Method and system for assessing severity of stenosis

Abstract

There is provided a method and system for assessing the severity of valve stenosis in a subject using echocardiographic images of a valve leaflet to obtain for the subject one or more of the valve leaflet displacement; valve leaflet velocity; valve leaflet acceleration; valve leaflet flexibility; valve leaflet displacement to blood flow ratio (D:Q); valve leaflet displacement to blood momentum ratio (D:M); (D:Q) to Energy Loss ratio ((D:Q):EL); and (D:M) to Energy Loss ratio ((D:M):EL); and comparing one or more of the metrics and or ratios to a reference value or reference range from a healthy subject to identify patients with moderate or severe valve stenosis.

Description

METHOD AND SYSTEM FOR ASSESSING SEVERITY OF STENOSISTechnical Field

[0001] The invention relates to methods for high-frequency tracking of angular valve leaflet motion, preferably with adjustment for flow state, to measure valve leaflet stiffness and hydraulic load on the leaflet. This information can be used to inform clinical decision making regarding valve replacement.Cross reference to related application

[0002] This application claims priority to Australian provisional patent application number 2023902641 filed 21 August 2023, the entire contents of which are incorporated herein by reference.Background

[0003] Aortic valve (AV) stenosis is one of the most common valvular diseases and is the third most common cardiovascular disease in developed countries. It occurs in around 3% of patients over 75 years of age and can be caused by degenerative calcification and congenital valvular defects such as bicuspid AVs or rheumatic disease.

[0004] Calcific aortic stenosis (AS) is associated with increased leaflet stiffness and a narrowed AV orifice which results in increased pressure gradients across the valve. The presence of a bicuspid AV significantly increases the risk of AS. The natural history of AS is a prolonged asymptomatic period, with progressive reduction of the AV orifice area due to sclerosis initially, culminating in calcific AS. This is accompanied by a corresponding increase in the transaortic pressure gradient (AP) and myocardial pressure overload.

[0005] Through the preload reserve, the left ventricle (LV) compensates for the increased workload until the preload reserve is exhausted after which increases in afterload are accompanied by a reduction in stroke volume (SV), resulting in afterload mismatch, ultimately, this causes LV hypertrophy.

[0006] Initial diagnosis of AS typically occurs during routine physical examination with the presence of a heart murmur, click, or other abnormal sounds, but undiagnosed patients may experience the onset of severe symptoms such as angina, syncope, and heart failure.Without intervention, patient mortality typically occurs within 5 years of the onset of symptoms.

[0007] While it has been long known that leaflet restriction creates hydraulic load in AS, this has never directly been demonstrated in vivo and thus indirect markers of valve stenosisseverity (e.g. Doppler velocity) are used in clinical practice, leading to uncertainties in diagnosis and therapeutic decision making.

[0008] The Doppler approach to measuring blood flow velocity through the valve works on the basis that as the valve becomes more diseased and motion of the valve leaflets are restricted the orifice become narrower and the blood flow velocity increases. This works for classical phenotypes of AS. More recently, however, the complexity of AS stenosis has become more apparent, and it is known that flow velocity through the valve can be reduced due to factors other than valve stiffness. Examples include impaired pumping capacity of the heart, stiffness of the heart muscle and stiffness of the arterial tree, which all alter blood flow. Thus, current metrics relying on blood flow velocity metrics can lead to inaccurate determination of valve disease severity, and there is consequently great difficulty in dealing with these patients.

[0009] Accordingly, there is a need for methods for measuring valve leaflet stiffness and hydraulic load on the leaflet, in particular to inform clinical decision making regarding treatment of AS.Summary

[0010] In a first aspect there is provided a method of assessing the severity of valve stenosis in a subject comprising a) obtaining from images of at least one valve leaflet of the subject, one or more of the following metrics for the subject: i. valve leaflet displacement; ii. valve leaflet velocity; iii. valve leaflet acceleration; and iv. valve leaflet flexibility b) comparing one or more of the metrics to a reference value or reference range for a healthy subject, a subject with mild valve stenosis, moderate valve stenosis, or a subject with severe valve stenosis to indicate the severity of valve stenosis in the subject.

[0011] In a second aspect there is provided a method of assessing the severity of valve stenosis in a subject comprising a) obtaining from images of at least one valve leaflet of the subject, one or more of the following ratios for the subject: i. valve leaflet displacement to blood flow ratio (D:Q); ii. valve leaflet displacement to blood momentum ratio (D:M);iii. (D:Q) to Energy Loss ratio ((D:Q):EL); iv. (D:M) to Energy Loss ratio ((D:M):EL); v. valve leaflet displacement to stroke volume (D:SV); vi. valve leaflet displacement to stroke volume index (D:SVI). b) comparing one or more of the ratios to a reference value or reference range for a healthy subject, a subject with moderate valve stenosis, or a subject with severe valve stenosis to indicate the severity of valve stenosis in the subject.

[0012] In one embodiment the valve leaflet displacement may be angular displacement.

[0013] The displacement may be calculated from a series of images of the at least one valve leaflet throughout the cardiac cycle.

[0014] In one embodiment a contour of the valve leaflet may be determined in each image in the series, a centroid of the valve leaflet may be identified in each image from the contour. For example, the centroid may be the average of the pixel coordinates of all the pixels in the contour.

[0015] The method may further comprise a) applying a boundary shape around the contour in a plurality of images in the series; and b) in each of the plurality of images applying reference lines extending from the centre of the boundary shape wherein a first reference line is generally parallel with the aortic wall; and a second reference line is generally parallel to the valve leaflet, c) measuring the maximum change in angle between the reference lines across a cardiac cycle to obtain the displacement.

[0016] The plurality of images may include an image when the valve is in a closed position and an images when the valve is in a fully open position, for example plurality of images may include images of the valve in multiple positions between and including the fully closed and fully open positions.

[0017] In another embodiment the method may further comprise selecting multiple points on the at least one leaflet and aortic wall in one of the images, wherein at least one of thepoints is at or near the leaflet base. The multiple points may include a point at or near the leaflet tip, and a point on the aortic wall.

[0018] The method may further comprise calculating the angle between the points on the leaflet and the points on the aortic wall wherein the point at or near the leaflet base is the vertex of the angle. The points may be manually positioned with reference the leaflet and / or aortic wall in at least some images.

[0019] In some embodiments the method includes positioning of additional points between the point on the leaflet and the point at or near the leaflet base; and / or the point on the aortic wall and the point at or near the leaflet base.

[0020] The points may be added in any of the images, for example between and including images showing the maximally open or maximally closed positions of the valve

[0021] The method may be performed manually, or by software in an automated or semiautomated or fashion.

[0022] In another embodiment the method may further comprise selecting multiple landmarks in one of the images, wherein the landmarks are the leaflet base, leaflet tip, aortic wall, and optionally mid leaflet, and opposing leaflet base.

[0023] The method may comprise calculating the magnitude of vectors between the landmarks to determine the distances individual landmarks travel between at least two timepoints, for example systole and diastole.

[0024] The landmarks may be the leaflet tip and the mid leaflet to provide the comparative displacement (leaflet flexibility) of the leaflet tip and the mid leaflet.

[0025] The energy loss may be total energy loss or Energy Loss Index.

[0026] The step of comparing the one or more metrics or ratios to a reference value or reference range from a healthy subject may be used to designate a patient as having no valvular disease, mild valvular disease, moderate valvular disease, severe valvular disease.

[0027] The valvular disease may be a disease of the tricuspid valve, the pulmonary valve, the mitral valve, or the aortic valve, for example stenosis.

[0028] In one embodiment the step of comparing the one or more metrics or ratios to a reference value or reference range from a healthy subject may be used to designate apatient as having no aortic stenosis, mild aortic stenosis, moderate aortic stenosis, severe aortic stenosis, low gradient severe aortic stenosis, or high gradient aortic stenosis.

[0029] In a third aspect there is provided a system for performing the method of the first or second aspects, comprising: a) a scanner configured to scan a valve of a subject, the scan providing a series of images of a valve leaflet across at least one cardiac cycle; b) a processor configured to calculate displacement of the leaflet as defined in the first or second aspects and wherein the processor is further configured to calculate one or more of the following metrics for the subject: valve leaflet displacement; valve leaflet velocity; valve leaflet acceleration; and valve leaflet flexibility and / or one or more of the following ratios for the subject: valve leaflet displacement to blood flow ratio (D:Q); valve leaflet displacement to blood momentum ratio (D:M);(D:Q) to Energy Loss ratio ((D:Q):EL); (D:M) to Energy Loss ratio ((D:M):EL); valve leaflet displacement to stroke volume (D:SV); and valve leaflet displacement to volume index (D:SVI).

[0030] The processor is further configured to obtain blood flow, blood momentum and Energy Loss measurements from the series of images; or is configured to receive blood flow, blood momentum and Energy Loss measurements.

[0031] The system may further comprise a display configured to generate a visualization of one or more of the images, contours and boundary lines over time and highlight the displacement of the leaflet.Definitions

[0032] The phrase 'and / or' as used herein should be understood to mean 'either or both' of the elements so conjoined, i.e. , elements that are conjunctively present in some cases and disjunctively present in other cases.

[0033] Throughout this specification, unless the context clearly requires otherwise, the word 'comprise', or variations such as 'comprises' or 'comprising', will be understood to imply the inclusion of a stated element, integer or step, or group of elements, integers or steps, butnot the exclusion of any other element, integer or step, or group of elements, integers or steps.

[0034] Unless the context requires otherwise or specifically stated to the contrary, integers, steps, or elements of the technology recited herein as singular integers, steps or elements clearly encompass both singular and plural forms of the recited integers, steps or elements.

[0035] In the context of the present specification, the terms 'a' and 'an' are used to refer to one or more than one (i.e., at least one) of the grammatical object of the article. By way of example, reference to 'an element' means one element, or more than one element.

[0036] In the context of the present specification, the term 'about' means that reference to a figure or value is not to be taken as an absolute figure or value, but includes margins of variation above or below the figure or value in line with what a skilled person would understand according to the art, including within typical margins of error or instrument limitation. In other words, use of the term 'about' is understood to refer to a range or approximation that a person or skilled in the art would consider to be equivalent to a recited value in the context of achieving the same function or result.

[0037] A 'subject' or 'patient' as used herein refers to a mammal, preferably a human. In addition to being useful for human treatment, the methods of the invention may also be useful for veterinary treatment of non-human mammals, including companion animals and farm animals, such as, but not limited to dogs, cats, horses, cows, sheep, and pigs. In preferred embodiments the subject is a human.

[0038] 'Displacement' as used herein in relation to valve leaflets is angular displacement or rotational displacement which is the angle through which the leaflet revolves or rotates around a centre or axis. As a point on the leaflet moves about an arc with length s and radius rthe angular displacement can be defined mathematically with the following equation: Equation 1: Displacement = 9 = -rad

[0039] The term 'administering' and variations of that term including 'administer' and 'administration', includes contacting, applying, delivering, or providing a therapy or a therapeutic agent to a subject by any appropriate means including, but not limited to, oral, intravenous, intra-arterial, intramuscular, intraperitoneal, subcutaneous, intramuscular, topically, or combinations thereof.

[0040] Any reference to or discussion of any document, act or item of knowledge in this specification is included solely for the purpose of providing a context for the presentinvention. It is not suggested or represented that any of these matters or any combination thereof formed, at the priority date, part of the common general knowledge, or was known to be relevant to an attempt to solve any problem with which this specification is concerned.

[0041] Those skilled in the art will appreciate that the technology described herein is susceptible to variations and modifications other than those specifically described. It is to be understood that the technology includes all such variations and modifications. For the avoidance of doubt, the technology also includes all of the steps, features, and compounds referred to or indicated in this specification, individually or collectively, and any and all combinations of any two or more of said steps, features and compounds.

[0042] In order that the present invention may be more clearly understood, preferred embodiments will be described with reference to the following drawings and examples.Brief Description of the Drawings

[0043] Figure 1. Contour detected revealing the aortic leaflet and the aortic wall

[0044] Figure 2. Contour showing closed leaflet (A) and Contour showing open leaflet (B)

[0045] Figure 3. Energy loss versus Leaflet displacement to momentum ratio in severe disease.

[0046] Figure 4. Energy loss versus Leaflet displacement to flow ratio in severe disease.

[0047] Figure 5. Energy loss versus Leaflet displacement (manual) to flow ratio in severe disease.

[0048] Figure 6. Energy loss versus Leaflet displacement (manual) to momentum ratio in severe disease.

[0049] Figure 7. Illustration of reference lines used to calculate metrics. In Frame 1 In this frame, the use of centrelines (dashed) across the leaflet and ipsilateral aortic wall allow measurement of angles between leaflet and aorta. And by using an orthogonal (dashed) line to the aortic wall, the closed reference position of the leaflet is known and angular displacement from closed position can be determined. In Frame 2, the valve leaflet has opened more, increasing the displacement angle from baseline closed position and reducing angle to aortic wall reference line.

[0050] Figure 8. Illustration of multi-point selection, in this case a 3-point selection

[0051] Figure 9. Illustration of an angle calculation from 3 points

[0052] Figure 10. Illustration of Landmarking the leaflet base (LB), leaflet tip (LT), the mid leaflet (LM) and the aortic wall (Ao), in addition to the opposing leaflet base (OB).

[0053] Figure 11. Leaflet displacement relative to mean gradient

[0054] Figure 12. Energy loss relative to leaflet displacement as a function of the severity of aortic stenosis..

[0055] Figure 13. Energy loss relative to flow adjusted leaflet displacement .

[0056] Figure 14. Leaflet tip motion, using the Euclidean approach.

[0057] Figure 15. Mid leaflet motion, using the Euclidean approach as a function of the severity of aortic stenosis..

[0058] Figure 16. Leaflet flexibility (comparative displacement of leaflet tip and midportion) as a function of the severity of aortic stenosis..

[0059] Figure 17: 2D cardiac movement, in mm, of a single fixed point (leaflet hinge) between the closed- and open-states of the aortic valve.

[0060] Figure 18: Distribution of cardiac rotation values, as determined by the change in LVOT angulation from baseline

[0061] Figure 19: Mid-leaflet linear displacement as a function of the severity of aortic stenosis. LGAS = Low gradient aortic stenosis. HGAS = High gradient aortic stenosis

[0062] Figure 20: Leaflet tip linear displacement as a function of the severity of aortic stenosis. LGAS = Low gradient aortic stenosis. HGAS = High gradient aortic stenosis

[0063] Figure 21 : Mid-leaflet angular displacement as a function of the severity of aortic stenosis. LGAS = Low gradient aortic stenosis. HGAS = High gradient aortic stenosis

[0064] Figure 22: Leaflet tip angular displacement as a function of the severity of aortic stenosis. LGAS = Low gradient aortic stenosis. HGAS = High gradient aortic stenosis

[0065] Figure 23: Mean angular displacement as a function of the severity of aortic stenosis. LGAS = Low gradient aortic stenosis. HGAS = High gradient aortic stenosis

[0066] Figure 24: Dynamic flexibility / deformation (Absolute) as a function of the severity of aortic stenosis. LGAS = Low gradient aortic stenosis. HGAS = High gradient aortic stenosis

[0067] Figure 25: Dynamic flexibility / deformation (Relative) as a function of the severity of aortic stenosis. LGAS = Low gradient aortic stenosis. HGAS = High gradient aortic stenosis

[0068] Figure 26: : Mid-leaflet angular displacement as a function of clinically adjudicated severity of aortic stenosis.

[0069] Figure 27: Mid-leaflet angular displacement as a function of severity of aortic stenosis assessed by haemodynamic parameters.

[0070] Figure 28: Flow adjusted angular displacement as a function of severity of aortic stenosis assessed by haemodynamic parameters.Description of EmbodimentsMethods

[0071] Conventional methods of assessing valve function utilize Doppler (for example) to solely observe the consequence of leaflet restriction on blood flow velocity metrics and their derivatives. In contrast the inventors have developed a technique to observe the motion of valve leaflets themselves and by analyzing images of valve leaflets to identify the aortic valve and track leaflet motion across the cardiac cycle. Preferably, the images are echocardiographic and the tracking is automated or semi-automated.

[0072] Metrics of leaflet dynamics can be quantified at high resolution, which can provide novel insights into the true restriction of valve disease. These metrics can be combined with existing methods and adjusted to blood flow rate to form an integrated marker of true valve restriction severity. The techniques disclosed herein permit a completely new diagnostic approach to aortic stenosis, a condition desperately in need of refined diagnosis and therapeutic decision making.

[0073] The techniques can also be used to, for example detect early wear on valve replacements.

[0074] In particular, in developing the techniques the present inventors have used high- frequency tracking of angular leaflet motion, with adjustment for flow state, to measure the stiffness of the aortic valve and consequent hydraulic load. In addition the inventors have developed a computer vision algorithm to analyse valve leaflet dynamics (particularly of the AV) and its association with markers of hydraulic load.

[0075] The flow adjusted leaflet excursion metrics were tested in either 89 patients undergoing transcatheter aortic valve insertion (TAVI), or patients with severe disease (n=89), moderate disease (n=69) or normal aortic valves (n=65). Metrics of leaflet dynamics were examined against markers of hydraulic load (energy loss).

[0076] The inventors observed a strong, inverse and linear correlation between energy loss (hydraulic load) and the flow adjusted leaflet excursion (both displacement / flow ratio and displacement / momentum ratio). This indicates that valve leaflet restriction leads to hydraulic load in aortic stenosis. Computer visualization of angular leaflet motion tracking was used to directly demonstrate that leaflet restriction corresponds to the degree of energy loss and therefore the hydraulic load conferred by a stenotic aortic valve.

[0077] The methods described herein provide metrics of angular motion of valve leaflets, including displacement, velocity and acceleration. These angular motion metrics are adjusted to transaortic volumetric flow rate and jet momentum to describe two metrics of flow adjusted leaflet excursion, namely:1) displacement / flow ratio (D:Q); and2) displacement / momentum ratio (D:M).

[0078] In particular the D:Q ratio can be calculated as follows:Equation 2: D ■■ Q =tro volLejection timeJwhere D is the vale leaflet displacement and Q is the blood flow. Valve leaflet displacement can be calculated according to Equation 1 (above) by using a centreline which represents the long axis of the valve leaflet or alternatively the centroid of the contour . Stroke volume and ejection time are calculated using conventional methods. Stroke volume is calculated as the left ventricular outflow tract pulsed wave Doppler velocity-time integral multiplied by the left ventricular outflow tract area (itself calculated as pi multiplied by the left ventricular outflow tract radius squared). The left ventricular ejection time is calculated as the time duration of left ventricular ejection using the left ventricular outflow tract velocity-time integral. Where stroke volume and ejection time cannot be calculated, Q may still be determined using the validated formula (Equation 3) from Namasivayam et al. (J Am Coll Cardiol. 2020 Apr 21 ;75(15): 1758-1769), that derives Q from the aortic valve area (A A), mean gradient (MG) and peak transvalvular velocity (Vp).Equation 3:

[0079] The D:M ratio can be calculated as follows:Equation 4:where momentum (M) can be calculated, for example, by the method of Thomas et al; Circulation. 1990;81 :247-259.

[0080] In one embodiment the displacement of a valve leaflet is calculated from a centerline or centroid of the leaflet.

[0081] In embodiments where a centerline is used, the centerline represents the long axis of the leaflet which forms an angle relative to the aortic wall.

[0082] A centroid, or center of an image is the arithmetic mean position of all the points in the surface of the image. In one embodiment, a centroid represents the center point of theimage and is calculated as the average of the pixel coordinates of all the pixels in the image. In embodiments that use a centroid (a single point) the arc length and radius is used to determine the angular displacement.

[0083] In this this embodiment at the jet origin, momentum can be calculated in a number of ways, for example by combining the fundamental definition (M=Qopo) with the continuity equation (Qo=Aopo) and the Bernoulli equation (Ap = 1 pp02) with pressure expressed in metric units (1 mm Hg = 1,333 dynes / cm2):Equation 5A M = QouoEquation 5B M = Aou2Equation 5CEquation 5D M = A^p / pEquation 5E M = Qo / ^p where:Ao = effective jet orifice area (cm2) u0= jet velocity at the orifice (cm / sec)Qo= orifice flow rate (cm3 / sec) p = pressure (mm Hg or dynes / cm2)Ap = pressure gradient producing jet (mm Hg or dynes / cm2) p = Blood density (1.05 g / cm3)

[0084] Equations 5A-E are valid only for plug flow (i.e. flow at the valve orifice) and so do not apply beyond the orifice where the velocity profile is not flat.

[0085] When blood is pumped across a valve it forms a jet, the jet is formed by a plug of flow with velocity u0entering through a round orifice (the valve) with area Ao. Within the jet blood velocity has an axial component u(x,r) and a radial component v(x,r). In the centre of the jet or column of blood the velocity profiles transverse to the jet axis (u vs r) at the orifice level (plug flow) and within the jet. The momentum flux crossing a given plane transverse to the jet axis can be calculated by integrating u2across the particular plane (see Equation 6 below). At the jet origin, the velocity profile is flat, and this integral is given simply by Equation 5B which is equivalent to flow x velocity, (Equation 5A). Within the free jet, the velocity profile is no longer flat, however, for a round orifice, it is axisymmetric, so velocity isa function of radial distance from the axis, and the momentum integral is presented in Equation 7 below).

[0086] That is, to calculate the momentum passing through an arbitrary plane orthogonal to the jet axis, Equation 5B can be generalised and p2integrated across the face of the jet:Equation 6: M = f u2dA

[0087] In cases where there is an axisymmetric jet, dA can be replaced with the circular rim 2TTT dr and integrated from the centre of the jet to the periphery:Equation 7:where:A = area of the plane orthogonal to the jet axis u = axial component of jet velocity (cm / sec) r = radial distance from the jet axis (cm)

[0088] In some embodiments the ratio of displacement / flow ratio (D:Q) to energy loss and / or the displacement / momentum ratio (D:M) to energy loss is used to measure the severity of valvular stenosis. These ratios can be calculated using the following equations:Equation 8Equation 9where EL is energy loss (either total energy loss (ELtot) or the energy loss index (ELI)).

[0089] In some embodiments energy loss can be measured according to the method of Pibarot et a / ; Circulation. 2013; 127:1101-1104. Conventional Doppler echocardiography can be used in Doppler mode to measure the maximum pressure drop through the valve from the maximum flow velocity recorded at the level of the vena contracta (the point in the blood flow through the valve where the diameter of the flow is the least, and the blood velocity is at its maximum), which is usually at the level of the valve. However, as blood flow decelerates between the valve and the ascending aorta, part of the kinetic energy is converted back to static energy due to pressure recovery, and hence, the maximum pressure gradient measured by Doppler overestimates the net gradient. Likewise, the Aortic Valve Area (A A) obtained by use of the Gorlin formula is derived from recovered pressures, is higher than the Doppler AVA derived by the continuity equation. The latter measures the actual area occupied by flow at the valvular level (i.e., the vena contracta),whereas the AVA calculated by the Gorlin formula is an estimate of the energy loss related to the stenosis rather than a true effective orifice area.

[0090] In some embodiments ratio of displacement / stroke volume index (D:SVI) and / or the displacement to stroke volume (D:SV) is used to measure the severity of valvular stenosis. Displacement can be calculated as discussed above and SVI is calculated as the strove volume divided by body surface area. The extent of pressure recovery is determined by the ratio between the valve's effective orifice area and the cross-sectional area of the ascending aorta, this becomes particularly relevant in patients with moderate to severe AS and small aortas, in whom measurement of AVA by Doppler echocardiography may lead to overestimation of severity. Conversely, patients with a dilation of the ascending aorta will have less or no pressure recovery and therefore a more important energy loss for a given valve effective orifice area.

[0091] Pressure recovery can be accounted for by calculating the Energy Loss Index (ELI) as follows:Equation 10: ELI =a~BSAAVA)where AA is the cross-sectional area of the aorta measured at the sinotubular junction and BSA is the body surface area. Hence, the ELI consists of an adjustment of the Doppler AVA for the size of the ascending aorta. From a physiological standpoint, the ELI is superior to the Doppler AVA or gradient in the sense that it better represents the actual energy loss caused by the stenosis and thus the increased burden imposed on the ventricle.

[0092] In some embodiments, for example if the body surface area is unavailable, the total energy loss (ELtot) can be used:Equation 11 :

[0093] Energy loss, may also be calculated from conventional imaging modalities that assess left ventricular (LV) strain which essentially assesses changes in myocardial fiber length relative to their resting phase. Strain rate indicates the speed at which this deformation occurs. Echocardiographic images can be used to myocardial deformation (strain) and assess for example echocardiography can be used to assess longitudinal strain, circumferential strain, radial strain, or twist and rotation.

[0094] Longitudinal strain refers to the percentage change in LV fiber length in the longitudinal axis. Global longitudinal strain (GLS) is the most common measure of longitudinal strain. GLS reflects contraction of longitudinally arranged subendocardial fibers.Severe AS hinders myocardial perfusion, particularly in the sub-endocardium, due to higher wall stress and impaired coronary blood flow and hence decreased longitudinal shortening is often the first impairment seen in patients with AS.

[0095] Circumferential strain is the percentage change in LV circumference in the short axis view, and reflects contraction of the circumferentially arranged mid-layer fibers. Following a decrease in longitudinal strain, circumferential fibers compensate for the loss in longitudinal function. Hence, impairment of both longitudinal and circumferential strain suggests more extensive myocardial damage. Global circumferential strain (GCS) is considered an important prognostic factor in patients with symptomatic AS.

[0096] Radial strain is the percentage change in wall thickness in the short axis view. Impaired subendocardial perfusion is an early feature of AS due to increased wall stress and impaired coronary blood flow. Impaired endocardial radial strain is observed in patients with AS and preserved LVEF and correlates with AS severity. Epicardial radial strain may, however, preserved.

[0097] Twist / torsion and rotation refers to the myofiber geometry in the LV myocardium changes from a right-handed helix in the sub-endocardium to a left-handed helix in the subepicardium. This configuration results in twisting during systole, with the apex rotating counter clockwise and the base in a clockwise direction. When both layers contract simultaneously, a larger radius of rotation for the outer epicardial myofiber layer results in mechanical predominance of the epicardial fibers in the overall direction of rotation

[0098] In one embodiment exemplified herein, left ventricular global longitudinal strain (GLS) analysis is performed. The GLS analysis can be performed using commercially available software, such as the TOMTEC Arena software (TOMTEC Imaging Systems, Germany), using the apical 2, 3 and 4-chamber views.Valve leaflet Imaging

[0099] The methods described herein use multiple images of the valve leaflet across the cardiac cycle. Typically conventional echocardiographic methods are used to acquire the images however it is envisaged that any imaging modality can be used to acquire images for use in the method including for example (transesophageal or stress echocardiography, computed tomography, cardiovascular magnetic resonance, positron emission tomography. Each of these modalities allow a detailed assessment of the valve leaflet (and the myocardial remodelling response).

[0100] In a preferred embodiment images can be obtained using 2D or 3D echocardiography of the type used in assessment of subjects with suspected or confirmedvalvular stenosis. The imaging modality selected should be appropriate for viewing valve leaflet anatomy and movement (so displacement can be calculated).

[0101] For example, a parasternal long-axis view, or a parasternal short-axis view can be used, optionally in zoom mode. When using these modalities it may be necessary to adjust the gain to optimise the blood tissue interface.Image processing (Denoising)

[0102] In some embodiments it is desirable to improve the quality by reducing image noise that may arise from the influence of patient tissues surrounding the heart, compression, transmission and low dose radiation that is typically used in medical imaging to reduce exposure. By removing noise from these low-dose images, denoising methods such can improve image quality and reduce the need for additional imaging tests

[0103] Noise reduction can be performed using any method known in the art. For example spatial domain methods aim to remove noise by calculating the grey value of each pixel based on the correlation between pixels / image patches in the original image. In general, spatial domain methods can be divided into two categories: spatial domain filtering and variational denoising methods.

[0104] Spatial domain filtering involves the application if linear filters or non-linear filters. Linear filters can be used to remove noise in the spatial domain, but may fail to preserve image textures. Mean filtering has been adopted for Gaussian noise reduction, however, it can over-smooth images with high noise. To overcome this Wiener filtering can be employed, but may blur sharp edges. By using non-linear filters, such as median filtering and weighted median filtering, noise can be suppressed without any identification. As a nonlinear, edge-preserving, and noise-reducing smoothing filter, Bilateral filtering is useful for image denoising. The intensity value of each pixel is replaced with a weighted average of intensity values from nearby pixels.

[0105] Spatial filters make use of low pass filtering on pixel groups with the statement that the noise occupies a higher region of the frequency spectrum. Normally, spatial filters eliminate noise to a reasonable extent but at the cost of image blurring, which in turn loses sharp edges.

[0106] Variational denoising methods use image priors and minimize an energy function (E) to calculate the denoised image. Useful prior models include gradient priors, non-local selfsimilarity (NSS) priors, sparse priors, and low-rank priors. There are a number of variational denoising methods including total variation regularisation, and non-local regularisation.

[0107] Total variation regularisation (TV) regularisation is based on the statistical position that natural images are locally smooth and the pixel intensity gradually varies in most regions. This method can not only effectively calculate the optimal solution but also retain sharp edges.

[0108] Non-local regularisation is useful with images with high noise levels in which correlations of neighbouring pixels are seriously disturbed by high level noise. Non-local means (NLM) uses the weighted filtering of the NSS prior to achieve image denoising. The basic idea is to build a pointwise estimation of the image, where each pixel is obtained as a weighted average of pixels centred at regions that are similar to the region centred at the estimated pixel. That is the weights are determined based on the similarity between the patch around the pixel and patches around other pixels in the image. The similarity is typically measured using the mean squared error (MSE) or Euclidean distance between the patches.

[0109] The frame can then be converted to grayscale and smoothed using a Gaussian blur filter. In some embodiments the grayscale frame is thresholded for example using the Otsu method to obtain a binary image. Morphological operations, including dilation and erosion, can also be applied to the binary image to remove noise and refine the contours.

[0110] NLM is different from local denoising methods as it can make full use of the information provided by the given images, which can contain a large amount of noise.

[0111] It is envisaged that considering the merits of TV and NLM methods, an adaptive regularization of NLM (R-NL) combining NLM with TV regularization may be used.

[0112] Other noise reduction methods can also be used. For example sparse representation, a technique that requires that each image patch can be represented as a linear combination of several patches from an over-complete dictionary (for example the same valve imaged over multiple cardiac cycles). As a dictionary learning method, the sparse representation model can be learned from a dataset. Since the learned dictionaries can more flexibly represent the image structures sparse representation models with learned dictionaries perform better than designed dictionaries.

[0113] Different from the sparse representation model, a low-rank-based model formats similar patches as a matrix. Each column of this matrix is a stretched patch vector. By exploiting the low-rank prior of the matrix, this model can effectively reduce the noise in an image Low-rank approaches for the reconstruction of noisy data can be grouped in two categories: methods based on low rank matrix factorization and those based on nuclearnorm minimization (NNM). Methods in the first category typically approximate a given data matrix as a product of two matrices of fixed low rank.

[0114] Alternatively, methods based on NNM aim to find the lowest rank approximation of an observed matrix. For NNM the weights of each singular value are equal, and the same threshold is applied to each singular value, however different singular values have different levels of importance. Hence, on the basis of the NNM, a weighted NNM model (WNNM) which can adaptively assign weights to singular values of different sizes and denoise them using a soft threshold method.

[0115] Although most low-rank minimization methods (especially the WNNM method) outperform previous denoising methods, the computational cost of the iterative boosting step is relatively high.

[0116] Other imaging techniques that may be used include transform techniques (originally developed from the Fourier transform) but may include cosine transform, wavelet domain methods, and block-matching and 3D filtering (BM3D). Transform domain methods rely on the fact that characteristics of image information and noise are different in the transform domain.

[0117] In contrast with spatial domain filtering methods, transform domain filtering methods first transform a noisy image to another domain, and then they apply a denoising procedure on the transformed image according to the different characteristics of the image and its noise. The transform domain filtering methods can be subdivided according to the chosen basis transform functions, which may be data adaptive or non-data adaptive.

[0118] An extension of the NLM approach, BM3D is a two-stage non-locally collaborative filtering method in the transform domain. In this method, similar patches are stacked into 3D groups by block matching, and the 3D groups are transformed into the wavelet domain. Then, hard thresholding or Wiener filtering with coefficients is employed in the wavelet domain. Finally, after an inverse transform of coefficients, all estimated patches are aggregated to reconstruct the whole image.

[0119] A further type of denoising that can be applied is convolutional neural network (CNN)-based denoising. The variational denoising methods are model-based optimisation schemes, which find optimal solutions to reconstruct a denoised image. However, such methods usually involve time-consuming iterative inference. In contrast CNN-based denoising methods attempt to learn a mapping function by optimizing a loss function on a training set that contains degraded and clean image pairs.

[0120] The so-called deep learning denoising methods may also be used and are typically based on CNNs and generally the model applies a residual learning formulation to learn a mapping function, and combines it with batch normalization to accelerate the training procedure while improving the denoising results.Tracking of leaflet motion

[0121] The methods described herein involve tracking of angular leaflet motion, with adjustment for flow state, to measure the stiffness of the aortic valve and consequent hydraulic load. The tracking can be performed in a number of ways such as by contouring, multipoint selection and tracking, and a Euclidean distance method. Each of which are described below.Contouring

[0122] In order to ensure only the valve leaflet is considered, the contour of the valve leaflet in each image is defined (see Figure 1). In this context the contour is an outline representing or bounding the shape or form of the valve leaflet.

[0123] The contours can be defined manually or can be defined using any means know in the rat such as those embodied in commercially available software.

[0124] A contour corresponds to a discontinuity in an image and contour detection may involve identification of luminescence change, texture change, perceptual grouping, and illusory contour, respectively. In the first and second cases, the contours arise from regions boundaries, whereas in the third and fourth cases, global relations give rise to the perception of a contour.

[0125] Contours may be detected using pixel-based approaches, edge-based approaches, region-based approaches or approaches using 'deep networks'.

[0126] In pixel-based approaches discontinuity in grey-scale intensity is the primary feature used. By convolving the image with local filters, this feature can be detected as the pixels with the highest gradient magnitude in their local neighbourhood. To extract the discontinuity feature arising from step edge, several linear filters are introduced, such as Sobel, Prewitt and Canny. In some embodiments quadratic filters are preferred. The discontinuity feature may be supplemented by approaches based on brain-inspired features and or natural features.

[0127] Brain-inspired features imitate human perception in contour detection are known in the art as are approaches based on natural features extracted from the images on the basisof, for example brightness and colour and which involve feature selection and determination.

[0128] Edge-based approaches are based on contour related edges or curves provided by edge detectors or human prior experience, aiming to determine whether they are contained in a certain contour. In many of these approaches, global optimization is conducted, whereby information from the entire image can be taken into consideration simultaneously.

[0129] In one embodiment contours are defined as a line joining all the points along the boundary of an image that have the same or similar intensity, for example the “find contour” function provided by OpenCV. This function analyses the image and identifies continuous regions of interest, for example the aortic valve. The resulting contours provide the necessary boundary information of the image.

[0130] In some embodiments the contour area and other relevant properties are also computed. In one embodiment contour area represents the total number of pixels enclosed by the contour, which can be used as a measure of the size of the object. The contour area value is then used to track and display only the valve leaflet across frames (i.e across the cardiac cycle). The contours can be provided with IDs to keep track of the valve across frames.

[0131] A bounding rectangle may also drawn to surround the boundary of the detected contours. Figure 1 shows the representation of the contour detection process.Leaflet displacement

[0132] As noted above displacement is change in position of an object such as a valve leaflet and can be defined by Equation 1. That is, displacement is the difference between the starting point of the valve leaflet (e.g. when closed) and its end point (e.g. when the valve is as open as possible).

[0133] In the methods described herein the displacement of one or more valve leaflets is determined from the images. Displacement can be determined form any point or combination of points on the leaflet. The point may be located on the free margin of the leaflet of may be a point. Alternatively, as valve leaflets are generally convex a centroid may be used as the point as centroid of a convex always lies in the object and a centroid can be calculated from data in the images of the leaflet. For example, a centroid, or center of an image is the arithmetic mean position of all the points in the surface of the image. In one embodiment, a centroid represents the center point of the image and is calculated as the average of the pixel coordinates of all the pixels in the image.

[0134] The moments of each contour are calculated to determine the centroid coordinates. Moments are statistical measures that describe the spatial distribution of the pixels in the contour. The moments can be used to determine various properties of the contour, including the centroid coordinates. In one embodiment the centroid represents the centre point of the contour and in one embodiment is calculated as the average of the pixel coordinates of all the pixels in the contour.

[0135] By analyzing the movement of the centroid across a cardiac cycle (in a series of images) the displacement of the centroid can be observed and measured. Accordingly, in one embodiment the displacement of a valve leaflet is calculated from a centerline of the leaflet or a centroid of the leaflet.

[0136] For example, a centerline or reference line representing the central position of the leaflet's major axis (i.e. the middle of the leaflet and in the direction of the longest axis of the leaflet) is used in one embodiment. The centerline's angle is compared to a static reference line. As exemplified herein, in one embodiment this is a line placed through the ipsilateral (nearest same side) aortic wall. A line orthogonal to that aortic wall is used to represent the closed valve position. In this regard Figure 7 shows a valve leaflet progressively opening from the closed position.

[0137] One embodiment of the methods disclosed herein involves detecting the contours, creating centerlines and calculating the angles automatically frame by frame across the image sequence, that is the method involves reference lines, or parallel equivalents, to determine angular displacement. By factoring in time between frames (which itself is calculated from image sampling frequency), angular velocity, and angular acceleration can be determined.

[0138] The centroid shown in Figure 2 is part of the contouring process and it is envisaged that in other embodiments the centroid and arc / radius and trigonometry could be used to determine displacement angle. In this embodiment the method would involve tracking the centroid's arc and reference to a fulcrum (also using centerlines).

[0139] Features such as angular displacement, cumulative displacement, velocity, and acceleration can also be calculated from the images.

[0140] For example the displacement or change in angle of each contour can be determined by calculating the vectors created by drawing lines from the centre of the bounding shape (such as a rectangle) to two or more reference points on the shape.

[0141] The reference points may correspond to the top left corner and the bottom right corner. The lines are generally parallel to the aortic wall and the aortic valve, are then used to compute the change in angle or displacement for each contour.

[0142] As the image series (and frame) progresses, the lines change based on the movement of the wall and the valve. Figures 2a, 2b and 7 provide visual representations of the vector points and reference lines in frames. The dot product and magnitudes of the vectors are then used to calculate the angle in radians. The angle is then converted to degrees and adjusted to the desired range.

[0143] The process of denoising, grayscale conversion, smoothing, thresholding, morphological operations, contour extraction, centroid calculation, and feature calculation steps are repeated for each frame in a series of images (or in a video).

[0144] The contour ID pertaining to the aortic valve can be recorded and stored with the respective values of the ratios described above.Multi-point selection and tracking method

[0145] An alternative to contouring is a multipoint (for example 3-point) point selection system illustrated in Figure 8, whereby multiple points on the leaflet and aortic wall (such as the leaflet tip, leaflet base and aortic wall) are selected. The angle between the points can then be calculated for example as illustrated in (Figure 9). The motion of the leaflets through the cardiac cycle is tracked by reference to the points to permit manual optimization of tracking, frame by frame, at a user's discretion.

[0146] In this approach, the angle between the leaflet and aortic wall is tracked over time and through the cardiac cycle. This permits calculation of leaflet displacement angle, displacement velocity and displacement acceleration.

[0147] In one embodiment, the right coronary cusp motion can be used to make the calculations. However, in other embodiments multiple cusps may be used. In the exemplified embodiments transthoracic echocardiographic images from the parasternal long axis view are used. However, this approach is equally applicable to other views (e.g. apical views or views obtained at transoesophageal echocardiogram, or views obtained by non- echocardiographic imaging modalities) that provide moving images of aortic valve leaflets.Euclidean distance method

[0148] Another alternative, and one that permits a complementary and more detailed approach to measuring angular leaflet metrics, involves the use of landmarks to track motion of the leaflet tip, leaflet base and aortic wall. Additional landmarks can optionally beused to track the mid leaflet and the opposing leaflet base (for example as illustrated in Figure 10). This permits additional capture of leaflet motion against a reference annular plane by Euclidean distance and permits assessment of intrinsic leaflet deformation or leaflet flexibility (base to mid to tip relative motion) which is shown herein to change with disease severity, and precedes overall leaflet restriction. In this regard it is noted that the first changes are reduction in the relative motion of the tip to the mid leaflet, due to leaflet stiffening, before the entire leaflet becomes restricted.

[0149] In this method vector geometry is used to track and compute relative motion of these key structures.Relative cardiac motion

[0150] The position of these landmarks, both intrinsically fixed (aortic wall) and mobile (leaflets), will change throughout the cardiac cycle owing to the overall motion (and marginal rotation) of the heart with ventricular contraction and relaxation, and the fixed position of the ultrasound probe.

[0151] Accordingly, when calculating the true 2D movement of landmarks (as a distance), the position / coordinates of the intrinsically mobile structures need to be adjusted for this motion. This is achieved by referencing the leaflet coordinates to a relatively fixed structure (aortic wall, to subtract the overall cardiac motion), and then adjusting for cardiac rotation, which can be estimated by tracking the change in angle of the vector transecting the upper and lower hinge points (i.e. coaxial to the left ventricular outflow tract).

[0152] First, the vectors between landmarks are calculated. These are simply calculated by subtracting the x- and y-coordinates of one landmark from another.Equation 12: BA = (Bx - Ax)

[0153] Since a vector represents both magnitude and angle / direction, if one of these components is known the other can be derived. The magnitude of a vector (in our case, the distance between two landmarks) can be determined by calculating the square root of the sum of the square of its components. This logic can also be applied to determine the distances individual landmarks travel between timepoints (e.g. systole and diastole).Equation 13: distance = (( 2 - I)2+ (Y2 - Yl)2)

[0154] In this context the dot product is a scalar output (i.e. not another vector). The dot product is equal to the length / magnitude of one vector multiplied by the length of the other vector, multiplied by the cosine of the angle between them.Equation 14: dp = |yi| |y2| * cos(0)Angle calculation

[0155] Angels may be calculated by any means known in the art. For example, in one embodiment the angles may be calculated as follows: calculate_angle <- function (xl , yl , x2 , y2 , x3 , y3 ) {BA <- c (xl - x2 , yl - y2 )BC <- c (x3 - x2 , y3 - y2 ) dot_product <- sum (BA * BC ) magnitude_BA <- sqrt ( sum (BA * BA) ) magnitude_BC <- sqrt ( sum (BC * BC ) ) angle_radians <- acos ( dot_product / (magnitude_BA * magnitude_BC) ) angle_degrees <- angle_radians * ( 180 / pi ) return ( angle_degrees )Leaflet flexibility

[0156] The intrinsic shape of the leaflet can also be quantified by measuring the internal leaflet angle, with the mid-leaflet acting as the apex of a triangle opposing the hinge and tip. The change in this angle, between the closed and open state, reflects how flexible the leaflet is during motion.

[0157] Pathophysiologically, aortic valve calcification and restriction tend to occur over many years in a 'base-to-tip' pattern and the inventors have found that the intrinsic flexibility of the leaflet deteriorates as the disease progresses. Specifically, leaflet flexibility inversely correlates with aortic stenosis severity, as quantified by conventional haemodynamic criteria.Method assessment

[0158] The consistency and reliability of the automated feature calculation (above) was assessed by conducting a comparative analysis with manually obtained measurements. In the manual process, a user selects specific points of interest on video frames, for examplethe open and closed position of the aortic valve. Subsequently, corresponding angles are calculated based on these user-selected points.

[0159] Typically following steps are performed manually:1. Initialization of variables and functions: A set of lists is created to store the coordinates of the selected points. A trigger event is set up as a callback function to enable users to select points on the video frames.2. Trigger function: When the user clicks on a video frame, the trigger function is activated. It saves the selected points' coordinates and visually represents them on the frame by drawing lines and circles3. Angle function: The gradient (slope) between two selected points is calculated. The angle function then determines the angle formed by the last three selected points. It utilizes the gradient values to compute the angle in radians and converts it to degrees.4. Video Processing Loop: After loading the video and displaying each frame, the angle function is invoked. If three points have been selected and the total number of selected points is a multiple of three, the program calculates the angle using the angle function. Finally, the frame number and angle values are subsequently stored in a separate file.

[0160] The frames used in the manual process capture the closure of the aortic valve and the highest point at which the valve opens. By calculating the respective displacements or changes in angles, the cumulative displacement can be derived by subtracting the angles.

[0161] As can be seen in the examples the methods described herein are consistent with the manual displacement analyses as seen in Figures 5 and 6 indicating that automation of the methods for example using software to calculate the D:Q ratio (Equation 2), the D:M ratio (Equation 3) , the (D:Q):EL ratio (Equation 7) or the (D:M):EL ratio (Equation 8) provides an accurate assessment of these ratios.

[0162] In the context of assessing the severity of valvular dysfunction (particularly aortic stenosis) the methods described herein provide insights on the motion dynamics and characteristics of the valve leaflets which can be used in the assessment of aortic stenosis severity. In this regard, one or more of the ratios can be compared to a reference ratios or ranges for healthy subjects, in much the same way that blood lipids or cholesterol are assessed against ranges for healthy patients.

[0163] For example, and with reference to Figures 11-25, the methods described herein can quickly identify a subject with mild, moderate, or severe valvular stenosis by comparingone or more metrics or ratios from the subject with reference values or reference ranges from a group of subjects known to have mild, moderate, or severe valvular stenosis, for example aortic stenosis.

[0164] The methods described herein can be performed using images from any valve, for example the tricuspid valve, the pulmonary valve, the mitral valve, or the aortic valve.

[0165] The methods described herein may be performed manually, or using software or both. For example, in a preferred embodiment the methods are performed on a system comprising software to instruct a processor to carry our the methods. In one embodiment a user may utilized the software to manually define or adjust the contours, points or landmarks in one or more images before the software causes a processor to perform the methods.

[0166] A skilled person will be able to implement the methods described herein by software that can be written by an ordinarily skilled programmer.Systems

[0167] The methods described herein can be performed on a system, such as a conventional echocardiography system, that is configured to provide a series of images of a valve leaflet across at least one cardiac cycle. The system will also include a user interface to allow the user to define contours, points or landmarks manually. Preferably, the system will provide a means to automatically contours, points or landmarks and these may optionally be adjusted manually by a user in one or more images of the series. The system will also include a processor to calculate the displacement of a leaflet as set out in any of the embodiments described herein. The processor can further configured to calculate one or more of the following metrics for the subject: valve leaflet displacement; valve leaflet velocity; valve leaflet acceleration; and valve leaflet flexibility. In addition or as an alternative the processor can calculate one or more of the following ratios for the subject: valve leaflet displacement to blood flow ratio (D:Q); valve leaflet displacement to blood momentum ratio (D:M); (D:Q) to Energy Loss ratio ((D:Q):EL); and (D:M) to Energy Loss ratio ((D:M):EL).

[0168] In some embodiments, the system or processer can also be configured to obtain blood flow, blood momentum and Energy Loss measurements from the series of images; or can be configured to receive blood flow, blood momentum and Energy Loss measurements from an external source.

[0169] The system may further comprise a display configured to generate a visualization of one or more of the images, contours and boundary lines over time and highlight the displacement of the leaflet.ExamplesExample 1 : Patient Cohort

[0170] The patient cohort consisted of 88 patients undergoing TAVI, details of the patient cohort are presented in Table 1.Table 1: Patient detailsCAD: coronary artery disease; CKD: chronic kidney disease, Cog: cognitive; HTN: hypertension;NYHA: New York Heart Association; OSA: obstructive sleep apnoea; PAD: peripheral arterialdisease; Revasc: revascularisation; STS: Society for Thoracic Surgeons Risk Score; IQR: Interquartile range.

[0171] Echocardiography databases were queried for patients undergoing TAVI. Analysis of raw echocardiographic images was performed at the Heart Valve Disease and Artificial Intelligence Laboratory at Victor Chang Cardiac Research Institute to determine conventional chamber and valvular measurements as per current guideline recommendations (Baumgartner H., et al. Journal of the American Society of Echocardiography 2017;30(4):372-92).

[0172] The present study exclusively employed the parasternal long axis view of the echocardiogram to investigate severe AS.Example 2: Features of Echocardiography

[0173] Left ventricular global longitudinal strain (GLS) analysis was performed using TOMTEC Arena software (TOMTEC Imaging Systems, Germany), using the apical 2, 3 and 4-chamber views (Amzulescu et al Eur Heart J Cardiovasc Imaging. 2019;20(6):605-19; and Yang et al JACC Cardiovasc Imaging. 2018;11(8):1196-201)

[0174] Studies were only included for analysis if the left ventricular endocardium could be visualized, and endocardial tracking was accurate in the apical 2,3 and 4-chamber views throughout the cardiac cycle. Aortic valve stenosis severity was assessed using aortic valve area (AVA) (as determined by continuity equation), mean and peak transvalvular gradient and dimensionless index (DI). Valvular hydraulic load was assessed using energy loss, as described by Garcia et al (Circulation 2000;101(7):765-71).

[0175] Body surface area was not available so total (rather than indexed) energy loss was evaluated. As flow conditions are important to metrics of aortic valve stenosis severity, transvalvular flow rate (Q) was also measured as the ratio of stroke volume to ejection time (Namasivayam et al Journal of the American Society of Echocardiography 2020; 33(4): 449- 51; and Namasivayam et al. J Am Coll Cardiol 2020;75(15): 1758-69))

[0176] The above were the baseline features upon which the algorithmic features were analysed and cross-examined to determine the similarities between the two methods.Example 3: Algorithm

[0177] The combination of computer vision and video processing techniques enable the extraction of meaningful information and insights from the echocardiography cases relevant to aortic valve motion dynamics. The development process involved several steps, including video pre-processing, contour extraction, centroid calculation, mathematical operations, along with manual feature extraction. These techniques assist with the automated detectionand tracking of the aortic valve based on various motion properties of the contour. Quantifiable features such as angular displacement, leaflet velocity, leaflet acceleration, peak cumulative displacement, peak leaflet velocity, and peak acceleration are computed across a single cardiac cycle. These derived values are employed for subsequent statistical and comparative analysis.3.1 Pre-processing

[0178] To enhance the quality of the frame and reduce noise, denoising techniques were applied using the fast non-local means (NLM) denoising algorithm (Buades et al. 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’05). IEEE; n.d. p. 60-5.). Each pixel in the image was replaced by a weighted average of nearby pixels, where the weights were determined based on the similarity between the patch around the pixel and patches around other pixels in the image. The similarity was measured using the mean squared error (MSE) or Euclidean distance between the patches. The frame was then converted to grayscale and smoothed using a Gaussian blur filter. The grayscale frame was thresholded using the Otsu method to obtain a binary image. Morphological operations, including dilation and erosion, were applied to the binary image to remove noise and refine the contours.3.2 Contour detection

[0179] Contours were extracted from the frame using the “find contour” function provided by OpenCV. This function analyses the image and identifies continuous regions of interest, representing the aortic valve in our case. The resulting contours provide the necessary boundary information of the image. The moments of each contour are calculated to determine the centroid coordinates. Moments are statistical measures that describe the spatial distribution of the pixels in the contour. The moments can be used to determine various properties of the contour, including the centroid coordinates. The centroid represents the centre point of the contour and is calculated as the average of the pixel coordinates of all the pixels in the contour. The contour area and other relevant properties are also computed. Contour area represents the total number of pixels enclosed by the contour, which can be used as a measure of the size of the object. The contour area value is then used to track and display only the aortic valve across frames. The contours are provided with IDs to keep track of the aortic valve across frames. Finally, a bounding rectangle is also drawn to surround the boundary of the detected contours. Figure 1 shows the representation of the contour detection process.3.3 Feature calculation

[0180] In this study, motion features such as angular displacement, cumulative displacement, velocity, and acceleration (determined using echocardiographic approaches) were identified as providing positive insights on the severity of aortic stenosis.

[0181] The displacement or change in angle of each contour is determined by calculating the vectors created by drawing lines from the centre of the bounding rectangle to two reference points on the rectangle. These reference points specifically correspond to the top left corner and the bottom right corner. These two lines which are parallel to the aortic wall and the aortic valve, are used to compute the change in angle or displacement for each contour. As the frame progresses, the lines change based on the movement of the wall and the valve. Figure 2a and 2b provides visual representations of the vector points in frames. The dot product and magnitudes of the vectors are then used to calculate the angle in radians. The angle is then converted to degrees and adjusted to the desired range.3.4 Frame-by-Frame Analysis:

[0182] The above methodology is then applied iteratively to each subsequent frame of the video. The process of denoising, grayscale conversion, smoothing, thresholding, morphological operations, contour extraction, centroid calculation, and feature calculation steps are repeated for each frame. The resulting frames are visualized to display the detected contours, bounding rectangles, and centroids. The contour ID pertaining to the aortic valve is recorded stored in conjunction with the respective values of the novel features. This collated information is intended for subsequent processing and analysis.3.5 Manual features calculation

[0183] The primary objective of this process is to determine the consistency and reliability of the automated feature calculation by conducting a comparative analysis with manually obtained measurements. It facilitates user interaction by enabling the manual selection of specific points of interest on video frames, specifically focusing on the opening and the closed position of the aortic valve. Subsequently, corresponding angles are calculated based on these user-selected points. The following steps outline the procedure:1. Initialization of variables and functions: A set of lists is created to store the coordinates of the selected points. A trigger event is set up as a callback function to enable users to select points on the video frames.2. Trigger function: When the user clicks on a video frame, the trigger function is activated. It saves the selected points' coordinates and visually represents them on the frame by drawing lines and circles3. Angle function: The gradient (slope) between two selected points is calculated. The angle function then determines the angle formed by the last three selected points. It utilizes the gradient values to compute the angle in radians and converts it to degrees.4. Video Processing Loop: After loading the video and displaying each frame, the angle function is invoked. If three points have been selected and the total number of selected points is a multiple of three, the program calculates the angle using the angle function. Finally, the frame number and angle values are subsequently stored in a separate file.

[0184] The frames used in this process specifically capture the closure of the aortic valve and the maximal valve opening. By calculating the respective displacements or changes in angles, the cumulative displacement can be derived by subtracting the angles. Similarly, multiple angles in between the minimal and maximal opening of the valve can be measured.Example 4: Results

[0185] The dynamics of the aortic leaflets were investigated to develop a metric that combines leaflet displacement and flow rate. This was based on the position that a higher displacement to flow ratio would result in increased energy loss, and leaflets experiencing higher energy loss would exhibit a lower displacement to momentum ratio.

[0186] The findings herein support both positions. It was observed that leaflets experiencing greater energy loss, which indicates higher hydraulic load and true severity, demonstrated a lower displacement to momentum ratio (Figure 3). This indicates that these leaflets had limited movement and showed reduced displacement for a given flow or opening force.

[0187] Additionally, it was found that there was a higher occurrence of energy loss when there was less displacement per unit flow (Q) (Figure 4), as evident in the momentum analysis.

[0188] The above results also hold true for the manual displacement analyses as seen in Figures 5 and 6.

[0189] These results illustrate a direct relationship between leaflet restriction and energy dissipation. The more restricted the leaflet movement, the higher the energy loss. This relationship was evident in both the displacement to momentum ratio and the displacement to flow ratio. Leaflets that were constrained in their motion experienced increased energy dissipation, indicating that the restriction of leaflet movement plays a crucial role in energy loss.Example 5: Expanded patient cohort

[0190] Example 2 describes patients with severe disease, i.e. those requiring TAVI. In contrast this example considers a total of 223 subjects (including the 89 subjects described in Example 2). The subjects are classified as normal (i.e. having normal aortic valves (n=65), those with severe disease (n=89), moderate stage disease (n=69).

[0191] Details of the patients having severe and moderate disease are summarized in Table 2Table 2: Patients with moderate and severe aortic stenosis assessed with the multi-point approachMG: mean transvalvular gradient (in mm Hg); PG: pressure gradient across the valve during the entire systolic ejection period (in mm Hg); AVA: Aortic valve area (in cm2); DI: dimensionless index, the ratio left ventricular out flow tract velocity and vena contractor velocity; q: transvalvular flow rate (in mL / sec); disp / q leaflet displacement adjusted to tansvalvular flow (in degrees / ml / sec)

[0192] A subset of 65 patients (of the 223 subjects) were assessed using the methods described herein and the results are presented in Figures 19-25. Details of these patients are shown in Table 3, noting that the two patients designated as having an 'unknown' haemodynamically determined phenotype are excluded from analysis.Table 3: Details of patients with nil, mild, moderate or severe aortic stenosisValues for each characteristic are median (IQR); n (%)

[0193] In Tables 2 and 3 the patients are a subset of the cohort of 223 discussed above. Those with clinically adjudicated severity were reclassified by the methods described herein according to conventional haemodynamic parameters.

[0194] Both the multi-point and Euclidean distance approaches (described above) are able to easily discriminate between these groups. Not only is leaflet flexibility described the methods were used to classify these patients into groupings in a matter of seconds rather than having to perform detailed manual calculations which can take up to 30 minutes per patient.

[0195] When compared to mean gradient, the 3-point method shows worsening leaflet motion with worsening disease stage (Figure 11).

[0196] There is worsening energy loss with worsening leaflet restriction (Figure 12), and there is clear discrimination when adjusting to transvalvular flow. That is, the more energy lost for a given degree of leaflet restriction per unit flow (as leaflet motion is dependent on intrinsic stiffness and flow conditions) indicates more severe disease (Figure 13).

[0197] The 5-point method also shows discriminatory ability to separate normal, moderate and severe cases based on angular motion of the leaflet tip and mid-portion (Figures 14 and 15).

[0198] Data show that in moderate stage disease, there is early loss of leaflet flexibility due to relative stiffness of the tip and mid portion (Figure 16).

[0199] The above approaches were used to assess normal subjects and those with previously diagnosed, mild, moderate, low gradient aortic stenosis (LGAS) or high gradient aortic stenosis (HGAS).

[0200] LGAS is defined as a concomitance of a small aortic valve area of <1.0 cm2 (which is consistent with severe stenosis) and a low gradient of <40 mm Hg (which is consistent with non-severe stenosis). HGAS is defined as a pressure gradient of >40 mm Hg.

[0201] In the Figures discussed below pHoim-adj indicates that the p values have been adjusted using Holm's procedure. In one embodiment the Holm procedure, in a stepwise way, computes the significance levels depending on the P value based rank of hypotheses. For the / 1hordered hypothesis H , the specifically adjusted significance level is computed:Equation 15:

[0202] The observed P value p^of hypothesis H is then compared with its corresponding »'(,) for statistical inference; and each hypothesis will be tested in order from the smallest to largest P values ( / 7^, ... , / 7(m)). The comparison will immediately stop when the first p(i) > a'^ is observed (1=1, ..., I) and hence all remaining hypotheses of H® (j=i, ...,rri) are directly declared non-significant without requiring individual comparison. Alternatively, the procedure can involve directly computing the adjusted P value for each hypothesis.

[0203] Both the multi-point and Euclidean distance approaches (described above) are able to easily discriminate between the moderate, severe and normal (nil disease) groups. The methods were used to classify these patients into groupings in a matter of seconds rather than having to perform detailed manual calculations which can take up to 30 minutes per patient.

[0204] When compared to mean gradient, the 3-point method shows worsening leaflet motion with worsening disease stage (Figure 11). There is worsening energy loss with worsening leaflet restriction (Figure 12), and there is clear discrimination when adjusting to transvalvular flow. That is, the more energy lost for a given degree of leaflet restriction per unit flow (as leaflet motion is dependent on intrinsic stiffness and flow conditions) indicates more severe disease (Figure 13).

[0205] The 5-point method also shows discriminatory ability to separate normal, moderate and severe cases based on angular motion of the leaflet tip and mid-portion (Figures 14 and 15).

[0206] Figure 14 shows the leaflet tip angular change using the Euclidian (5 point) method in a subset of patients. Groups of subjects were compared using the Games-Howell pairwise test and the difference in leaflet tip angular motion is shown, fl.mean refers to the mean angular change at the leaflet tip in degrees. Leaflet angular motion for each subject represented by a dot. The bars shown are significant. For this figure loge(BF01) = -1.7, 0.71 and FWelch(2, 15.54) = 6.12, p = 0.01 ,

[0207] Figure 15 shows the mid leaflet angular motion using the Euclidian (5 point) method in a subset of patients. Groups of subjects were compared using the Games-Howell pairwise test and the difference in leaflet tip angular motion is shown, fl.mean refers to the mean angular change at the mid-leaflet in degrees. Leaflet angular motion for each subject represented by a dot. The bars shown are significant. For this figure loge(BF01) = -7.28, R^ aysean = 0.55, CI™ [0.31 , 0.71], r^chy= 0.71 and FWelch(2, 15.78) = 18.88, p = 6.51e-05, 0.66, Cl95% [0.36, 1.00], nofts= 27

[0208] The data show that in moderate stage disease, there is early loss of leaflet flexibility due to relative stiffness of the tip and mid portion (Figure 16). Figure 16 shows the change in intrinsic leaflet flexibility in a subset of patients. Groups of subjects were compared using the Games-Howell pairwise test and the difference in leaflet flexibility is shown. Relative flexibility is calculated by dividing the change in intrinsic angle (from closed to open states) by the baseline intrinsic angle (closed state), multiplied by 100 (to become a percentage), mean refers to the mean percentage. Leaflet flexibility for each subject represented by a dot.. — posteriorHnTThe bars shown are significant. For this figure loge(BFOi) = -12.35, R2Baysean = 0.48, CIg5o / o[0.30, 0.61], razuschy= 0.71 and F Welch (2, 23.38) = 11.26, p = 3.75e-04, a)2= 0.44, Cl95% [0.16, 1.00], nofts= 51

[0209] Figure 17 show the 2D movement, in mm, of a single fixed point (leaflet hinge) between the closed- and open-states of the aortic valve. This does not represent total cardiac motion throughout the entire cardiac cycle, simply two dimensional movement of the annulus. For this Figure tstudent^Q) ~ 12.39, p — 7.03e-19, g Hedges ~ 1.5, Cl 95% [1.15, 1.84],0.71

[0210] Figure 18 shows the distribution of cardiac rotation values, as determined by the change in LVOT angulation from baseline. For this Figure tstudent(6Q) = 9.36, p = 9.88e-14, g>*^ = 1.13,-25.07,Of J* P-30. 5.16], r<!“hy= 0.71

[0211] Figure 19 shows the mid leaflet linear displacement using the landmark method in a subset of patients. Groups of subjects were compared using the Games-Howell pairwise test and the difference in mean linear displacement at the mid leaflet is shown, fl.mean refers to the mean displacement at the mid leaflet in mm. Mid leaflet displacement for each subject represented by a dot. The bars shown are significant. For this figure loge(BF01) = -28.8, ^ ysean = 0.675, CI™ [0.579, 0.745], r^chy= 0.707

[0212] Figure 20 shows leaflet tip linear displacement using the landmark method in a subset of patients. Groups of subjects were compared using the Games-Howell pairwise test and the difference in mean linear displacement at the leaflet tip is shown, (imean refers to the mean displacement at the leaflet tip in mm. Mid leaflet displacement for each subject represented by a dot. The bars shown are significant. For this figure loge(BF01) = -16, ^ ysean = 0.484, CI™ [0.337, 0.609], r^chy= 0.707

[0213] Figure 21 shows mid-leaflet angular displacement as assessed using the landmark method in a subset of patients. Groups of subjects were compared using the Games-Howell pairwise test and the difference in mean angular displacement at the mid-leaflet is shown, mean refers to the mean angular displacement at the mid-leaflet in degrees, relative to the aortic wall . The bars shown are significant. For this Figure loge(BF01) = -31.7,^ ysean = 0.706, CI™ [0.625, 0.769], r^chy= 0.707

[0214] Figure 22 shows leaflet tip angular displacement as assessed using the landmark method in a subset of patients. Groups of subjects were compared using the Games-Howell pairwise test and the difference in mean angular displacement at the leaflet tip is shown, mean refers to the mean angular displacement at the leaflet tip in degrees, relative to theaortic wall . The bars shown are significant. For this Figure loge(BF01) = -15.9,^Zysean = 0.486, CI™ [0.341, 0.608], r^chy= 0.707

[0215] Figure 23 shows mean angular displacement as assessed using the landmark method in a subset of patients. Groups of subjects were compared using the Games-Howell pairwise test and the difference in mean angular displacement at the leaflet tip and midleaflet in degrees, relative to the aortic wall is shown. The bars shown are significant. For this Figure loge(BF01) = -27, R^Bay^an = 0.654, CI™ [0.551 , 0.728], r^chy= 0.707

[0216] Figure 24 shows the absolute dynamic flexibility / deformation of the leaflet assessed in a subset of patients. The internal leaflet angle was measured, with the mid-leaflet acting as the apex of a triangle opposing the hinge and tip. Groups of subjects were compared using the Games-Howell pairwise test and the difference in angular displacement in degrees, is shown. The bars shown are significant. For this Figure loge(BF01) = -29.5, ^Baysean = 0-683, CI™ [0.589, 0.747], r^chy= 0.707

[0217] Figure 25 shows the relative dynamic flexibility / deformation of the leaflet assessed in a subset of patients. The the internal leaflet angle was measured, with the mid-leaflet acting as the apex of a triangle opposing the hinge and tip. Groups of subjects were compared using the Games-Howell pairwise test and the difference in angular displacement in degrees, is shown. The bars shown are significant. For this Figure loge(BF01) = -25.3, ^Baysean = 0-633, CI™ [0.525, 0.718], r^chy= 0.707

[0218] In relation to Figures 24 and 25 flexibility is defined by the change in internal angle of the leaflet between resting (closed) and open states. Absolute flexibility is the angle in degrees (i.e. the difference between resting and open states). Relative flexibility is the absolute flexibility as a proportion of the resting / closed angle.

[0219] Figure 26 shows the mid-leaflet angular displacement as a function of clinically adjudicated severity of aortic stenosis. Groups of subjects were compared using the Games-Howell pairwise test and the difference in angular displacement in degrees, is• — posterior shown. The bars shown are significant. For this Figure loge(BF01) = -226, R2Baysean = 0.822,0.707

[0220] Figure 27 shows mid-leaflet angular displacement as a function of severity of aortic stenosis assessed by haemodynamic parameters. Groups of subjects were compared using the Games-Howell pairwise test and the difference in angular displacement in degrees, is—^posteriorl7c shown. For this Figure loge(BF01) = -188, R2Baysean ~ 0.848, CI™ [0.83, 0.86], r^chy=0.707

[0221] Figure 28 shows flow adjusted angular displacement as a function of severity of aortic stenosis assessed by haemodynamic parameters. Groups of subjects were compared using the Games-Howell pairwise test and the difference flow adjusted angular. — posterior displacement in degrees / mL / sec is shown. For this Figure loge(BFOi) = -116, R2Baysean = 0.722,0.707

[0222] Figures 27 and 28 show that in a large expanded cohort of patients the methods described herein can classify patients according to the American Society of Echocardiography classification. This conventional classification requires the use of Doppler echocardiography whereas the present methods do not. That is, not only can the methods described herein classify based on clinical adjudication but can also rapidly classify into guideline based clinical groupings which is an advancement beyond clinical adjudication.

[0223] Accordingly, users of the methods described herein (especially those without detailed valve expertise) can classify into guideline based groups and then help determine optimal therapy, all within seconds, which is advantageous for cardiologists, general practitioners, emergency doctors, anaesthesia doctors, rural doctors etc. who do not have the expertise of those that specialise in the treatment and care of valvular disease.Example 4: Reliability

[0224] In the above examples the methods were performed by different readers who manually identified landmarks or points on an image before the remaining images were characterized automatically. Analysis of intra-class coefficient for intra- and inter-rater reliability between 2 readers who read 60 studies independently twice are set out in Table 4.Table 4. Inter and Inter-rater reliability statistics

[0225] These data indicate that the methods are highly reliable and reproducible and are not dependent on the reader (user).Example 5: Discussion

[0226] Accurate classification of AS severity has been a perpetual challenge for clinicians. Part of the difficulty lies in the inability to determine to what extent leaflet restriction (and therefore, hemodynamic severity) is driven by alterations in flow state, intrinsically related the function of the left ventricle, or by true disease of the aortic valve. Currently, leaflet motion dynamics do not inform diagnostic criteria or treatment decisions. Herein the inventors demonstrate a method for assessing the severity of valvular stenosis using the relationship between aortic leaflet motion and valvular hydraulic load. The pertinent findings the study include:• leaflet motion dynamics correlate with conventional metrics of aortic stenosis severity• Computer vision determination of aortic leaflet motion dynamics is feasible and reproducible

[0227] Based on these findings the inventors have employed computer vision techniques to investigate leaflet restriction and its implications in AS. By quantifying leaflet mobility and exploring its correlation with hydraulic load, AS diagnosis and patient outcomes can be improved.

[0228] The methods utilize analysis echocardiographic images to precisely measure the degree of leaflet restriction. This approach enables a more accurate assessment of the mechanical impairment caused by leaflet restriction, thereby enhancing the diagnostic precision and reliability of AS.

[0229] The findings presented herein illustrate that leaflet restriction contributes to increased hydraulic load in AS patients. This discovery fundamentally enhances understanding of AS pathophysiology and offers a more direct and accurate diagnostic approach. Unlike traditional methods that rely on indirect measurements, the vision analysis provides an objective assessment of the mechanical impairment caused by leaflet restriction.

[0230] The methods involve leaflet tracking which can be performed in a number of ways including use of angular displacement metrics, for example the 3-point selection system with automated and semi-automated tracking options (frame by frame assessment); or the 5-point Euclidean distance vector approach to also measure leaflet motion and intrinsic leaflet flexibility.

[0231] Importantly, the methods described herein can be used to identify moderate stage disease, instead of severe disease alone, compared to data from patients with normal valves (see Figures 11-25).

[0232] Moreover, the methods described herein allow retrospective analysis of existing echocardiographic images, enabling a re-evaluation of previous patient cases and contributing to the advancement of AS research and patient care.

[0233] In summary, the present study demonstrates the direct relationship between leaflet restriction and hydraulic load in AS patients. The methods not only provide new insights into aortic stenosis severity, but are able to easily classify patients into severity groups within seconds, instead of requiring detailed manual assessment, which can take up to 30 minutes per patient, which forms the current standard of care. Therefore, the methods provide a number of advantages including:• a substantial improvement in workflow efficiency for echocardiographic analyses and clinical assessments, in particular the present methods can be carried out far more quickly than conventional analyses (i.e. within seconds or minutes) .• an additional clinical metric to complement current criteria and aid diagnostic workup.• an ability to perform remote analyses, the images can be obtained remotely and sent elsewhere for analysis, alternatively the methods can be performed remotely• simplicity and reliability. The methods described herein can be simply implemented allowing relatively unskilled users to assess the severity of valvular disease, accelerating referral for patients needing prompt intervention.• rapid identification of patients that do and do not need intervention.• improved phenotyping and severity classification leading to improved patient outcomes.• a more accurate measure of true disease severity.• the current thresholds for what constitutes ‘severe’ valvular disease are associated with advanced disease state, often associated with irreversible left ventricular dysfunction. The methods described herein provide a physician with information to allow appropriate interventions and improve a patient's prognosis.

Claims

Claims:

1. A method of assessing the severity of valve stenosis in a subject comprising a) obtaining from images of at least one valve leaflet of the subject, one or more of the following metrics for the subject: i. valve leaflet displacement; ii. valve leaflet velocity; iii. valve leaflet acceleration; and iv. valve leaflet flexibility b) comparing one or more of the metrics to a reference value or reference range for a healthy subject, a subject with moderate valve stenosis, or a subject with severe valve stenosis to indicate the severity of valve stenosis in the subject.

2. A method of assessing the severity of valve stenosis in a subject comprising a) obtaining from images of at least one valve leaflet of the subject, one or more of the following ratios for the subject: i. valve leaflet displacement to blood flow ratio (D:Q); ii. valve leaflet displacement to blood momentum ratio (D:M); iii. (D:Q) to Energy Loss ratio ((D:Q):EL); iv. (D:M) to Energy Loss ratio ((D:M):EL); v. valve leaflet displacement to stroke volume (D:SV); and vi. valve leaflet displacement to volume index (D:SVI). b) comparing one or more of the ratios to a reference value or reference range for a healthy subject, a subject with moderate valve stenosis, or a subject with severe valve stenosis to indicate the severity of valve stenosis in the subject.

3. The method of claim 1 or 2, wherein valve leaflet displacement is angular displacement.

4. The method of any one of claims 1 to 3, wherein the displacement is calculated from a series of images of the at least one valve leaflet throughout the cardiac cycle.

5. The method of claim 4, wherein a contour of the valve leaflet is determined in each image in the series.

6. The method of claim 5, wherein a centroid of the valve leaflet is identified in each image from the contour.

7. The method of claim 6, wherein the centroid is the average of the pixel coordinates of all the pixels in the contour.

8. The method of any one of claims 1 to 7, further comprising a) applying a boundary shape around the contour in a plurality of images in the series; and b) in each of the plurality of images applying reference lines extending from the centre of the boundary shape wherein a first reference line is generally parallel with the aortic wall; and a second reference line is generally parallel to the valve leaflet, c) measuring the maximum change in angle between the reference lines across a cardiac cycle to obtain the displacement.

9. The method of claim 8, wherein the plurality of images include an image when the valve is in a closed position and an images when the valve is in a fully open position.

10. The method of claim 9, wherein the plurality of images include images of the valve in multiple positions between and including the fully closed and fully open positions.

11. The method of claim 4, further comprising selecting multiple points on the at least one leaflet and aortic wall in one of the images, wherein at least one of the points is at or near the leaflet base.

12. The method of claim 11, wherein the multiple points include a point at or near the leaflet tip, and a point on the aortic wall.

13. The method of claim 11 or 12, further comprising calculating the angle between the points on the leaflet and the points on the aortic wall wherein the point at or near the leaflet base is the vertex of the angle.

14. The method of claim 13, wherein the points are manually positioned with reference the leaflet and / or aortic wall in at least some images.

15. The method of claim 14, further comprising positioning of addition points between the points on the leaflet and the point at or near the leaflet base; and / or between the points on the aortic wall and the point at or near the leaflet base.

16. The method fo any one of claims 11 to 15, wherein the points may be added in any of the images17. The method of claim 4, further comprising selecting multiple landmarks in one of the images, wherein the landmarks are the leaflet base, leaflet tip, aortic wall, and optionally mid leaflet, and opposing leaflet base.

18. The method of claim 17, further comprising calculating the magnitude of vectors between the landmarks to determine the distances individual landmarks travel between at least two timepoints19. The method of claim 18, wherein the at least two time points are systole and diastole.

20. The method of claim 17 to 19, wherein the landmarks are the leaflet tip and the mid leaflet to provide the comparative displacement of the leaflet tip and the mid leaflet.

21. The method of any one of claims 2 to 20, wherein the energy loss is total energy loss or Energy Loss Index.

22. The method of any one of claims 1 to 21 , wherein the step of comparing the one or more metrics or ratios to a reference value or reference range from a healthy subject is used to designate a patient as having no valvular disease, mild valvular disease, moderate valvular disease, severe valvular disease.

23. The method of any one of claims 1 to 22, wherein the valvular disease is a disease of the tricuspid valve, the pulmonary valve, the mitral valve, or the aortic valve.

24. The method of claim 23 wherein the disease is stenosis.

25. The method of claim 24 wherein the stenosis is mild aortic stenosis, moderate aortic stenosis, severe aortic stenosis, low gradient severe aortic stenosis, or high gradient aortic stenosis.

26. The method of any one of claims 1 to 25 when performed manually, or by software in an automated or semi-automated or fashion.

27. A system for performing the method of any one of claims 1 to 25, comprising: a) a scanner configured to scan a valve of a subject, the scan providing a series of images of a valve leaflet across at least one cardiac cycle; b) a processor configured to calculate displacement of the leaflet as defined in any one of claims 5 to 20 and wherein the processor is further configured to calculate one or more of the following metrics for the subject: valve leaflet displacement; valve leaflet velocity; valve leaflet acceleration; and valve leaflet flexibility and / or one or more of the following ratios for the subject: valve leaflet displacement to blood flow ratio (D:Q); valve leaflet displacement to blood momentum ratio (D:M);(D:Q) to Energy Loss ratio ((D:Q):EL);(D:M) to Energy Loss ratio ((D:M):EL); valve leaflet displacement to stroke volume (D:SV); and valve leaflet displacement to volume index (D:SVI).

28. The system of claim 27, wherein the processor is further configured to obtain blood flow, blood momentum and Energy Loss measurements from the series of images; or is configured to receive blood flow, blood momentum and Energy Loss measurements.

29. The system of claim 27 or 28 further comprising a display configured to generate a visualization of one or more of the images, contours and boundary lines over time and highlight the displacement of the leaflet.

Images