be used herein to describe various features / elements (including steps), these features / elements should not be limited by these terms, unless the context indicates otherwise. These terms may be used to distinguish one feature / element from another feature / element. Thus, a first feature / element discussed below could be termed a second feature / element, and similarly, a second feature / element discussed below could be termed a first feature / element without departing from the teachings of the present invention.

[0112] In general, any of the apparatuses and methods described herein should be understood to be inclusive, but all or a sub-set of the components and / or steps may alternatively be exclusive and may be expressed as “consisting of’ or alternatively “consisting essentially of’ the various components, steps, sub-components or sub-steps.

[0113] As used herein in the specification and claims, including as used in the examples and unless otherwise expressly specified, all numbers may be read as if prefaced by the word "about" or “approximately,” even if the term does not expressly appear. The phrase “about” or “approximately” may be used when describing magnitude and / or position to indicate that the value and / or position described is within a reasonable expected range of values and / or positions. For example, a numeric value may have a value that is + / - 0.1% of the stated value (or range of values), + / - 1% of the stated value (or range of values), + / - 2% of the stated value (or range of values), + / - 5% of the stated value (or range of values), + / - 10% of the stated value (or range of values), etc. Any numerical values given herein should also be understood to include about or approximately that value, unless the context indicates otherwise. For example, if the value " 10" is disclosed, then "about 10" is also disclosed. Any numerical range recited herein is intended to include all sub-ranges subsumed therein. It is also understood that when a value is disclosed that "less than or equal to" the value, "greater than or equal to the value" and possible ranges between values are also disclosed, as appropriately understood by the skilled artisan. For example, if the value "X" is disclosed the "less than or equal to X" as well as "greater than or equal to X" (e.g., where X is a numerical value) is also disclosed. It is also understood that the throughout the application, data is provided in a number of different formats, and that this data represents endpoints and starting points, andranges for any combination of the data points. For example, if a particular data point “10” and a particular data point “15” are disclosed, it is understood that greater than, greater than or equal to, less than, less than or equal to, and equal to 10 and 15 are considered disclosed as well as between 10 and 15. It is also understood that each unit between two particular units are also disclosed. For example, if 10 and 15 are disclosed, then 11, 12, 13, and 14 are also disclosed.

[0114] Although various illustrative embodiments are described above, any of a number of changes may be made to various embodiments without departing from the scope of the invention as described by the claims. Optional features of various device and system embodiments may be included in some embodiments and not in others. Therefore, the foregoing description is provided primarily for exemplary purposes and should not be interpreted to limit the scope of the invention as it is set forth in the claims.

[0115] The examples and illustrations included herein show, by way of illustration and not of limitation, specific embodiments in which the subject matter may be practiced. As mentioned, other embodiments may be utilized and derived there from, such that structural and logical substitutions and changes may be made without departing from the scope of this disclosure. Such embodiments of the inventive subject matter may be referred to herein individually or collectively by the term “invention” merely for convenience and without intending to voluntarily limit the scope of this application to any single invention or inventive concept, if more than one is, in fact, disclosed. Thus, although specific embodiments have been illustrated and described herein, any arrangement calculated to achieve the same purpose may be substituted for the specific embodiments shown. This disclosure is intended to cover any and all adaptations or variations of various embodiments. Combinations of the above embodiments, and other embodiments not specifically described herein, will be apparent to those of skill in the art upon reviewing the above description.

Claims

CLAIMSWhat is claimed is:

1. A method of identifying a pain generator disc, the method comprising: receiving a scan volume data including one or more spinal discs; processing the scan volume data to generate a plurality of virtual slices through the scan volume data, comprising virtual axial slices and virtual sagittal slices; segmenting the plurality of virtual slices to identify the one or more spinal discs; detecting beaking in any of the one or more spinal discs by estimating a beaking score for each disc of the one or more spinal discs and generating a first subset of the plurality of virtual slices in which the beaking score corresponds to a likelihood of herniation of the disc greater than a beaking threshold; detecting a nuclear tail in any of the one or more spinal discs by estimating a nuclear tail score for each disc of the one or more spinal discs and generating a second subset of the plurality of virtual slices in which the nuclear tail score corresponds to a likelihood of nuclear degeneration of the disc greater than a nuclear tail threshold; detecting an internal disruption in any of the one or more spinal discs by estimating an internal disruption score for each disc of the one or more spinal discs and generating a third subset of the plurality of virtual slices in which the internal disruption score corresponds to a likelihood of degeneration of the disc greater than an internal disruption threshold; modifying the one or more virtual slices of the first subset of the plurality of virtual slices to indicate a beaking including a representation of the corresponding beaking score; modifying the one or more virtual slices of the second subset of the plurality of virtual slices to indicate a nuclear tail region including a representation of the corresponding a nuclear tail score; modifying the one or more virtual slices of the third subset of the plurality of virtual slices to indicate internal disruption in a region of a disc of the one or more virtual slices including a representation of the corresponding internal disruption score for each region that has an internal disruption greater than an internal disruption threshold; andoutputting, on a user interface, the modified one or more virtual slices of the first subset, the modified one or more virtual slices of the second subset and the modified one or more virtual slices of the third subset.

2. The method of claim 1, wherein estimating a beaking score for each disc of the one or more spinal discs comprises estimating a change in a slope around a circumference of each disc.

3. The method of claim 2, wherein the change in slope around the circumference of the disc is determined as a radial difference between two consecutive spans extending between a centroid of the disc and a perimeter of the disc, wherein consecutive spans are separated by between about 2 and 30 degrees.

4. The method of claim 2, wherein estimating a beaking score for each disc of the one or more spinal discs comprises estimating a deviation of the slope from a smooth contour fitted over each disc from a segmented contour of the disc.

5. The method of any of claims 1-4, wherein estimating a nuclear tail score for each disc of the one or more spinal discs comprises determining a width of a region of a nucleus of the disc having a height that is greater than a minimum threshold height of the nucleus when height is measured perpendicular to a long axis of the spinal disc.

6. The method of claim 5, further comprising normalizing the plurality of virtual slices to align and / or resize the virtual slices.

7. The method of claim 5, further comprising segmenting the plurality of virtual slices to identify a nucleus in each of the spinal discs.

8. The method of any of claims 1-7, wherein estimating the internal disruption score for each disc of the one or more spinal discs comprises further segmenting each virtual slice of the plurality of virtual slices to determine a boundary between an annulus and a nucleus, dividing each virtual slice in a plurality of regions comprising nucleus region and annulus regions and estimating disruption rations from an amount of nucleus outside of a nucleus boundary and / or an amount of annulus region within the nucleus boundary for each region of the plurality of regions.

9. The method of claim 8, wherein dividing each virtual slice comprises dividing into 12 or more radial regions.

10. The method of claim 8, wherein estimating the amount of nucleus outside of the nucleus boundary and / or the amount of annulus region within the nucleus boundary comprises estimating a number of nucleus pixels outside of the boundary and the / or a number of annulus pixels inside the boundary.

11. The method of claim 8, wherein determining the boundary comprises receiving user input to refine the boundary.

12. The method of any of claims 1-11, wherein estimating the internal disruption score for each disc of the one or more spinal discs comprises generating a digital twin for each disc of the one or more spinal discs and comparing each disc of the one or more spinal discs to its corresponding digital twin.

13. The method of claim 12, wherein generating the digital twin comprises algorithmically determining the digital twin based on an elliptical curve using an axial view of the disc.

14. The method of any of claims 1-13, further comprising receiving one or more user inputs of the user interface to switch adjust one or more of: the beaking threshold, the nuclear tail threshold, and / or the internal disruption threshold.

15. The method of any of claims 1-14, wherein outputting comprises outputting an anatomical identifier for each disc of the modified one or more virtual slices of the first subset, the modified one or more virtual slices of the second subset and the modified one or more virtual slices of the third subset.

16. The method of any of claims 1-15, wherein processing the scan volume data to generate the plurality of virtual slices through the scan volume data comprises identifying a representative slice for each disc of the one or more spinal discs and including only the representative slices in the plurality of virtual slices.

17. The method of any of claims 1-16, wherein processing the scan volume data to generate the plurality of virtual slices comprises one or more of: re-sizing the one or more virtual slices, histogram equalizing the one or more virtual slices, re-orienting the one or more virtual slices, and / or cropping the one or more virtual slices.

18. The method of any of claims 1-17, wherein segmenting comprises generating a mask of the spinal disc for each of the one or more virtual slices.

19. The method of any of claims 1-18, further comprising aligning the discs from the segmented virtual slices to a standard alignment.

20. The method of any of claims 1-19, further comprising outputting one or more candidates for the pain generator disc based on the modified one or more virtual slices of the first subset, the modified one or more virtual slices of the second subset and the modified one or more virtual slices of the third subset.

21. An apparatus, the apparatus comprising: a processor; a memory coupled to the one or more processors, the memory storing computerprogram instructions, that, when executed by the one or more processors, perform a computer-implemented method comprising: receiving a scan volume data including one or more spinal discs; processing the scan volume data to generate a plurality of virtual slices through the scan volume data, comprising virtual axial slices and virtual sagittal slices; segmenting the plurality of virtual slices to identify the one or more spinal discs; detecting beaking in any of the one or more spinal discs by estimating a beaking score for each disc of the one or more spinal discs and generating a first subset of the plurality of virtual slices in which the beaking score corresponds to a likelihood of herniation of the disc greater than a beaking threshold; detecting a nuclear tail in any of the one or more spinal discs by estimating a nuclear tail score for each disc of the one or more spinal discs and generating a second subset of the plurality of virtual slices in which the nuclear tail score corresponds to a likelihood of nuclear degeneration of the disc greater than a nuclear tail threshold; detecting an internal disruption in any of the one or more spinal discs by estimating an internal disruption score for each disc of the one or more spinal discs and generating a third subset of the plurality of virtual slices in which the internal disruption score corresponds to a likelihood of degeneration of the disc greater than an internal disruption threshold;modifying the one or more virtual slices of the first subset of the plurality of virtual slices to indicate a beaking including a representation of the corresponding beaking score; modifying the one or more virtual slices of the second subset of the plurality of virtual slices to indicate a nuclear tail region including a representation of the corresponding a nuclear tail score; modifying the one or more virtual slices of the third subset of the plurality of virtual slices to indicate internal disruption in a region of a disc of the one or more virtual slices including a representation of the corresponding internal disruption score for each region that has an internal disruption greater than an internal disruption threshold; and outputting, on a user interface, the modified one or more virtual slices of the first subset, the modified one or more virtual slices of the second subset and the modified one or more virtual slices of the third subset.

22. The system of claim 21, wherein estimating a beaking score for each disc of the one or more spinal discs comprises estimating a change in a slope around a circumference of each disc.

23. The system of claim 22, wherein the change in slope around the circumference of the disc is determined as a radial difference between two consecutive spans extending between a centroid of the disc and a perimeter of the disc, wherein consecutive spans are separated by between about 2 and 30 degrees.

24. The system of claim 22, wherein estimating a beaking score for each disc of the one or more spinal discs comprises estimating a deviation of the slope from a smooth contour fitted over each disc from a segmented contour of the disc.

25. The system of any of claims 21-24, wherein estimating a nuclear tail score for each disc of the one or more spinal discs comprises determining a width of a region of a nucleus of the disc having a height that is greater than a minimum threshold height of the nucleus when height is measured perpendicular to a long axis of the spinal disc.

26. The system of claim 25, further comprising normalizing the plurality of virtual slices to align and / or resize the virtual slices.

27. The system of claim 25, further comprising segmenting the plurality of virtual slices to identify a nucleus in each of the spinal discs.

28. The system of any of claims 21-27, wherein estimating the internal disruption score for each disc of the one or more spinal discs comprises further segmenting each virtual slice of the plurality of virtual slices to determine a boundary between an annulus and a nucleus, dividing each virtual slice in a plurality of regions comprising nucleus region and annulus regions and estimating disruption rations from an amount of nucleus outside of a nucleus boundary and / or an amount of annulus region within the nucleus boundary for each region of the plurality of regions.

29. The system of claim 28, wherein dividing each virtual slice comprises dividing into 12 or more radial regions.

30. The system of claim 28, wherein estimating the amount of nucleus outside of the nucleus boundary and / or the amount of annulus region within the nucleus boundary comprises estimating a number of nucleus pixels outside of the boundary and the / or a number of annulus pixels inside the boundary.

31. The system of claim 28, wherein determining the boundary comprises receiving user input to refine the boundary.

32. The system of any of claims 21-31, wherein estimating the internal disruption score for each disc of the one or more spinal discs comprises generating a digital twin for each disc of the one or more spinal discs and comparing each disc of the one or more spinal discs to its corresponding digital twin.

33. The system of claim 32, wherein generating the digital twin comprises algorithmically determining the digital twin based on an elliptical curve using an axial view of the disc.

34. The system of any of claims 21-33, further comprising receiving one or more user inputs of the user interface to switch adjust one or more of: the beaking threshold, the nuclear tail threshold, and / or the internal disruption threshold35. The system of any of claims 21-34, wherein outputting comprises outputting an anatomical identifier for each disc of the modified one or more virtual slices of thefirst subset, the modified one or more virtual slices of the second subset and the modified one or more virtual slices of the third subset.

36. The system of any of claims 21-35, wherein processing the scan volume data to generate the plurality of virtual slices through the scan volume data comprises identifying a representative slice for each disc of the one or more spinal discs and including only the representative slices in the plurality of virtual slices.

37. The system of any of claims 21-36, wherein processing the scan volume data to generate the plurality of virtual slices comprises one or more of re-sizing the one or more virtual slices, histogram equalizing the one or more virtual slices, re-orienting the one or more virtual slices, and / or cropping the one or more virtual slices.

38. The system of any of claims 21-37, wherein segmenting comprises generating a mask of the spinal disc for each of the one or more virtual slices.

39. The system of any of claims 21-38, further comprising aligning the discs from the segmented virtual slices to a standard alignment.

40. The system of any of claims 21-39, further comprising outputting one or more candidates for the pain generator disc based on the modified one or more virtual slices of the first subset, the modified one or more virtual slices of the second subset and the modified one or more virtual slices of the third subset.

41. A method of identifying a pain generator disc, the method comprising: receiving a scan volume data including one or more spinal discs; processing the scan volume data to generate a plurality of virtual slices through the scan volume data, comprising virtual axial slices and virtual sagittal slices; segmenting the plurality of virtual slices to identify the one or more spinal discs; detecting, in any of the one or more spinal discs, one or more of breaking, a nuclear tail, and / or an internal disruption, wherein: detecting beaking comprises estimating a beaking score for each disc of the one or more spinal discs and generating a first subset of the plurality of virtual slices in which the beaking score corresponds to a likelihood of herniation of the disc greater than a beaking threshold, detecting a nuclear tail comprises estimating a nuclear tail score for each disc of the one or more spinal discs and generating a second subset ofthe plurality of virtual slices in which the nuclear tail score corresponds to a likelihood of nuclear degeneration of the disc greater than a nuclear tail threshold, and detecting an internal disruption comprises estimating an internal disruption score for each disc of the one or more spinal discs and generating a third subset of the plurality of virtual slices in which the internal disruption score corresponds to a likelihood of degeneration of the disc greater than an internal disruption threshold; modifying the one or more virtual slices of the first subset of the plurality of virtual slices to indicate a beaking including a representation of the corresponding beaking score; modifying the one or more virtual slices of the second subset of the plurality of virtual slices to indicate a nuclear tail region including a representation of the corresponding a nuclear tail score; modifying the one or more virtual slices of the third subset of the plurality of virtual slices to indicate internal disruption in a region of a disc of the one or more virtual slices including a representation of the corresponding internal disruption score for each region that has an internal disruption greater than an internal disruption threshold; and outputting, on a user interface, the modified one or more virtual slices of the first subset, the modified one or more virtual slices of the second subset, and / or the modified one or more virtual slices of the third subset.

42. An apparatus, the apparatus comprising: a processor; a memory coupled to the one or more processors, the memory storing computerprogram instructions, that, when executed by the one or more processors, perform a computer-implemented method comprising: receiving a scan volume data including one or more spinal discs; processing the scan volume data to generate a plurality of virtual slices through the scan volume data, comprising virtual axial slices and virtual sagittal slices; segmenting the plurality of virtual slices to identify the one or more spinal discs;detecting, in any of the one or more spinal discs, one or more of: breaking, a nuclear tail, and / or an internal disruption, wherein: detecting beaking comprises estimating a beaking score for each disc of the one or more spinal discs and generating a first subset of the plurality of virtual slices in which the beaking score corresponds to a likelihood of herniation of the disc greater than a beaking threshold, detecting a nuclear tail comprises estimating a nuclear tail score for each disc of the one or more spinal discs and generating a second subset of the plurality of virtual slices in which the nuclear tail score corresponds to a likelihood of nuclear degeneration of the disc greater than a nuclear tail threshold, and detecting an internal disruption comprises estimating an internal disruption score for each disc of the one or more spinal discs and generating a third subset of the plurality of virtual slices in which the internal disruption score corresponds to a likelihood of degeneration of the disc greater than an internal disruption threshold; modifying the one or more virtual slices of the first subset of the plurality of virtual slices to indicate a beaking including a representation of the corresponding beaking score; modifying the one or more virtual slices of the second subset of the plurality of virtual slices to indicate a nuclear tail region including a representation of the corresponding a nuclear tail score; modifying the one or more virtual slices of the third subset of the plurality of virtual slices to indicate internal disruption in a region of a disc of the one or more virtual slices including a representation of the corresponding internal disruption score for each region that has an internal disruption greater than an internal disruption threshold; and outputting, on a user interface, the modified one or more virtual slices of the first subset, the modified one or more virtual slices of the second subset, and / or the modified one or more virtual slices of the third subset.

Images