Patent Application
20250031982
The present invention relates to the estimation of blood pressure inside the heart, and in particular to a system and method for estimation of filling pressure in the left side of the heart using non-invasive imaging data.
Heart failure is one of the most prevalent heart diseases. In about one half of all heart failure patients, left ventricular (LV) pump function is normal when measured as LV ejection fraction. This condition is named heart failure with preserved ejection fraction, abbreviated as HFpEF. In this large group of heart failure patients, the assessment of cardiac function is challenging and there is a need for better methods to determine blood pressure inside the heart.
A clear indicator of HFpEF is obtained by the demonstration of elevated LV filling pressure, which is done in the prior art via a pressure catheter inside the left ventricle or indirectly by a catheter in the right ventricle and the pulmonary artery. These examinations are referred to as heart catheterization, and are done in hospitals, usually only when planning heart surgery or cardiac interventions.
If LV filling pressure could be measured without the need for invasive techniques needing a hospital visit, the doctor could more easily confirm the HFpEF diagnosis and hence prescribe drugs to reduce LV filling pressure earlier, thereby preventing hospitalizations. It was recently demonstrated that patients with HFpEF can be treated successfully with empagliflozin, which is an oral medication. Entresto is now indicated to reduce the risk of cardiovascular death and hospitalization for heart failure in adult patients with chronic heart failure, i.e. also covering HFpEF.
Viewed from a first aspect, the present invention provides a system for estimating the LV filling pressure of a human patient, the system comprising:
The system of the first aspect is considered to provide an accurate and reliable non-invasive method for assessment of LV filling pressure, which can then be used to improve diagnosis of heart failure. The estimates of LV filling pressure can be applied in HFpEF to improve a diagnosis. Furthermore, the estimates of LV filling pressure can be used in both HFpEF and in heart failure with reduced ejection fraction to determine need for therapy to reduce symptoms caused by high filling pressure.
The measurement of LV filling pressure can be done as left atrial (LA) mean pressure, as LV diastolic pressure prior to LA contraction (LV pre-A pressure), or as LV pressure at the end of diastole (LV end-diastolic pressure). LA mean pressure and LV pre-A pressure have essentially similar value, whereas LV end-diastolic pressure is in most cases slightly higher. In this document, the term LV filling pressure refers collectively to LA mean pressure and LV pre-A pressure. Estimation of LV end-diastolic pressure is also described and implemented in this document.
The estimate of LV filling pressure provided by the claimed system can be used for assessing the condition of the patient's heart, and hence to provide a medical practitioner with information useful to make the HFpEF diagnosis. Alternative diagnoses typically need to be excluded before it can be concluded that a patient suffers from HFpEF. In relation to other disorders the demonstration of elevated LV filling pressure may be used as one of several criteria for the diagnosis or as indication that further diagnostic tests are needed.
The ability of the proposed method to estimate LV filling pressure in patients with atrial fibrillation, severe valvular heart disease and some other conditions may require modification of the algorithm with inclusion of other imaging-based parameters in the algorithm.
In step (a) the patient specific parameters may be used to determine other markers including BMI and/or systolic pressure. Alternatively or additionally the patient specific parameters may be used directly. For example, the patient specific parameters may include patient height and/or weight with these being used directly rather than being converted into a BMI value. The systolic pressure may be used along with the original patient specific parameters (e.g. height and/or weight) or along with BMI.
The system may include an output device for providing an indication of the estimated LV filling pressure, such as a display device for visual feedback to a user or a data transfer device for communicating data to another system, such as an external computer system or network, or a mobile device. The data processing apparatus may be configured to compare the estimate of the LV filling pressure to a threshold value, such as a value considered to identify healthy or unhealthy patients. In that case an output device may be configured to indicate to the user if the patient is healthy or unhealthy, in particular with regard to an elevated LV filling pressure. The patient may be considered to potentially suffer from HFpEF, i.e. possible stiffening of the heart muscle, if the estimate of the LV filling pressure is equal to or in excess of 15 mmHg, or optionally the threshold value may be set lower, for example in excess of 12 mmHg.
The summation at step (b) may be of the magnitude of the cardiac markers multiplied by the associated constant. Alternatively step (b) may use some other value derived from the magnitude of the cardiac marker, e.g. a square thereof. The system makes use of a constant value and various constants of proportionality that have been derived from a statistical analysis of historic patient data for a plurality of patients. The statistical analysis may for example comprise a multivariate linear regression analysis accounting for the different cardiac markers. Alternatively, the statistical analysis may be based on non-linear analysis, multiple logistic regression, discriminant analysis, or artificial neural networks, including physics-informed neural networks (PINNs), amongst others. The inventors have found that by making appropriate use of a suitably comprehensive dataset of a given patient population it is possible to estimate the minimum LV pressure via a statistical analysis, such as the multivariate linear regression described in more detail herein or alternative techniques based on the same cardiac markers.
As well as the markers discussed above, the cardiac markers used in the statistical analysis may also include other measures and/or estimates, such as one or more of: end systolic or end diastolic LV volume normalized to Body mass index (BMI) or body surface area (respectively, ESVi or EDVi); the ratio between peak early diastolic velocity and peak late diastolic velocity across mitral valve (EoverA); maximal mitral annular tissue velocity at early diastole filling (e′) and/or LA volume indexed to body surface area (LAVi).
The systolic pressure, which can refer to arterial, aortic or LV systolic pressure, is a peak system pressure and can be measured via suitable means, such as via a blood pressure cuff. The system can receive this measure via the data input device, e.g. as a numeric value. BMI can be obtained in known ways, i.e. from height and weight measurements and once again the system can receive this measure via the data input device. That may be done as the source height and weight data, or as input of a pre-calculated BMI. The imaging data is used to obtain an estimate of LA reservoir strain and an estimate of the time constant of LV isovolumic pressure decay. In prior art invasive methods this is done according to known principles documented with pressure catheter inside the left ventricle, but in this method we use a non-invasive imaging based method to obtain the time constant of LV isovolumic pressure decay. The same imaging data may also be used for the estimation of peak pressure drop at step (c).
The LA reservoir strain may be any suitable indicator of strain in this context, such as a long axis reservoir strain, a short axis reservoir strain, a mean area strain or a volumetric measure of strain.
The summation of step (b) may be considered as:
where each of A, B, C, D and E are constants obtained via the statistical analysis. Optionally, there may be further terms in the sum for ESVi, EDVi, EoverA, e′, LV mass and/or LAVi as mentioned above, each with their own constant of proportionality.
In one example, the constant A is in the range −24.1 to 0.3, for example it may be in the range −14.4 to −6.5. Optionally it is about −6.9, for example −6.885.
In one example, the constant B, i.e. the constant of proportionality associated with the systolic artic pressure (peak system pressure) is in the range 0.001 to 0.056, for example it may be in the range 0.02 to 0.03. Optionally it is about 0.025.
In one example, the constant C, i.e. the constant of proportionality associated with the BMI is in the range −0.085 to 0.432, for example it may be in the range 0.10 to 0.30. Optionally it is about 0.243.
In one example, the constant D, i.e. the constant of proportionality associated with the estimate of LA reservoir strain is in the range −0.126 to 0.040, for example it may be in the range −0.10 to 0.03. Optionally it is about −0.0973.
In one example, the constant E, i.e. the constant of proportionality associated with the estimate of the time constant of LV isovolumic pressure decay is in the range 0.097 to 0.275, for example it may be in the range 0.12 to 0.20. Optionally it is about 0.1575.
Thus, in one particular example the summation of step (b) may be expressed as:
Where one or more of ESVi, EDVi, EoverA, e′, and/or LAVi are included then the relevant constants may be in the ranges tabulated below:
The data processing apparatus may be configured to provide an estimated LV diastolic pressure curve showing estimated pressure throughout diastole. This may be done by obtaining key markers for LV pressure in relation to time and/or pressure for points in the cycle and fitting a normalised curve shape to the key markers. The key markers for the estimated LV pressure curve will include the estimated LV filling pressure as well as one or more of: time of aortic valve closure; time of mitral valve (MV) opening; minimum LV pressure and respective time; time of peak E-wave; time of end diastasis, time of peak A-wave, and time of end diastole (corresponding to MV closure). In addition, the pressure may be estimated at some or all of these time points, such that the system may be configured to provide a non-invasive estimate of at least seven patient-specific pressure-time points thus enabling close fitting of a normalised curve. The normalised curve may be an average of previously obtained sample patient data. The data processing apparatus may be configured to interpolate curve segments of the normalised curve in order to fit to the key markers of pressure and time to thereby obtain the estimated LV pressure curve. Furthermore, the increase in LV pressure caused by LA contraction will be calculated from the estimated LA to LV pressure difference and estimated diastolic stiffness of the left ventricle.
The imaging system is for non-invasively obtaining images of the patient's heart and may be, for example, an ultrasound imaging system, a magnetic resonance imaging (MRI) system, an x-ray imaging system such as computed tomography (CT), or any other system able to obtain suitable images of the heart. In example embodiments echocardiography is used and hence the imaging system may comprise an ultrasound probe and associated computer devices as are already known for use in echocardiography.
The data processing apparatus may comprise a computer processor and may for example be a computer device. It may include a self-contained computer device, such as a computer system provided with the imaging system for controlling the imaging system and/or for image processing functions. Alternatively or additionally, the computer device may comprise multiple distributed computer processor devices within a network, including data storage and/or computer processing within a cloud network.
The data input device may be any suitable device for receiving the data from a user, such as by entry of numeric information, and transmitting it to the data processing apparatus. It may be provided by one or more parts of the data processing apparatus or a peripheral thereof, such as a keyboard, touch screen, or other user interface. The data input device may be provided by a mobile computing device such as a smartphone or tablet.
Viewed from a second aspect, the present invention provides a computer programme product comprising instructions, which when executed on a data processing apparatus will configure the data processing apparatus to receive imaging data from an imaging system; receive patient specific parameters relating to BMI and systolic blood pressure of the patient; and then:
The second aspect may be considered as a computer programme product with instructions that when executed will configure a data processing device to carry out the method of the third aspect, as discussed below. The computer programme device may also include any of the further features set out below and/or may configure the data processing device to perform as discussed above in connection with the first aspect and optional features thereof. The computer programme product may be provided for a computer system of an imaging system, such as the imaging system of the first aspect.
Viewed from a third aspect the invention provides a method of estimating the LV filling pressure of the heart of a human patient, the method comprising:
The estimate of the LV filling pressure may be used for assessing the condition of the patient's heart, and hence the method may provide a medical practitioner with information useful to make the HFpEF diagnosis as mentioned above. The demonstration of elevated LV filling pressure may be used as one of several criteria for the diagnosis or as indication that further diagnostic tests are needed.
The method for assessment of the LV filling pressure can thus be used to improve diagnosis of heart failure. The estimates of LV filling pressure may be applied in HFpEF to improve a diagnosis. Furthermore the estimates of LV filling pressure can be used in both HFpEF and in heart failure with reduced ejection fraction to determine need for therapy to reduce symptoms caused by high filling pressure.
The method may involve use of any or all features of the system discussed above. For example, the method may include using an output device for providing an indication of the estimated LV filling pressure, such as using a display device for visual feedback to a user or a data transfer device for communicating data to another system, such as an external computer system or network, or a mobile device.
The method may include comparing the estimate of the LV filling pressure to a threshold value, such as a value considered to identify healthy or unhealthy patients. This may be done via the data processing apparatus for example. In that case an output device may be used to indicate to the user if the patient is healthy or unhealthy, in particular with regard to an elevated LV filling pressure. The patient may be considered to potentially suffer from HFpEF, i.e. possible stiffening of the heart muscle, if the estimate of the LV filling pressure is equal to or in excess of 15 mmHg, or optionally the threshold value may be set lower, for example in excess of 12 mmHg.
The summation at step (b) may be of the magnitude of the cardiac markers multiplied by the associated constant. Alternatively, it may use some other value derived from the magnitude of the cardiac marker, e.g. a square thereof. The system makes use of a constant value and various constants of proportionality that have been derived from a statistical analysis of historic patient data for a plurality of patients. The method of this aspect may include such a statistical analysis. The statistical analysis may for example comprise a multivariate linear regression analysis accounting for the different cardiac markers. Alternatively, the statistical analysis may be based on non-linear analysis, multiple logistic regression, discriminant analysis, or neural network analysis amongst others. The inventors have found that by making appropriate use of a suitably comprehensive dataset of a given patient population it is possible to estimate the minimum LV pressure via a statistical analysis, such as the multivariate linear regression described in more detail herein or alternative techniques based on the same cardiac markers.
As well as the markers discussed above, the cardiac markers used in the statistical analysis may also include other measures and/or estimates, such as one or more of the following: end systolic or end diastolic LV volume normalized to BMI or body surface area (respectively, ESVi or EDVi); the ratio between peak early diastolic velocity and peak late diastolic velocity across MV (EoverA); maximal mitral annular velocity at early diastole filling (e′) and/or LA volume indexed to body surface area (LAVi).
The systolic pressure, which can refer to arterial, aortic or LV systolic pressure, is a peak system pressure and can be measured via suitable means, such as via a blood pressure cuff. The method may include measuring the systolic pressure. The method may include using the data input device for input of systolic pressure data, e.g. as a numeric value. BMI can be obtained in known ways, i.e. from height and weight measurements and once again the method may include receiving this measure via the data input device. That may be done as the source height and weight data, or as input of a pre-calculated BMI. The imaging data is used to obtain an estimate of LA reservoir strain and an estimate of the time constant of LV isovolumic pressure decay. In this method we use a non-invasive imaging based method to obtain the time constant of LV isovolumic pressure decay. The same imaging data may also be used for the estimation of peak pressure drop at step (c).
The LA reservoir strain may be any suitable indicator of strain in this context, such as a long axis reservoir strain, a short axis reservoir strain, a mean area strain or a volumetric measure of strain.
The summation of step (b) may be considered as:
where each of A, B, C, D and E are constants obtained via the statistical analysis. Optionally, there may be further terms in the sum for ESVi, EDVi, EoverA, e′, LV mass and/or LAVi as mentioned above, each with their own constant of proportionality.
In one example, the constant A is in the range −24.1 to 0.3, for example it may be in the range −14.4 to −6.5. Optionally it is about −6.9, for example −6.885.
In one example, the constant B, i.e. the constant of proportionality associated with the systolic artic pressure (peak system pressure) is in the range 0.001 to 0.056, for example it may be in the range 0.02 to 0.03. Optionally it is about 0.025.
In one example, the constant C, i.e. the constant of proportionality associated with the BMI is in the range −0.085 to 0.432, for example it may be in the range 0.10 to 0.30. Optionally it is about 0.243.
In one example, the constant D, i.e. the constant of proportionality associated with the estimate of LA reservoir strain is in the range −0.126 to 0.040, for example it may be in the range −0.10 to 0.03. Optionally it is about −0.0973.
In one example, the constant E, i.e. the constant of proportionality associated with the estimate of the time constant of LV isovolumic pressure decay is in the range 0.097 to 0.275, for example it may be in the range 0.12 to 0.20. Optionally it is about 0.1575.
Thus, in one particular example the summation of step (b) may be expressed as:
Where one or more of ESVi, EDVi, EoverA, e′, and/or LAVi are included then the relevant constants may be in the ranges tabulated below:
The method may include determining an estimated LV diastolic pressure curve showing estimated pressure throughout diastole. This may be done by obtaining key markers for LV pressure in relation to time and/or pressure for points in the cycle and fitting a normalised curve shape to the key markers. The key markers for the estimated LV pressure curve will include the estimated LV filling pressure as well as one or more of: time of aortic valve closure; time of MV opening; minimum LV pressure and respective time; time of peak E-wave; time of end diastasis, time of peak A-wave, and time of end diastole (corresponding to MV closure). In addition, the pressure may be estimated at some or all of these time points, such that the method may comprise a non-invasive estimate of at least seven patient-specific pressure-time points thus enabling close fitting of a normalised curve. The normalised curve may be an average of previously obtained sample patient data. The data processing apparatus may be configured to interpolate curve segments of the normalised curve in order to fit to the key markers of pressure and time to thereby obtain the estimated LV pressure curve. Furthermore, in some examples, the increase in LV pressure caused by LA contraction will be calculated from the estimated LA to LV pressure difference and estimated diastolic stiffness of the left ventricle.
The method may include non-invasively obtaining images of the patient's heart using an imaging system. The imaging system may be, for example, an ultrasound imaging system, a MRI system, an x-ray imaging system such as CT, or any other system able to obtain suitable images of the heart. In example embodiments echocardiography is used and hence the imaging system may comprise an ultrasound probe and associated computer devices as are already known for use in echocardiography. The method may be used as a part of a method for diagnosing and treating patients with HFpEF or other cardiac conditions, within which method the LV filling pressure of a patient is estimated using a method as in the third aspect. The patient with heart failure is determined to require treatment if the estimated LV filling pressure is equal to or in excess of 15 mmHg, optionally in excess of 12 mmHg, and if the patient is determined to require HFpEF treatment then a drug indicated for treatment of HFpEF is administered. The drug may for example be Empagliflozin, Entresto or another SGLT inhibitor.
Viewed from a still further aspect the invention provides an HFpEF drug for use as a medicament for treatment of HFpEF via a method as defined above. Thus, the invention may be embodied as a HFpEF drug for use as a medicament for treatment of HFpEF via a method for diagnosing and treating patients with HFpEF, within which method the LV filling pressure of a patient is estimated using a method as in the third aspect, the patient is determined to require treatment if the estimated LV filling pressure is in excess of 15 mmHg, optionally in excess of 12 mmHg, and if the patient is determined to require HFpEF treatment then a drug indicated for treatment of HFpEF is administered, such as Empagliflozin or Entresto. The invention may alternatively or additionally comprise a method of use of a HFpEF drug for the manufacture of a medicament for treatment of HFpEF via a method as defined above.
Certain example embodiments of the invention will now be described by way of example only and with reference to the accompanying drawings in which:
LV diastolic pressure can be measured as the average pressure during diastole, as pressure prior to atrial contraction (pre-A pressure), and as end-diastolic pressure, and all these pressures provide information of essentially similar clinical value.
Because left atrium and left ventricle are directly connected in diastole, and the left atrium is directly connected with the pulmonary vasculature, elevation of LV diastolic pressure is transmitted to the pulmonary vasculature. Thus, when LV diastolic pressure becomes elevated, pressure in the pulmonary capillaries increase with about similar magnitude. This leads to pulmonary congestion and breathing problems. Therefore, LV diastolic pressure is so important to know and to target when treating heart failure patients.
Direct measurement of LA pressure is not feasible clinically, but LV pre-A pressure is a good approximation of LA mean pressure. Therefore, LV pre-A pressure is an important parameter to measure. In the present innovation, LV pre-A pressure is employed as an indirect measure of LA mean pressure. There are rare cases with MV stenosis when this assumption is not valid. Pulmonary capillary wedge (PCWP) pressure measured during right heart catheterization, is used as an indirect measure of LA mean pressure.
In this document explicit mention of physical unit is sometimes omitted. Unless stated otherwise, all pressures are given in mmHg. Some equations show apparently incompatible units, for example equation (3) where ΔPA≈4vmax2. It is understood that vmax is given in m/s, and ΔPA is in mmHg, while the calculation is carried out in a unit less fashion.
A basic assumption in our innovation is that LA mean pressure can be approximated by the sum of the minimum LV diastolic pressure, PLV, and the peak LA to LV pressure drop during early diastole, PLA−PLV. This assumption has been supported by several studies in conscious and anaesthetized dogs under a wide range of hemodynamic conditions and by observations in patients.
The method consists of three steps: estimation of the maximum pressure drop, estimation of the minimum LV pressure, and estimation of (PLA) as detailed in the following sections.
The algorithm for calculating the maximum pressure drop, max(PLA−PLV), requires the inputs listed in Table 3 below.
The pressure drop estimate is based on the Navier-Stokes momentum conservation equation:
Where Q is the flow rate computed across the MV over the diastolic period, t is the temporal derivative of the kinetic energy within MV inflow jet, A is the advective energy rate describing the energy transfer because of the physical movement of a fluid in and out of the domain, and V is the rate of viscous dissipation describing energy losses because of friction.
The viscous term, ΔPV, can be neglected in the case of relatively high velocities, such as in flow across the MV.
The advective term is approximated using the simplified Bernoulli equation:
The kinetic term is calculated as:
where the term ρ=1060 kg/m3 is the density of blood.
The time of minimum LV pressure, tmin P
The resulting estimate for the pressure drop is:
The algorithm for calculating the minimum LV pressure, min PLV, requires the following inputs (Table 4 below). In some cases the algorithm can be adapted to use differing parameters/markers, such as by using height and/or weight directly in place of using BMI.
Statistical analysis of clinical data has shown that the min PLV can be approximated by a regression on several other clinical parameters, so that min PLV is almost equal to
The statistical analysis of the clinical data yielded the following as the best linear approximation:
It is important to note that other regression coefficients can be used, also including further variables including end-diastolic volume (absolute or indexed to BSA), E/A transmitral velocity ratio, E/e′ velocity ratio, or e′ tissue velocity amongst other. The units of the clinical parameters are specified in Table 2, while the isovolumic relaxation time constant, τ, is given in seconds.
The calculation of min PLV is presented as a function of t, since these are interdependent variables which require a numerical solver to find the roots ({circumflex over (τ)} and min PLV (τ{circumflex over ( )})). The time constant t describes the rate of decay of the LV pressure during the period where both the LV valves, the aortic valve (AV) and the MV, are closed; that is, a time interval between the closing of the aortic valve, tAVC, and the opening of the MV, tMVO. During this time interval the LV pressure can be approximated by the expression
where t0, t1 are time instances during the isovolumic relaxation phase, PLV(t) is the LV pressure at time t, PLV(t0) is the LV pressure at t0. From this we get that
Where ln denotes the natural logarithmic function.
Several different choices for t0 and t1 are possible. One possible choice is
Based on statistical analysis of a suitable dataset, the following approximations are used:
The motivation for the approximation of PLV(t1) is to neglect the PLV variation due to hemodynamic changes at the time of onset LV filling, PLV(t1) is set 5 mmHg above PMVO, and PMVO is defined as:
Finally, in the proposed PLV curve estimation, the IVR has an exponential behavior down to PMVO+5 mmHg and a linear behavior afterwards.
Values for min PLV and τ can be obtained by combining equations (7) and (9) and solving for the unknown quantities. For example, substituting the right-hand side of (9) for τ in (7) results in the following equation:
Where
The estimated value, τ{circumflex over ( )}, for the isovolumic relaxation time constant, τ, is found by numerically solving equation (13). This can be done with any standard root-finding software, and an initial guess for the value of τ{circumflex over ( )}. The initial guess for the value of τ{circumflex over ( )}can e.g., be set by inserting an initial value of PLV(t1), following equation (11), into equation (9), one example of this being 20 mmHg.
After finding the root, τ{circumflex over ( )}, of equation (13) the LV filling pressure described in equation (1) can be presented as the sum of equation (7) in function of τ{circumflex over ( )} and equation (5):
The LA contraction induced PLV rise accounts for three independent PLV variations that are computed over two periods of atrial contraction:
The maximal MV dP is obtained as the sum of the convective and kinetic components at the instant of maximal A-wave acceleration. The A-acceleration is calculated from transmitral velocities by pulsed Doppler. The A-wave upslope can be approximated to a sinusoidal behavior since at the beginning of atrial contraction the acceleration is lower, the maximal acceleration is reached at the middle of A-wave upslope and then the acceleration is reduced until the peak A velocity is reached. Inherently, the maximal A-acceleration is considered to occur at ½ of A-wave acceleration duration, and the velocity at this instant is also ½ of the peak A velocity.
Furthermore, the peak A-acceleration is given by the mean A-acceleration slope multiplied by the acceleration constant: Kacc=PI/2˜1.57. This constant Kacc is obtained from the relationship between the maximal and mean derivative of any sinusoidal upslope trace. This value is also very similar to the value found experimentally across the current cohort when dividing maximal A-acceleration by mean A-acceleration.
resulting in the constant change in pressure of: dP/dt=(dV/dt)/Cn. The equation (15), considering that all relations are constant, can be extrapolated to:
Finally, the EDP is obtained as the sum of A-deceleration ΔP with the peak-A PLV.
In order to validate the atrial contraction induced PLV, a dataset of n=7 patients undergoing simultaneously echocardiography and catheterization examination were considered. The results show a good agreement between measured and non-invasive estimative of atrial contraction induced PLV increase, as well as a good agreement for the estimation of end-diastolic PLV. The results are presented in Table 5 below.
The final step of the provided methodology is the generation of the fully non-invasive and patient specific LV diastolic pressure curve. This curve is tabulated based on a reference curve for LV diastolic pressure, previously generated from data collected by Smiseth et al (1998). The method uses the calculated times and pressures for a set of key events during diastole, and subsequently fits the reference curve to those events using a combination of scaling and interpolation.
The reference LV diastolic pressure curve was collected invasively from several heart patients. The duration between each event was normalized, by interpolating each subsection of each recorded pressure curve onto a fixed set of time samples, and then averaging the pressure curves across all patients.
After normalization, the reference curve is specified as a sequence of values along with labels pointing to key events.
The following inputs are used to calculate the patient specific diastolic LV pressure curve. They consist of the times and pressures of several diastolic events, in addition to the reference pressure curve, sampled at a larger number of time points. Other events, more or fewer, can be used. The units of the reference pressure curve, Pref, is given as “arbitrary” or “unitless”.
For each time range between two events, the estimated patient specific curve is fitted to the corresponding samples from the reference curve, Pref. In order to accomplish this, both the sample times and the sample pressure values can be interpolated onto the reference curve; for example, to calculate time- and pressure-sample number k, respectively equations (18) and (19), for samples occurring between the first and second key event (evt):
This procedure is repeated for the pressure curve segments between each pair of key events.
A specific implementation of this algorithm, requires a set of diastolic markers to serve as the key events. For example, the events listed in Table 7 can be used.
In order to calculate pressures and times listed in Table 4, the measured systolic PsAO and the LV volume and blood flow traces are used as input parameters (Table 8).
Quantities V and Q are related through temporal differentiation, Q=dV/dt, so one can be obtained from the other, but because V and Q are often obtained from different measurements, they are considered separate inputs in this case. The quantities dV/dt and dQ/dt, used below, can be obtained through any standard numerical differentiation method.
Calculate (PAVC,tAVC): The estimates are PAVC≈0.65·max(PsLV) where PsLV is assumed to be equal to PsAO in cases without aortic stenosis, and tAVC is approximately the time of the negative peak of the [dP/dt] trace applied in the estimation of isovolumic relaxation time constant, τ, as described in equation (9).
Calculate (PMVO,tMVO): The estimates of PMVO are given by equation (12), and tMVO is the first time point after tAVC when both the volume trace is increasing and transmitral dQ/dt is positive.
Calculate (min PLV, tmin P
Calculate (PE-wave, tE-wave): The estimate of PE-wave is based on the end-diastolic pressure-volume relationship that characterizes ventricle stiffness, for example it can be assumed to be linear from E-wave peak velocity until A-wave-start: PE-wave≈a·V(tE-wave)+b, and tE-wave is the time of end of the deceleration of flow rate dQ/dt. The stiffness constants (a, b) are computed based on the estimated incomplete relaxation PLV. Exemplary values are a=0.19, b=−8.08.
Calculate (PAPLA
, and tA
Calculate (PA
The maximal MV pressure difference max(PLA−PLV) is given by equation (5), at the time when A-wave acceleration is maximum (tA-wave-max-acceleration≈(tAPLA
computation. tA
Calculate (PMVC,tMVC): The pressure at end-diastole/MV closure is estimated based on the LV atrioventricular compliance across MV (Flachskampf 1992) during A-wave deceleration as PMVC=PA-wave-peak+(V−(tMVC)−V(tA-wave-peak))/Cn, where the compliance is given by Cn=AreaMV/(ρ·(vMV(tA-wa-e-peak)−vMV(tMVC) and effective orifice MV area during A-wave is AreaMV=Q(tA-wave-peak)/vMV(tA-wave-peak). tMVC is at the end of the diastolic volume curve, when dV/dt is close to zero, i.e., the volume curve is flat.
An example estimated patient specific curve is displayed in
In order to validate the full diastolic PLV curve, a dataset of n=7 patients undergoing simultaneously echocardiography and catheterization examination were considered. The results show a very good agreement between measured and non-invasive PLV curves, both visually (
One possible product implementation is a software product that will provide a measurement of LV filling pressure as LA mean pressure and display/save/export the image of the LV diastolic pressure curve. The software product might be integrated with a commercial ultrasound scanner such that some input parameters do not need to be entered, or a post-processing tool that loads stored examination data.
In
As summarized in the most recent guidelines on imaging of LV diastolic function from the Association of Cardiovascular Imaging in the European Society of Cardiology, there is currently no clinically applicable non-invasive method to measure LV filling pressure. Our innovation is the first non-invasive method to provide an estimate of LV filling pressure with good accuracy and which provides a measure of the LV diastolic pressure curve.
The main applications of this innovation will be to improve diagnosis of heart disease and to provide filling pressure as a guide when making therapeutic decisions.
It should be apparent that the foregoing relates only to the preferred embodiments of the present application and the resultant patent. Numerous changes and modification may be made herein by one of ordinary skill in the art without departing from the general spirit and scope of the invention as defined by the following claims and the equivalents thereof.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2115123.8 | Oct 2021 | GB | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/EP2022/079293 | 10/20/2022 | WO |
