[go: up one dir, main page]

WO2015188279A1 - Mesure non invasive du changement de volume choroïdien et de la rigidité oculaire par imagerie oct - Google Patents

Mesure non invasive du changement de volume choroïdien et de la rigidité oculaire par imagerie oct Download PDF

Info

Publication number
WO2015188279A1
WO2015188279A1 PCT/CA2015/050546 CA2015050546W WO2015188279A1 WO 2015188279 A1 WO2015188279 A1 WO 2015188279A1 CA 2015050546 W CA2015050546 W CA 2015050546W WO 2015188279 A1 WO2015188279 A1 WO 2015188279A1
Authority
WO
WIPO (PCT)
Prior art keywords
choroidal
nodes
ocular
pixels
csi
Prior art date
Application number
PCT/CA2015/050546
Other languages
English (en)
Inventor
Santiago Costantino
Mark R. LESK
Luke Beaton
Original Assignee
Rsem, Limited Partnership
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Rsem, Limited Partnership filed Critical Rsem, Limited Partnership
Publication of WO2015188279A1 publication Critical patent/WO2015188279A1/fr

Links

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B3/00—Apparatus for testing the eyes; Instruments for examining the eyes
    • A61B3/10—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions
    • A61B3/102—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions for optical coherence tomography [OCT]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0012—Biomedical image inspection
    • G06T7/0014—Biomedical image inspection using an image reference approach
    • G06T7/0016—Biomedical image inspection using an image reference approach involving temporal comparison
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B3/00—Apparatus for testing the eyes; Instruments for examining the eyes
    • A61B3/10—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions
    • A61B3/16—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions for measuring intraocular pressure, e.g. tonometers
    • A61B3/165—Non-contacting tonometers
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/02—Detecting, measuring or recording for evaluating the cardiovascular system, e.g. pulse, heart rate, blood pressure or blood flow
    • A61B5/024—Measuring pulse rate or heart rate
    • A61B5/02416—Measuring pulse rate or heart rate using photoplethysmograph signals, e.g. generated by infrared radiation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10072—Tomographic images
    • G06T2207/10101—Optical tomography; Optical coherence tomography [OCT]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30004—Biomedical image processing
    • G06T2207/30041—Eye; Retina; Ophthalmic

Definitions

  • the present invention relates to imaging methods to assess ocular elasticity and tissue deformation and more particularly, to methods and tools for measuring the volumetric changes of the eye due to choroidal pulsations.
  • Glaucoma remains a major cause of blindness due to daunting challenges in both its diagnosis and treatment.
  • the pathogenesis of the disease is poorly understood, making early detection more difficult and by the time patients present with symptoms, they may have already suffered major irreversible vision loss.
  • early detection a serious challenge, but once diagnosed it is also not possible to give an accurate prognosis about the speed of evolution of the disease, i.e. whether a patient is in need of more aggressive treatment options.
  • IOP intraocular pressure
  • OAG open angle glaucoma
  • OCT Optical Coherence Tomography
  • a non-invasive ocular assessment method comprises obtaining a plurality of images of a choroid over a duration of at least one cardiac cycle of a patient using optical coherence tomography (OCT) to record pulsatile choroidal filling; determining a change in choroidal thickness over time (ACT) from the images by delineating a Bruchs Membrane (BM) and a choroidal-scleral interface to compute a BM-CSI distance; and deriving an ocular volume change from the change in choroidal thickness over time (ACT) and predetermined ocular shape and size parameters of the patient.
  • OCT optical coherence tomography
  • the method further comprises determining ocular rigidity using the ocular volume change and measurements of intraocular pressure of the patient.
  • choroidal thickness in each image is obtained using a segmentation algorithm based on graph search.
  • an edge- probability weighting scheme is also used to measure choroidal thickness.
  • the images are acquired using Spectral Domain OCT, with Enhanced Depth Imaging (EDI).
  • EDI Enhanced Depth Imaging
  • a system for non-invasive ocular assessment comprises a memory; a processor coupled to the memory; and an application stored in the memory.
  • the application is executable by the processor for obtaining a plurality of images of a choroid over a duration of at least one cardiac cycle of a patient using optical coherence tomography (OCT) to record pulsatile choroidal filling; determining a change in choroidal thickness over time (ACT) from the images by delineating a Bruchs Membrane (BM) and a choroidal-scleral interface to compute a BM-CSI distance; and deriving an ocular volume change from the change in choroidal thickness over time (ACT) and predetermined ocular shape and size parameters of the patient.
  • OCT optical coherence tomography
  • the application is further configured for determining ocular rigidity using the ocular volume change and measurements of intraocular pressure of the patient.
  • choroidal thickness in each image is obtained using a segmentation algorithm based on graph search.
  • an edge- probability weighting scheme is also used to measure choroidal thickness.
  • the images are acquired using Spectral Domain OCT, with Enhanced Depth Imaging (EDI).
  • EDI Enhanced Depth Imaging
  • the system comprises an OCT apparatus for acquiring the images.
  • the system also comprises a Dynamic Contour Tonometry (DCT) apparatus for obtaining measurements of intraocular pressure (IOP).
  • DCT Dynamic Contour Tonometry
  • a method for choroidal mapping comprises obtaining at least one image of a choroid using optical coherence tomography (OCT); delineating a Bruchs Membrane (BM) in the at least one image; delineating a choroidal-scleral interface (CSI) in the at least one image using a graph search having weighted connections between nodes, the weighted connections comprising weight components according to a Euclidean distance squared, a vertical distance in pixels between nodes, and a horizontal distance in pixels between nodes; and generating a map of the choroid comprising pixel content found between the BM and the CSI.
  • OCT optical coherence tomography
  • BM Bruchs Membrane
  • CSI choroidal-scleral interface
  • the methods for choroidal mapping may also be embodied in a system having a processor, a memory accessible by the processor, and an application stored in the memory and executable by the processor to perform the steps of the method.
  • All of the methods described herein may also be embodied on a non- transitory computer readable medium having stored thereon program code executable by a processor for performing the steps of the methods.
  • the images are acquired over a plurality of cardiac cycles.
  • the images are validated by confirming that high frequency components of a frequency spectrum of the choroidal thickness (CT) coincide with first and second harmonics of a heart rate frequency measured independently from an oximeter signal.
  • CT choroidal thickness
  • FIG. 1 is an exemplary flowchart of a method for determining ocular rigidity
  • Fig. 2a is an exemplary image failing to meet a Quality > 20 metric
  • Fig. 2b is an exemplary image having a portion of A-scans that will be removed;
  • Fig. 3a is an exemplary image showing the vertical gradient of a smoothed image
  • Fig. 3b is an exemplary image showing the retina, RPE, and BM;
  • Fig. 4a is an exemplary image showing a curved and slanted BM, pre-shift
  • FIG. 4b is the image of figure 4a showing the flattened BM, post-shift
  • Fig. 5 is an exemplary image showing a sub-BM region
  • Fig. 6a is an exemplary image of an uncompensated sub-BM region
  • Fig. 6b is the image of fig. 6a compensated and contrast enhanced
  • Fig. 6c is a heat map composed of a combination of X 2 images of different scales and orientations
  • Fig. 7a is a histogram for the upper half of the circle of figure 6b;
  • Fig. 7b is a histogram for the lower half of the circle of figure 6b;
  • Fig. 8a illustrates regions where a maximum horizontal connection distance is not significantly larger than the column-wise spacing of nodes
  • Fig. 8b illustrates regions where C max is many times the horizontal spacing of nodes
  • Fig. 9a is an exemplary image of an overlay of the heat mpa, node locations, and the CSI on a flattened b-scan;
  • Fig. 9b is the original b-scan overlaid with the BM and the mean BM-CSI distance;
  • Fig. 10a is a graph of CT raw with the signal CT overlaid thereon;
  • Fig. 10b is a graph of the Fourier spectrum of CT overlaid on the Fourier spectrum of CT raw ;
  • Fig. 1 1 a is a graph of the final trace CT and corresponding time points of the oximeter reading
  • Fig. 1 1 b is a graph of the final trace CT with specific peaks and valleys identified
  • Fig. 12 is a plot of axial length versus mean CT;
  • Fig. 13 is a plot of axial length versus ACT;
  • Fig. 14 is a plot of OPA versus OR;
  • Fig. 15 is a block diagram of an exemplary system for determining ocular rigidity
  • Fig. 16 is a block diagram of an exemplary application running on the computing device of figure 15.
  • FIG. 1 is a flowchart of an exemplary embodiment of a method.
  • a plurality of images in time such as a video, are acquired using Optical Coherence Tomography (OCT). While the images may be acquired using Time Domain OCT, Spectral Domain OCT, and Swept Source OCT, the examples of the present description are presented using Spectral Domain OCT, with Enhanced Depth Imaging (EDI) at the fundus.
  • OCT Optical Coherence Tomography
  • a segmentation algorithm may be used to determine frame-by-frame choroidal thickness (CT) 104.
  • CT may be converted to global ocular volume change 106 with an approximate model for the eye.
  • IOP Intraocular Pressure
  • OPA ocular pulse amplitude
  • DCT Dynamic Contour Tonometry
  • OR Ocular Rigidity
  • k is the constant of OR, accounting for the combined mechanical properties of the retina, choroid and sclera.
  • the sclera is responsible for the majority of the stiffness of the ocular globe, so while Friedenwald's formula was developed experimentally from data collected from cadaver eyes, it can be derived through a simplification of the collagen-like stress-strain behaviour exhibited by the sclera and by considering the eye to be a thin-shelled sphere. Therefore, it is used in reverse to investigate variations in rigidity by measuring IOP changes caused by ocular volume change. For a given volume change, more rigid eyes will have a correspondingly larger increase in IOP, and vice versa for less rigid eyes.
  • ocular volume changes may be modeled by estimating the total choroidal volume change over time.
  • the choroidal-scleral interface (CSI) is identified in one or more images of the eye. Manual labelling of the CSI may be used, however in some embodiments, it is not desirable due to its lack of objectivity, nor is it feasible when the number of images to be segmented is very large.
  • a segmentation algorithm based on graph search and in some embodiments, with an edge probability weighting scheme is thus provided.
  • the images may be obtained using a Spectralis OCT Plus system from the Heidelberg Engineering company in Germany.
  • the system may be configured to provide raw image time series at a user-specified number of scans averaged.
  • the acquisition parameters of this machine may be set to high-speed mode (such as 496 x 768 pixel images), enhanced depth imaging (EDI), and 30° width (8-9 mm depending on individual participant's biometry). With these settings, scans can be acquired at a maximum of 40Hz, and at 400 images the memory buffer is filled. The more scans that are averaged together, the higher the quality of a single image, but the lower the resulting effective rate of sampling.
  • averaging 5 scans per image may be used as an acceptable compromise between visibility/detectability of the CSI and sufficiently high temporal sampling to track the pulsatile volume change of the choroid.
  • Scanning may be performed using the 'star' scanning protocol centered on the fovea, giving 6 B- scans at 30° increments. The orientation with the most visible and continuous CSI is selected, and a full 400-image movie is then acquired at this orientation.
  • an oximeter may be placed on the index finger to record the heart rate. The raw images and the oximeter readings may be exported for analysis.
  • Imaging data may be provided in an extensible markup language (XML) file and the oximeter readings may be provided in a text file.
  • XML extensible markup language
  • Other file formats may also be used.
  • an XML file contains information regarding the acquisition of each image. This includes the number of scans averaged, the time of the acquisition, and a SpectralisTM proprietary metric of quality. Due to involuntary eye movements, the OCT system may temporarily lose its position. When this occurs, an eye-tracking algorithm may automatically halt acquisition to avoid artifacts, and resume when it has relocated the proper position. This means that images may be acquired at uneven time intervals and thus, they may not all be averages of the same number of scans (but will always be equal to or less than the number specified). However, a time-stamp for each image may be contained in the XML file and provides the temporal information required for frequency analysis of CSI movement.
  • pressure measurements may be taken with a Pascal Dynamic Contour Tonometer.
  • This device has a contoured head designed to match corneal curvature and thus measure pressure changes independently of biomechanical properties such as corneal rigidity.
  • Other known pressure measurement devices may also be used.
  • three measurements are taken for each individual, and the average of all measurements having a Quality Index of 3 or greater (ZiemerTM proprietary algorithm) is obtained. More or less measurements may be taken and averaged or not. This provides two values, the IOP at diastole (i.e. the minimum pressure), and the ocular pulse amplitude (OPA), the pressure change from diastole to systole.
  • OPA ocular pulse amplitude
  • every image in the time- series may be aligned (or registered) to the first, for example using the imregister function of MatlabTM, to ensure that no error in volume change is introduced through lateral movement of the location of the B-Scan.
  • Two inputs to this function are the user's choice of metric and optimizer.
  • the metric is a quantitative measure of the similarity between two images, and the optimizer is the method by which the function converges to minimum or maximum value of the metric.
  • Figure 2a is an exemplary image failing to meet the Quality > 20 metric.
  • Figure 2b is an exemplary image meeting the Quality > 20 metric. A portion 200 of the image may be removed from all images in the series as the choroid is not present.
  • the segmentation algorithm uses a sequential process of layer identification.
  • the search region for a given layer can be appropriately limited.
  • the retina-vitreous interface (RVI) is found first.
  • the large standard deviation of the filter simply ensures that the image is sufficiently blurred for the RVI to be continuous across the image, even in the presence of large amounts of noise.
  • the edges detected form a binary image, with pixels whose value is zero constituting the background and pixels valued at 1 lying on detected edges. This is then converted to a label image by setting the value of the pixels in each distinct 8-connected edge to a unique integer value, with the pixels in the first object labelled 1 , the second 2, etc. Then, at an interval of 5 columns, the label of the first edge found by starting from the top of the image and descending the column is recorded. Any label which is found more than 10 times is considered part of the RVI, and all segments are joined with cubic splines before smoothing.
  • the RPE is a very bright band in every B-scan (or image scan).
  • the maximum intensity pixel of each A-scan in images acquired at high speed are not guaranteed to fall in this band, and differences in the shape of the RPE from individuals do not lend themselves to robust delineation using a polynomial fit of fixed degree. Therefore, assuming that both the retina above and the choroid below this bright band have lesser intensity, then the top and bottom can be found by searching in each column for extrema of the vertical intensity gradient.
  • FIG. 3a is an example of the vertical gradient of the smoothed image.
  • Figure 3b illustrates the retina 302, the RPE 304, and the BM 306.
  • FIG. 4a shows the image pre-shift with the curved and slanted BM 306.
  • Figure 4b shows the shifted image with a flattened BM 306.
  • a graph search approach is used with a combined geometric and contour-detection approach.
  • the graph consists of nodes at various positions throughout the B-scan, which are connected to one another based on chosen criteria. These connections are assigned a weight, and Dijkstra's algorithm is used to find the minimum-weight path between any two specified nodes. Therefore the two main inputs selected for each frame are the location of nodes and the weight of connections.
  • the first input is the actual location of the nodes in the graph.
  • the nodes of the graph are found using variations in image intensity along each A-scan.
  • nodes are found using the smoothed first and second gradient of image intensity in each A-scan.
  • a positive threshold on the 1 st derivative ensures identification of dark-to-light transitions only, and a near-zero threshold on the 2nd derivative localizes identified points at the inflection point of intensity gradient, correctly placing them on the lower extremity of the transition. This gives a binary image of regions meeting both thresholds, which then undergoes a number of morphological operations (cleaned, thinned, etc) to leave contours of single pixel thickness.
  • FIG. 5 illustrates the sub-BM region, showing nodes comprised between the BM 306 and a line 502 representing a lower limit of 150 pixels below the BM. This zone is where nodes are allowed to exist.
  • the nodes are very dense along the bottom edge of dark blood vessels and sparse in the sclera, which is largely noise pixels in most images.
  • the second input plays a role in the weight of connections in the graph.
  • the likelihood of a boundary is provided by using a multi-scale, multi-orientation approach.
  • the output of the contour-detection algorithm is the oriented gradient X 2 (i,j,o,0) that provides the posterior likelihood of a boundary at each pixel (i,j) in an image. This is computed for each intensity image, in the search region from the BM to no deeper than 150 pixels below it, as the difference in histograms between the two halves of a disk of radius o, centered on (i,j), and divided along its diameter at an angle ⁇ , according to the formula:
  • Figure 6a illustrates an example of the uncompensated sub-BM region.
  • Figure 6b illustrates the same region, compensated and contrast enhanced, to which the oriented gradient algorithm will be applied. Overlaid is an example disk of radius 30 pixels. The top 602 and bottom 604 halves of the disk correspond to the histograms in figures 7a and 7b, respectively.
  • the response at each orientation over different scales may be added together, and the maximum response at each pixel over the resulting combinations, normalized between 0 and 1 , becomes the final output image Pb.
  • the oriented gradient image composed of the combination of the X 2 images of different scales and orientation is illustrated in figure 6c.
  • the heat map shows pixels which are very likely to lie on a boundary. Even weak boundaries can be detected while excluding noisy regions with this method.
  • the signal coming from deeper sections of the eye are often very weak as well as being highly inhomogeneous due to the presence of highly attenuating structures above them, including blood vessels.
  • the brightness of a section of the shifted image extending from 5 pixels below the BM to the bottom of the image may be enhanced to improve the contrast of the CSI, prior to applying the contour-detection (as shown in figure 6b).
  • the graph is constructed by connecting each vertex in nodes to all vertices Cmax columns to the right and [-R ma x, Rmax] rows above and below it.
  • C ma x should be sufficiently large to allow a path to be found through regions of the CSI that were too dim to be included in the set of nodes.
  • the first term w Eu ciid represents the Euclidean distance squared, ⁇ and Ay being respectively the horizontal and vertical distance in pixels between nodes a and b. This penalty is logical for a geometrically motivated shortest path search from left to right in the image, and will force the path to include the largest number of closely spaced nodes as possible when ⁇ and Ay are below their respective thresholds T H and T v .
  • the second term wv er t is the vertical jump penalty, with H being the Heaviside function, w v some large penalty, and a the rate of growth of the sigmoid function. When Ay is larger than some threshold T v , this large, smoothed penalty is applied to discourage large vertical jumps in the boundary.
  • the third term WHoriz is similar to wvert but applied in the horizontal direction. This was added to discourage large erroneous horizontal jumps in regions where the true boundary was increasing or decreasing in depth. The path found by Dijkstra's algorithm when Ay is less than T v should pass through any intermediary vertex V2 located within a circle with vi and V3 on its diameters, as shown in figure 8a.
  • an additional term w A ffj n is added, and it uses the oriented gradient image Pb.
  • This term represents the probability that a pixel corresponds to an edge.
  • the value of A is computed as the sum of the intensity of the pixels of P b that lie on the straight line connecting nodes a and b, and w A is some large value, therefore the penalty is inversely proportional to the edge strength.
  • This term allows for a low threshold of gradient and many nodes, while ensuring that the path is guided to nodes on even faint edges. It also heavily penalizes connections through noise in both the blood vessel shadows and in the sclera below the CSI, as the intermediary pixels of Pb have low intensity.
  • two 'virtual' nodes may be added in the zero" 1 and (n+1 ) th columns, and are connected to the nodes inside the image as per the restrictions on C ma x-
  • the graph search then uses Dijkstra's algorithm to find the minimum-weight path from the zero" 1 node of each image to the (n+1 ) th , which is then interpolated and smoothed to give the CSI boundary.
  • This search is run once to find the CSI in every frame.
  • a mean CSI curve is computed for all frames, and the correlation of each individual frame CSI to the mean CSI is computed.
  • the mean BM-CSI distance in each frame is computed, giving a time series of macular choroidal thickness (CT).
  • CT macular choroidal thickness
  • Figure 9a is an example of the gradient image (i.e. the heatmap), the node locations, and the CSI 902 found using two inputs to the graph search, overlaid onto the flattened b-scan.
  • Figure 9b shows the original b-scan overlaid with the BM 306, the CSI 902 and the mean BM-CSI distance 902 (or CT). This distance CT is what is tracked from frame to frame.
  • the raw signal CT raw may be filtered using the Lomb- Scargle reconstruction method.
  • the Lomb-Scargle periodigram allows for the retrieval of the amplitude Fourier spectrum F of the time series CT raw , even though it is sampled at uneven time points. Because the pulsations of the choroid are expected to have large periodic components related to the heart frequency, a modified Fourier spectrum F m can be obtained to exclude frequency components significantly above or below the heart rate recorded using the oximeter (low and high-frequency noise), or which have very low amplitude.
  • Figure 10a illustrates an exemplary raw signal CT raw 1004. Overlaid on top is the result of the inverse Fourier transform applied, giving the final signal CT 1002.
  • Figure 10b illustrates the Fourier spectrum of the BM-CSI distance.
  • Signal F 1006 corresponds to the unmodified signal CT raw .
  • Overlaid on top is the modified spectrum F m 1008 showing the frequency components that remain after the thresholds are applied.
  • the dotted lines show the locations of the 1 st and 2 nd amplitude peaks in the Fourier spectrum of the oximeter readings, and the frequency of the maximum amplitude of the choroid can be seen to lie at the same locations. This correlation shows that CT fluctuations in time are, at least in part, due to the pulsatile blood flow.
  • a windowed peak-to-valley algorithm may be applied to determine the value of the pulsatile change in CT. Sequential peaks or valleys may be removed to leave a strictly peak-valley-peak-valley signal, and the peak-to-valley distance may be computed for each peak to a set of neighboring valleys. Distances of less than 4 m (the vertical resolution of the SpectralisTM OCT for the acquisition parameters) may be excluded, and the final value of ACT is computed as the mean of the windowed peak-to-valley distances.
  • Figure 1 1 a illustrates an example overlay of the final trace CT 1 102 and the corresponding time points of the oximeter readings 1 104.
  • Figure 1 1 b shows the same trace CT 1 102.
  • the specific peaks and valleys which are used for the windowed calculation of peak-to-valley-distance are shown as triangles and squares, respectively. Minor or peak-peak/valley-valley extrema can be seen to have been removed.
  • dV volume change of the choroid
  • the choroid is modeled as the volume between two spheres shifted by ACT, as:
  • R is the axial length half distance, measured with usual biometric tools. This model accounts for the tapering of the choroid towards the anterior section of the eye. Using this volume change, and computing the IOP fraction as:
  • Equation (1 ) may be solved for k, the value of OR.
  • the mean CT was 296.85 (71 .31 ) ⁇ .
  • the mean magnitude of thickness change at the macula ACT was 19.87 (10.07) m.
  • the pulse volume dV was 6.17 (2.68) ⁇ _, and this was used to estimate the pulsatile ocular blood flow (POBF) by multiplying with heart rate (HR).
  • POBF was 468.30 (268.7158) MlJmin.
  • the mean the OR constant was 0.0270 (0.0123) 1/ ⁇ _.
  • the method used to calculate ocular volume change from a change of choroidal thickness at the macula is a first order approximation.
  • the fluctuation ACT is found in a relatively small ( ⁇ 9mm) section at the fundus, which amounts to between 9-1 1 % of the total circumference of the eyes.
  • Choroidal thickness has been shown to vary not only in the temporal-nasal directions, but more generally in all directions.
  • the pulse volume dV and POBF agree with estimated values obtained both using commercial devices that assume a given ocular rigidity, and using the slope of pressure-volume curves and OPA. Together these results provide good evidence for not only the proper segmentation of the choroid, but also that the dynamic change in volume is being well computed.
  • the results obtained with the present method based on OCT segmentation is neither sensitive nor biased by axial head movement.
  • ⁇ is the stress in the material, in units of pressure
  • ⁇ is the engineering strain, computed as a change in length over the original unstressed length AL/L 0 or analogously a change in volume over the original unstressed volume AV/V 0 .
  • the full form pressure-volume (P-V) relations derived from these assumptions expresses pressure as a function of volumetric strain, and incorporates not only material properties such as small-strain elastic modulus (Aa), strain-hardening characteristics (a) and Poisson ratio (v), but also geometric and structural variables including shell thickness t and eye radius R 9 .
  • the assumptions required to arrive at Equation 1 have implications for the dependencies of the resulting parameter OR.
  • FIG 15 there is illustrated an exemplary embodiment for a system to measure choroidal blood flow and ocular rigidity.
  • the system allows calculation of the true OR parameter as defined by Friedenwald, as it is based on directly quantifying ocular volume changes.
  • the system uses measurements for IOP and OPA combined with ACT to obtain a value for OR.
  • An OCT apparatus 1500 is operatively connected to a computing device 1502.
  • the images captured by the OCT apparatus 1500 are transmitted to the computing device 1502. Transmission can occur in real time, i.e. at the time of capture, or at a later time after having saved the captured images on a memory device (not shown).
  • the connection means may be wired or wireless, via a network such as the Internet, a WAN, a LAN, or a cellular network, or using anyone of many known wireless transmission technologies such as Bluetooth, RF, and infrared.
  • the computing device 1502 comprises, amongst other things, a plurality of applications 1508a ... 1508n running on a processor 1506, the processor being coupled to a memory 1504. It should be understood that while the applications 1508a ... 1508n presented herein are illustrated and described as separate entities, they may be combined or separated in a variety of ways.
  • the memory 1504 accessible by the processor 1506 receives and stores data.
  • the memory 1504 may be a main memory, such as a high speed Random Access Memory (RAM), or an auxiliary storage unit, such as a hard disk, a floppy disk, or a magnetic tape drive.
  • the memory may be any other type of memory, such as a Read-Only Memory (ROM), or optical storage media such as a videodisc and a compact disc.
  • the processor 1506 may access the memory 1504 to retrieve data.
  • the processor 1506 may be any device that can perform operations on data.
  • Examples are a central processing unit (CPU), a front-end processor, a microprocessor, a graphics processing unit (GPU/VPU), a physics processing unit (PPU), a digital signal processor, and a network processor.
  • the applications 1508a ... 1508n are coupled to the processor 1506 and configured to perform steps 104 to 108 as illustrated in figure 1 and described above.
  • An output may be transmitted to any type of output device.
  • Figure 16 is an exemplary embodiment for an application 1508A running on the processor 1506 for determining ocular rigidity.
  • a data receiving module 1602 receives a plurality of images of the choroid over a duration of at least one cardiac cycle of a patient from the OCT apparatus 1500. Other data, such as the IOP and OPA measurements may also be received by this module. Any sets of data used by the other modules in the system 1502 may be received by the data receiving module 1602. Data used for the segmentation, namely the raw images, the oximeter readings, the number of scans averaged, the time stamps for the images, and the quality metric of the apparatus, are provided to a segmentation module 1604. The segmentation module 1604 will obtain a time series of choroidal thickness by delineating the BM and the CS and determining the BM-CSI distance, as described above.
  • the BM-CSI distance, or choroidal thickness CT, found in the series of images, is converted to ocular volume change by the ocular volume change module 1606, to generate ACT, which is then used by the ocular rigidity module 1608 in combination with measurements of intraocular pressure of the patient to obtain a value indicative of ocular rigidity.
  • the segmentation module 1604 may be configured to delineate the CSI using a graph search having connections between nodes assigned a weight according to Euclidean distance squared, vertical jump, and horizontal jump. In some embodiments, a weight is also assigned according to edge strength.
  • the segmentation module 1604 may be configured for filtering the choroidal thickness by excluding frequency components significantly above or below a heart rate of the patient using oximeter readings taken concurrently with (or before/after) the images acquired.
  • the segmentation method described above may be used to perform choroidal mapping by performing the steps of acquiring at least one image of a choroid using OCT, delineating the BM, delineating the CSI using the search graph method, and generating a map of the choroid comprising pixel content found between the BM and the CSI.
  • This map may be output directly from the segmentation module 1604, as shown in figure 16, and displayed on a display device (not shown). Therefore in some embodiments, the segmentation module 1604 may operate in a first mode for obtaining a time series of CT or in a second mode for generating a choroidal map. When in the second mode, the ocular rigidity module 1608 and ocular volume change module 1606 are not required.

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Radiology & Medical Imaging (AREA)
  • Theoretical Computer Science (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Quality & Reliability (AREA)
  • Biophysics (AREA)
  • Ophthalmology & Optometry (AREA)
  • Biomedical Technology (AREA)
  • General Physics & Mathematics (AREA)
  • Molecular Biology (AREA)
  • Surgery (AREA)
  • Animal Behavior & Ethology (AREA)
  • Public Health (AREA)
  • Veterinary Medicine (AREA)
  • Investigating Or Analysing Materials By Optical Means (AREA)

Abstract

La présente invention concerne des procédés et des systèmes pour déterminer le changement de volume oculaire pulsatile au moyen d'images en fonction du temps de la choroïde par tomographie par cohérence optique (OCT). Les images sont obtenues sur une durée d'au moins un cycle cardiaque d'un patient pour enregistrer le remplissage choroïdien pulsatile. Le changement de volume oculaire peut être combiné avec des mesures de pression intraoculaire du patient afin de déterminer la rigidité oculaire. Les images peuvent également être utilisées pour la cartographie choroïdienne.
PCT/CA2015/050546 2014-06-12 2015-06-12 Mesure non invasive du changement de volume choroïdien et de la rigidité oculaire par imagerie oct WO2015188279A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US201462011199P 2014-06-12 2014-06-12
US62/011,199 2014-06-12

Publications (1)

Publication Number Publication Date
WO2015188279A1 true WO2015188279A1 (fr) 2015-12-17

Family

ID=54832666

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CA2015/050546 WO2015188279A1 (fr) 2014-06-12 2015-06-12 Mesure non invasive du changement de volume choroïdien et de la rigidité oculaire par imagerie oct

Country Status (1)

Country Link
WO (1) WO2015188279A1 (fr)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111655151A (zh) * 2018-01-25 2020-09-11 国立大学法人大阪大学 压力状态的检测方法以及压力检测装置
EP3763280A1 (fr) * 2019-07-11 2021-01-13 Carl Zeiss Vision International GmbH Détermination d'un changement d'une erreur de réfection d'un il
US20220254011A1 (en) * 2015-03-26 2022-08-11 Eyekor, Llc Image analysis

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7971998B2 (en) * 2006-11-16 2011-07-05 Rsem, Limited Partnership Apparatus and method for measuring a displacement within an eye in vivo in situ, and method of assessment

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7971998B2 (en) * 2006-11-16 2011-07-05 Rsem, Limited Partnership Apparatus and method for measuring a displacement within an eye in vivo in situ, and method of assessment

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
MARGOLIS, R ET AL.: "A pilot study of enhanced depth imaging optical coherence tomography of the choroid in normal eyes.", AMERICAN JOURNAL OF OPHTHALMOLOGY., vol. 147, no. Issue 5, May 2009 (2009-05-01), pages 811 - 815, XP026053446 *
MIURA, M. ET AL.: "An approach to measure blood flow in single choroidal vessel using Doppler optical coherence tomography.", IOVS., vol. 53, no. 11, October 2012 (2012-10-01), pages 7137 - 7141, XP055244337 *

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20220254011A1 (en) * 2015-03-26 2022-08-11 Eyekor, Llc Image analysis
US11861830B2 (en) * 2015-03-26 2024-01-02 Merit Cro, Inc. Image analysis
CN111655151A (zh) * 2018-01-25 2020-09-11 国立大学法人大阪大学 压力状态的检测方法以及压力检测装置
US20210038077A1 (en) * 2018-01-25 2021-02-11 Osaka University Method for detecting stressed state and stress detection apparatus
EP3744254A4 (fr) * 2018-01-25 2021-11-03 Osaka University Procédé de détection d'état de stress et dispositif de détection de stress
CN111655151B (zh) * 2018-01-25 2024-03-22 国立大学法人大阪大学 压力状态的检测方法以及压力检测装置
US12035973B2 (en) 2018-01-25 2024-07-16 Osaka University Method for detecting stressed state and stress detection apparatus
EP3763280A1 (fr) * 2019-07-11 2021-01-13 Carl Zeiss Vision International GmbH Détermination d'un changement d'une erreur de réfection d'un il
WO2021005213A1 (fr) * 2019-07-11 2021-01-14 Carl Zeiss Vision International Gmbh Détermination d'une modification d'un défaut de réfraction d'un œil
CN112740098A (zh) * 2019-07-11 2021-04-30 卡尔蔡司光学国际有限公司 眼睛的屈光不正的变化的确定
US11185223B2 (en) 2019-07-11 2021-11-30 Carl Zeiss Vision International Gmbh Determination of a change in a refractive error of an eye
CN112740098B (zh) * 2019-07-11 2023-09-05 卡尔蔡司光学国际有限公司 眼睛的屈光不正的变化的确定

Similar Documents

Publication Publication Date Title
Besenczi et al. A review on automatic analysis techniques for color fundus photographs
Abràmoff et al. Retinal imaging and image analysis
Beaton et al. Non-invasive measurement of choroidal volume change and ocular rigidity through automated segmentation of high-speed OCT imaging
US9418423B2 (en) Motion correction and normalization of features in optical coherence tomography
JP4909377B2 (ja) 画像処理装置及びその制御方法、コンピュータプログラム
AU2019340215B2 (en) Methods and systems for ocular imaging, diagnosis and prognosis
Novosel et al. Locally-adaptive loosely-coupled level sets for retinal layer and fluid segmentation in subjects with central serous retinopathy
US9299139B2 (en) Volumetric analysis of pathologies
US11883099B2 (en) Noninvasive techniques for identifying choroidal neovascularization in retinal scans
JP5631339B2 (ja) 画像処理装置、画像処理方法、眼科装置、眼科システム及びコンピュータプログラム
US20180353066A1 (en) Image processing apparatus and image processing method
Nath et al. Techniques of glaucoma detection from color fundus images: A review
JP7648056B2 (ja) 時空間データに基づき医用画像を解析するシステム及び方法
Twa et al. Validation of macular choroidal thickness measurements from automated SD-OCT image segmentation
EP3417401B1 (fr) Procédé de réduction d'artefacts dans une oct grâce à des techniques d'apprentissage automatique
WO2015188279A1 (fr) Mesure non invasive du changement de volume choroïdien et de la rigidité oculaire par imagerie oct
Zheng et al. Automatic and efficient detection of the fovea center in retinal images
Deshmukh et al. Machine learning based approach for lesion segmentation and severity level classification of diabetic retinopathy
Mehmood et al. EDDense-Net: Fully Dense Encoder Decoder Network for Joint Segmentation of Optic Cup and Disc
Noronha et al. Automated diagnosis of diabetes maculopathy: a survey
Sathananthavathi et al. Case studies of cognitive computing in healthcare systems: Disease prediction, genomics studies, medical image analysis, patient care, medical diagnostics, drug discovery
Pathan et al. Segmentation techniques for computer-aided diagnosis of glaucoma: A review
Kim Dynamic biomechanical characteristics of the optic nerve head by OCT imaging
Richer et al. Noise-free one-cardiac-cycle OCT videos for local assessment of retinal tissue deformation
US20220151482A1 (en) Biometric ocular measurements using deep learning

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 15807067

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 15807067

Country of ref document: EP

Kind code of ref document: A1