Method for evaluating hemodynamic response to stimulation of adenosis receptor agonists, system for evaluating hemodynamic

Abstract

A method, computer-readable medium, and system for assessing a hemodynamic response of a human patient to adenosine receptor agonist stimulation are disclosed. The starting point of the method is non-invasive pressure data in a resting state. The data is used for pharmacokinetic / pharmacodynamic modeling to obtain an individual patient-specific response to an agonist-stimulated adenosine receptor. The reaction is used for calibrating a lumped parameter Windkessel type model of blood circulation. The calibrated model allows for the calculation of hemodynamic parameters of the hyperemic state, in particular for the calculation of flow and pressure values in the two coronary inlets. The hemodynamic parameters allow computer simulation of coronary artery vascular diagnostics. The effectiveness of the developed method has been proved in multi-center clinical trials.

Description

Field of the Invention

[0001] The present invention relates to: a method for evaluating the hemodynamic response of a human patient to stimulation with an adenosine receptor agonist; a computer-readable medium comprising instructions which, when executed by a computer, cause the computer to perform the steps of the mentioned method; and a system for evaluating the hemodynamic response of a human patient to stimulation with an adenosine receptor agonist. In particular, the present invention relates to estimating hemodynamic parameters of the coronary circulation of a patient in a hyperemic state by modeling the individual patient-specific response to stimulation of adenosine receptors by an agonist. The present invention allows the diagnosis and treatment of coronary insufficiency, in particular of coronary heart disease.

[0002] background

[0003] According to the National Center for Health Statistics, 5.6% of adults in the United States (defined as those aged 18 years or older) had coronary heart disease (CHD) in 2018. Men are more likely to have CHD than women—7.4% of adult men vs. 4.1% of adult women in the United States. The risk increases with age—15.5% of people aged 65-74 years had CHD, while 23.9% of people aged 75 years and older had CHD (Tables of Summary Health Statistics for USAdults: 2018 National Health Interview Survey, Villarroel MA, Blackwell DL, Jen A, 2019, published in the National Center for Health Statistics (accessed December 17, 2021)). In 2018, coronary heart disease caused 365,744 deaths in the United States - almost 13% of all deaths recorded in 2018 (see the public use data file documentation: mortality multiple cause-of-death micro-data files published by the Centers for Disease Control and Prevention, the National Center for Health Statistics, and the National Vital Statistics System (accessed December 17, 2021)). In 2017, ischemic heart disease caused 623,276 deaths in the European Union, almost 12% of all deaths recorded in 2017 (according to Causes of death-deaths by country of residence and occurrence [HLTH_CD_ARO] published by Eurostat in 2021 (accessed December 17, 2021)).

[0004] The diagnostic pathway for coronary insufficiency is very complex and needs to be considered from many aspects. In addition to the initial stage, each of the following main stages has gradually increasing reliability, which may be related to the increased invasiveness of the following main stages and the increased risk of patients undergoing the diagnostic pathway (see Scanlon PJ et al., 1999, Circulation, 99 (17) 2345-2357 pages, ACC / AHA Guidelines for Coronary Angiography: Executive Summary and Recommendations, ESC Scientific Literature Group, 2020, Eur Heart J, 41 (3) 407-477 pages, 2019 ESC Guidelines for the diagnosis and management of chronic coronary syndromes). The diagnostic accuracy pyramid comprehensively demonstrates that its apex and gold standard is the coronary artery blood flow reserve ratio in the form of fractional blood flow reserve (FFR) (see Johnson NP et al., Continuum of Vasodilator Stress from Rest to Contrast Medium to Adenosine Hyperemia for Fractional Flow Reserve Assessment, 2016, JACC Cardiovasc Interv, 9 (8), pp. 757-767). FFR is described as a basic ratio of distal and proximal pressures, where each pressure measurement must be performed in a hyperemic state. As the discoverers of the FFR ratio, Nico Pijls and Bernard De Bruyne, wrote in an excellent monograph: "A meaningful measurement of coronary pressure can only be made in a state of maximal hyperemia" (see Pijls NH and Bruyne BD, Coronary Pressure, Springer-Science+Business Media, 1997). This statement applies not only to FFR, but also to the accuracy of every CHD diagnostic method. There is nothing wrong with the accuracy of a ratio (such as FFR or hyperemic ratio) exceeding the 95% threshold. At the same time, calculated rates in the resting state are only slightly over 80%, and visual assessments, including angiographic methods, are below 70%. As a general principle, reliable functional assessment is only possible in a hyperemic state.

[0005] The hyperemic state in the functional assessment is achieved by stimulating adenosine receptors (AR) with a pharmacological agent. There are many marketed drugs that act as adenosine receptor agonists. Adenosine in pharmaceutical form is sold under the following trade names: Adenocard, Adenocor, Adenic, Adenoco, Adeno-Jec, Adenoscan, etc. An example of the dosage and administration of Adenoscan according to the FDA is: The recommended dose for adults is 140 μg / kg / min when administered as a continuous peripheral intravenous infusion for 6 minutes (excerpted from the U.S. Food and Drug Administration's FDA-Approved Drugs–Labels for NDA020059. (Accessed on December 17, 2021)). Regadenoson is sold under the trade name Lexiscan or Rapiscan, Binodenoson is sold under the trade name CorVue, and Apadenoson is sold as Stedivaze. An example of dosing for these products is: 5 ml of Lexiscan (0.4 mg of Reganoxan) is injected intravenously over 10 seconds, followed by a radiopharmaceutical and saline flush (from the U.S. Food and Drug Administration's FDA-Approved Drugs–Labels for NDA 022161. (Accessed on December 17, 2021)). According to the DrugBank database, the Phase III clinical trial of Apadenoson was terminated, and the Phase III clinical trial of Binodixon was completed (see DrugBank Online (Accessed on December 17, 2021)).

[0006] Adenosine receptors are integral membrane proteins that belong to the P1 class of purinergic receptors coupled to G proteins. G proteins are heterotrimeric protein complexes constructed from three subunits (α, β, and γ). The stable dimeric complex of this protein is constructed from the α and β subunits. Adenosine receptors are divided into four types: A1, A 2A , A 2B and A3. Each receptor is built of seven transmembrane domains. Each receptor has a different number of amino acids, and A 2AAdenosine receptors are the longest receptors (see Pharmacology of Adenosine Receptors: The State of the Art, published by Borea et al. in Physiological Reviews, 98 (3) pp. 1591-1625 in 2018, and Adenosine receptors: expression, function and regulation, published by Sheth S et al. in International Journal of Molecular Sciences, 15 (2) pp. 2024-2052 in 2014). Each transmembrane domain is constructed of 21-28 amino acids that form a helix. The C-terminal side of the adenosine receptor is located in the cytoplasmic region, while the N-terminal side is located on the extracellular side of the cell membrane (see Vincenzi M, Bednarska K and ZJ, Comparative Study of Carborane-and Phenyl-Modified Adenosine Derivatives as Ligands for the A2A and A3 Adenosine Receptors Based on a Rigid in SilicoDocking and Radioligand Replacement Assay, Molecules, Basel, Switzerland, 23(8), p. 1846, 2018.

[0007] Adenosine receptors can be found in the nervous, cardiac, pulmonary, digestive, urinary, and immune systems (see Borea et al., 2018, Physiological Reviews, 98(3), pp. 1591-1625, Pharmacology of Adenosine Receptors: The State of the Art). Each type of adenosine receptor (i.e., A1, A2, A3, A4, A5, A6, A7, A8, A9, A10, A110, A12, A13, A14, A15, A16, A17, A18, A19, A20, A21, A22, A23, A24, A25, A26, A27, A28, A29, A30, A31, A32, A43, A54, A35, A36, A37, A38, A39, A40, A 2A , A 2B The activation of A1 adenosine receptors (A1 and A3) can affect the function of the heart. A1 adenosine receptors have the highest affinity for adenosine. Activation of this receptor will slow the heart rate and will cause vasoconstriction and inhibition of neurotransmitter release. 2A Has the largest molecular weight of all adenosine receptors. A is found in the greatest number in the heart. 2A AR mRNA. 2A Activation of the AR leads to an increase in cAMP levels, vasodilation, and inhibition of platelet aggregation. 2B, high concentrations of adenosine are required. 2B The expression level of adenosine receptors is low. 2B AR activation leads to an increase in cAMP levels, vasodilation and cardiac contraction. The largest number of A3 adenosine receptors appear in the liver, less in the aorta or heart. Although the exact location in the heart is unknown, the activation of these receptors has a protective effect on the heart. A3AR activation also leads to inhibition of adenylate cyclase (see Shryock JC and Belardinelli L. in 1997 in Am J Cardiol. 19; 79 (12A): 2-10 published Adenosine and adenosine receptors in the cardiovascular system: biochemistry, physiology, and pharmacology.; Wilson C and Mustafa SJ published Adenosine Receptors in Health and Disease in Springer Nature in 2009.; Guieu R et al. in 2020 in J. Clin. Med. 9 (1366): 1-21 published Adenosine and the cardiovascular system: The good and the bad.).

[0008] Most adenosine receptor agonists are purine nucleoside derivatives: adenosine or xanthine. Alkylxanthine derivatives are the prototypical antagonists for adenosine receptors. A few parts of adenosine may interact with the amino acids of the receptor: the five nitrogen atoms of the adenine group (i.e., N1, N3, N6, N7, and N9) and the three hydroxyl groups of the ribose moiety (i.e., 2', 3', and 5'). There are many different adenosine receptor agonists. Substitutions on adenosine can alter the selectivity of the receptor. Substitution at the N6 position by arylalkyl, cycloalkyl, and alkyl groups alters the selectivity of the A1 adenosine receptor. Substitution reactions at the 2-position of adenosine with (thio)ethers, alkynes, and secondary amines lead to the formation of new synthetic A 2A Adenosine selective analogs. Substitution of the N6 position of adenosine with selected aromatic groups increases the affinity for A 2B Adenosine receptor affinity. Increased selectivity for A3 adenosine receptors can be achieved by substitution with N6-benzyl or substituted benzyl. Adenosine receptor antagonists can be obtained in a variety of ways. High affinity and selectivity for A1 adenosine receptors can be achieved by modifying the eighth position of the xanthine core structure with an aryl or cycloalkyl group. Adenosine receptor selectivity can be achieved by modifying the eighth position of xanthine with an olefin. 2AAdenosine receptor selectivity can be achieved by modifying the xanthine core structure with an aromatic group at the eighth position. 2B Selectivity for adenosine receptors. Many non-xanthines have been defined as A3 adenosine receptor antagonists (see Recent developments in adenosine receptor ligands and their potential as novel drugs, published by CE and Jacobson KA in Biochimica et Biophysica Acta (BBA)-Biomembranes, 1808 (5) pp. 1290-1308 in 2011). 2A It has an orthosteric binding site. This binding site can be found by the residues involved in binding agonists and antagonists (according to Human Adenosine A2A Receptor: Molecular Mechanism of Ligand Binding and Activation published by Carpenter B and Lebon G in Frontiers in Pharmacology, 8, p. 898, 2017). Caffeine is A 2A A natural antagonist of adenosine receptors. Studies have shown that caffeine affects the cardiovascular system, however, the results of these studies are inconclusive. The cardiovascular system's response to caffeine intake depends on a variety of factors: the amount consumed, the time of consumption, the frequency, the level of liver metabolism, and the level of absorption. The effects of caffeine also depend on gender, and the difference may be related to steroid hormone concentrations (see Recent developments in adenosine receptor ligands and their potential as novel drugs published by Müller CE and Jacobson KA in 2011 in Biochimica et Biophysica Acta (BBA)-Biomembranes, 1808 (5) pp. 1290-1308). Based on information from the Protein Data Bank, the O13 atom and the N7 atom and hydrogen bonds were used, and the N3 atom and cation-π interactions were used to make caffeine bind to A 2A Adenosine receptor binding. 2AAdenosine receptor agonist. Reganoxan is metabolized more slowly in plasma than adenosine. Reganoxan can cause a dose-dependent increase in coronary blood flow (according to Recent developments in adenosine receptor ligands and their potential as novel drugs published by M. R. CE and Jacobson KA in Biochimica et Biophysica Acta (BBA)-Biomembranes, 1808 (5) pp. 1290-1308 in 2011).

[0009] Additional selective A 2A Adenosine receptor agonists are binodixon (BND) and apadenoson (APA); apadenoson is more selective than binodixon. In this comparison, regadenoson is the least selective. APA and BND have equal affinity for adenosine receptors. The average onset of binding action is 1 to 2 minutes. Apadenoson has a longer duration of action than binodixon: 10-20 minutes. The duration of binodixon is less than 5 minutes. The dose of APA and BND is calculated based on the patient's weight. Both adenosine agonists are introduced by intravenous push (see Elkholy KO et al., 2021, Cureus, 13(1), page e12940, Regadenoson Stress Testing: A Comprehensive Review with a Focused Update).

[0010] Adenosine receptors are involved in the regulation of vascular tone, although this is not fully understood. 2A Adenosine receptor activation regulates vascular tone, but it has also been shown that A 2B Adenosine receptors are involved in arteriolar relaxation. Adenosine receptors interact to regulate coronary vasculature. Regulation of coronary microcirculation mediated by adenosine can be endothelium-dependent or endothelium-independent. 2A Reactive hyperemic nitric oxide is involved in the control of basal tension mediated by adenosine receptors. It is also adenosine-mediated A 2A AR activation and reactive hyperemia 2A AR-K ATP Part of the pathway (see Adenosine and adenosine receptor-mediated action in coronary microcirculation, published by Zhang Y et al. in Basic Research in Cardiology, 12(116) in 2021).

[0011] Due to the importance and role of adenosine receptors, the pharmacokinetics and pharmacodynamics of their endogenous and exogenous agonists and antagonists are the subject of research. The most studied and investigated to date is adenosine. Kern et al. are testing adenosine as an agent for determining coronary vasodilator reserve. During the trial, adenosine was infused intravenously at doses of 50, 100, and 150 μg / kg / min for 3 minutes. 34 patients participated in these trials: 17 patients had coronary artery disease (CAD) and 17 patients did not have coronary artery disease (CAD) (see Kern MJ et al., Intravenous adenosine: continuous infusion and low dose bolus administration for determination of coronary vasodilator reserve in patients with and without coronary artery disease, Journal of the American College of Cardiology, 18(3), pp. 718-29, 1991). Alexopoulos et al. conducted a trial to investigate the effects of increasing adenosine doses on P d / P a Thirty-eight patients were treated with a standard adenosine dose (140 μg / kg / min) introduced via the femoral vein and repeated P d / P a Eight patients refused to participate in the high-dose adenosine trial. Thirty patients received high-dose adenosine (200 μg / kg / min) and repeated P d / P aEvaluation. Patients had stable coronary artery disease or, in some cases, presented with an acute coronary syndrome. The mean systolic blood pressure dropped more after a high adenosine dose of 200 μg / kg / min compared with a mean adenosine dose of 140 μg / kg / min. The mean diastolic blood pressure dropped to lower values ​​after a high adenosine dose compared with treatment with a mean adenosine dose. The heart rate (HR) increased more after a high adenosine dose (see Effect of High (200 μffeg per Minute) Adenosine Dose Infusion on Fractional Flow Reserve Variability, 2016, by Alexopoulos D et al., Journal of the American Heart Association, 5). In a trial conducted by Thomas et al., cardiovascular magnetic resonance imaging was used to measure the effects of adenosine on hemodynamics. Adenosine was administered intravenously to 25 healthy patients at a dose of 140 μg / kg / min for 6 minutes. The results of the trial showed that there was a small decrease in mean systolic blood pressure after adenosine injection, but there was no change in mean diastolic blood pressure after adenosine injection, and there was a higher increase in heart rate after adenosine injection (see Thomas D et al., 2017, Journal of Cardiovascular Magnetic Resonance, 19, Effects of adenosine and regadenoson on hemodynamics measured using cardiovascular magnetic resonance imaging). Mishra et al. conducted another trial in which researchers studied the infusion of a normal dose of adenosine for 6 minutes and its effects on hemodynamics. 348 healthy volunteers participated in this trial. The results obtained showed that the mean systolic and mean diastolic blood pressures changed very little. Only the heart rate increased significantly (according to Mishra R et al., 2005, Journal of the American College of Cardiology, 45, pp. 553-8, Quantitative relation between hemodynamic changes during intravenous adenosine infusion and the magnitude of coronary hyperemia-Implications for myocardial perfusion imaging).Lee et al. are testing the use of adenosine as an alternative to exercise testing (called pharmacological stress testing) in 15 healthy volunteers. Adenosine was infused intravenously at a dose of 140 μg / kg / min for 6 minutes. In agreement with previous trials, the mean systolic blood pressure decreased by a small amount. The mean diastolic blood pressure decreased by a small amount, and the heart rate increased significantly (for more details, see Lee JH et al., Biokinetics of thallium-201 in normal subjects: comparison between adenosine, dipyridamole, dobutamine and exercise, Journal of nuclear medicine: official publication, 1994, pp. 535-41, in Society of Nuclear Medicine). The results of the experiments showed that it was not possible to mathematically model the effects of adenosine receptor stimulation based on simple relationships, which is necessary to estimate the effects of adenosine on cardiac hemodynamics. In each of the tests performed, the heart rate increased, but by different amounts. Also, the HR did not differ significantly between the group with coronary artery disease (CAD) and the group without CAD. The results for systolic blood pressure were equivocal. In each test, there was a difference between the systolic blood pressure value at baseline and the systolic blood pressure value during peak hyperemia after adenosine infusion. These results show that systolic blood pressure decreased after adenosine infusion, but in the experiment conducted by Brown et al. (see Brown L et al., 2021, Journal of Cardiovascular Magnetic Resonance, 23, A comparison of standard and high dose adenosine protocols in routine vasodilator stress cardiovascular magnetic resonance: dosage affects hyperemicmyocardial blood flow in patients with severe left ventricular systolic impairment), there was no change in systolic blood pressure after adenosine infusion in a group of 16 heart failure (HF) patients with left ventricular ejection fraction (LVEF) ≥ 40% (124 ± 16 at rest and 124 ± 11 during stress), while in another group of 20 HF patients with LVEF < 40%, systolic blood pressure increased (119 ± 16 at rest and 122 ± 18 during stress). The test results describe either a decrease in diastolic blood pressure or no change in diastolic blood pressure values.Vasu et al are testing adenosine, reganoxan, and dipyridamole in 15 healthy volunteers. The relative potency of these vasodilators was assessed using quantitative stress and resting myocardial perfusion. Results were obtained using cardiovascular magnetic resonance (CMR). Stress tests were performed over several days after bolus infusions of 400 μg reganoxan, 0.56 mg / kg dipyridamole, and 140 μg / kg / min adenosine. Each trial began with recording of resting perfusion imaging. After 20 minutes, vasodilation stress imaging was recorded during the peak period (i.e., 3-4 minutes after adenosine infusion, 4 minutes after dipyridamole infusion, and 70 seconds after reganoxan infusion). Myocardial blood flow (MBF) and myocardial perfusion reserve (MPR) were quantified by fully quantitative model-constrained deconvolution. Stress MBF and heart rate (HR) values ​​recorded after reganosine injection were higher compared to those recorded after either dipyridamole or adenosine injection. After adjusting stress MBF values ​​to HR values, there was no difference between adenosine and reganosine, but a difference was still observed between dipyridamole and reganosine. The unadjusted MPR for reganosine was higher than that for adenosine and dipyridamole. After adjusting the stress MPR values ​​to HR values, there was no difference between regadenoson and adenosine, but differences were still observed between regadenoson and dipyridamole and between adenosine and dipyridamole (see Vasu S et al., 2013, Journal of cardiovascular magnetic resonance: official journal of the Society for Cardiovascular Magnetic Resonance, 15, p. 85, Regadenoson and adenosine are equivalent vasodilators and are superior than dipyridamole-a study of first pass quantitative perfusion cardiovascular magnetic resonance). Average values ​​of systolic blood pressure (SYS), diastolic blood pressure (DIA) and heart rate (HR) as well as coronary artery disease (CAD) from different clinical trials can be found in the cited references.

[0012] Despite some difficulties, non-invasive diagnostic methods for circulatory failure based on computational fluid dynamics (CFD) technology are slowly trying to gain trust in the medical environment. Current results of validation studies do not provide clear recommendations (according to Nakanishi R and Budoff MJ, 2016, Vasc Health Risk Manager, 12, pp. 269-78, Noninvasive FFR derived from coronary CT angiography in the management of coronary artery disease: technology and clinical update, Kruk M et al., 2016, JACC Cardiovasc Imagin, 9(6), pp. 690-9, Workstation-based calculation of CTA-based FFR for intermediate stenosis, Chung JH et al., 2017, Am J Cardiol, 120(3), pp. 362-368, Diagnostic performance of a novel method for fractional flow reserve computed from noninvasive computed tomography (NOVEL-FLOW Study), Wang ZQ et al., 2019, J Geriatr Cardiol, 16(1), pp. 42-48, accuracy of a deep learning approach to calculate FFR from coronary CT angiography, Fujimoto S et al. in 2019 in Eur Heart J CardiovascImaging,Diagnostic performance of on-site computed CT-fractional flow reserve based on fluid structure interactions: comparison with invasive fractional flow reserve and instantaneous wave-free ratio, published in JACC Cardiovasc Imaging, 13(4), pp. 980-990, in 2020; Comparison of diagnostic performance in on-site based CT-derived fractional flow reserve measurements, published in Int J Cardiol Heart Vasc, 35, pp. 1-5, published by Nozaki YO et al., in 2021.

[0013] Even if limited to a single in silico evaluation platform, satisfactory sensitivity (understood as true positive rate (TPN)) and specificity (understood as true negative rate (TNR)) values ​​can be achieved for a single clinical trial (TPR / TNR = 93.2% / 82.9%, then 74% / 67% and 85% / 79%). Any result obtained in similar trials can become a problem of the so-called "grey area". This can be better understood when using examples. For example, a few years ago, a team from Imperial College London demonstrated in a review paper that “the diagnostic accuracy of FFR-CT was 86.4% (95% CI, 78.0%-94.0%), 74.7% (95% CI, 71.9%-77.5%), 46.1% (95% CI, 42.9%-49.3%), 87.3% (95% CI, 85.1%-89.5%), and 97.9% (95% CI, 97.9%-98.8%) for vessels with FFR-CT values ​​below 0.60, 0.60 to 0.70, 0.70 to 0.80, 0.80 to 0.90, and above 0.90, respectively” (from Cook CM et al., Diagnostic Accuracy of Computed Tomography-Derived Tomography, JAMA Cardiol, 2(7), pp. 803-810, 2017). Fractional Flow Reserve:A Systematic Review).

[0014] This review paper has caused controversy and debate. It can be concluded that the authors' assessment is not very precise, after all, a diagnostic accuracy of 50% is equivalent to flipping a coin. This level of accuracy is considered unsatisfactory.

[0015] It can be speculated that this ambiguous view of computer simulation technology is caused by its relatively early stage of development. Most of the developments in computer simulation technology are still in the research stage and more time is needed to translate them into commercially developed products. Of course, there are still many factors that need to be improved in these developments. The most promising methods are those concerning the invasive coronary flow reserve fraction (FFR), as FFR is currently described as the gold standard test. Nico Pijls and Bernard De Bruyne, the authors of the concept of the FFR index mentioned above, wrote in their research summary monograph that "…coronary pressure measurement has not been useful for the assessment of blood flow in the past for three reasons: first: the first prerequisite for reliable intracoronary pressure recordings is the use of ultrathin pressure monitoring guidewires, second: pressure measurements are only meaningful during maximal coronary congestion, and third: it is not the gradient that determines myocardial perfusion, but the residual distal coronary pressure." (Quoted from Pijls NHJ and de BruyneB Coronary Pressure, Springer-Science+Business Media, 2000, p. 52). Most likely, in the coming years we will be able to observe the development of proposals for improvements in computer simulation techniques. Today, the first proposal concerns the measuring instrument, which for computer techniques should be interpreted as an improvement in the resolution of medical imaging and discrete geometric models. This proposal does not raise major questions and is well-founded. In addition, the second proposal concerns the improvement of the description of very important physiological circumstances. In particular, those that can make any measurement more or less reliable and, therefore, lead to an incomplete display of the state of coronary insufficiency. The third proposal seems to be expected and does not raise any objections. This proposal concerns what should be evaluated or measured as a diagnostic index. In computer simulations, the fractional flow reserve ratio is calculated by correlating the distal and proximal pressures, just as in invasive procedures. Curiously, the proximal pressure, although mandatory for every clinician, is treated differently by everyone who deals with computer simulation techniques.

[0016] The above in-depth prior art review demonstrates that there is a need for a reliable computer simulation method to assess the condition of the coronary artery system of a human patient. The invention disclosed and claimed herein addresses this need.

[0017] Brief description of the invention

[0018] The following description is for a better understanding of the principles and advantages of the present invention as claimed in the accompanying claims. Therefore, it is not meant to be limiting in any sense.

[0019] The present invention solves the problem of assessing the coronary system by using an appropriate description of the hyperemic state. The inventors have concluded that a reliable functional assessment of the coronary system can only be made if it is known how the coronary system will behave when it is in a hyperemic state. Since the hyperemic state is induced by stimulating adenosine receptors with agonists of said receptors, for said reliable functional assessment it is necessary to assess the hemodynamic state of the patient after exposure to an adenosine receptor agonist. The present invention simulates the patient's response to stimulation with an adenosine receptor agonist.

[0020] In one aspect, the present invention relates to a method for assessing the hemodynamic response of a human patient to stimulation with an adenosine receptor agonist. The method may be implemented as a computer-implemented invention.

[0021] The method according to the invention involves estimating the response of one or more adenosine receptors to stimulation with one or more agonists of these receptors. The estimation comprises obtaining the parameters of systolic and diastolic blood pressure and heart rate in the hyperemic state using a pharmacokinetic / pharmacodynamic (PK / PD) model. This is done in a step called mapping.

[0022] The use of the pharmacokinetic / pharmacodynamic model to map the resting hemodynamic parameters describing a human patient at rest to the hyperemic hemodynamic parameters describing a human patient at hyperemia requires that the pharmacokinetic / pharmacodynamic model be a nonlinear model. The pharmacokinetic / pharmacodynamic model must also include Michaelis-Menten, Hill-Langmuir, Black-Leff, multi-pathway and / or multi-species blood-tissue exchange models (for more details, see, for example, Bassingthwaighte JB, Wang CY and Chan IS in 1989, Circ Res, 71 (3): 590-604, Blood-tissue exchange via transport and transformation by capillaryendothelial cells, Circ Res, 65 (4): 997-1020; Kroll K, Deussen A, Sweet IR (1992) Comprehensive model of transport and metabolism of adenosine and S-adenosylhomocysteine ​​in the guinea pig heart.). The use of at least one model (eg, the Hill-Langmuir model) is sufficient for the present invention. In some cases, the use of a combination of models may provide better results.

[0023] The resting hemodynamic parameters include the patient's systolic blood pressure, the patient's diastolic blood pressure, and / or the patient's heart rate. In one embodiment, the resting hemodynamic parameters consist only of the patient's systolic blood pressure, the patient's diastolic blood pressure, and the patient's heart rate. For example, if the resting hemodynamic parameters consist only of the patient's systolic blood pressure at rest, the patient's diastolic blood pressure at rest, and the patient's heart rate at rest, then the hyperemic hemodynamic parameters consist only of the patient's systolic blood pressure at hyperemia, the patient's diastolic blood pressure at hyperemia, and the patient's heart rate at hyperemia.

[0024] These resting parameters may be measured directly from the patient in any non-invasive manner, or taken from earlier non-invasive measurements made on the patient. Alternatively, a non-invasively recorded continuous patient pressure waveform may be used to assume resting hemodynamic parameters. The waveform may or may not be recorded simultaneously with the resting hemodynamic parameters. In order to reliably assume resting hemodynamic parameters, the waveform must be recorded for a time window that includes at least one complete cycle of the patient's heart. More of the time windows will provide improved results. In order to be able to assume the resting hemodynamic parameters, the waveform must represent a resting state and cannot be a waveform recorded for a hyperemic state. In various embodiments, the resting hemodynamic parameters also include a non-invasively recorded continuous resting patient pressure waveform.

[0025] In one embodiment, the PK / PD model also uses the non-invasively recorded continuous resting patient pressure waveform to calculate the patient pressure waveform in the hyperemic state. This is not beneficial in terms of calculation speed, but it is beneficial because the patient pressure waveform in the hyperemic state can be used during the parameter identification step of the method.

[0026] The method according to the invention then involves obtaining a model describing a patient in a hyperemic state. This is achieved by performing parameter identification of the lumped parameter model using hyperemic hemodynamic parameters describing a human patient in a hyperemic state. In this step, the hyperemic hemodynamic parameters describing a human patient in a hyperemic state are used to adjust the empirical parameters of the lumped parameter model. Therefore, the empirical parameters of the lumped parameter model and the entire model describe the patient in a hyperemic state. In an embodiment, the empirical parameters may include parameters of resistance, compliance, inertia, time-varying elasticity concepts and / or parameters of myocardial fiber strain / stress concepts. In other embodiments, the empirical parameters may also include parameters of myocardial vascular interaction, blood circulatory system (BCS) components, cardiac chamber pressure-volume (HPV) components and / or coronary blood flow (CBF) components. The empirical parameters may include any other additional parameters for adjusting the model for the patient and / or things that are beneficial in terms of model stability.

[0027] The lumped parameter model must be of the Windkessel type, which means that the model must be built using a variant of the blocks developed by Windkessel. The lumped parameter model must at least partially describe the patient's hemodynamics, which is to be understood as involving a description of the relevant part of the blood circulatory system or the entire blood circulatory system of a human patient. The level of description of the hemodynamics is adjusted to the information required to be provided by the present invention.

[0028] The use of a lumped parameter model representing the hyperemic state obtained using the method according to the invention allows a computer simulation diagnosis of a patient. In particular, it allows the calculation of blood flow parameters, which may include pressure, flow velocity and / or cardiac time intervals, including but not limited to cardiac cycle duration and ejection time. The blood flow parameters mentioned are hyperemic transients, and they can be used to obtain a reliable diagnosis of a human patient. In particular, the transients can be used for the diagnosis and treatment of coronary heart disease. Computer simulation diagnosis is beneficial because it allows to bypass all the disadvantages of invasive surgery on the patient.

[0029] The method may include collecting or using demographic and health data that affects the blood flow dynamics of the patient's body circulation. The data may specifically cover data that affects the propagation of pulses in the patient's body. The demographic and health data may include the patient's gender, age, height, general health assessment and / or current medication. Current medications include, but are not limited to, beta-adrenergic blockers, angiotensin converting enzyme inhibitors and / or antiarrhythmic agents. Current medications may include all medications that may affect the human heart. The data may be of any patient, including data taken from a database. The data is used to obtain a first approximation of the empirical parameter values ​​of the lumped parameter model. This shortens the time required to obtain the results of the method. It may also be beneficial in terms of the stability of the calculation. In one embodiment, this is done by finding a relationship between the empirical parameters and the data. The first approximation allows faster calculations, and the method more quickly and stably derives a lumped parameter model that describes a human patient in a congestive state. In another embodiment, when data is collected or obtained from a patient with coronary heart disease, the method obtains a lumped parameter model that describes a human patient in a congestive state even faster and even more stably. The reason for this is that the empirical parameters of the model are closer to the parameters describing a patient in the congestive state that needs to be diagnosed.

[0030] The method according to the present invention may also include recording the patient's pressure waveform in a non-invasive manner. The waveform must represent the patient's resting state. The patient's pressure waveform is recorded in a selected time window that includes at least one cycle of the patient's heart. The patient's pressure waveform can be recorded in a continuous manner, preferably for the radial artery. The preferred technology for recording the patient's pressure waveform is finger-cuff photoplethysmography and / or applanation tonometer. Non-invasive recording refers to recording that does not involve any type of surgery and / or significant health risks; in some embodiments, the recording does not include inserting any probe into the patient's body.

[0031] The key development provided by the present invention is a reliable method to model the myocardial vascular interaction (MVI) which forms part of the three-component model of the present invention. This interaction is required for improving the calculation of the blood flow present in the coronary arteries. The three-component model of the present invention provides a reliable description of the hemodynamics under hyperemic conditions. This also covers the description of the hemodynamics of patients with complex diseases. The three-component model of the present invention provides adaptability and good computational performance.

[0032] In another aspect, the present invention relates to a computer-readable [storage] medium comprising instructions which, when executed by a computer, cause the computer to perform the steps of a method according to the embodiments disclosed herein.

[0033] In another aspect, the present invention relates to a system for assessing the hemodynamic response of a human patient to adenosine receptor agonist stimulation, wherein the system comprises a measuring device for non-invasively measuring a resting hemodynamic parameter of a human patient, wherein the resting hemodynamic parameter comprises the patient's systolic blood pressure, the patient's diastolic blood pressure and / or the patient's heart rate, and a computer system; the computer system is adapted to perform the steps of the method defined in any embodiment of the invention. This aspect allows simultaneous measurement and diagnosis. This is beneficial for the patient to arrive at a diagnosis more quickly. Alternatively, this allows the place where the patient is diagnosed to be separated from the place where the calculations are performed. This is beneficial in many ways, for example, some calculations may require more computer power, which is not available at the place of diagnosis. Another benefit is that it is possible to use a device with less computing power (e.g., a handheld device) to perform the measurements compared to a setup where the same device is used for calculations and measurements / recording.

[0034] In one embodiment, a system for assessing the hemodynamic response of a human patient to adenosine receptor agonist stimulation as defined in the present disclosure comprises a computer system as defined in the present disclosure, the computer system also comprising a recording device for non-invasive continuous recording of the resting pressure waveform of the human patient. This aspect allows even more extensive simultaneous measurement / recording and diagnosis. This is even more advantageous for the above-described embodiments.

[0035] In one embodiment, a system for assessing the hemodynamic response of a human patient to adenosine receptor agonist stimulation as defined in the present disclosure, wherein the computer system as defined in the present disclosure comprises at least one computer adapted to perform a mapping of resting hemodynamic parameters describing a human patient at rest to hyperemic hemodynamic parameters describing a human patient at hyperemia using a pharmacokinetic / pharmacodynamic model, and at least one computer adapted to perform parameter identification of a lumped parameter model using the hyperemic hemodynamic parameters describing a human patient at hyperemia to derive a lumped parameter model describing a human patient at hyperemia, and wherein the at least one computer adapted to perform a mapping of resting hemodynamic parameters describing a human patient at rest to hyperemic hemodynamic parameters describing a human patient at hyperemia using a pharmacokinetic / pharmacodynamic model is configured to communicate directly or indirectly with the at least one computer adapted to perform parameter identification of a lumped parameter model using the hyperemic hemodynamic parameters describing a human patient at hyperemia to derive a lumped parameter model describing a human patient at hyperemia. The system of the present invention allows for separation of where various calculations are performed. This allows for greater flexibility and adaptability to the situation where there is a patient to be diagnosed. In some cases, this may facilitate faster diagnosis.

[0036] Other features and advantages of the present invention will become more apparent from the detailed description and accompanying drawings of non-limiting embodiments. It should be understood that, unless otherwise stated, the embodiments presented in the specification and claims can be combined in any order and quantity to produce new embodiments that form a part of the present disclosure. This description includes a large number of references to prior art documents (particularly scientific journals). The entire disclosure of the references is a part of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The present invention will now be described in more detail with reference to the accompanying drawings, in which:

[0039] Figure 1 An embodiment of the invention is presented which also implements optional steps,

[0040] Figure 2 A block diagram of the generalized form of the three-component model (A) and the three basic building blocks for the lumped parameter model are presented, which can be used to build the blood circulation system (B), cardiac chamber pressure-volume (C), and coronary blood flow (D).

[0041] Figure 3 A block diagram detailing one embodiment of the present invention for determining a three-component model for hemodynamics is presented,

[0042] Figure 4The PK / PD approximation results of the transition from resting state to hyperemia state in patients with typical responses obtained using embodiments of the present invention are presented.

[0043] Figure 5 Presents PK / PD approximation results for the transition from resting state to hyperemia state in patients with heave-type responses obtained using embodiments of the present invention,

[0044] Figure 6 presents results of a regression model of selected internal empirical parameters of a three-component model for describing a patient in a hyperemic state according to an embodiment of the present invention, and

[0045] Figure 7 A comparison of in vivo measured pressure values ​​under hyperemic conditions according to an embodiment of the present invention and pressure values ​​calculated in computer simulation using a three-component model is presented.

[0046] Detailed Description

[0047] The present invention is intended to diagnose coronary insufficiency and provides developments in models and methods for specifying descriptions of hyperemia in silico. As already described, achieving a hyperemic state involves stimulation of specific pharmacological adenosine receptors (usually, the ADORA2 receptor is the target) in clinical practice. Stimulation is performed using adenosine, reganoxan, bisoprolol or apadison administered by intracoronary bolus (ic) or intravenous infusion (iv). Agonist binding takes the form of a cascade of adenosine receptor signaling pathways, which ultimately leads to coronary vasodilation. This results in reducing coronary vascular resistance (CVR) to a minimum and increasing coronary blood flow to the maximum achievable value. The present invention achieves this by being divided into two aspects: achieving coronary blood flow itself, and tracking changes that occur during the hyperemic state.

[0048] In one embodiment of the present invention, the hemodynamic model is defined by a three-component model (3CM) of coronary hemodynamics (see Figure 2 The model includes the following coupled components: blood circulatory system (BCS) component, cardiac chamber pressure-volume (HPV) component, and coronary blood flow (CBF) component. Each component is created using one or more functional blocks, where a functional block must be specific to a given component (for BCS, HPV, and CBF, we use CRL, ERv, and RCpRp, respectively. CRL, ERv, and RCpRp are depicted in Figure 2 Figure B in Figure 2 Figure C and Figure 2The ERv functional block includes a valve that is a diode modeled as a heart valve. A diode is understood to model a unidirectional flow, in accordance with the terminology used in the art.

[0049] The following description will provide detailed relationships in equation form that can be used for each of the blocks. It should be noted that, for example, the pressure parameter (p) in one equation may or may not correspond to the pressure parameter (p) in another equation. The pressure that must be used in each equation can be obtained, for example, from Figure 3 The equation form used is reasonable because each block can be used multiple times in a given component. Listening to all possible configurations with corresponding relationships would unnecessarily extend the description and, due to its complexity, could impair the understanding of the invention.

[0050] In the case of the CRL functional block, the blood flow parameter relationship between input and output can be presented using the following basic equation:

[0051]

[0052] Among them, C, R and L are compliance, resistance and inertia respectively.

[0053] In one embodiment, the systemic (left heart circuit) and pulmonary (right heart circuit) models have the form of at least two CRL functional blocks connected in series for each circuit (so the nth compartment has the nth CRL, where n=2, 3, 4, ...). This is beneficial in terms of obtaining a model that provides a reliable description while not requiring too high a computing power to provide the results. Further, the heart model consists of a 2-chamber type block ( Figure 2 C), these blocks are used for the left and right blood circulation sides respectively. The cardiac chamber hemodynamics can be described by the differential equations given below:

[0054]

[0055] where the symbol <> denotes a Macaulay bracket. In order to close this system of equations, it is desirable and sometimes necessary to know the variation of blood pressure (p) in all four heart chambers during systole and diastole as the volume (V) varies periodically. Unlike in the vascular tree, where compliance is assumed to satisfy C(t)≠const, in the present invention the variation is significant and not of a passive character. In a preferred embodiment of the present invention, the heart chamber fragility is described as the inverse of the time-varying compliance by the concept of time-varying elasticity (E) (see Time course of left ventricular pressure-volume relationship under various enddiastolic volumes, Suga H, 1969, Jpn Heart J, 10 (6), pp. 509-15), which has the following form:

[0056]

[0057] Where V0 is the unloaded volume defined by the relationship p=f(V) of the cardiac chamber at the end of systole.

[0058] The actual realization of relation (3) requires that E(t) approximates the use of cardiac physiology. In a preferred embodiment of the present invention, a periodic double-Hill function is used for approximation (see Determinants of stroke volume and systolic and diastolic aortic pressure published by Stergiopulos N et al. in Am J Physiol, 270 (6 Pt 2) pp. H2050-9, 1996), and the equation is as follows:

[0059]

[0060] and

[0061]

[0062] Where: E has the value defined by equation (3), E min 、E max ——the minimum and maximum values ​​of equation (3), A——the dimensionless proportionality factor, n——the coefficient of the S-shaped characteristic, and t n =(t%T) / t max ——Normalized value of time (%——modulus operator, T——cardiac cycle, t max ——the moment when the elasticity function reaches its maximum value).

[0063] A proper description of the circulation requires a model to complement the description of the pV relationship of the myocardium. In one embodiment, the relationship is expressed using time-varying elasticity (E), and in another embodiment, it is expressed by implementing the myocardial fiber strain / stress concept (see Mirota K, Constitutive Models of Vascular Tissue, Solid State Phenomena, 144, pp. 100-105, 2008, Avazmohammadi R et al., A Contemporary Look at Biomechanical Models of Myocardium, Annual Review of Biomedical Engineering, 21, pp. 417-442, 2019, Voigt JU and Cvijic M, 2-and 3-Dimensional Myocardial Strain in Cardiac Health and Disease, JACC Cardiovasc Imaging, 12 (9), pp. 1849-1863, 2019). The fiber strain / stress concept assumes that for all thin-walled and nearly rotationally symmetric geometries, the fiber stress (s f ) changes with bulk modulus 1 / V·dp / dV~s f , so after integration, the cavity pressure (p) and fiber stress (σ f ) ratio is expressed by the following formula:

[0064]

[0065] The myocardial fiber strain / stress concept (MF) is much more complex compared to the time-varying elasticity concept (E). The reason is that it covers both deformation and strain in the myocardial muscle fibers. On the other hand, the time-varying elasticity concept (E) is beneficial in terms of computational speed.

[0066] (exist Figure 2 The last component of the three-component model used to determine the coronary blood flow dynamics (depicted in Figure A) is the coronary blood flow (CBF) component. This component includes Figure 2 The RCpRp function block is shown in Figure D. The relationship between its inputs and outputs can be presented by the general equation given below:

[0067]

[0068] where p zf is the zero-flow pressure in coronary blood flow.

[0069] In a preferred embodiment of the present invention, the zero flow pressure p zf It is assumed to be equal to 14.8±7 mmHg for patients with stable angina, 22.5±9.1 mmHg for patients with non-Q wave myocardial infarction, and 37.1±1.9 mmHg for patients with Q wave myocardial infarction (see Nanto S et al., Angiology, 47(2), pp. 115-22, 1996, Zero flow pressure in human coronary circulation, Van Herck PL et al., Heart, 93(10), pp. 1231-7, 2007, Coronary microvascular dysfunction after myocardial infarction: increased coronary zero flow pressure both in the infarcted and in the remote myocardium is mainly related to left ventricular filling pressure). The assumption is to facilitate calculation speed and shorten the time required to provide a diagnosis.

[0070] Myocardial vascular interaction (MVI) is used to characterize the coronary circulation. Flow in the coronary arteries is a function of the arterial inlet conditions and arterial outlet conditions, as well as a function of MVI. In order to obtain more reliable values ​​of coronary blood flow, MVI is needed, and there is no clear and reliable MVI modeling strategy known in the art. The development of a modeling strategy for MVI is a very important component of the system developed by the inventors of the present invention. Pressure p C and p RThe role of MVI is described, which is the basis of the modeling (see Boileau E et al., Computational and Experimental Biomedical Sciences: Methods and Applications, 21, pp. 137-155, 2015, One-dimensional modelling of the coronary circulation. Application to noninvasive quantification of fractional flow reserve (FFR), Bruinsma T et al., Basic Res Cardiol, 83, pp. 510-524, 1988, Model of the coronary circulation based on pressure dependence of coronary resistance and compliance, Epstein S et al., American Journal of Physiology, Heart and Circulatory Physiology, 309 (1), pp. H222-H234, 2015, Reducing the number of parameters in 1D arterial blood flow modeling: less is more for–patient-specific simulations, Mohrman D et al., Cardiovascular Medicine by McGraw-Hill, 2013). physiology, Lange NY et al. Cardiovascular physiology, Elsevier, 2013, Zamir M. The physics of coronary blood flow, Springer-Verlag, 2005). If we focus only on passive effects inherent to the hyperemic state, the pressure value is controlled by two mechanisms: (i) interstitial, cavity-induced extracellular pressure CEP = μ1p V (t), and (ii) shortening-induced intracellular pressure SIP = μ2E n(t) (as described in Mechanisms of myocardium–coronary vessel interaction published by Algranati D et al. in American Journal of Physiology. Heart and Circulatory Physiology, 298(3), pages H861-H873 in 2010, Scalability and in vivo validation of a multiscale numerical model of the left coronary circulation published by Mynard JP et al. in American Journal of Physiology. Heart and Circulatory Physiology, 306(4), pages H517-H528 in 2014, and Cross-talk between cardiac muscle and coronary vasculature published by Westerhof N et al. in Physiological Reviews, 86(4), pages 1263-1308 in 2006). Therefore, in a preferred embodiment of the present invention, the coronary blood flow throttling pressure is calculated as follows:

[0071] p R,C =k R,C (CEP+SIP)=k R,C (μ1p(t)+μ2E n (t)), (7)

[0072] Where: p——interstitial intraluminal pressure, k R,C ——resistance and compliance parts of the crown tree coefficient, CEP——cavity-induced extracellular pressure, SIP——shortening-induced intracellular pressure, μ1——cavity-induced extracellular pressure coefficient, μ2——shortening-induced intracellular pressure, and E n - Normal elastance of Eq. 4a This embodiment provides a reliable description required for reliable diagnosis while having balanced computational requirements.

[0073] By using the functional blocks listed above: CRL, ERv and RCpRp (respectively Figure 2 Figure B in Figure 2 Figure C in Figure 2 As shown in Figure D in Figure 2, a complete hemodynamic model can be constructed. Figure 2A summary of the complete hemodynamic model is given in Figure A of Figure 3 A more detailed structure of the model according to a preferred embodiment of the present invention is presented in . In order for the model to be useful in a method as its actual implementation, it must be completed by a set of empirical parameters. The most direct way to achieve this goal is to consult a large number of publications. Research on blood flow modeling over the past few decades has not only produced thousands of articles in scientific journals, but also many monographs on the subject (for example: Mathematical and computer modeling of physiological systems by Rideout VC, published in 1991 through Prentice Hall, Coronary blood flow mechanics, distribution, and control by Spaan JAE, published in 1991 through Springer Science+Business Media Dordrecht, Modeling and Simulation in Medicine and the Life Sciences by Hoppensteadt FC and Peskin CH, published in 1992 through Springer-Verlag; Applied Mathematical Models in Human Physiology by Ottesen JT, Olufsen MS, and Larsen JK, published in 2004 through SIAM, The Physics of Coronary Blood Flow by Zamir M, published in 2005 through Springer Science+Business Media, van Meurs WL (Willem van Meurs) published in 2011 through McGraw-Hill. Modeling and Simulation in Biomedical Engineering: Applications in Cardiorespiratory Physiology) published by Hill.

[0074] Insofar as this approach successfully solves the problem of explicit method formulation, it gives the infinitesimal possibility of taking into account patient-specific variability.

[0075] In an embodiment of the invention, one can individually determine or measure selected (preferably those that are easily measurable) anatomical or hemodynamic parameters for each analysis case, and then use the anatomical or hemodynamic parameters to calibrate the empirical parameters of the model (see Kirk JA et al., Cardiovasc Eng Technol, 4 (4) pp. 500-512, 2013, A priori identifiability analysis of cardiovascular models, Stergiopulos N et al., Am J Physiol, 270 (6 Pt 2) pp. H2050-9, 1996). In a preferred embodiment of the invention, it is a continuous non-invasive measurement using photoplethysmography.

[0076] The problem solved by the present invention is handled differently in the prior art. The following briefly describes the strategies that are not used in the present invention and form the prior art. This is relevant to a correct understanding of the present invention.

[0077] In the prior art, there are strategies based on statistical methods. In the simplest attempt, one can try to estimate patient-specific empirical parameter variability by using population-based statistical rules suitable for demographic or routine medical data (according to Burattini R and Di Salvia PO, Development of systemic arterial mechanical properties from infancy to adulthood interpreted by four-element windkessel models, J Appl Physiol (1985), 103(1), pp. 66-79, 2007; McVeigh GE et al., Age-related abnormalities in arterial compliance identified by pressure pulse contour analysis: aging and arterial compliance, Hypertension, 33(6), pp. 1392-8, 1999; Kassab GS, Coronary Circulation, Anatomy, Mechanical Properties, and Biomechanics, Springer Science+Business Media, 2019; Westerhof BE et al., Pressure and Flow Relations in the Front Pediatr, 8, p. 251, 2020). Systemic Arterial Tree Development Throughout From Newborn to Adult).

[0078] A very easy to measure hemodynamic parameter is the pressure at the aorta level. At the same time, pressure is a direct reference data for the model prediction results. In this regard, available are the auscultatory method using a stethoscope and a sphygmomanometer or the oscillometric method.

[0079] Next, cardiac output estimation is achieved based on echocardiographic assessment (attempts in this regard can be found in Lang RM et al., 2015, J Am Soc Echocardiogr, 28(1), pp. 1-39, Recommendations for cardiac chamber quantification by echocardiography in adults: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging, and Dumesnil JG et al., 1995, Can J Cardiol, 11(11), pp. 1007-14, A new, simple and accurate method for determining ejection fraction by Dopple rechocardiography) or by using computed tomography to assess cardiac anatomy.

[0080] Based on knowledge of the mass or volume of the cardiac chambers, cardiac output is estimated from myocardial blood flow (MBF) [ml / min 100 g] (see Anderson HV et al., 2000, Circulation, 102(1), pp. 48-54, Coronary artery flow velocity is related to lumen area and regional left ventricular mass), which is expressed as follows:

[0081] MBF=8·(7·10 -4 HR·SYS-0.4), (8)

[0082] Alternatively, cardiac output can be estimated directly by estimating myocardial oxygen consumption MVO2 [ml O2 / min 100g] (Baller D et al., 1979, Clin Cardiol, 2(5) pp. 317-27, Validity of myocardial oxygen consumption parameters provides a very extensive review of potential available estimates). Figure 3The strategy of using empirical parameters of the model (preferred embodiment) comprises: selecting (calibrating) empirical parameters of the model according to available partial characteristics of hemodynamics identified as used (such as pressure, myocardial blood flow, etc.). This method allows to directly relate to the real and actual hemodynamic state, but has significant limitations.

[0083] Unfortunately, the strategy presented above is flawed in the context of functional assessment for the diagnosis of coronary heart disease using the above strategy. Its disadvantage is the accessibility of some parameters of the target hemodynamic state - if the model prediction is related to the hyperemic state, its parameters must be established for the hyperemic state and not for the resting state. In fact, it is possible to measure pressure (SYS, DIA) and heart rate (HR) in the resting state (or even continuously using pressure waveforms without difficulty or complexity). However, it is not reasonable to administer adenosine or another vasodilator used in pharmacological stress tests during the said measurement of SYS, DIA and HR.

[0084] The prior art provides various other strategies to solve the problems of the present invention. In the simplest approach, one can try to reconstruct hyperemia by assuming some arbitrary changes in the main hemodynamic parameters. For example, one can assume that the coronary resistance is reduced by a constant factor of 4, while the aortic blood pressure is reduced by about 10% and the heart rate is increased by about 10% (as disclosed in U.S. Pat. No. 8157742B2, entitled Method and system for patient-specific modeling of bloodflow, published by Taylor CA in 2012; U.S. Pat. No. 20210282860A1, entitled Method and system for image processing to determine blood flow, published by Taylor CA in 2021). As shown by an extensive review of clinical trial results, there is a very high risk that the results obtained with those prior art methods will be far from reality. Mainly because the patient-specific variability here is huge, and what is missing here is a direct correlation with demographic or conventional medical data.

[0085] The observation of high patient-specific variability in adenosine receptor stimulation was made long before any model was developed to quantify said variability. Therefore, in looking for suggestions of mechanisms to explain this phenomenon, it is worth mentioning earlier prior art methods. This will help to better understand the details of the present invention.

[0086] To test the dose-response kinetics of intravenous and intracoronary adenosine, Wilson et al. performed a test in 39 patients. Initially, coronary angiography was performed to diagnose the patient's chest pain syndrome. The 39 patients were divided into two groups, 8 patients with microcirculatory abnormalities related to ventricular hypertrophy of unknown etiology and 31 patients with atypical chest pain due to angina, mild or normal stenosis of the epicardial coronary arteries, normal coronary flow reserve (CFR), and normal left ventricular function. The researchers are testing the kinetic response to adenosine doses administered by intravenous infusion, intracoronary infusion, and intracoronary bolus. Single doses for intracoronary bolus ranged from 2 to 16 μg, intracoronary infusions ranged from 10-240 μg / min, and intravenous infusions ranged from 35-140 μg / kg / min. A Doppler coronary catheter 3F was used to measure coronary blood flow velocity (CBFV). The maximum change in CBFV was calculated as the ratio of peak CBFV to resting CBFV after intracoronary bolus of adenosine or papaverine (ΔCBFV).The index of change in total coronary artery resistance was calculated as follows.

[0087]

[0088] The change in CBFV during intracoronary adenosine infusion was calculated as:

[0089]

[0090] The change in CBFV during intravenous adenosine infusion was calculated as follows:

[0091]

[0092] Intracoronary adenosine safely achieves peak coronary vasodilation and almost peak hyperemia after intravenous injection of 140 μg / kg / min of adenosine (see Wilson RF et al., Effects of adenosine on human coronary arterial circulation, Circulation, 82(5), pp. 1595-1606, 1990). Another important observation is that different doses of adenosine have different effects on test results.

[0093] Researchers led by Sparv are investigating the effects of increasing adenosine doses on fractional flow reserve (FFR), mean arterial pressure (MAP), heart rate (HR), and discomfort (measured by visual analog scale (VAS)). Eighty-seven patients participated in the trial (10 patients developed atrioventricular block after adenosine infusion and were excluded from the trial, and two patients withdrew due to great discomfort). During the assessment of FFR, patients received a standard dose of 140 μg / kg / min of intravenous adenosine, followed by a high dose of 220 μg / kg / min of intravenous adenosine. Another group of patients was also analyzed, but this group included 43 patients who consumed caffeine less than 6 hours before FFR. The results showed that there were no significant changes in FFR results, MAP, and HR between standard and high doses of adenosine. The only difference observed between standard and high doses of adenosine was higher discomfort in patients after high doses of adenosine. During the trial, MAP and HR were monitored and recorded at 30-second intervals. Statistical analysis was performed using the Wilcoxon paired signed rank test and linear regression model correlation of adenosine dose. Bland-Altman plots were created to show agreement, and the area under the curve (AUC) was analyzed for the continuous hemodynamic variables of MAP and HR (see Assessment of increasing intravenous adenosine dose in fractional flow reserve, Sparv D et al., 2017, BMC Cardiovascular Disorders, 17, pp. 1-9).

[0094] Another existing approach was developed by Robert Wilson and his team. Robert Wilson et al. from the University of Minnesota Medical School not only provided a multilateral adenosine dynamics assessment, but also a total coronary resistance index TCRI (see equation (9a)). It does parallel the linear resistance laws that appear in many branches of physics (such as Ohm's law or Hagen-Poiseuille's law). It can be assumed that his work was an attempt to gain a deeper understanding of the problem, but it did not lead to anything close to a breakthrough. However, it is encouraging that this work shows a slightly different strategy for the selection of empirical parameters of contemporary hemodynamic circulation models for hyperemic states. That is, instead of making questionable ad hoc assumptions, because flow or heart rate may change in specific circumstances, one can try to establish a hypothesized structural relationship between them (see Sharma P et al., A framework for personalization of coronary flow computations during rest and hyperemia, Annu Int Conf IEEE Eng Med Biol Soc, 6665-8, 2012; Sharma et al., U.S. Patent No. 10373700B2, Non-invasive functional assessment of coronary artery stenosis including simulation of hyperemia by changing resting microvascular resistance, published in 2019; Itu LM et al., U.S. Patent No. 10622110B2, Framework for personalization of coronary flow computations during rest and hyperemia, published in 2020).

[0095] By analogy with systemic circulatory resistance (usually 9…20 mmHg min / l), the following formula defines coronary circulatory resistance:

[0096]

[0097] And the TRCI index as the quotient of the determined resistances - for the resting and congested states, is expressed here as:

[0098]

[0099] This equation allows obtaining the resistance parameters in the hyperemic state based on the resistance parameters in the resting state. This is because R hiper =TRCI·R rest Obviously, the effectiveness of this strategy will depend on the determination of the TRCI index value.

[0100] TRCI can also be determined so that it will include a correction for HR. See equation (11) (as shown in Sharma P et al., 2012, which is based on the clinical trial results of McGinn et al., 1990, Circulation, 81 (4) pp. 1319-30, Interstudy variability of coronary flow reserve. Influence of heart rate, arterial pressure, and ventricular preload), which is in the following form:

[0101]

[0102] In some cases, a strategy based on the concept of the TRCI index can provide flexibility and can even be effective in reconstructing the impact of patient-specific variability. However, this strategy can only be effectively applied when the set of empirical parameters in the model is small. In the example cited above (specifically, shown by Sharma P et al., 2012), the recalculation associated with the transition from resting state to hyperemia was related only to resistance. In other words, this strategy is only valid for a selected model, and it will fail for any model that aims to describe more complex medical cases. One cannot ignore the individual medical parameters of the patient.

[0103] When compared to the previous description, various embodiments of the present invention utilize a different strategy, in particular different from the use of population-based statistics or TRCI. The strategy of the present invention is based on a mapping between resting and hyperemic states, which is accomplished by selected hemodynamic parameters (e.g., SYS, DIA, etc.) that are fitted to the results provided by a pharmacokinetic / pharmacodynamic (PK / PD) model of adenosine metabolism.

[0104] In the following paragraphs, important methods of PK / PD modeling of adenosine metabolism will be described. In the field of PK / PD modeling of adenosine metabolism, a base point is the classic work of Blardi et al., from the team of the Institute of Clinical Medicine of the University of Siena, in which experiments were performed after infusion of adenosine at doses of 2.5, 5 or 10 mg (38, 79 and 48 μg / kg, respectively), and after administration of an infusion of 200 μg / kg over a period of 10 minutes, and then one minute after administration of a dose of 400 μg / kg over a period of 10 minutes. Twenty-two healthy volunteers were divided into 4 groups: Group A (6 volunteers, intravenously administered 2.5 mg adenosine via 1-minute infusion), Group B (6 volunteers, intravenously administered 5 mg adenosine via 1-minute infusion), Group C (5 volunteers, intravenously administered 10 mg adenosine via 1-minute infusion), and Group D (5 volunteers, intravenously administered 200 μg / kg of adenosine infusion over a 10-minute period, followed by an intravenous adenosine infusion of 400 μg / kg over a 10-minute period). Blood samples were collected at the time of adenosine infusion (time 0) and at selected time points (times 1, 3, 5, 7, 10, and 15 minutes for groups A, B, and C, and time 0, 5, 10, 15, 20, 21, 25, and 30 minutes for group D). One minute after adenosine infusion, adenosine concentration increased, mean clearance decreased, and half-life increased. After 20 minutes, adenosine plasma concentrations reached a peak value compared to the concentrations achieved after a 5 mg adenosine infusion administered over a 1 minute period, but the mean clearance and half-life were different. Adenosine concentrations returned to baseline 5-15 minutes after the infusion was stopped. Adenosine plasma concentrations were checked by HPLC. Pharmacokinetic analysis of adenosine concentrations over time compared to baseline concentrations was performed. The adenosine concentrations C(t) for groups A, B, and C were calculated as

[0105]

[0106] Where: t is time, A and β are parameters estimated by nonlinear regression (see Blardi P et al., Pharmacokinetics of exogenous adenosine in man after infusion, European journal of clinical pharmacology, 44 (5), pp. 505-507, 1993).

[0107] Although Blardi's work differs from contemporary PK / PD models used in the field, it is from the early 1990s and is a very important reference point.

[0108] Another very interesting prior art attempt to quantify the dynamics of adenosine is the work of Davis et al. The researchers attempted to determine the ATP channel) and N ω -Effect of nitro-L-arginine methyl ester (L-NAME) (blocks nitric oxide synthesis). Adenosine was administered by coronary artery infusion (using a dose of 10-1000 μg / min). The dog was first appropriately prepared and then anesthetized using mechanical ventilation. A left-sided tracheotomy was performed and catheters were placed in the descending aorta and left atrium. A Konigsberg micromanometer was placed in the left ventricle through the left ventricular apex. A Transonic Systems transit time ultrasound blood flow probe and a hydraulic occluder were placed in the proximal left circumflex coronary artery or in the left anterior descending coronary artery. A 22-gauge angiocatheter connected to a small-bore tube was placed in the artery, distal to the blood flow probe. A bipolar pacing lead was sutured on the left atrial appendage. Under basal conditions, a 50% maximum increase in conductivity (ED) was achieved after administration of 1 mg / kg or 10 mg / kg L-NAME. 50 ) requires a 3-fold increase in adenosine plasma concentration. After administration of glibenclamide, ED 50 The dose-dependent increase in ED 50 Increased 15 times, added K ATP Channel and NO synthesis blockade. Glyburide and L-NAME lead to increased aortic pressure and decreased coronary conductance. During the experiment, hemodynamics were monitored in a simulated form with a sampling rate of 30 seconds. For each individual, the variable mean was recorded every 30 seconds. Adenosine plasma concentration was calculated using the Olsson method (see equation 14). The dose-response data of adenosine were fitted to a sigmoid model, expressed as follows:

[0109]

[0110] Where: x-logarithmic adenosine plasma concentration; f(x)-conductivity; A, B, C, D-constants (see Coronary vascular responsiveness to adenosine is impaired additively by blockade of nitric oxide synthesis and a sulfonylurea, published by Davis CA et al. in 1998 in Journal of the American College of Cardiology, 31(4) pp. 816-822). In this review, the typical S-shaped adenosine receptor response model is directly mentioned.

[0111] Tune et al. did further prior art work. They were analyzing the role of adenosine in exercise hyperemia in the coronary arteries of ten dogs. Catheters were placed in the aorta and coronary sinus. A blood flow probe was placed in the circumflex coronary artery. Adenosine concentrations in the cardiac interstitium were estimated from coronary and arteriovenous plasma concentrations using a mathematical model. Coronary blood flow (CBF), myocardial oxygen consumption, heart rate (HR), and aortic pressure were measured at rest and during exercise on a treadmill with and without adenosine receptor blockade. During exercise, oxygen consumption increased 4.2-fold, CBF increased 3.8-fold, HR increased 2.5-fold in the control dogs, and mean blood pressure (MBP) remained unchanged. Adenosine plasma concentrations in the coronary veins changed little during exercise. The estimated interstitial adenosine concentrations were well below the threshold for coronary vasodilation. After adenosine receptor blockade, oxygen consumption or CBF at rest and during exercise were almost stable. Adenosine concentrations in the coronary veins and estimated interstitial adenosine concentrations did not change to compensate for receptor blockade. Adenosine arterial plasma concentration [ADO] was calculated using the following equation:

[0112]

[0113] Where: [Ado] - adenosine concentration, Hct - hematocrit. Subsequently, the Hill equation dose-response adenosine data were fitted using the following equation:

[0114]

[0115] Where: q——coronary blood flow, q min ——Measured resting coronary blood flow, q max ——measured peak coronary blood flow, [A] ——calculated arterial adenosine concentration, ED 50 ——50% of the effective dose, H——Hilby (see Tune JD et al., Adenosine is not responsible for local metabolic control of coronary blood flow in dogs during exercise, American Journal of Physiology-Heart and Circulatory Physiology, 278 (1), pp. H74-H84, 2000).

[0116] The present invention approaches the PK / PD model in a different way, which will be presented below. To achieve biochemical control, nature has developed very precise regulatory mechanisms in the form of enzyme-catalyzed reactions. In addition to the substrate S and the product P, a biocatalyst E is present in the device, whose activity is subtly regulated, adapted to the needs and circumstances. This relationship can be summarized in its simplest form as the chemical equation: It is reasonable to assume that the concentration of the enzyme-substrate complex (ES) does not change significantly during the reaction. Similarly, it can be assumed that an equilibrium occurs between the synergistic effects of the complex and the substrate: k1[E][S]-k -1 [ES] = k2[ES]. By combining all of this, we obtain the simple relationship that describes the rate of product formation as a function of substrate accessibility:

[0117]

[0118] The equation presented above was discovered by Leonor Michaelis and Maud Menten in their studies of fermentation kinetics. Over time, this equation has become very useful for describing widely understood enzyme-catalyzed processes (see RodwellVW et al., Harpers Illustrated Biochemistry, McGraw-Hill Medical, 2018). In the case of complex biochemical systems, the number of reactants can be large and the reactions often form cascade systems. A good example is purine receptors (especially adenosine receptors), whose reactions often proceed in parallel with the use of n catalytic centers, so (according to Hemodynamic aspects of atherogenesis, Mirota K, 2013) we have the following equation:

[0119]

[0120] Among them, K n =K1·K2·…·K n When evaluating the effect of an agonist A binding to a receptor R at one of its n catalytic centers, we must conclude that [S] = [AR], so according to the Michaelis-Menten model we will have the following equation:

[0121]

[0122] By combining these two equations and introducing the transducer ratio (defined as τ = [R0] / K A), we obtained a new equation describing the contractual catalytic effect (based on the original work published by Black JW and Leff P in Proc R Soc Lond B Biol Sci. 220(1219):141-62 in 1983: Operational models of pharmacological agonism):

[0123]

[0124] This is a generalization of equations (16a) and (16b).

[0125] The transition between resting and hyperemic states is not limited to regulating active coronary vasomotor tone. In fact, the observed and investigated changes in coronary circulatory resistance are only a symptom of some of the effects produced by different forms of cardiac tropism (including heart rate, myocardial contractility, myocardial vasomotor tone, etc.).

[0126] In a preferred embodiment, the transition from baseline to hyperemic state is considered a net cardiotropic effect of purinergic receptor agonist binding and is modeled by a cooperative binding relationship of the form:

[0127]

[0128] Where: X is an arbitrary measurement reflecting changes in ligand concentration, and K represents the apparent dissociation constant. This equation represents one approach to PK / PD modeling.

[0129] In another embodiment of the present invention, the net cardiac tropism is modeled by the Black-Leff operational model of pharmacological agonism (based on Black JW and Leff P, Proc R Soc Lond B Biol Sci, 220 (1219), pp. 141-62, 1983, and Leff P, Martin GR and Morse JM, Application of the operational model of agonism to establish conditions when functional antagonism may be used to estimate agonist dissociation constants, Br J Pharmacol, 85 (3), pp. 655-63, 1985), and has the following form:

[0130]

[0131] where: X and K are also conceptual values ​​for the concentration change and dissociation constant, and τ is a ratio that defines the agonist's efficacy (see Jambhekar SS and Breen PJ, Basic Pharmacokinetics, Pharmaceutical Press, 2012, Frigyesi A and O, Estimating the parameters of the operational model of pharmacologicalagonism, Stat Med, 25(17), pp. 2932-45, 2006.

[0132] All embodiments of the present invention have been tested. The method according to the present invention is verified based on medical experimental results from multicenter, non-randomized clinical trials. It is planned to perform this study on 60 participants. Finally, 67 patients were screened and 48 patients were included. The collected patient demographic and health data affecting the hemodynamics of the body circulation are presented in Table 1 (parameters are expressed as mean ± standard deviation). For this purpose, data affecting the propagation of pressure pulses in the patient's body are presented by category.

[0133]

[0134]

[0135] For each included patient, continuous noninvasive blood pressure measurement (CNBP) was performed in the radial artery at a sampling rate of 100 Hz. In addition, independent noninvasive measurements were performed in the brachial artery using the oscillometric method. Since the patients underwent invasive diagnosis, the results of the intra-aortic catheter pressure measurement were used as reference data. The pressure signal from the intra-aortic catheter pressure measurement was sampled at a rate of 200 Hz.

[0136] Every patient included in a clinical trial is diagnosed with CHD, however the trial itself must confirm CHD via functional assessment. This is presented in Table 2.

[0137]

[0138]

[0139] This specific selection of the validation group (only humans diagnosed with coronary heart disease) reflects the intended application focus of the present invention.

[0140] Figure 4Results of an embodiment of a method according to the invention are presented which incorporates a model based on the Hill-Langmuir or Black-Leff equations. Both types of equations produce similar results and can be used interchangeably. In the background, there is a 3-minute recording of a left coronary artery catheterization, and the predictions of the changes in systolic pressure (triangle down) and diastolic pressure (triangle up) are plotted with solid lines. At rest, the mean systolic pressure is at SYS=122.7±3.0 mmHg (which gives a coefficient of variation cv=2.4%), and the diastolic pressure is at DIA=57.9±1.51 mmHg (while cv=2.6%). Adenosine was administered at exactly 43.7 seconds, and stable hyperemia was reached at 108.44 seconds. Afterwards, the pressure dropped—SYS=92.5±1.5 mmHg (cv=1.6%), and DIA=32.1±0.6 mmHg (cv=1.8%). Basically, whether model (8) or model (9) is used, the predictions follow the same pattern (coincidence) as shown by the SYS and DIA pressure lines. The mean residual for the systolic value area for model (8) is only 1.6 mmHg, not exceeding the range of -5.2-12.2 mmHg, and for diastolic pressure, the mean residual is 0.4, within the range of -3.7-3.3 mmHg. Similarly, for model (9), the mean residual for systolic pressure is 1.3 mmHg, within the range of -6.0-13.1 mmHg, and for diastolic pressure is 0.3, within the range of -3.6-3.2 mmHg. It should also be noted that Figure 4 The case A shown expresses the response to adenosine receptor stimulation in a classic S-shaped form. According to statistics, this form is observed in about 57% of clinical cases, so there is a little more than one in every two patients. A completely different reaction pattern is observed in about 39% of patients (i.e., two out of five patients enter a hyperemic state). For these patients, the reaction pattern is a protuberance type, so it is an S-shape with superimposed variant protuberances (see Johnson NP in 2015, JACC Cardiovasc Interv, 8 (8) pp. 1018-1027, Repeatability of Fractional Flow Reserve Despite Variations in Systemic and Coronary Hemodynamics).

[0141] Figure 5A second response pattern to adenosine is shown. On the resting state side, there are pressures SYS = 125.5 ± 2.1 mmHg (cv = 1.7%) and DIA = 71.9 ± 1.2 mmHg (cv = 1.7%). In fact, in the transient phase, higher residual values ​​can be clearly seen and the model used has no significant impact on these residual values. For the first model (Equation 18), the systolic pressure is -3.7 (-13.9 ... 9.5) mmHg and the diastolic pressure is -1.6 (-7.4 ... 5.8) mmHg. For the second model (Equation 19), they are -3.6 (-13.9 ... 10.0) mmHg and -1.6 (-7.5 ... 6.7) mmHg, respectively. The residual values ​​of the residuals of the protuberance type are higher in the transient phase, but this does not have a negative impact on the prediction quality of both models for the maximum stable hyperemic state.

[0142] A method comprising the model developed above using (18) and (19) provides the possibility of using the results of in vivo measurements made during the patient's resting state (e.g., central aortic pressure measured non-invasively using photoplethysmography) to efficiently map from the resting state to the expected state under hyperemic conditions.

[0143] In a preferred embodiment of the present invention, a regression model can be used, which provides the starting values ​​of the internal empirical parameters of the three-component model (3CM) (according to Heijmans RDH and Magnus JR, published in J Econ, 32, pp. 253-285 in 1986, Consistent maximum-likelihood estimation with dependent observations. The general (non-normal) case and the normal case, Escobar J, published in J Math Control Inform, 29 (2), pp. 235-258 in 2012, Time-varying parameter estimation under stochastic perturbations using LSM, Banks HT, Hu S and Thompson WC, published in 2014 by Springer International Publishing. Model validation and uncertainty quantification). The model is used to process data from experiments.

[0144] As mentioned above, clinical trial population data are used for parameter identification. They are used to estimate the empirical parameter values ​​of 3CM. The one-time completion of parameter identification provides a model that can be used for patient-specific calculations. This is particularly preferred in conventional usage when it is necessary to provide fast and accurate starting values ​​for calculations of specific cases.

[0145] Figure 6 The prediction results for selected empirical parameters of 3CM are shown, where 3CM has Figure 3 Structure shown. In each graph, the horizontal axis represents the value of the model parameter obtained as a result of calibration (parameter identification) with respect to the values ​​obtained using invasive measurements (performed in vivo for the hyperemic state). The circle symbols mark the exact position of the value obtained using the in vivo measurements. Calibration was performed separately for each patient participating in the clinical trial within a given window of five cardiac cycles. When evaluating the quality of reproduction of the invasive measurement results, the L2 norm of the pressure difference (in vivo vs. computer simulation) was calculated. For all cases and cycles, the L2 norm value was 2.3 mmHg (interval 2.1…2.5 at 95% confidence level). The empirical parameter values ​​resulting from the calibration are statistically likely to represent the hyperemic state of a particular patient.

[0146] Subsequently, for each parameter, a regression model was constructed using demographic and general medical data. The results of this model are plotted in Figure 6 The vertical axes are plotted as shown in Figure 1, where asterisks mark the values ​​of a given empirical parameter obtained using the corresponding demographic and general medical data. In addition, each graph contains a simple regression represented by a solid line and a range with a 95% confidence level represented by a dashed line. Figure 6 Only ten selected empirical parameters out of dozens of empirical parameters included in 3CM are presented. Figure 3 and Figure 6 The same type of markers are used in the 3CM parameter identification using demographic and medical data. It is not difficult to notice that the results of the continuous non-invasive measurement and the results obtained by the calibration using invasive data have a very high correlation (correlation values ​​are given in percentages in the caption of each figure). The adaptability of the method according to the invention to specific patients and their variability (e.g., variability described by the patient's demographic and medical data) is shown in Figure 6 The use of a regression model is completely optional for the present invention and does not allow obtaining more precise results. It should be noted that a regression model implementing statistically invasive data is able to provide much better starting values ​​for the method of the present invention. This leads to a conclusion of the parameter identification of the model several times faster and, as a result, allows obtaining patient-specific parameters calculated by the model much faster.

[0147] Figure 7Selected results obtained with an embodiment of the method of the present invention are presented. Results are given using two extreme borderline cases. This selection allows comparison. Two selected cases of hyperemic states: Case A (left) and Case B (right). The first case presents the most typical and common form of response of the systemic circulation to adenosine administration, such as known from textbooks. The second (Case B) is extremely unusual and peculiar. It should be noted that this is the only such sample in the sample library selected for clinical trials. This case was deliberately selected for comparison to benchmark the present invention in the most adverse environment. Nevertheless, for both cases, very good agreement was achieved between the results obtained using the method of the present invention and the results obtained using the results obtained from invasive measurements. The agreement was evaluated by calculating the Perason correlation coefficient reaching 99% (Case A) and 98% (Case B) values. The variability diagram presented as a function of time of the calculated pressure (solid line in the figure) and the invasively measured pressure (points in the form of circles) also shows excellent agreement. In the fourth cycle of condition B, a large deviation (about 10 mmHg) of the maximum value range was observed. This deviation was expected due to the extreme features of case B.

[0148] In one embodiment, the method of the present invention comprises Figure 1The five steps shown. In the first step, the patient's demographic and general medical data are collected. As mentioned above, this step is not required for the method, but it may be beneficial for calculation speed. Although this data is only used in step 4 of the method (which is related to the parameter identification of the model), the importance of this data is related to the efficiency of the method. First, data related to gender, age, weight and / or height are collected (in one embodiment, according to Influence of body height on pulsatile arterial hemodynamic data published by Smulyan H et al. in 1998 in Journal of the American College of Cardiology. 31(5):1103-9, Relationship between Resting Heart Rate, Blood Pressure and Pulse Pressure in Adolescents published by Christofaro DGD et al. in 2017 in Arquivos Brasileiros de Cardiologia. 108(5):405-410, Body Size Predicts Cardiac and Vascular Resistance Effects on Men's and Women's Blood Pressure published by Evans JM et al. in 2017 in Front Physiol. 9; 8:561, Testing of Cardiovascular Diseases published by Gallo C et al. in 2021 in Cardiovascular Engineering and Technology. 12(2):144-157). aPatient-Specific In-Silico Model to Noninvasively Estimate Central Blood Pressure estimates its impact). Drug therapy can also affect pulse wave propagation. In particular, drug therapy involving beta-adrenergic blockers (BBLOCK), angiotensin converting enzyme inhibitors (ACE) and / or antiarrhythmic agents (AARR) is relevant (in one embodiment, according to Harris WS, Schoenfeld CD and Weissler AM in 1967 in Journal of Clinical Investigation.Effects of adrenergic receptor activation and blockade on the systolic preejection period, heart rate, and arterial pressure in man, published in 1974 in Postgraduate Medical Journal. 46(11):1704-14; A comparison of beta adrenergic blocking drugs in the treatment of hypertension, published in 1974 in Postgraduate Medical Journal. 50(583):253-259; Effect of beta-adrenoreceptor blockers on heart rate and blood pressure indynamic and isometric exercise, published in 1976 in Drugs. 11SUPPL 1:185-95 by Nyberg G; Hemodynamic effects of angiotensin-converting enzyme inhibitors in essential hypertension: a review, published in 1985 in Journal of Cardiovascular Pharmacology. 7Suppl 1:S35-9 by Fitzpatrick MA and Julius S; CT et al., Arterial hemodynamics in human hypertension. Effects of angiotensin converting enzyme inhibition., Hypertension. 22(6): 839-46, 1993; Jobs A et al., Angiotensin-converting-enzyme inhibitors in hemodynamic congestion: a meta-analysis of early studies., Clinical Research in Cardiology. 108(11): 1240-1248, 2019; Block PJ and Winkle RA, American Journal of Cardiology.52(6):14C-23C, Weiner B, 1991, Journal of Cardiovascular Nursing.5(4):39-48. to estimate its effects). In some embodiments, drug therapy is included in the method, and in other embodiments, the drug therapy includes the drugs listed above. Other drugs can also affect the propagation of pulse waves in the patient, and therefore, other embodiments include other drug therapies. It should be noted that drug therapy also includes different dosing regimens and any further medical effects of drug therapy. Each of the above factors can be used alone or in combination in any way to obtain a better starting value (start) for the model identification process for a specific patient case. A better starting value directly affects the efficiency of the method according to the present invention. It is expected that in some embodiments, only selected data, such as gender, age, or selected drugs, will be collected. The more data used, the better the starting value can be obtained. .

[0149] In the second step of the method of the present invention, Figure 1As shown, a pressure waveform of a patient at rest is obtained. This step is also not necessary for the present invention, but can be used for the parameter identification step. Alternatively or additionally, this step can be used to assume the main hemodynamic parameters for the purpose of the third step. Taking into account the limitations and preferences of clinical practice when measuring continuous non-invasive blood pressure (CNBP), in a preferred embodiment of the present invention, pressure measurement is performed in the radial artery. In one embodiment, pressure measurement is performed using a finger cuff photoplethysmography and / or an applanation tonometer. The benefit of using the measurement techniques is that they are easy to use, and the present invention will work even for these techniques. Recording and further analysis of continuous pressure waveforms are performed within a window including signal blocks covering one or more complete cycles of the patient's heart. Each cycle corresponds to the contraction-relaxation action of the heart. In a preferred embodiment of the present invention, a window with a width corresponding to the length of the respiratory cycle can be analyzed (according to Normal respiratory rate and peripheral blood oxygen saturation in the elderly population. published by Rodríguez-Molinero A in Journal of the American Geriatrics Society. 61 (12): 2238-2240 in 2013, Real-time estimation of respiratory rate from a photoplethysmogram using an adaptive lattice notch filter. published by Park C and Lee B in Biomedical Engineering Online. 17 in 2014; 13: 170, Scholkmann F and Wolf U in Front Physiol. 9 in 2019; The Pulse-Respiration Quotient: A Powerful but Untapped Parameter for Modern Studies About Human Physiology and Pathophysiology. published in 10: 371). This embodiment is beneficial because it allows more accurate assumptions about parameters.

[0150] In the third step of the method of the present invention, Figure 1As shown, the main hemodynamic parameters of the patient in the hyperemic state are determined. This is done using a mapping process, during which the parameters describing the resting state are transposed to the parameters describing the hyperemic state. The mapping of the parameters obtained using the results of non-invasive measurements performed in the resting state is done using the PK / PD model and equations (18) and (19). In one embodiment, only the main hemodynamic parameters including systolic blood pressure (SYS) and diastolic blood pressure (DIA) and heart rate (HR) can be mapped. In another embodiment, additional hemodynamic parameters can be used. In one embodiment, the PK / PD model is not directly used for 3CM (with Figure 3 In another embodiment, the waveform is not used in the PK / PD model, and the waveform of the hyperemia state is provided by 3CM. Compared with the embodiment in which the waveform is calculated using the PK / PD model, this embodiment is more time-saving.

[0151] In the fourth step, parameter identification of the hemodynamic lumped parameter 3CM is performed ( Figure 3 ). To this end, a regression model can be used. The regression model provides a random quantitative description. The regression model can be used to provide a starting value for 3CM. It should be noted that in one embodiment, the regression model is not used in a method that requires additional time to complete the calculation. This additional time is close to tens of hours. The embodiment with the regression model obtains results faster, and the parameter identification in this embodiment is a more numerically stable process. In another embodiment, parameter identification is performed independently and in parallel with adenosine receptor stimulation. Any suitable mathematical method can be used to refine the empirical parameter values ​​using hyperemic hemodynamic parameters.

[0152] In a fifth step, one or more blood flow parameters in the systemic, pulmonary and / or coronary circulations are calculated by means of a three-component model (3CM). The blood flow parameters include flow velocity and pressure in locations corresponding to the used structure of the 3CM. The specific structure of the 3CM is of secondary importance for the heart chamber blood flow parameters (in each case two atria and ventricles), but for the blood circulatory system (BCS), the differences may be important. Figure 3 The embodiment of the BCS model with two compartments for systemic circulation and two compartments for pulmonary circulation only provides the flow rate and pressure related to these compartments (specifically, for arterial and venous blood reservoirs). In different embodiments, there is an n-compartment model with n higher than 2. These embodiments provide a more detailed description at each level of the circulatory system. In a preferred embodiment of the invention, the coronary blood flow parameters of the left (LCA) and right (RCA) arterial inlets are calculated.

[0153] Any computational step or sub-step of the method according to the present invention may be implemented using a computer or computer program. In some specific embodiments, some or all of the computations are performed using a computer program stored on a computer or on any type of memory device or both. In another embodiment, some or all of the computations for the purposes of the method may be performed remotely, for example, using a cloud-based infrastructure (which may include the use of the Internet or a local network) at a location different from where the patient is located.

[0154] Although all the measurement methods and algorithms described herein are intended to define the parameters of the present invention and provide tangible results compared with clinical trials, they are by no means restrictive, but exemplary. Words such as "including" or "comprising" are not restrictive, and for example, when element A includes another element B, in addition to element B, element A may also include one or more other elements. The use of singular or plural forms does not limit the scope of the present disclosure, and for example, a part of the description indicating that element A includes element B, which also discloses that multiple element B is included in one element A and multiple element A is included in one element B and wherein multiple element A includes multiple element B. Embodiment. After reviewing the contents of the present disclosure, many other embodiments will be obvious and clear to those skilled in the art. Therefore, the scope of the present invention should be determined with reference to the entire scope of the appended claims together with the equivalents to which the claims are entitled.

Claims

1. A method for evaluating the hemodynamic response of a human patient to stimulation with an adenosine receptor agonist, wherein: The method comprises the following steps: mapping a resting hemodynamic parameter of a human patient to a hyperemic hemodynamic parameter of a human patient using a pharmacokinetic / pharmacodynamic model, wherein the pharmacokinetic / pharmacodynamic model is a nonlinear model, wherein the pharmacokinetic / pharmacodynamic model comprises a Michaelis-Menten, Hill-Langmuir, Black-Leff, multi-pathway and / or multi-species blood-tissue exchange model, wherein the resting hemodynamic parameter comprises the patient's systolic blood pressure, the patient's diastolic blood pressure and / or the patient's heart rate, and the resting hemodynamic parameter is measured non-invasively in a resting state, or is assumed using a continuous patient pressure waveform recorded non-invasively, wherein the non-invasively recorded continuous patient pressure waveform is a pressure waveform of a human patient in the resting state, wherein the non-invasively recorded continuous patient pressure waveform includes a time window, and the time window includes at least one complete cycle of the heart of the human patient, and Parameter identification of a lumped parameter model of the human patient is performed using the hyperemic hemodynamic parameters of the human patient to obtain the lumped parameter model of the human patient in a hyperemic state, wherein the lumped parameter model at least partially describes the hemodynamics of the human patient and is of Windkessel type.

2. The method according to claim 1, wherein: The method also includes calculating blood flow parameters using the lumped parameter model of the human patient in the hyperemic state, wherein the blood flow parameters include pressure, flow rate, waveform and / or cardiac time intervals, including but not limited to cardiac cycle duration and ejection time.

3. The method according to any one of claims 1 to 2, wherein: The resting hemodynamic parameters of a human patient also include a non-invasively recorded continuous resting patient pressure waveform.

4. The method according to any one of claims 1 to 3, wherein: The parameter identification comprises the following steps: estimating values ​​of empirical parameters of the lumped parameter model for a human patient using demographic and health data that affects pressure pulse propagation in the human patient, and The hyperemic hemodynamic parameters of the human patient are used to refine values ​​of the empirical parameters of the lumped parameter model of the human patient.

5. The method according to claim 4, wherein: The demographic and health data were collected for patients diagnosed with CHD.

6. The method of any one of claims 4-5, wherein the demographic and health data includes gender, age, height, general health assessments and / or current medications, wherein, The current medications include, but are not limited to, beta-adrenergic blockers, angiotensin converting enzyme inhibitors, and / or antiarrhythmic agents.

7. The method of any one of claims 1 to 6, wherein the Black-Leff model comprises the following equation: in, Z is the pharmacological effect; Z m is the maximal response; τ is the ratio defining the agonist efficacy; [A] is the concentration of the agonist; K A is the agonist-receptor dissociation constant; and n is the factor that determines the steepness of the curve.

8. The method according to any one of claims 1 to 7, wherein: The Hill-Langmuir model includes the following equations: Where Z is the pharmacological effect; Z max is the maximum response; Z min is the minimum response; X is an arbitrary measurement reflecting a change in agonist concentration; K is the apparent dissociation constant; and n is a factor determining the steepness of the curve.

9. The method according to any one of claims 1 to 8, wherein: The Black-Leff model includes the following equations: Where Z is the pharmacological effect; Z max is the maximum response; Z min is the minimum response; X is an arbitrary measurement reflecting the change in agonist concentration; K is the apparent dissociation constant; n is a factor that determines the steepness of the curve; and τ is a ratio that defines the efficacy of the agonist.

10. The method according to any one of claims 1 to 9, wherein: The non-invasively recorded continuous patient pressure waveform is recorded on the radial artery.

11. The method according to any one of claims 3 to 10, wherein: The non-invasively recorded continuous resting patient pressure waveform is recorded on the radial artery.

12. The method according to any one of claims 1 to 11, wherein: The non-invasively recorded continuous patient pressure waveform is recorded using a method selected from photoplethysmography and / or applanation tonometer.

13. The method according to any one of claims 3 to 12, wherein: The non-invasively recorded continuous resting patient pressure waveform is recorded using a method selected from photoplethysmography and / or applanation tonometer.

14. The method according to any one of claims 1 to 13, wherein: The time window includes at least one complete breathing cycle.

15. The method according to any one of claims 1 to 14, wherein: The lumped parameter model of a human patient is a three-component model comprising: a blood circulatory system (BCS) component, wherein the BCS component includes one, two or more CRL functional blocks, a cardiac chamber pressure-volume (HPV) component, wherein the HPV component includes one, two or more ERv functional blocks, and a coronary blood flow (CBF) component, wherein the CBF component comprises one, two or more RCpRp functional blocks, and Wherein, the CRL functional block includes the following structure: and The ERv functional block includes the following structure: and Wherein, the RCpRp functional block includes the following structure: and Among them, p in ,q in is the pressure and flow rate at the compartment inlet; p out ,q out are the pressure and flow rate at the compartment outlet; R0, R, L, C are the proximal and distal resistances, inertia and compliance; Valve is the heart valve modeling diode; p C 、p R is the myocardial vascular interaction (MVI) pressure, where C is the compliant artery and R is the resistance artery; and X is E or MF, where E is the time-varying elasticity concept and MF is the myocardial fiber stress and strain concept.

16. The method according to claim 15, wherein: The myocardial vascular interaction (MVI) includes the calculation of the coronary blood flow throttling pressure: p R,C =k R,C (CEP+SIP)=k R,C (μ1p(t)+μ2E n (t)), Where μ1 is the intraluminal pressure of the interstitial cavity; k R,C is the resistance and compliance part of the crown tree coefficient; μ1 is the cavity-induced extracellular pressure coefficient; μ2 is the shortening-induced intracellular pressure; and E n is the normal elasticity.

17. The method according to any one of claims 15-16, wherein: Each of the RCpRp function blocks satisfies the following equation: Among them, p zf is the zero-flow pressure in coronary blood flow, and / or Wherein, each of the CRL functional blocks satisfies the following equation: Among them, C, R and L are compliance, resistance and inertia respectively.

18. The method according to claim 17, wherein: For patients with stable angina, the zero flow pressure is equal to 14.8±7 mmHg, or for patients with non-Q wave myocardial infarction, the zero flow pressure is equal to 22.5±9.1 mmHg, or for patients with Q wave myocardial infarction, the zero flow pressure is equal to 37.1±1.9 mmHg.

19. The method according to any one of claims 15 to 18, wherein: The lumped parameter model of a human patient comprises a systemic circulation and a pulmonary circulation model, wherein the systemic circulation comprises a left heart circuit and the pulmonary circulation comprises a right heart circuit, and wherein each circuit has the form of at least two of the CRL functional blocks connected in series.

20. The method according to any one of claims 15 to 19, wherein: The lumped parameter model of a human patient includes the following structure: Where p is pressure; q is flow rate; pa is artery; pr is reservoir; pv is vein; sa is aorta; sp is proximal artery; sd is distal artery; sr is reservoir; sv is vein; L is inertia; R is resistance; C is compliance; X is E or MF, where E is the time-varying elasticity concept and MF is the myocardial fiber stress and strain concept; RA, RV are right atrium and ventricle; LA, LV are left atrium and ventricle; tv is tricuspid (atrioventricular) valve; pv is pulmonary (ventricular) valve; mv is mitral (atrioventricular) valve; and av is aortic (ventricular) valve.

21. A computer-readable [storage] medium comprising instructions which, when executed by a computer, cause the computer to perform the steps of the method defined in any one of claims 1 to 20.

22. A system for evaluating the hemodynamic response of a human patient to stimulation with an adenosine receptor agonist, wherein: The system comprises: A measuring device for non-invasively measuring a resting hemodynamic parameter of a human patient, wherein the resting hemodynamic parameter comprises the patient's systolic blood pressure, the patient's diastolic blood pressure and / or the patient's heart rate, and A computer system adapted to perform the steps of the method defined in any one of claims 1-20.

23. A system for assessing the hemodynamic response of a human patient to adenosine receptor agonist stimulation according to claim 22, wherein: The computer system also includes a recording device for non-invasively and continuously recording a resting pressure waveform of a human patient.

24. A system for assessing the hemodynamic response of a human patient to adenosine receptor agonist stimulation according to any one of claims 22-23, wherein: The computer system comprises: at least one computer adapted to perform a mapping of resting hemodynamic parameters of a human patient to hyperemic hemodynamic parameters of a human patient using a pharmacokinetic / pharmacodynamic model, and at least one computer adapted to perform parameter identification of a lumped parameter model of the human patient using the hyperemic hemodynamic parameters of the human patient to derive the lumped parameter model of the human patient in a hyperemic state, and Wherein, the at least one computer adapted to perform a mapping of resting hemodynamic parameters of a human patient to hyperemic hemodynamic parameters of the human patient using a pharmacokinetic / pharmacodynamic model is configured to communicate directly or indirectly with the at least one computer adapted to perform parameter identification of a lumped parameter model of the human patient using the hyperemic hemodynamic parameters of the human patient to derive the lumped parameter model of the human patient in a hyperemic state.

Images