System and method for assessing coronary function using pulsed-wave doppler analysis

Abstract

A system, an apparatus and a method for non-invasive evaluation of coronary circulatory dysfunction is disclosed. Spectral Doppler ultrasound data representing coronary blood flow is obtained at one or more anatomical locations during resting and stress states. A plurality of flow-related features is extracted from the Doppler data and processed using an artificial intelligence (AI)-based model to evaluate coronary flow values. A non-invasive coronary flow reserve (CFR) value is computed based on a ratio between stress-state and rest-state flow values. The AI-based model may comprise a dynamic model configured to derive time-varying flow characteristics and a static model configured to determine non-linear relationships between the characteristics and coronary pathology. A physiological assessment output is generated indicating presence and / or severity of coronary circulatory dysfunction, including coronary artery disease and microvascular dysfunction, optionally based on retrograde flow patterns and multi-parametric flow analysis.

Description

ISUS Ref: 53516-001.2PCTSYSTEM AND METHOD FOR ASSESSING CORONARY FUNCTION USING PULSED-WAVE DOPPLER ANALYSISCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims priority to U.S. Provisional Application No. 63 / 777,690, titled AUTOMATIC DETECTION, DIAGNOSTICS, EVALUATION AND MONITORING OF CARDIOVASCULAR DISEASES, filed March 26, 2025, which is hereby incorporated by reference in its entirety.FIELD OF INVENTION

[0002] The present disclosure relates to cardiovascular diagnostics, and more particularly to a system and method for non-invasive assessment of coronary flow reserve using spectral Doppler ultrasound data and artificial intelligence-based analysis to evaluate coronary circulatory dysfunction.BACKGROUND

[0003] Coronary artery disease remains a leading cause of morbidity and mortality worldwide, affecting both men and women across diverse populations. The condition involves reduction of blood flow to the cardiac muscle, which can result in cumulative damage to heart tissues and lead to complications such as heart failure, arrhythmias, and adverse cardiovascular events. Early detection and ongoing monitoring of coronary circulatory dysfunction can facilitate timely intervention and improved patient outcomes.

[0004] Coronary microvascular dysfunction represents another form of coronary circulatory impairment that affects the small coronary arterioles rather than the larger epicardial coronary arteries. Unlike obstructive coronary artery disease where atherosclerotic plaques lead to artery narrowing, coronary microvascular dysfunction affects the ability of small blood vessels to dilate properly, restricting blood flow to the heart muscle even in the absence of visible stenosis on conventional angiography. This condition presents diagnostic challenges because conventional coronary angiography may appear normal despite patients experiencing symptoms of ischemia.ISUS Ref: 53516-001.2PCT

[0005] Various diagnostic approaches exist for evaluating coronary circulatory function, ranging from invasive catheter-based techniques to non-invasive imaging modalities. Each approach presents different tradeoffs in terms of accessibility, cost, radiation exposure, and diagnostic accuracy. There remains interest in developing non-invasive methods capable of providing physiological assessment of coronary function with improved diagnostic precision while avoiding the risks and costs associated with invasive procedures.ISUS Ref: 53516-001.2PCTBRIEF SUMMARY

[0006] This summary is provided to introduce a selection of concepts in a simplified form that are further described below in the detailed description. This summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used as an aid in determining the scope of the claimed subject matter.

[0007] One exemplary embodiment of the disclosed subject matter is a computer-implemented method for non-invasive detection and evaluation of coronary circulatory dysfunction. The method comprises obtaining spectral Doppler ultrasound data representing coronary blood flow at one or more anatomical locations, the spectral Doppler ultrasound data being acquired during a resting state and during a stress state. The method further comprises extracting, by at least one processor, a plurality of flow-related features from the spectral Doppler ultrasound data. The method further comprises processing the extracted flow-related features using an artificial intelligence (Al)-based model to evaluate coronary flow values corresponding to the resting state and the stress state. The method further comprises computing a non-invasive Coronary Flow Reserve (CFR) value based on a ratio between the coronary flow value during the stress state and the coronary flow value during the resting state. The method further comprises generating, based on the processing and the computed non-invasive CFR value, a physiological assessment output indicative of the presence and / or severity of coronary circulatory dysfunction.

[0008] Optionally, the plurality of flow-related features comprises at least one of: flow velocity parameters, flow intensity parameters, diastolic flow characteristics, and flow direction components.

[0009] Optionally, processing the extracted features using the Al-based model comprises applying a dynamic model configured to derive time-varying characteristics of coronary blood flow from the extracted flow-related features; and applying a static model configured to determine non-linear relationships between the derived characteristics and coronary circulatory dysfunction indicators.

[0010] Optionally, extracting the plurality of flow-related features comprises computing one or more of: diastolic acceleration time; diastolic deceleration time; systolic time; diastolic time; end-diastolic velocity; end-systolic velocity; mean diastolicISUS Ref: 53516-001.2PCTvelocity; mean systolic velocity; peak diastolic velocity; peak systolic velocity; time-averaged velocity; diastolic time-velocity integral; systolic time-velocity integral; a resistive index; and a pulsatility index.

[0011] Optionally, the Al-based model is configured to process spectral Doppler ultrasound data synchronized with cardiac cycle phases based on electrocardiogram (ECG) signals.

[0012] Optionally, the Al-based model is further configured, during a training or a validation phase, to be refined based on a comparison between physiological assessment outputs and invasive coronary physiological measurements including at least one of: invasive CFR and Index of Microcirculatory Resistance (IMR).

[0013] Optionally, computing the non-invasive CFR value comprises converting Doppler-derived velocity measurements during the resting state and the stress state into estimated volumetric coronary flow values.

[0014] Optionally, computing the non-invasive CFR value comprises analyzing the Doppler-derived velocity measurements to determine at least one of: net positive flow; net negative flow; and a ratio of net positive flow to net negative flow.

[0015] Optionally, the method further comprises synchronizing extraction of the plurality of flow-related features with cardiac cycle phases based on electrocardiogram signals.

[0016] Optionally, the spectral Doppler ultrasound data comprises Pulsed Wave Doppler (PWD) data acquired from at least one of a left anterior descending artery, a right coronary artery, and a left circumflex artery.

[0017] Optionally, the coronary circulatory dysfunction is selected from a group consisting of: Coronary Artery Disease (CAD) and Coronary Microvascular Dysfunction (CMD).

[0018] Optionally, identifying CMD comprises determining that the computed non-invasive CFR value is below a pathological threshold; and determining an absence of epicardial CAD indicators based on additional diagnostic input data; wherein the coronary circulatory dysfunction is classified as CMD when the pathological CFR value is present in the absence of epicardial coronary artery disease indicators.ISUS Ref: 53516-001.2PCT

[0019] Optionally, the additional diagnostic input data comprise at least one of coronary angiography or Coronary Angiography Computed Tomography (CTCA).

[0020] Optionally, identifying CMD comprises detecting retrograde coronary bloodflow components below a Doppler baseline and / or quantifying reverse perfusion patterns in distal segments of primary coronary arteries.

[0021] Optionally, the Al-based model is trained using training data comprising spectral Doppler ultrasound data and corresponding reference physiological measurements.

[0022] Optionally, the reference physiological measurements comprise invasive coronary measurements including at least one of CFR and IMR.

[0023] Optionally, the Al-based model is trained using training data further comprising ECG signals aligned with cardiac cycle phases.

[0024] Optionally, the Al-based model is trained using training data further comprising patient-specific data including at least one of: demographic data, physiological measurements, medical history, and laboratory test results.

[0025] Optionally, the Al-based model is trained to learn relationships between the plurality of flow-related features and volumetric coronary flow characteristics.

[0026] Optionally, the Al-based model comprises a dynamic model trained to derive time-varying characteristics of coronary blood flow; and a static model trained to determine non-linear relationships between the derived characteristics and coronary circulatory dysfunction indicators.

[0027] Optionally, the Al-based model is trained to identify retrograde coronary bloodflow patterns and associate the patterns with microvascular dysfunction.

[0028] Optionally, the Al-based model is trained using spectral Doppler data reconstructed from Digital Imaging and Communications in Medicine (DICOM) images using image processing.

[0029] Optionally, the Al-based model is iteratively refined based on comparison between predicted outputs and clinical reference measurements.

[0030] Optionally, the Al-based model is trained using a Reinforcement Learning (RL) framework configured to optimize ultrasound probe positioning or signal acquisition.ISUS Ref: 53516-001.2PCT

[0031] Optionally, the RL framework is trained based on reward signals derived from Doppler signal quality metrics, using human feedback indicating preferred probe positioning or acquisition states or using training data annotated by expert operators with respect to signal quality or probe positioning.

[0032] Optionally, generating the physiological assessment output comprises jointly evaluating the computed non-invasive CFR value and retrograde coronary flow magnitude.

[0033] Optionally, evaluating CAD further comprising computing at least one of: a pre-stenotic to stenotic flow velocity ratio; a distal to proximal diastolic peak velocity ratio; a post-stenotic damping indicator; and a retrograde coronary flow indicator.

[0034] Optionally, evaluating CAD comprises determining stenosis severity based on at least one of peak diastolic velocity, pre-stenotic to stenotic flow velocity ratio, poststenotic damping, and diastolic-to-systolic flow ratio.

[0035] Optionally, determining stenosis severity further comprises identifying flow disturbances including spectral broadening or aliasing within a coronary artery segment.

[0036] Optionally, the physiological assessment output includes an estimated anatomical location and degree of arterial narrowing.

[0037] Optionally, in evaluating CAD, the physiological assessment output comprises a virtual coronary evaluation including at least one of: an anatomical location of vascular impairment, a severity classification of dysfunction, and a predicted plaque composition category.

[0038] Optionally, the virtual coronary evaluation further comprises a risk stratification output based on the computed CFR value.

[0039] Optionally, the virtual coronary evaluation further comprises at least one of: a recommended follow-up imaging or monitoring plan; and a recommendation regarding invasive diagnostic procedures based on the physiological assessment output.

[0040] Another exemplary embodiment of the disclosed subject matter is a system for non-invasive detection and evaluation of coronary circulatory dysfunction, comprising an ultrasound data acquisition interface configured to receive spectral Doppler ultrasound data representing coronary blood flow acquired at one or more anatomical locationsISUS Ref: 53516-001.2PCTduring a resting state and during a stress state. The system further comprises at least one processor; and a memory storing instructions which, when executed by the at least one processor, cause the system to extract a plurality of flow-related features from the spectral Doppler ultrasound data; process the extracted features using an Al-based model to determine coronary flow values corresponding to the resting state and the stress state; compute a non-invasive CFR value based on a ratio between the coronary flow value during the stress state and the coronary flow value during the resting state; and generate a physiological assessment output indicative of a presence or severity of coronary circulatory dysfunction, based on the computed non-invasive CFR.

[0041] Optionally, the Al-based model comprises a dynamic model configured to derive time-varying characteristics of coronary blood flow from the extracted features and a static model configured to determine non-linear relationships between the derived characteristics and coronary circulatory dysfunction indicators.

[0042] Optionally, identifying coronary circulatory dysfunction comprises jointly evaluating the computed non-invasive CFR value and retrograde coronary flow magnitude.

[0043] Optionally, the instructions further cause the system to synchronize extraction of the plurality of flow-related features with cardiac cycle phases based on electrocardiogram signals.

[0044] Optionally, the spectral Doppler ultrasound data comprises Pulsed Wave Doppler data acquired from at least one of a left anterior descending artery, a right coronary artery, and a left circumflex artery.

[0045] Optionally, coronary circulatory dysfunction comprises at least one of CAD and CMD.

[0046] Optionally, the physiological assessment output comprises a virtual coronary evaluation including at least one of: an anatomical location of vascular impairment, a severity classification of dysfunction, and a predicted plaque composition category.

[0047] Optionally, the virtual coronary evaluation further comprises a risk stratification output based on the computed non-invasive CFR value.ISUS Ref: 53516-001.2PCT

[0048] Optionally, the system is further configured to compute at least one of: a pre-stenotic to stenotic flow velocity ratio; a distal to proximal diastolic peak velocity ratio; a post-stenotic damping indicator; and a retrograde coronary flow indicator.

[0049] Yet another exemplary embodiment of the disclosed subject matter is a non-transitory computer-readable medium storing instructions which, when executed by at least one processor, cause the processor to obtain spectral Doppler ultrasound data representing coronary blood flow at one or more anatomical locations, the spectral Doppler ultrasound data being acquired during a resting state and during a stress state; extract by at least one processor, a plurality of flow-related features from the spectral Doppler ultrasound data; process the extracted flow-related features using an Al-based model to evaluate coronary flow values corresponding to the resting state and the stress state; compute a non-invasive CFR value based on a ratio between the coronary flow value during the stress state and the coronary flow value during the resting state; and generate, based on the processing and the computed non-invasive CFR value, a physiological assessment output indicative of the presence and / or severity of coronary circulatory dysfunction.

[0050] Optionally, the operations further comprise integrating patient- specific data with the spectral Doppler ultrasound data, the patient-specific data comprises at least one of physiological measurements, demographic information, medical history, and laboratory test results.

[0051] Optionally, the physiological assessment output comprises a virtual coronary evaluation including at least one of an anatomical location of vascular impairment and a severity classification of dysfunction.

[0052] Optionally, the operations further comprise synchronizing extraction of the plurality of flow-related features with cardiac cycle phases based on electrocardiogram signals.

[0053] Optionally, the spectral Doppler ultrasound data comprises Pulsed Wave Doppler data reconstructed from DICOM images using image processing to extract spectral and flow-related information.

[0054] Yet another exemplary embodiment of the disclosed subject matter is a computer-implemented method for training an Al-based model for evaluating coronaryISUS Ref: 53516-001.2PCTcirculatory dysfunction, comprising: receiving training data comprising spectral Doppler ultrasound data representing coronary blood flow acquired at one or more anatomical locations; extracting, by at least one processor, a plurality of flow-related features from the spectral Doppler ultrasound data; training the Al-based model to map the extracted flow-related features to at least one of: coronary flow values, CFR, and physiological assessment outputs indicative of coronary circulatory dysfunction; and updating one or more parameters of the Al-based model based on comparison between model outputs and reference physiological measurements.

[0055] Optionally, the reference physiological measurements comprise invasive coronary measurements including at least one of: invasive CFR and IMR.

[0056] Optionally, updating the one or more parameters comprises minimizing a difference between model-predicted outputs and the reference physiological measurements.

[0057] Optionally, the training data further comprises ECG signals, and wherein extracting the plurality of flow-related features is synchronized with cardiac cycle phases based on the ECG signals.

[0058] Optionally, the training data further comprises patient-specific data including at least one of: demographic data, physiological measurements, medical history, and laboratory test results.

[0059] Optionally, training the Al-based model comprises training a dynamic model configured to derive time-varying characteristics of coronary blood flow from the extracted flow-related features; and training a static model configured to determine nonlinear relationships between the derived characteristics and coronary circulatory dysfunction indicators.

[0060] Optionally, the plurality of flow-related features comprises at least one of: diastolic flow parameters, systolic flow parameters, velocity-based parameters, intensitybased parameters, and directional flow components including antegrade and retrograde flow.

[0061] Optionally, training the Al-based model comprises associating retrograde coronary blood-flow patterns with microvascular dysfunction labels.ISUS Ref: 53516-001.2PCT

[0062] Optionally, the spectral Doppler ultrasound data comprises data reconstructed from DICOM images using image processing to extract spectral and flow-related information.

[0063] Optionally, the method further comprises iteratively refining the Al-based model based on repeated comparison between predicted outputs and clinical reference measurements.

[0064] Optionally, the method further comprises training a Reinforcement Learning (RL) model configured to generate probe positioning or ultrasound acquisition guidance. The RL model is trained based on reward signals derived from signal quality metrics associated with acquired Doppler signals. Additionally, or alternatively, the RL model is trained using human feedback indicating preferred probe positioning or acquisition states.

[0065] Optionally, the training data further comprises expert annotations associated with probe positioning, signal quality, or diagnostic interpretation. Additionally, or alternatively, the training data comprises time-series Doppler waveform data representing multiple cardiac cycles.

[0066] Yet another exemplary embodiment of the disclosed subject matter is a system for training an Al-based model for evaluating coronary circulatory dysfunction, comprising at least one processor; and a memory storing instructions which, when executed by the at least one processor, cause the system to receive training data comprising spectral Doppler ultrasound data representing coronary blood flow acquired at one or more anatomical locations; extract, by at least one processor, a plurality of flow-related features from the spectral Doppler ultrasound data; train the Al-based model to map the extracted flow-related features to at least one of: coronary flow values, CFR, and physiological assessment outputs indicative of coronary circulatory dysfunction; and update one or more parameters of the Al-based model based on comparison between model outputs and reference physiological measurements.

[0067] Yet another exemplary embodiment of the disclosed subject matter is a non-transitory computer-readable medium storing instructions which, when executed by at least one processor, cause the processor to receive training data comprising spectral Doppler ultrasound data representing coronary blood flow acquired at one or more anatomical locations; extract, by at least one processor, a plurality of flow-related featuresISUS Ref: 53516-001.2PCTfrom the spectral Doppler ultrasound data; train the Al-based model to map the extracted flow-related features to at least one of: coronary flow values, CFR, and physiological assessment outputs indicative of coronary circulatory dysfunction; and update one or more parameters of the Al-based model based on comparison between model outputs and reference physiological measurements.

[0068] The foregoing general description of the illustrative embodiments and the following detailed description thereof are merely exemplary aspects of the teachings of this disclosure and are not restrictive.ISUS Ref: 53516-001.2PCTTHE BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

[0069] The present disclosed subject matter will be understood and appreciated more fully from the following detailed description taken in conjunction with the drawings in which corresponding or like numerals or characters indicate corresponding or like components. Unless indicated otherwise, the drawings provide exemplary embodiments or aspects of the disclosure and do not limit the scope of the disclosure. In the drawings:

[0070] Figure 1 shows a clinical environment in which the disclosed subject matter is used, in accordance with some exemplary embodiments of the subject matter;

[0071] Figure 2 shows a block diagram of an Al-based diagnostic system for non-invasive evaluation of coronary circulatory dysfunction, in accordance with some exemplary embodiments of the disclosed subject matter;

[0072] Figure 3 shows a flowchart diagram of a method for non-invasive evaluation of coronary circulatory dysfunction using spectral Doppler ultrasound data, in accordance with some exemplary embodiments of the disclosed subject matter;

[0073] Figure 4 shows a schematic representation of a data processing pipeline for transforming spectral Doppler ultrasound data into flow-related representations, in accordance with some exemplary embodiments of the disclosed subject matter;

[0074] Figure 5 shows an exemplary PWD ultrasound imaging display, in accordance with some exemplary embodiments of the disclosed subject matter;

[0075] Figure 6 shows a schematic flow diagram of an integrated processing and analysis architecture for non-invasive evaluation of coronary circulatory dysfunction, in accordance with some exemplary embodiments of the disclosed subject matter; and

[0076] Figure 7 shows an exemplary results display of Doppler-derived coronary flow analysis, in accordance with some exemplary embodiments of the disclosed subject matter.ISUS Ref: 53516-001.2PCTDETAILED DESCRIPTION

[0077] The following description sets forth exemplary aspects of the present disclosure. It should be recognized, however, that such description is not intended as a limitation on the scope of the present disclosure. Rather, the description also encompasses combinations and modifications to those exemplary aspects described herein.

[0078] One technical problem addressed by the disclosed subject matter relates to the lack of non-invasive methods for accurately assessing coronary flow reserve with diagnostic precision comparable to invasive catheter-based techniques. CFR is a physiological metric defined as a ratio between maximal coronary blood flow during stress and coronary blood flow at rest, reflecting functional integrity of both epicardial arteries and the microcirculation. A reduced CFR, typically below a threshold of approximately 2.0, may be indicative of pathological coronary physiology. Currently, CFR may be most accurately measured using invasive techniques, such as pressure or flow guidewires during cardiac catheterization. The risks, costs, and limited accessibility associated with invasive procedures create barriers to widespread screening and routine monitoring of coronary physiology.

[0079] Coronary Flow Velocity Reserve (CFVR) may be derived from Doppler velocity measurements obtained non-invasively. However, CFVR may not fully represent volumetric coronary flow, and non-linear relationships between Doppler signal characteristics and coronary flow dynamics may reduce diagnostic accuracy. While CFVR, derived from Doppler velocity measurements, may be obtained non-invasively, CFVR alone may lack sufficient accuracy and may not fully represent volumetric coronary flow, resulting in inferior diagnostic precision compared to invasive CFR measurements. Conventional CFVR approaches rely solely on velocity-based measurements obtained during rest and stress conditions and do not account for absolute or inferred volumetric flow characteristics. In addition, velocity -based indices such as CFVR inherently assume a proportional relationship between measured velocity and volumetric blood flow, which may not hold under conditions of vessel narrowing, flow disturbance, or microvascular regulation. As a result, reliance on velocity measurements alone may lead to inaccurate estimation of true coronary flow reserve.ISUS Ref: 53516-001.2PCT

[0080] Another technical problem addressed by the disclosed subject matter relates to the difficulty in differentiating between epicardial CAD and CMD (also referred to as Microvascular Disease or MVD). using conventional diagnostic approaches. CMD is a type of heart disease that involves dysfunction of small coronary arterioles rather than the larger epicardial coronary arteries. Unlike obstructive CAD, where atheromatous plaques lead to artery narrowing, CMD affects the ability of small blood vessels to dilate properly, restricting blood flow to the heart muscle. CMD may result from endothelial dysfunction, impaired vasodilation, increased microvascular resistance, and altered autoregulation of myocardial blood flow, even in the absence of obstructive coronary artery disease. CMD may often be underdiagnosed because conventional coronary angiography may appear normal, despite patients experiencing symptoms of ischemia, leading to misdiagnosis or delayed diagnosis. When epicardial coronary stenosis has been excluded, for example using coronary CT angiography or invasive angiography, a pathological CFR may be commonly attributed to CMD. However, existing non-invasive diagnostic approaches may not reliably distinguish between these two conditions. In addition, velocity -based indices such as CFVR inherently assume a proportional relationship between measured velocity and volumetric blood flow, which may not hold under conditions of vessel narrowing, flow disturbance, or microvascular regulation. As a result, reliance on velocity measurements alone may lead to inaccurate estimation of true coronary flow reserve.

[0081] Yet another technical problem addressed by the disclosed subject matter relates to the complexity and operator-dependency of acquiring reliable spectral Doppler signals from coronary arteries. Capturing reliable and high-quality physiological signals, such as blood flow patterns or other cardiovascular indicators, may be challenging due to variability in patient anatomy, motion artifacts, and environmental conditions. Visualizing coronary arteries may be inherently difficult due to their small size, particularly when contrast agents are not used. Acoustic window limitations may arise in certain patient populations, such as those with obesity or interference from lung tissue, making imaging less reliable.

[0082] Variations in coronary anatomy and flow patterns may add to the complexity of acquiring and interpreting signals. While imaging proximal segments of the left anterior descending artery (LAD) may be relatively feasible, visualizing distal segments orISUS Ref: 53516-001.2PCTsmaller branches may present challenges. Successful transthoracic Doppler echocardiography (TTDE) imaging for CAD diagnosis may require a high level of expertise to correctly identify coronary arteries and measure flow velocities. Misalignment of the Doppler beam with the coronary flow may lead to inaccurate velocity measurements, reducing diagnostic reliability. Patient-related factors, including obesity, pulmonary disease, or inadequate acoustic windows, may reduce imaging success rates. Artifacts may be more prevalent in larger or less cooperative patients, further complicating the acquisition process.

[0083] Yet another technical problem addressed by the disclosed subject matter relates to the limitations of velocity-based measurements (such as CFVR) in representing volumetric coronary flow due to complex non-linear relationships between Doppler signal characteristics and vascular pathology. Existing non-invasive Doppler-based approaches may rely primarily on velocity measurements and may fail to account for complex, non-linear relationships between Doppler signal characteristics, coronary flow dynamics, and vascular pathology. The relationship between flow velocity and volumetric flow depends on the assumption that the cross-sectional area of the vessel remains constant, which may not hold true in all clinical scenarios. Coronary artery stenosis may create disruptions in blood flow, such as turbulent flow, increased velocity, or changes in flow patterns downstream of a narrowed segment. These hemodynamic changes may introduce non-linearities that velocity-based measurements alone may not adequately capture.

[0084] Converting velocity measurements to accurate volumetric flow estimates may require accounting for these complex relationships. Doppler ultrasound measurements primarily provide blood flow velocity rather than direct volumetric flow. Changes in vessel geometry, such as stenosis, may result in increased local velocity without directly reflecting volumetric flow. Accordingly, characterization of coronary flow may require analysis of velocity-derived flow profiles rather than reliance on velocity magnitude alone. Accordingly, accurate characterization of coronary physiology may require analysis of multiple Doppler-derived features and their interrelationships, rather than reliance on velocity magnitude or simple ratios alone.

[0085] Yet another technical problem addressed by the disclosed subject matter relates to the challenge of extracting clinically actionable insights from acquired Doppler signalsISUS Ref: 53516-001.2PCTacross diverse ultrasound platforms and patient conditions. Even when high-quality signals are obtained, deriving actionable insights to enable early and accurate diagnosis of coronary pathologies, such as partial or complete blockages, may involve computational challenges. Processing and analyzing the data in a way that ensures clinical relevance, accuracy, and ease of interpretation for healthcare providers may be difficult, particularly in environments with limited access to specialized expertise. The acquired signals or available data may be limited in resolution, completeness, or quality. Integration of diagnostic systems across diverse existing imaging platforms may present challenges due to differences in device architecture, signal processing capabilities, and hardware specifications.

[0086] In some exemplary embodiments, each system may require calibration of preset settings to optimize imaging parameters, including gain, depth, frequency, dynamic range, and pulse repetition frequency, all of which may need to be tuned to match the anatomical and physiological characteristics of specific arterial segments. Variability in transducer performance, signal processing algorithms, and imaging techniques may further complicate standardization efforts, leading to inconsistency in diagnostic accuracy. Some ultrasound systems may store data as DICOM images rather than raw signal data, requiring additional processing to extract spectral and flow-related information from the image data.

[0087] One technical solution provided by the disclosed subject matter relates to an AI-based approach for enabling non-invasive physiological assessment of coronary circulatory function using spectral Doppler ultrasound data. The disclosed approach allows transformation of Doppler-derived measurements into clinically meaningful indicators of coronary blood flow and function, thereby supporting evaluation of coronary circulatory dysfunction without reliance on invasive procedures.

[0088] In some exemplary embodiments, the disclosed approach utilizes spectral Doppler ultrasound data acquired under different physiological conditions, including resting and stress states, to capture variations in coronary flow behavior. Rather than relying solely on conventional velocity-based indices, the disclosed approach leverages Al-based analysis to interpret complex relationships between Doppler signal characteristics and underlying coronary flow dynamics. In some exemplary embodiments, the Al-based analysis may comprise a plurality of models, including aISUS Ref: 53516-001.2PCTdynamic model configured to extract time-varying physical characteristics of coronary blood flow, and a static model configured to determine non-linear relationships between the extracted characteristics and coronary circulatory dysfunction indicators.

[0089] In some exemplary embodiments, spectral Doppler measurements may be acquired at one or more predefined sampling locations along selected coronary arteries. The disclosed approach may not require full imaging of the entire coronary vascular tree. Instead, a targeted sampling methodology may be employed, wherein diagnostic measurements are obtained at specific anatomical points sufficient to characterize coronary flow dynamics and detect pathological conditions. The selected sampling locations may include proximal and / or distal segments of coronary arteries, enabling inference of both local epicardial flow behavior and downstream microvascular effects.

[0090] The disclosed subject matter enables derivation of flow-related representations that reflect the functional state of the coronary circulation, including both epicardial vessels and microvasculature. These representations may account for temporal, spectral, and directional aspects of blood flow, and may capture non-linear relationships that are not directly observable using traditional Doppler analysis techniques. In some embodiments, the system may extract a plurality of flow-related features, for example on the order of tens or more (e.g., 10-100 features), representing different aspects of coronary blood flow dynamics. The extracted features may be organized into a feature map capturing temporal, spectral, and directional characteristics of flow. The disclosed approach may derive a flow profile representing the temporal and spectral evolution of coronary blood flow throughout the cardiac cycle. In some exemplary embodiments, the extracted features may represent physical characteristics of blood flow, including temporal flow evolution, directional flow components, and spectral intensity distributions. The extracted features may be provided to one or more Al-based models, including a dynamic model and a static model, for further analysis.

[0091] Based on these Al-derived representations, the disclosed approach enables estimation of CFR in a manner that more closely approximates true physiological flow reserve. This is achieved by overcoming limitations associated with velocity-only measurements and by incorporating learned relationships between Doppler data and volumetric flow behavior. In some exemplary embodiments, the estimation of CFR is notISUS Ref: 53516-001.2PCTbased solely on velocity ratios, but on multi-parametric analysis of Doppler-derived features reflecting underlying flow behavior.

[0092] The disclosed approach further enables generation of a physiological assessment indicative of the presence and / or severity of coronary circulatory dysfunction. Such assessment may provide clinically actionable insights, including differentiation between epicardial coronary disease and microvascular dysfunction, and may be presented in a form comparable to invasive diagnostic outputs. Accordingly, the disclosed subject matter provides a non-invasive, Al-driven framework for functional coronary assessment that improves diagnostic capability while reducing the risks, costs, and accessibility limitations associated with conventional invasive techniques.

[0093] One technical effect of utilizing the disclosed subject matter is enabling non-invasive estimation of CFR with diagnostic accuracy approaching invasive catheterbased techniques. The Al-based diagnostic system may process spectral Doppler ultrasound data to compute non-invasive CFR values that approximate volumetric coronary flow reserve rather than relying solely on velocity-based measurements. By applying the dynamic model to derive time-varying characteristics of coronary blood flow and the static model to determine non-linear relationships between the derived characteristics and coronary circulatory dysfunction indicators, the system may achieve diagnostic precision that approaches the accuracy of invasive CFR measurements obtained using pressure or flow guidewires during cardiac catheterization. The training of the Al-based model on ground truth data from cardiac catheterization measurements may enable the system to learn complex relationships between Doppler-derived features and true volumetric flow, thereby improving the accuracy of non-invasive CFR estimation beyond what may be achieved with conventional CFVR calculations alone.

[0094] Another technical effect of utilizing the disclosed subject matter is eliminating the need for guidewires and intracoronary instrumentation while providing physiological assessment of coronary circulatory function. The disclosed system may enable clinicians to obtain physiological assessment of coronary flow reserve using transthoracic Doppler echocardiography without requiring invasive cardiac catheterization procedures. By processing spectral Doppler ultrasound data acquired non-invasively through the chest wall, the system may generate a virtual diagnostic catheterization report that provides clinicians with physiological assessment comparable to invasive catheter-basedISUS Ref: 53516-001.2PCTdiagnostics. The elimination of invasive instrumentation may reduce procedural risks associated with catheterization, including bleeding, infection, arterial damage, and contrast-induced nephropathy. The non-invasive approach may also reduce the costs associated with catheterization laboratory time, specialized personnel, and postprocedural monitoring. The physiological assessment may be presented visually, including graphical representations of coronary flow dynamics or anatomical mapping of affected regions.

[0095] Y et another technical effect of utilizing the disclosed subject matter is improving differentiation between CAD and CMD through joint evaluation of CFR values and retrograde flow patterns. The system may generate the physiological assessment output by jointly evaluating the computed non-invasive CFR value and retrograde coronary flow magnitude. Identifying CMD may comprise determining that the computed non-invasive CFR value is below a pathological threshold and determining an absence of epicardial CAD indicators based on additional diagnostic input data. The coronary circulatory dysfunction may be classified as CMD when the pathological CFR value is present in the absence of epicardial coronary artery disease indicators. Identifying CMD may further comprise detecting retrograde coronary blood-flow components below a Doppler baseline and quantifying reverse perfusion patterns in distal segments of primary coronary arteries. The detection and quantification of retrograde flow may provide a strong indication of CMD, as the microvascular arteries of the heart may exhibit reverse perfusion patterns when compensating for restricted flow in the epicardial arteries. By analyzing both CFR values and retrograde flow characteristics, the system may provide improved differentiation between epicardial stenosis and microvascular dysfunction compared to approaches that rely on CFR values alone. In patients with non-obstructive coronary arteries (NOCA), reduced CFR in the absence of epicardial stenosis may be indicative of microvascular dysfunction. The ability to detect microvascular dysfunction non-invasively may reduce unnecessary invasive diagnostic procedures in patients without epicardial disease.

[0096] Yet another technical effect of utilizing the disclosed subject matter is reducing false-negative and false-positive diagnoses through Al-driven analysis of complex flow dynamics. The Al-based model may analyze non-linear relationships between diverse signals, including spectral Doppler data, electrocardiogram signals, demographic data,ISUS Ref: 53516-001.2PCTvital signs, and medical history, to evaluate the localization and severity of coronary stenosis. The static model may be configured to determine non-linear relationships between the derived characteristics and coronary circulatory dysfunction indicators, enabling the system to account for complex interactions between Doppler signal characteristics, coronary flow dynamics, and vascular pathology that may not be captured by conventional diagnostic approaches. The system may identify flow disturbances including spectral broadening or aliasing within a coronary artery segment, which may indicate the presence of stenosis.

[0097] By computing parameters such as pre-stenotic to stenotic flow velocity ratio, distal to proximal diastolic peak velocity ratio, post-stenotic damping indicator, and retrograde coronary flow indicator, the system may detect subtle hemodynamic changes associated with coronary pathology. The Al-driven analysis may reduce false-positive diagnoses that lead to unnecessary catheterization procedures in patients with normal coronary arteriograms and may reduce false-negative diagnoses that result in missed detection of clinically significant coronary disease. The Al-based model may further analyze relationships between measurements obtained from different coronary arteries or segments, enabling detection of interdependent hemodynamic patterns across the coronary vascular tree. The Al-based model may analyze spectral Doppler output and optionally additional ultrasound or medical imaging data.

[0098] Yet another technical effect of utilizing the disclosed subject matter is enabling routine screening and follow-up monitoring with reduced risk and cost. The non-invasive nature of the disclosed system may enable frequent routine monitoring of previously diagnosed patients without the risks and costs associated with repeated invasive procedures. The system may facilitate periodic evaluations of disease progression, reducing the need for repetitive invasive diagnostics in follow-up care. The system may also enable rapid diagnosis in cases of acute coronary syndrome and may provide ongoing follow-up post-coronary catheterization or cardiac CT. The reduced risk profile of non-invasive spectral Doppler ultrasound compared to invasive catheterization may make the disclosed system suitable for screening applications in diverse healthcare settings, including primary care clinics, emergency departments, and outpatient cardiology clinics. The cost-effectiveness of ultrasound-based assessment compared to invasive catheterization or advanced imaging modalities such as positron emission tomographyISUS Ref: 53516-001.2PCT(PET) may enable broader deployment of coronary flow reserve assessment across healthcare systems with varying resource availability. In some embodiments, the system may provide guided clinical indications based on the physiological assessment output, assisting clinicians in determining appropriate diagnostic or therapeutic pathways.

[0099] Yet another technical effect of utilizing the disclosed subject matter is providing compatibility with standard Doppler ultrasound systems across diverse clinical environments. The disclosed system may integrate with standard ultrasound hardware from established vendors without requiring additional hardware development. The system may be configured to receive spectral Doppler ultrasound data from various ultrasound platforms, including both cart-based ultrasound machines and point-of-care ultrasound (POCUS) devices. The system may incorporate an advanced image processing module designed to handle cases where ultrasound systems store data as DICOM images rather than raw signal data. The image processing module may leverage computer vision algorithms and Al-driven analysis to extract spectral and flow-related information directly from the images, enabling the system to reconstruct flow velocity profiles and other diagnostic markers from image data. The system may feature adaptive preset optimization capability designed to dynamically adjust ultrasound settings based on the specific ultrasound machine in use, the targeted blood vessel, and the precise vascular segment being measured. Compatibility with diverse ultrasound platforms may enable deployment of the disclosed system across various clinical settings, including private clinics, hospitals, health insurance providers, and imaging centers, facilitating improved diagnostic capabilities and patient care without requiring replacement of existing ultrasound equipment.

[0100] Referring now to Figure 1, showing a clinical environment in which the disclosed subject matter is used, in accordance with some exemplary embodiments of the subject matter.

[0101] Al-based Diagnostic System 100 may be configured for non-invasive evaluation of coronary circulatory dysfunction in a clinical environment. In some exemplary embodiments, Al-based Diagnostic System 100 may comprise Ultrasound Device 115, Probe 102, Display 110, and User Interface 112. Ultrasound Device 115 may be configured to acquire spectral Doppler ultrasound data from Patient 120 and process the acquired data to generate diagnostic outputs indicative of presence or severity of coronaryISUS Ref: 53516-001.2PCTcirculatory dysfunction. In some exemplary embodiments, Al-based Diagnostic System 100 may further comprise one or more Al-based guidance and signal evaluation components configured to assist in acquisition of Doppler signals, optimize acquisition parameters, and verify signal quality prior to diagnostic analysis.

[0102] In some exemplary embodiments, Probe 102 may be coupled to Ultrasound Device 115 and may be positioned by Operator 125 on chest of Patient 120 to acquire Ultrasound Signal 104 representing coronary blood flow. Probe 102 may be a component of Ultrasound Device 115, or may be an additional component operatively coupled thereto. In some exemplary embodiments, Ultrasound Device 115 may comprise a conventional or commercially available ultrasound system, wherein one or more existing hardware components are utilized without modification. Additionally, or alternatively, one or more components of Ultrasound Device 115 may be replaced, enhanced, or supplemented. Such components may include Probe 102, which may be configured as an improved probe, a specialized coronary imaging probe, or a probe integrated with one or more additional sensors.

[0103] In some exemplary embodiments, Ultrasound Device 115 may be augmented with additional components (including Processing Unit 240 represented in Figure 2), enhanced signal processing capabilities, and / or an improved User Interface 112 configured to support Al-based analysis and presentation of diagnostic outputs. In some exemplary embodiments, positioning of Probe 102 and acquisition of Ultrasound Signal 104 may be assisted by one or more trained models configured to analyze real-time ultrasound data and generate guidance instructions for probe positioning and Doppler sampling location selection. The guidance instructions may comprise visual, audio, or automated control signals.

[0104] Additionally, or alternatively, Al-based Diagnostic System 100 may be implemented as an add-on module configured to interface with an existing ultrasound system, thereby enabling non-invasive coronary flow analysis without requiring replacement of the underlying imaging hardware.

[0105] In some exemplary embodiments, Patient 120 may be lying in a supine position on an examination table during the diagnostic procedure, although other clinically appropriate positions may be used as needed. Operator 125 may manipulate Probe 102 to obtain spectral Doppler measurements from one or more anatomical locations withinISUS Ref: 53516-001.2PCTcoronary vasculature of Patient 120. Additionally, or alternatively, Probe 102 may be manipulated automatically using a robotic mechanism (such as Robotic Mechanism 214 presented in Figure 2). The robotic mechanism may be configured to control positioning, orientation, and movement of Probe 102 relative to Patient 120, optionally based on predefined scanning protocols, real-time imaging feedback, and / or control signals. In some exemplary embodiments, the robotic mechanism may assist in stabilizing Probe 102, maintaining optimal acoustic contact, and guiding Probe 102 to anatomical locations associated with acquisition of coronary Doppler signals. In some exemplary embodiments, the robotic mechanism or operator guidance may be driven by a guidance model configured to predict probe movement along multiple degrees of freedom, including translation, rotation, tilt, and contact pressure, based on real-time analysis of ultrasound data.

[0106] In some exemplary embodiments, the anatomical locations accessed using Probe 102 may include left anterior descending artery, right coronary artery, and left circumflex artery. Probe 102 may comprise a high-frequency phased-array transducer configured to provide adequate penetration and resolution for imaging coronary arteries. Ultrasound Signal 104 may be transmitted from Probe 102 to Ultrasound Device 115 for processing. In some exemplary embodiments, Doppler measurements may be obtained at one or more predefined sampling locations within the coronary arteries, without requiring full imaging of the entire coronary vascular tree, wherein selected sampling points may be sufficient to characterize coronary flow dynamics and detect pathological conditions.

[0107] In some exemplary embodiments, Display 110 may be mounted on Ultrasound Device 115 and may present PWD Signal 106 representing blood flow velocity over time. PWD Signal 106 may comprise a two-dimensional grayscale ultrasound scan showing a vascular or cardiac structure in an upper portion of Display 110 and a spectral Doppler waveform graph in a lower portion of Display 110. The spectral Doppler waveform graph may display blood flow velocity measurements derived from Doppler shift frequency, with horizontal axis representing time and vertical axis representing blood flow velocity. The spectral Doppler waveform may show characteristic diastolic-dominant flow patterns associated with coronary artery hemodynamics. In some exemplary embodiments, Display 110 may further present B-mode ultrasound data and colorISUS Ref: 53516-001.2PCTDoppler flow data, optionally overlaid, to provide anatomical and flow context for placement of a Doppler sampling cursor defining a sample volume for PWD acquisition.

[0108] A User Interface 112 may be positioned on Ultrasound Device 115 below, on or as a component of Display 110, and may comprise controls that enable Operator 125 to adjust imaging parameters, select anatomical segments for signal acquisition, and interact with Al-based Diagnostic System 100. User Interface 112 may enable Operator 125 to input patient data, initiate signal acquisition sequences, and review diagnostic outputs generated by Al-based Diagnostic System 100. Ultrasound Device 115 may be mounted on a wheeled cart that enables positioning of system within clinical environment. In some exemplary embodiments, User Interface 112 may further present real-time guidance instructions, signal quality indicators, and acquisition recommendations generated by AI-based models, enabling interactive or automated optimization of signal acquisition.

[0109] Additionally, or alternatively, Al-based Diagnostic System 100 may be configured to communicate with external systems for data exchange and integration with clinical workflows. Al-based Diagnostic System 100 may be configured to receive patient-specific data from electronic health records via application programming interface (API) connections. The patient- specific data may comprise demographic information such as age, gender, weight, and height, medical history including diabetes and dyslipidemia, vital signs such as pulse and blood pressure, family medical history, and medications that Patient 120 is chronically using. Al-based Diagnostic System 100 may integrate patient- specific data with spectral Doppler ultrasound data to enhance accuracy of physiological assessment output.

[0110] In some exemplary embodiments, Al-based Diagnostic System 100 may further comprise a signal evaluation module configured to assess whether an acquired Doppler signal satisfies predefined clinical validity and technical quality criteria, including signal-to-noise ratio, stability, and physiological consistency. When one or more criteria are not satisfied, the system may initiate iterative optimization by updating probe position and / or Doppler sampling parameters.

[0111] In some exemplary embodiments, Al-based Diagnostic System 100 may be configured to process spectral Doppler ultrasound data acquired via Probe 102 to extract flow-related features, compute non-invasive coronary flow reserve values, and generate physiological assessment outputs indicative of presence or severity of coronaryISUS Ref: 53516-001.2PCTcirculatory dysfunction including coronary artery disease and microvascular disease. AI-based Diagnostic System 100 may receive spectral Doppler ultrasound data acquired during both a resting state and a stress state to enable computation of non-invasive CFR value based on ratio between coronary flow values during stress state and resting state. The stress state may be induced pharmacologically through administration of vasodilator agents or physiologically through exercise.

[0112] Al-based Diagnostic System 100 may be further configured to generate a virtual diagnostic catheterization report that provides clinicians with physiological assessment comparable to invasive catheter-based diagnostics. The report may be displayed on Display 110 or transmitted to external systems. Additionally, or alternatively, Al-based Diagnostic System 100 may store acquired data, extracted features, computed CFR values, and generated reports locally or transmit them to remote or cloud-based systems, thereby enabling telemedicine applications and remote consultation. In some exemplary embodiments, the generated outputs may comprise clinical insights derived from combined analysis of flow velocity, flow distribution, and temporal flow characteristics, enabling identification of stenosis severity, flow abnormalities, and microvascular dysfunction.

[0113] Referring now to Figure 2 showing an Al-based diagnostic system for non-invasive evaluation of coronary circulatory dysfunction, in accordance with some exemplary embodiments of the disclosed subject matter.

[0114] System 200, such as Al-based Diagnostic System 100 of Figure 1, may be configured to provide non-invasive physiological assessment of coronary circulation based on spectral Doppler ultrasound data, in accordance with the disclosed subject matter. In some exemplary embodiments, System 200 may be configured to transform Doppler-derived signal data into a functional physiological representation of coronary circulation, enabling evaluation of CFR, coronary artery disease, and microvascular dysfunction with diagnostic capabilities comparable to invasive catheter-based techniques.

[0115] In some exemplary embodiments, System 200 may comprise Ultrasound Imaging Device 210, Processing Unit 240, Communication Module 220, and User Device 230. Ultrasound Imaging Device 210 may comprise Probe 212 configured to emit and receive ultrasound signals, and additionally, or alternatively, Robotic Mechanism 214ISUS Ref: 53516-001.2PCTconfigured for automated positioning and manipulation of Probe 212. In some exemplary embodiments, Ultrasound Imaging Device 210 may comprise a conventional ultrasound system, and System 200 may be implemented as an add-on module operatively coupled thereto, or as an integrated system.

[0116] In some exemplary embodiments, Ultrasound Imaging Device 210 may be configured to acquire spectral Doppler ultrasound data from a subject. Ultrasound Imaging Device 210 may comprise Probe 212 configured to emit and receive ultrasound signals for capturing coronary blood flow data. Probe 212 may comprise a high-frequency phased-array transducer configured to provide adequate penetration and resolution for imaging coronary arteries. In some exemplary embodiments, Probe 212 may be one of a plurality of transducers having different configurations, shapes, sizes, or frequencies, enabling adaptation to different anatomical regions and acquisition conditions. Ultrasound Imaging Device 210 may further comprise Robotic Mechanism 214 coupled to Probe 212. Robotic Mechanism 214 may be configured for automated positioning and manipulation of Probe 212 during signal acquisition. In some cases, Robotic Mechanism 214 may receive movement instructions from Processing Unit 240 to guide Probe 212 to anatomical locations that enable acquisition of spectral Doppler signals from coronary arteries including the left anterior descending artery, the right coronary artery, and the left circumflex artery.

[0117] Additionally, or alternatively, positioning of Probe 212 may be guided based on learned acquisition patterns, reinforcement learning models, or operator feedback. In some exemplary embodiments, acquisition of the spectral Doppler ultrasound data may be performed in a closed-loop manner, wherein Processing Unit 240 analyzes real-time ultrasound data and iteratively generates guidance instructions for adjusting probe position, Doppler sampling location, and acquisition parameters until a target coronary signal is obtained.

[0118] Processing Unit 240 may be configured to receive and process the ultrasound data from Ultrasound Imaging Device 210 and generate diagnostic outputs. Processing Unit 240 may comprise one or more processors configured to perform computations associated with feature extraction, Al-based processing, CFR computation, and assessment output generation. The one or more processors may comprise a Central Processing Unit (CPU), a microprocessor, an electronic circuit, an Integrated Circuit (IC),ISUS Ref: 53516-001.2PCTa Graphics Processing Unit (GPU), a Tensor Processing Unit (TPU), or combinations thereof. Processing Unit 240 may be configured to execute program instructions stored in Memory 250 to perform the operations associated with non-invasive detection and evaluation of coronary circulatory dysfunction, and may operate in real-time, near realtime, or offline processing modes. In some exemplary embodiments, Processing Unit 240 may implement a multi-stage processing pipeline comprising preprocessing, signal processing, feature extraction, Al-based analysis, and validation stages arranged in a sequential and / or iterative architecture.

[0119] Memory 250 may be coupled to Processing Unit 240 and may be configured to store program instructions, trained Al models, acquired data, extracted features, and results. Memory 250 may comprise a hard disk drive, a flash disk, a Random Access Memory (RAM), a read-only memory (ROM), a memory chip, a solid-state drive, or combinations thereof. Memory 250 may retain program code operative to cause the one or more processors of Processing Unit 240 to perform acts associated with feature extraction from spectral Doppler ultrasound data, Al-based processing of extracted features, computation of non-invasive CFR values, and generation of physiological assessment outputs. Memory 250 may store trained parameters for Dynamic Model Module 272 and Static Model Module 274, enabling Processing Unit 240 to apply the trained models during diagnostic operation without requiring retraining.

[0120] Processing Unit 240 may further comprise Data Acquisition Interface 262 configured to receive ultrasound data from Ultrasound Imaging Device 210. Data Acquisition Interface 262 may be configured to receive spectral Doppler ultrasound data representing coronary blood flow acquired at one or more anatomical locations during a resting state and during a stress state. Data Acquisition Interface 262 may also be configured to receive ECG signals from external ECG monitoring equipment or from ECG sensors integrated with Ultrasound Imaging Device 210. The ECG signals may be used to synchronize extraction of flow-related features with cardiac cycle phases. Data Acquisition Interface 262 may further be configured to receive additional ultrasound modalities, including B-mode imaging data and color Doppler data, enabling combined anatomical and flow-based analysis.

[0121] Processing Unit 240 may further comprise Preprocessing Module 264 configured to process incoming ultrasound data. Preprocessing Module 264 may performISUS Ref: 53516-001.2PCTfiltering to remove noise and artifacts from the spectral Doppler data, normalization to standardize signal amplitudes across different acquisitions, segmentation to identify individual cardiac cycles, and temporal alignment between rest and stress acquisitions. In some cases, Preprocessing Module 264 may be configured to process DICOM images to extract spectral and flow-related information when ultrasound systems store data as images rather than raw signal data. Additionally, or alternatively, Preprocessing Module 264 may be configured to reconstruct Doppler signal characteristics from image-based data, compensate for incomplete or low-quality data, and enhance signal interpretability. In some exemplary embodiments, Preprocessing Module 264 may further perform signal enhancement operations including noise filtering, contrast enhancement, and signal stabilization to improve downstream feature extraction and analysis.

[0122] Processing Unit 240 may further comprise Feature Extraction Module 266 configured to extract flow-related features from the spectral Doppler data. Feature Extraction Module 266 may extract velocity parameters including peak diastolic velocity, peak systolic velocity, mean diastolic velocity, mean systolic velocity, end-diastolic velocity, end-systolic velocity, and time-averaged velocity. Feature Extraction Module 266 may extract spectral intensity parameters including spectral energy distribution and signal amplitude characteristics. Feature Extraction Module 266 may extract diastolic characteristics including diastolic acceleration time, diastolic deceleration time, diastolic time, and diastolic time-velocity integral. Feature Extraction Module 266 may extract directional flow features including antegrade and retrograde flow components. Feature Extraction Module 266 may compute derived parameters including resistive index, pulsatility index, net positive flow, net negative flow, and ratios between diastolic and systolic flow measurements. Additionally, or alternatively, Feature Extraction Module 266 may generate a multi-parametric representation of coronary flow based on combinations of temporal, spectral, and directional features. In some exemplary embodiments, the extracted features may be derived from measurements obtained at one or more predefined sampling locations along coronary arteries, enabling inference of both local epicardial flow behavior and downstream microvascular effects.

[0123] In some exemplary embodiments, Processing Unit 240 may comprise Al Analysis Module 260 configured to perform Al-based interpretation of extracted features. Al Analysis Module 260 may be configured to employ Al-based models trained toISUS Ref: 53516-001.2PCTanalyze the ultrasound data (such as directly analyzing PWD data) or the features extracted by Feature Extraction Module 266, and evaluate blood flow and stenosis severity based on the analysis. Al Analysis Module 260 may comprise Dynamic Model Module 272 configured to derive time-varying characteristics of coronary blood flow from extracted features. Dynamic Model Module 272 may operate on Doppler-derived data to model time-varying flow behavior including diastolic and systolic flow patterns, flow velocity profiles, flow intensity, spectral energy distribution, retrograde flow, and temporal variability, or to extract other physical characteristics of blood flow. Dynamic Model Module 272 may translate Doppler signal information into physiologically meaningful flow descriptors. In some exemplary embodiments, Dynamic Model Module 272 may be configured to model physical flow behavior from Doppler-derived signals, including temporal evolution of flow, directional flow components, and flow intensity patterns, thereby generating physiologically meaningful representations of blood flow.

[0124] Al Analysis Module 260 may further comprise Static Model Module 274 configured to determine non-linear relationships or connections between the derived characteristics and coronary circulatory dysfunction indicators. Static Model Module 274 may map combinations of Doppler-derived features to clinical outcomes including presence and severity of epicardial stenosis, microvascular dysfunction, and global coronary flow impairment. The combination of Dynamic Model Module 272 and Static Model Module 274 may provide a comprehensive framework for analyzing spectral Doppler data that accounts for both temporal flow dynamics and pathophysiological correlations. Additionally, or alternatively, Al Analysis Module 260 may be configured to derive physiologically meaningful flow representations from Doppler-derived measurements based on learned relationships between signal characteristics and underlying hemodynamic behavior. In some exemplary embodiments, Static Model Module 274 may further be configured to differentiate between epicardial coronary artery disease and coronary microvascular dysfunction based on the learned relationships between derived flow characteristics and pathological conditions.

[0125] Processing Unit 240 may further comprise Flow Estimation Module 280 configured to derive coronary flow values from the Al-processed features. Flow Estimation Module 280 may perform inference of coronary flow characteristics from Doppler-derived inputs based on learned relationships between Doppler signal featuresISUS Ref: 53516-001.2PCTand physiological flow behavior. Flow Estimation Module 280 may generate coronary flow values corresponding to the resting state and the stress state for use in CFR computation. In some exemplary embodiments, the derived coronary flow values may represent inferred volumetric flow characteristics that are not directly measurable from velocity data alone.

[0126] Additionally, or alternatively, Processing Unit 240 may comprise CFR Calculation Module 285 configured to compute a non-invasive coronary flow reserve value based on rest-state and stress-state flow values. CFR Calculation Module 285 may compute the non-invasive CFR value based on a ratio between the coronary flow value during the stress state and the coronary flow value during the resting state.

[0127] Processing Unit 240 may further comprise Assessment Module 290 configured to generate a physiological assessment indicative of the presence and severity of coronary circulatory dysfunction. Assessment Module 290 may evaluate the computed CFR value against pathological thresholds and may jointly evaluate the CFR value and retrograde coronary flow magnitude to differentiate between coronary artery disease and microvascular disease. Assessment Module 290 may classify coronary circulatory dysfunction and may compute parameters including pre-stenotic to stenotic flow velocity ratio, distal to proximal diastolic peak velocity ratio, post-stenotic damping indicator, and retrograde coronary flow indicator. Additionally, or alternatively, Assessment Module 290 may generate decision-support outputs including diagnostic classification, risk stratification, and confidence measures. In some exemplary embodiments, Processing Unit 240 may further comprise a Signal Evaluation Module configured to assess whether acquired Doppler signals satisfy predefined technical and clinical validity criteria. When the acquired signal does not meet the criteria, the system may trigger iterative acquisition refinement by updating probe positioning and / or acquisition parameters.

[0128] Processing Unit 240 may further comprise Report Generation Module 292 configured to generate an output report such as a virtual diagnostic catheterization report. Report Generation Module 292 may generate a virtual coronary evaluation including anatomical location of vascular impairment, severity classification of dysfunction, predicted plaque composition category, and recommended clinical actions including follow-up or intervention strategies. In some exemplary embodiments, the generated report may further comprise clinical insights derived from analysis of flow velocity, flowISUS Ref: 53516-001.2PCTdistribution, and temporal flow characteristics, including identification of stenosis severity, flow abnormalities, and microvascular dysfunction indicators.

[0129] Communication Module 220 may be coupled between Processing Unit 240 and User Device 230. Communication Module 220 may be configured to facilitate communication between modules of System 200 and / or with external systems including electronic health record systems, databases, and remote servers. Communication Module 220 may support wired and wireless communication and may enable integration of patient-specific data into the diagnostic process.

[0130] User Device 230 may comprise User Interface 232 configured to receive user inputs and present acquired Doppler data, computed CFR values, and physiological assessment outputs. User Device 230 may further comprise Display 234 configured to present diagnostic information and reports. User Device 230 may comprise a workstation, tablet, smartphone, or other computing device configured to interface with Processing Unit 240.

[0131] Referring now to Figure 3 showing a flowchart diagram of a method for non-invasive evaluation of coronary circulatory dysfunction using spectral Doppler ultrasound data, in accordance with some exemplary embodiments of the disclosed subject matter.

[0132] Method 300 may be performed by a system comprising at least one processor and Memory storing instructions, such as System 200 in Figure 2.

[0133] In Step 310, spectral Doppler ultrasound data representing coronary blood flow may be obtained. The spectral Doppler ultrasound data may be acquired at one or more anatomical locations within coronary vasculature. In some exemplary embodiments, the anatomical locations may include left anterior descending artery, right coronary artery, and left circumflex artery. The acquisition may be performed using transthoracic Doppler echocardiography, wherein a probe positioned on chest of a subject captures Doppler signals representing coronary blood flow through chest wall. The spectral Doppler ultrasound data may be acquired during a resting state and during a stress state. The resting state may correspond to baseline physiological conditions without pharmacological or physiological stress. The stress state may be induced pharmacologically, for example using vasodilator agents such as adenosine orISUS Ref: 53516-001.2PCTdipyridamole, or physiologically, for example through exercise or other stress-inducing conditions.

[0134] In some exemplary embodiments, acquisition of the spectral Doppler ultrasound data may be performed using a guided acquisition process, wherein real-time ultrasound data is analyzed to determine a patient- specific coronary artery segment suitable for Doppler-based flow assessment, and guidance instructions are generated for positioning and orienting the probe toward the identified segment. In some exemplary embodiments, acquisition may be performed using a targeted sampling methodology, wherein Doppler measurements are obtained at one or more predefined anatomical sampling locations without requiring full imaging of the coronary vascular tree.

[0135] In some exemplary embodiments, the spectral Doppler ultrasound data may comprise PWD data representing blood flow velocity over time. Additionally, or alternatively, the spectral Doppler ultrasound data may be reconstructed from imagebased formats such as DICOM images using image processing techniques configured to extract spectral and flow-related information. Such reconstruction may include identification of spectral envelopes, velocity distributions, and temporal flow characteristics embedded within the image data. Additionally, or alternatively, data acquisition may further comprise obtaining ECG signals. The ECG signals may be used to provide temporal reference for cardiac cycle phases, enabling synchronization of acquired Doppler signals with diastolic and systolic phases of cardiac cycle. Additionally, or alternatively, the PWD data may be processed to extract time-resolved Doppler waveforms synchronized with cardiac cycle phases. In some exemplary embodiments, the spectral Doppler ultrasound data may comprise pulsed-wave Doppler data reconstructed from DICOM images using image processing techniques to extract spectral and flow-related information. Additionally, or alternatively, the spectral Doppler ultrasound data may be acquired in combination with B-mode imaging data and color Doppler data, enabling anatomical localization and flow visualization to support Doppler signal acquisition.

[0136] In Step 320, a plurality of flow-related features may be extracted. The plurality of flow-related features may comprise at least one of flow velocity parameters, flow intensity parameters, diastolic flow characteristics, and flow direction components. The extracted features may be derived from spectral Doppler waveform data representingISUS Ref: 53516-001.2PCTcoronary blood flow dynamics. In some exemplary embodiments, the plurality of flow-related features may comprise tens or more features representing temporal, spectral, and directional aspects of coronary blood flow dynamics.

[0137] In some exemplary embodiments, extracting the plurality of flow-related features may comprise computing one or more of diastolic acceleration time, diastolic deceleration time, systolic time, diastolic time, end-diastolic velocity, end-systolic velocity, mean diastolic velocity, mean systolic velocity, peak diastolic velocity, peak systolic velocity, time-averaged velocity, diastolic time-velocity integral, and systolic time-velocity integral. Additionally, derived indices such as resistive index and pulsatility index may be computed.

[0138] Additionally, or alternatively, the extraction may further comprise analyzing Doppler-derived velocity measurements to determine net positive flow, net negative flow, and ratios between forward and reverse flow components. Directional flow features may include antegrade flow components and retrograde flow components, including detection of flow below Doppler baseline. In some exemplary embodiments, identification of coronary circulatory dysfunction may comprise analyzing net positive and net negative flow components across one or more cardiac cycles.

[0139] In some exemplary embodiments, feature extraction may further comprise generating a multi-parametric representation of coronary flow, combining temporal, spectral, and directional features to characterize coronary flow dynamics across cardiac cycles. Such representation may capture waveform morphology, temporal variability, and flow distribution patterns. Additionally, or alternatively, extraction of the plurality of flow-related features may be synchronized with cardiac cycle phases based on ECG signals, enabling alignment of extracted features with corresponding systolic and diastolic phases.

[0140] In some exemplary embodiments, Doppler waveform analysis may be synchronized with cardiac cycle phases based on ECG signals to align diastolic and systolic measurements. In some exemplary embodiments, coronary circulatory dysfunction may be evaluated based on a ratio between diastolic time-velocity integral and systolic time-velocity integral. In some exemplary embodiments, the multiparametric representation may comprise a flow profile describing temporal evolution,ISUS Ref: 53516-001.2PCTspectral distribution, and directional components of coronary blood flow across one or more cardiac cycles.

[0141] In Step 330, the extracted flow-related features may be processed using one or more Al-based models to evaluate coronary flow values corresponding to the resting state and the stress state. The processing of the extracted flow-related features may comprise applying a dynamic Al-based model configured to derive time-varying characteristics of coronary blood flow from the extracted features. The dynamic Al-based model may model temporal flow behavior including diastolic-dominant flow patterns, systolic components, flow velocity profiles, spectral energy distribution, retrograde flow magnitude, and beat-to-beat variability.

[0142] Additionally, or alternatively, processing of the extracted flow-related features may comprise applying a static Al-based model configured to determine non-linear relationships between the derived flow characteristics and coronary circulatory dysfunction indicators. The static Al-based model may map combinations of Doppler-derived features to clinical outcomes including presence and severity of epicardial stenosis, microvascular dysfunction, and global coronary flow impairment.

[0143] In some exemplary embodiments, the Al-based models may be trained based on spectral Doppler ultrasound data synchronized with cardiac cycle phases using ECG signals. Additionally, or alternatively, the Al-based models may be configured to derive physiologically meaningful flow representations from Doppler-derived measurements based on learned relationships between signal characteristics and underlying hemodynamic behavior.

[0144] In Step 340, a non-invasive CFR value may be computed based on a ratio between stress-state and rest-state flow values. In some exemplary embodiments, Computing the non-invasive CFR value may comprise converting Doppler-derived velocity measurements obtained during resting state and stress state into estimated coronary flow values based on relationships learned by the Al-based model. The conversion may account for relationships between velocity measurements and physiological flow behavior. The non-invasive CFR value may be computed according to a ratio between flow during the stress state and flow during the resting state, thereby providing a measure of functional capacity of coronary circulation. In some exemplary embodiments, additional flow-related indices may be computed, including ratios of netISUS Ref: 53516-001.2PCTpositive to net negative flow, or other derived parameters reflecting coronary flow dynamics. In some exemplary embodiments, the computed CFR value is derived from Al-inferred flow representations and is not limited to a ratio of measured velocities, thereby improving correspondence with true volumetric coronary flow.

[0145] Additionally, or alternatively, the non-invasive CFR value may be automatically and directly computed using one or more Al-based models configured to infer the CFR value from Doppler-derived features without explicitly computing intermediate flow values. In some exemplary embodiments, the Al-based model may comprise an end-to-end model configured to receive spectral Doppler data or extracted features as input and generate a CFR value as output.

[0146] In Step 350, a physiological assessment output may be generated based on the computed CFR value and processed flow-related features. The physiological assessment output may be indicative of presence and / or severity of coronary circulatory dysfunction. The coronary circulatory dysfunction may be selected from a group consisting of CAD and CMD. Generating the physiological assessment output may comprise jointly evaluating the computed CFR value and retrograde coronary flow magnitude. In some exemplary embodiments, identification of microvascular disease may comprise determining that the computed CFR value is below a pathological threshold and determining absence of epicardial coronary artery disease indicators based on additional diagnostic input data. The additional diagnostic input data may comprise coronary angiography or CTCA. Identification of microvascular dysfunction may further comprise detecting retrograde coronary flow components and quantifying reverse perfusion patterns in distal coronary segments. In some exemplary embodiments, identification of microvascular disease may further comprise identifying a reduced diastolic flow response during the stress state relative to the resting state. In some exemplary embodiments, determination of coronary circulatory dysfunction may be based on a combination of the computed CFR value and retrograde flow magnitude. In some exemplary embodiments, Method 300 may further comprise evaluating a quality and / or validity of the acquired Doppler signal based on predefined technical and clinical criteria, and, when the signal does not satisfy the criteria, iteratively updating probe positioning, Doppler sampling location, and / or acquisition parameters until a valid signal is obtained.ISUS Ref: 53516-001.2PCT

[0147] In some exemplary embodiments, evaluation of coronary artery disease may comprise computing parameters including pre-stenotic to stenotic flow velocity ratio, distal to proximal diastolic peak velocity ratio, post-stenotic damping indicator, and retrograde coronary flow indicator. Additionally, or alternatively, evaluation of coronary artery disease may be performed for one or more coronary arteries including Left Main (LM), Left Anterior Descending (LAD), Left Circumflex (LCx), and Right Coronary Artery (RCA). Evaluation may further comprise determining stenosis severity based on flow velocities, flow ratios, and diastolic-to-systolic flow relationships, and identifying flow disturbances including spectral broadening or aliasing. In some exemplary embodiments, stenosis severity may be determined based on at least one of peak diastolic velocity, pre-stenotic to stenotic flow velocity ratio, post-stenotic damping, and diastolic-to-systolic flow ratio. In some exemplary embodiments, evaluation may further comprise analyzing relationships between measurements obtained from different coronary arteries or segments to identify interdependent hemodynamic patterns across the coronary vascular tree.

[0148] The physiological assessment output may comprise a virtual coronary evaluation including anatomical localization of vascular impairment including an estimated anatomical location and degree of arterial narrowing, severity classification, predicted plaque composition category, and risk stratification based on the computed CFR value. The physiological assessment output may further comprise recommended follow-up imaging, monitoring plans, and recommendations regarding invasive diagnostic procedures. The output may be presented as a virtual diagnostic catheterization report. In some exemplary embodiments, the physiological assessment output may further comprise clinically actionable insights derived from analysis of flow velocity, flow distribution, and temporal flow characteristics.

[0149] Additionally, or alternatively, Method 300 may comprise integrating patientspecific data with the spectral Doppler ultrasound data. The patient- specific data may comprise physiological measurements, demographic information, medical history, and laboratory test results, enabling improved diagnostic accuracy. Additionally, or alternatively, Method 300 may comprise integrating patient- specific data with the spectral Doppler ultrasound data. The patient-specific data may comprise one or more of physiological measurements, risk factors, medical background including medications andISUS Ref: 53516-001.2PCTchronic conditions, demographic information, laboratory test results, and family medical history. The patient-specific data may be incorporated into the Al-based models to enable personalized assessment and improved diagnostic accuracy.

[0150] Additionally, or alternatively, Method 300 may comprise, during a training or validation phase, comparing generated physiological assessment outputs with invasive coronary physiological measurements including invasive CFR and IMR, and refining model performance based on the comparison. In some exemplary embodiments, model refinement may further comprise updating model parameters based on discrepancies between predicted and measured physiological outcomes.

[0151] Referring now to Figure 4, showing a schematic representation of a data processing pipeline for transforming spectral Doppler ultrasound data into flow-related representations, in accordance with some exemplary embodiments of the disclosed subject matter.

[0152] Figure 4 illustrates a progression of data representations derived from spectral Doppler ultrasound data, beginning with raw Doppler data and proceeding through successive processing stages to generate time-resolved and intensity-based flow representations.

[0153] Spectral Doppler Ultrasound Data 410 may represent Doppler-based data, such as pulsed-wave Doppler (PWD) data obtained from an ultrasound imaging system. In some exemplary embodiments, Spectral Doppler Ultrasound Data 410 may correspond to raw Doppler data, including data reconstructed from DICOM images. The data may represent blood flow velocity over time and may include noise, background signal components, and unprocessed spectral information.

[0154] Spectral Doppler Ultrasound Data 410 may represent Doppler-based data, such as PWD data obtained from an ultrasound imaging system. In some exemplary embodiments, Spectral Doppler Ultrasound Data 410 may comprise a spectral Doppler waveform graph representing blood flow velocity over time, optionally presented together with an ultrasound scan providing anatomical context. The spectral Doppler waveform may represent blood flow velocity derived from Doppler shift frequency, where the horizontal axis represents time and the vertical axis represents velocity, and intensity indicates the distribution of blood cells moving at a given velocity. The dataISUS Ref: 53516-001.2PCTmay include raw or minimally processed Doppler information, including noise, background signal components, and unprocessed spectral information.

[0155] Additionally, or alternatively, Spectral Doppler Ultrasound Data 410 may comprise data reconstructed from DICOM images using image-processing techniques configured to extract spectral and flow-related information. In some exemplary embodiments, the reconstruction may include identification of spectral envelopes, velocity distributions, and temporal flow characteristics. Additionally, or alternatively, Spectral Doppler Ultrasound Data 410 may comprise other forms of flow-related data, including continuous-wave Doppler data, color Doppler data, duplex ultrasound data, or flow-related data derived from other imaging modalities capable of representing blood flow dynamics.

[0156] The data of Spectral Doppler Ultrasound Data 410 may be processed to generate Processed Doppler Waveform 420, which may represent a refined Doppler waveform. In some exemplary embodiments, the transition from Spectral Doppler Ultrasound Data 410 to Processed Doppler Waveform 420 may comprise initial signal processing operations including noise reduction, signal enhancement, and extraction of a velocity envelope from the Doppler signal, such as identification of peak velocity contours corresponding to systolic and diastolic flow components. Additionally, or alternatively, the processing may comprise segmentation of the Doppler waveform into cardiac cycle phases, including identification of systolic and diastolic intervals, optionally based on temporal waveform patterns or synchronized cardiac-cycle information. Processed Doppler Waveform 420 may therefore represent a structured waveform in which flow-related features, waveform contours, and temporal phase information are explicitly defined. In some exemplary embodiments, segmentation and temporal alignment may be further refined using synchronized ECG signals, enabling accurate identification of systolic and diastolic phases and improving robustness of waveform analysis across cardiac cycles.

[0157] In some exemplary embodiments, Processed Doppler Waveform 420 may further be refined through optimization of phase segmentation and identification of cardiac cycles and characteristic peaks. Such processing may include selection of representative cardiac cycles, alignment of waveform segments, and identification of peak systolic velocity (PSV), end-diastolic velocity (EDV), and additional waveform-ISUS Ref: 53516-001.2PCTderived parameters, thereby improving robustness and consistency of subsequent flow analysis.

[0158] From Processed Doppler Waveform 420, the data may be further processed along one or more paths to generate flow-related representations, including Time-Resolved Velocity Representation 430 and Intensity-Based Flow Representation 440. In some exemplary embodiments, the parallel generation of velocity-based and intensitybased representations enables complementary characterization of coronary flow, capturing both peak velocity behavior and distributed flow dynamics.

[0159] Time-Resolved Velocity Representation 430 may correspond to a representation of a graph representing extracted positive and negative maximum flow velocities over time. In some exemplary embodiments, Time-Resolved Velocity Representation 430 may comprise velocity curves segmented into cardiac cycle phases such as systole and diastole. The representation may include forward flow components corresponding to positive velocities and reverse flow components corresponding to negative velocities. The data may represent peak velocities extracted from the Doppler waveform and may include temporal variation of velocity across cardiac cycles. Time-Resolved Velocity Representation 430 may therefore provide a structured temporal representation of coronary blood flow dynamics suitable for extraction of velocity-based features.

[0160] Intensity-Based Flow Representation 440 may correspond to a representation of a graph representing positive, negative, and total flow components over time. In some exemplary embodiments, Intensity-Based Flow Representation 440 may be derived from integration of velocity and intensity information across spectral Doppler data, including pixel-wise or region-based analysis of Doppler signal intensity. The representation may comprise total flow, positive flow, and negative flow components, where total flow may be computed as a combination of forward and reverse flow contributions. The representation may exhibit periodic variations corresponding to pulsatile flow behavior and may enable identification of flow magnitude, flow distribution, and temporal variability. In some exemplary embodiments, intensity-based flow representation may provide an indirect estimation of flow magnitude by integrating spectral energy distribution, thereby enabling analysis of flow characteristics beyond peak velocity measurements.ISUS Ref: 53516-001.2PCT

[0161] In some exemplary embodiments, the transition from Processed Doppler Waveform 420 to Time-Resolved Velocity Representation 430 and Intensity-Based Flow Representation 440 may comprise parallel or alternative processing paths. One path may focus on extraction of velocity-based temporal characteristics, including peak velocity detection and phase segmentation, while another path may focus on extraction of intensity-based flow characteristics, including spectral energy distribution and flow integration across time. The resulting representations may correspond to different aspects of coronary flow dynamics and may be used jointly or separately for feature extraction.

[0162] In some exemplary embodiments, Time-Resolved Velocity Representation 430 and Intensity-Based Flow Representation 440 may be utilized to extract flow-related features, including velocity parameters, temporal characteristics, intensity-based measures, and directional flow components. The extracted features may include peak velocities, mean velocities, time-velocity integrals, diastolic and systolic ratios, net positive flow, net negative flow, and ratios therebetween, as well as identification of retrograde flow patterns and flow disturbances. In some exemplary embodiments, the extracted features may comprise tens or more parameters representing multiple aspects of coronary flow dynamics, including temporal evolution, spectral distribution, and directional flow behavior.

[0163] In some exemplary embodiments, the combined representations may provide a multi-parametric characterization of coronary flow dynamics, integrating temporal, spectral, and directional information. Such characterization may support subsequent evaluation of coronary circulatory dysfunction, including identification of abnormal flow patterns, stenosis-related effects, and microvascular flow behavior.

[0164] Referring now to Figure 5, showing an exemplary PWD ultrasound imaging display, in accordance with some exemplary embodiments of the disclosed subject matter.

[0165] Ultrasound Imaging Display 500a illustrates a realistic ultrasound image, while Ultrasound Imaging Display 500b illustrates a corresponding schematic representation of the image of Ultrasound Imaging Display 500a.

[0166] In some exemplary embodiments, Ultrasound Imaging Display 500a, 500b may be generated by an ultrasound imaging system configured to acquire spectral DopplerISUS Ref: 53516-001.2PCTultrasound data representing blood flow within a cardiovascular structure. The displayed data may comprise PWD data, and additionally, or alternatively, other spectral Doppler modalities or reconstructed Doppler data derived from image-based formats such as DICOM. The illustrated display may represent coronary blood flow, although similar representations may be obtained from other vascular regions.

[0167] In some exemplary embodiments, Ultrasound Imaging Display 500a may be obtained using an automated or semi-automated acquisition process. Such acquisition may be performed using one or more trained models configured to analyze real-time ultrasound data and determine a patient-specific coronary artery segment suitable for Doppler-based flow assessment. The system may generate guidance instructions for positioning and orienting a probe toward the identified coronary segment and for placing a Doppler sampling cursor at a patient- specific final acquisition position optimized for PWD signal acquisition. Additionally, or alternatively, the guidance may be provided to an operator as visual or audio instructions, or may be used to control a robotic or autonomous positioning mechanism configured to adjust probe position and orientation. In some exemplary embodiments, the acquisition process may further comprise verification of the acquired PWD signal based on clinical validity and technical quality criteria, and iterative refinement of probe positioning and / or cursor placement until a valid coronary flow signal is obtained. In some exemplary embodiments, acquisition may be performed at one or more predefined sampling locations within coronary arteries, enabling targeted evaluation without requiring full visualization of the coronary vascular tree.

[0168] In some exemplary embodiments, Ultrasound Imaging Display 500a, 500b may provide a visualization of spectral Doppler data configured to represent blood-flow velocity and flow patterns within coronary arteries. The displayed PWD data may be analyzed to evaluate coronary flow characteristics, including assessment of stenosis severity and identification of coronary circulatory dysfunction. The visualization may therefore serve as a non-invasive basis for evaluating both epicardial coronary artery disease and microvascular dysfunction.

[0169] In some exemplary embodiments, Ultrasound Scan 510 may comprise a two-dimensional anatomical imaging region, such as a B-mode ultrasound image, providing structural context for Doppler signal acquisition. Ultrasound Scan 510 may enableISUS Ref: 53516-001.2PCTidentification of a target vessel and positioning of a sampling region within the vessel lumen. A magnified inset View 540 may be provided, illustrating an enlarged portion of a sector in Ultrasound Scan 510, for example to enhance visualization thereof.

[0170] In some exemplary embodiments, Sample Volume Indicator 520 may represent a selected spatial location within Ultrasound Scan 510 from which Doppler measurements are obtained. Sample Volume Indicator 520 may define a region of interest corresponding to a Doppler sampling gate. Placement of Sample Volume Indicator 520 may be performed manually by an operator, or additionally, or alternatively, automatically using image analysis or Al-based positioning techniques.

[0171] In some exemplary embodiments, Cursor 530 may represent a positioning marker associated with Doppler interrogation within Ultrasound Scan 510 and / or Inset View 540. Cursor 530 may indicate an alignment direction, interrogation line, or selected placement position used in association with Sample Volume Indicator 520 for obtaining Doppler measurements from a target vessel region. In some exemplary embodiments, Cursor 530 may be positioned manually by an operator. Additionally, or alternatively, Cursor 530 may be positioned or adjusted automatically using image analysis or Al-based guidance techniques configured to identify a patient- specific acquisition position suitable for PWD signal acquisition. In some exemplary embodiments, Cursor 530 may be guided toward a patient- specific final acquisition position based on real-time analysis of ultrasound data and / or Doppler flow data.

[0172] In some exemplary embodiments, Spectral Doppler Waveform 570 may represent blood flow velocity over time, derived from Doppler shift measurements. Spectral Doppler Waveform 570 may comprise a time axis and a velocity axis, and may include both positive and negative velocity components corresponding to forward and reverse flow, respectively. The intensity of the waveform may reflect the distribution of velocities within the sampled volume.

[0173] In some exemplary embodiments, Velocity Envelope 550 may represent a boundary of peak velocities within Spectral Doppler Waveform 570. Velocity Envelope 550 may be determined using signal processing, image analysis, or Al-based techniques configured to identify maximum velocity trajectories over time. The extracted envelope may be used to derive flow-related parameters including peak systolic velocity, end-diastolic velocity, and temporal flow characteristics.ISUS Ref: 53516-001.2PCT

[0174] In some exemplary embodiments, Spectral Doppler Waveform 570 may further exhibit flow disturbances including spectral broadening or aliasing, which may be indicative of abnormal hemodynamic conditions. Such waveform characteristics may be utilized in assessing coronary pathology and identifying flow irregularities associated with stenotic or dysfunctional vascular regions.

[0175] In some exemplary embodiments, Spectral Doppler Waveform 570 may further comprise Positive Flow Component 574 and Negative Flow Component 572, representing antegrade and retrograde flow, respectively. Analysis of Positive Flow Component 574 and Negative Flow Component 572 may enable identification of flow direction, reverse perfusion patterns, and flow disturbances associated with coronary pathology.

[0176] In some exemplary embodiments, analysis of Positive Flow Component 574 and Negative Flow Component 572 may enable quantification of net positive flow, net negative flow, and ratios therebetween across one or more cardiac cycles. Such analysis may provide insight into flow balance, retrograde perfusion patterns, and dynamic flow redistribution, including identification of reverse flow phenomena associated with microvascular dysfunction.

[0177] In some exemplary embodiments, Auxiliary Signal 580 may represent an additional physiological signal, such as an ECG, synchronized with Spectral Doppler Waveform 570. Auxiliary Signal 580 may provide temporal reference for cardiac cycle phases, enabling alignment of Doppler-derived measurements with systolic and diastolic phases.

[0178] In some exemplary embodiments, Ultrasound Imaging Display 500b may provide a schematic abstraction of the signal shown in Ultrasound Imaging Display 500a, emphasizing features relevant for computational analysis. The schematic representation may highlight structures including Cursor 530, Velocity Envelope 550, Positive Flow Component 574, Negative Flow Component 572, and temporal variations within Spectral Doppler Waveform 570, thereby facilitating automated processing and feature extraction.

[0179] In some exemplary embodiments, the extracted flow-related features may comprise parameters including diastolic acceleration time, diastolic deceleration time, diastolic time, systolic time, end-diastolic velocity, end-systolic velocity, mean diastolicISUS Ref: 53516-001.2PCTvelocity, mean systolic velocity, peak diastolic velocity, peak systolic velocity, time-averaged velocity, diastolic time-velocity integral, systolic time-velocity integral, pulsatility indices, resistive indices, beat duration, and spectral window characteristics. Additionally, or alternatively, flow-derived measures may include total net flow across a cardiac cycle, maximum and mean diastolic and systolic flow values, and ratios between diastolic and systolic flow components.

[0180] In some exemplary embodiments, Ultrasound Imaging Display 500a, 500b may serve as an input to one or more Al-based models configured to analyze Doppler-derived data. The Al-based models may be configured to extract flow-related features from Spectral Doppler Waveform 570, including velocity parameters, temporal characteristics, intensity-based measures, and directional flow components. The extracted features may include peak velocities, mean velocities, time-velocity integrals, ratios between systolic and diastolic flow, net positive flow, net negative flow, and ratios therebetween.

[0181] In some exemplary embodiments, the extracted features may be utilized to generate representations such as those described with reference to Figure 4, including time-resolved velocity representations and intensity -based flow representations. Additionally, or alternatively, the features may be provided directly to Al-based diagnostic models configured to evaluate coronary flow characteristics and identify coronary circulatory dysfunction.

[0182] In some exemplary embodiments, processing of Ultrasound Imaging Display 500a, 500b may be performed using an Al architecture comprising a dynamic model configured to extract time-varying flow characteristics and a static model configured to determine non-linear relationships between the extracted characteristics and coronary pathology. The dynamic model may analyze temporal flow behavior reflected in Spectral Doppler Waveform 570, while the static model may map derived features to diagnostic outcomes including presence and severity of coronary artery disease and microvascular dysfunction.

[0183] In some exemplary embodiments, analysis of Spectral Doppler Waveform 570 may enable identification of coronary artery disease and microvascular dysfunction based on flow dynamics. Such analysis may include detection of reduced diastolic flow, abnormal velocity ratios, retrograde flow patterns, spectral broadening, or other deviations from expected physiological behavior.ISUS Ref: 53516-001.2PCT

[0184] In some exemplary embodiments, analysis of Spectral Doppler Waveform 570 and derived features may support computation of coronary flow reserve-related measures and additional hemodynamic indicators, including ratios between diastolic and systolic flow, pre-stenotic to stenotic flow velocity ratios, post-stenotic damping indicators, and retrograde flow metrics. Such measures may be utilized to assess functional capacity of coronary circulation and to distinguish between epicardial and microvascular contributions to coronary dysfunction.

[0185] In some exemplary embodiments, detection and quantification of retrograde flow within Negative Flow Component 572 may indicate adaptive microvascular behavior in response to upstream flow limitation. Such reverse perfusion patterns, particularly when identified in distal segments of coronary arteries, may be indicative of microvascular dysfunction and may be evaluated in conjunction with other flow-derived parameters.

[0186] In some exemplary embodiments, the analysis may be performed across one or more coronary arteries including left anterior descending artery, right coronary artery, and left circumflex artery, and additionally, or alternatively, left main coronary artery. The resulting assessment may identify affected anatomical segments and characterize severity and location of coronary impairment.

[0187] In some exemplary embodiments, the analysis of Ultrasound Imaging Display 500a, 500b may be combined with patient-specific data including demographic information, physiological measurements, medical history, laboratory results, and risk factors. Such data may be obtained from electronic health records, external systems, user input, or connected devices, and may be integrated with Doppler-derived features to enhance diagnostic accuracy and contextual interpretation.

[0188] In some exemplary embodiments, the processed outputs derived from Ultrasound Imaging Display 500a, 500b may be utilized to generate a virtual coronary evaluation or virtual diagnostic catheterization output. The output may include anatomical localization of vascular impairment, severity classification of dysfunction, predicted plaque composition, and recommendations for further evaluation or intervention. Additionally, or alternatively, model outputs may be compared with invasive coronary measurements, including invasive coronary flow reserve or index ofISUS Ref: 53516-001.2PCTmicrocirculatory resistance, to refine model performance and improve diagnostic reliability.

[0189] Referring now to Figure 6, showing a schematic flow diagram of an integrated processing and analysis architecture for non-invasive evaluation of coronary circulatory dysfunction, in accordance with some exemplary embodiments of the disclosed subject matter. Figure 6 illustrates a pipeline configured to process spectral Doppler ultrasound data and derive physiological assessments indicative of CAD and CMD.

[0190] In some exemplary embodiments, Ultrasound Data 610 may comprise spectral Doppler ultrasound data representing coronary blood flow, including PWD data acquired at one or more selected anatomical sampling locations within coronary vasculature. The acquisition may be performed using TTDE and may be obtained from one or more coronary arteries, including left main (LM), left anterior descending (LAD), right coronary artery (RCA), and left circumflex (LCx). In some cases, the acquisition may be performed at predefined sampling points, without requiring full scanning of an entire vascular tree, thereby enabling efficient and targeted evaluation of coronary flow dynamics. The selected sampling locations may include proximal and / or distal segments, enabling inference of both epicardial flow characteristics and downstream microvascular effects.

[0191] Ultrasound Data 610 may reflect blood flow velocity over time, wherein Doppler-derived signals provide velocity-based representations rather than direct volumetric flow measurements. In some exemplary embodiments, flow characteristics may be inferred from velocity profiles, including peak velocity behavior, diastolic-dominant flow patterns, and ratios between systolic and diastolic components. Localized increases in velocity may be indicative of stenotic regions, while distal flow behavior may reflect downstream microvascular regulation. In some exemplary embodiments, the system does not rely solely on direct velocity measurements but derives flow-related characteristics through multi-parametric analysis of Doppler signals.

[0192] Ultrasound Data 610 may be provided to Noise Filtering and Signal Enhancement Module 612, configured to reduce noise and artifacts and enhance signal quality. Noise Filtering and Signal Enhancement Module 612 may remove artifacts arising from surrounding anatomical structures, unrelated vascular signals, or transient physiological effects, thereby improving robustness and reliability of subsequent featureISUS Ref: 53516-001.2PCTextraction. In some exemplary embodiments, the module may further enhance signal quality to support subsequent evaluation of signal validity and clinical usability.

[0193] Ultrasound Data 610 may further be processed by Preprocessing Module 620, configured to prepare Doppler waveform data for analysis. Preprocessing Module 620 may include segmentation of Doppler signals into cardiac cycles, temporal alignment of signals, and synchronization with cardiac cycle phases. Preprocessing Module 620 may be coupled to Data Normalization Module 622, configured to standardize signal amplitudes and temporal characteristics across acquisitions, subjects, and imaging systems. Preprocessing may further account for variability across different ultrasound systems and acquisition conditions to ensure consistent downstream analysis.

[0194] Output of Preprocessing Module 620 may be provided to Signal Processing Module 630, configured to extract a plurality of flow-related features from Doppler waveform data. In some exemplary embodiments, Signal Processing Module 630 may extract dozens of flow-related features, including velocity-based parameters, temporal characteristics, intensity-based measures, and directional flow components. Extracted features may include diastolic peak velocity, systolic-to-diastolic flow ratios, turbulence indicators, and additional hemodynamic parameters derived from Doppler waveform analysis. The extracted features may comprise tens or more parameters characterizing coronary flow dynamics.

[0195] Signal Processing Module 630 may generate multiple representations, including Signal Morphology Module 634, configured to characterize waveform shape, diastolic and systolic structure, spectral envelope properties, and temporal evolution of flow patterns; and Flow Velocity Analysis Module 636, configured to extract peak velocities, velocity ratios, and forward and reverse flow components across cardiac cycles.

[0196] In some exemplary embodiments, extracted features may further include CFR, peak diastolic velocity (PDV), diastolic-to-systolic flow ratio, distal-to-proximal diastolic peak velocity ratio, pre-stenotic to stenotic flow velocity ratio, retrograde coronary flow indicators, post-stenotic damping patterns, flow disturbances including spectral broadening or aliasing, or the like. Signal Processing Module 630 may further apply flow equations and predefined or learned parameters to evaluate flow behavior and support localization and severity assessment of coronary stenosis. In some exemplaryISUS Ref: 53516-001.2PCTembodiments, extracted features may reflect distal flow behavior indicative of microvascular regulation and compensatory flow redistribution.

[0197] In some exemplary embodiments, Signal Processing Module 630 may operate in conjunction with ECG Signal 632, enabling synchronization of Doppler waveform analysis with systolic and diastolic phases, thereby improving temporal alignment and accuracy of extracted features.

[0198] Signal Processing Module 630 may further receive auxiliary inputs including Physiological Measurements 642 and Demographic and Clinical Data 644, enabling integration of Doppler-derived features with patient- specific context. Physiological Measurements 642 may comprise parameters such as blood pressure, heart rate, respiratory rate, and oxygen saturation. Demographic and Clinical Data 644 may comprise age, sex, medical history, risk factors, medications, and laboratory results. Such integration may enable multi-modal analysis combining Doppler-derived flow characteristics, electrophysiological signals, and patient- specific clinical data to enhance diagnostic interpretation. Such integration may enable personalized evaluation of coronary physiology and improved differentiation between pathological conditions.

[0199] Output of Signal Processing Module 630 may be provided to Al Analysis Module 640, configured to analyze extracted features using one or more Al-based models. In some exemplary embodiments, Al Analysis Module 640 may comprise a dynamic model configured to derive physical and time-varying characteristics of blood flow, including diastolic dominance, velocity distributions, and retrograde flow behavior, and a static model configured to determine non-linear relationships between the extracted characteristics and coronary pathology, including CAD and CMD.

[0200] In some exemplary embodiments, Al Analysis Module 640 may model complex, non-linear interactions between features, including relationships across different arterial segments and distal flow regions and between measurements obtained from different coronary arteries or segments. Al Analysis Module 640 may further refine extracted features through iterative processing and optimization, enabling improved sensitivity and accuracy in detection of coronary abnormalities.

[0201] In some exemplary embodiments, analysis of distal flow characteristics and retrograde flow components may enable identification of microvascular dysfunction.ISUS Ref: 53516-001.2PCTReverse or retrograde flow patterns in distal segments of major coronary arteries may reflect compensatory microvascular regulation in response to upstream flow limitation, thereby providing an indication of CMD even in the absence of significant epicardial stenosis.

[0202] Al Analysis Module 640 may generate outputs provided to Diagnostic Output Module 650, configured to produce a physiological assessment indicative of coronary circulatory dysfunction. Diagnostic Output Module 650 may identify coronary stenosis, occlusions, and abnormal flow dynamics based on the analyzed features and may evaluate both epicardial and microvascular components of coronary disease. The output may further comprise clinically actionable insights derived from analysis of flow patterns and hemodynamic features.

[0203] In some exemplary embodiments, Diagnostic Output Module 650 may generate a virtual coronary evaluation including identification of affected coronary arteries, estimation of stenosis severity, characterization of flow abnormalities, and assessment of coronary functional capacity. Diagnostic Output Module 650 may further provide risk stratification and recommendations for further diagnostic evaluation or therapeutic intervention.

[0204] In some exemplary embodiments, the system may derive a non-invasive CFR measure based on Al-inferred flow values. Unlike conventional CFVR approaches that rely solely on velocity ratios between rest and stress states, the disclosed approach may infer flow-related characteristics from multiple Doppler-derived features and incorporate additional parameters, including retrograde flow and multi-feature relationships, thereby providing improved physiological accuracy. The derived CFR measure may be based on Al-inferred flow representations rather than direct velocity ratios alone.

[0205] Diagnostic Output Module 650 may further provide outputs to Validation Module 660, configured to evaluate model performance and diagnostic reliability. Validation Module 660 may receive Comparison Data 662, including gold-standard coronary angiography (CAG) measurements and other invasive coronary physiological measurements, enabling comparison between non-invasive outputs and reference standards, thereby enabling continuous learning and adaptive improvement of model performance over time.ISUS Ref: 53516-001.2PCT

[0206] Validation Module 660 may be coupled to Model Refinement Module 664, configured to update model parameters based on comparison results, thereby enabling continuous learning and improvement of diagnostic accuracy.

[0207] In some exemplary embodiments, the architecture illustrated in Figure 6 may provide an end-to-end, non-invasive alternative to catheter-based diagnostics, enabling comprehensive evaluation of coronary physiology, including both epicardial and microvascular components, based on spectral Doppler ultrasound data acquired at selected sampling locations and integrated with patient-specific clinical data.

[0208] Referring now to Figure 7, showing an exemplary results display of Doppler-derived coronary flow analysis, in accordance with some exemplary embodiments of the disclosed subject matter.

[0209] In some exemplary embodiments, Results Display 700 may present a consolidated visualization of coronary flow analysis derived from spectral Doppler ultrasound data acquired during resting and stress conditions. Results Display 700 may comprise a CFR indicator 710 configured to present a quantitative measure of coronary flow reserve. The CFR indicator may include a visual scale and a value marker 712 indicating a computed CFR value based on analysis of Doppler-derived flow characteristics.

[0210] In some exemplary embodiments, Results Display 700 may further comprise a rest-state Doppler analysis panel 720 presenting spectral Doppler data acquired during a resting condition. The rest-state panel may include a velocity waveform representing blood flow velocity over time, along with associated quantitative parameters derived from the waveform. Such parameters may include peak diastolic velocity, peak systolic velocity, mean velocity measures, velocity ratios, and temporal flow characteristics.

[0211] In some exemplary embodiments, Results Display 700 may further comprise a stress-state Doppler analysis panel 750 presenting spectral Doppler data acquired during a stress condition. The stress-state panel may similarly include a velocity waveform and corresponding quantitative parameters derived from the waveform, enabling evaluation of flow behavior under increased physiological demand.

[0212] In some exemplary embodiments, the results display may enable direct comparison between resting and stress flow characteristics to support evaluation ofISUS Ref: 53516-001.2PCTcoronary flow reserve and identification of coronary circulatory dysfunction. The displayed parameters and waveform characteristics may be derived from automated signal processing and Al-based analysis of Doppler data.

[0213] In some exemplary embodiments, the computed CFR value presented by CFR indicator 710 may be determined based on relationships between flow characteristics extracted from the rest-state panel 720 and the stress-state panel 750. The resulting output may provide a physiological assessment indicative of presence and / or severity of coronary artery disease and / or microvascular dysfunction.

[0214] The present invention may be a system, a method, and / or a computer program product. The computer program product may include a computer readable storage medium (or media) having computer readable program instructions thereon for causing a processor to carry out aspects of the present invention.

[0215] The computer readable storage medium can be a tangible device that can retain and store instructions for use by an instruction execution device. The computer readable storage medium may be, for example, but is not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. A non-exhaustive list of more specific examples of the computer readable storage medium includes the following: a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanically encoded device such as punch-cards or raised structures in a groove having instructions recorded thereon, and any suitable combination of the foregoing. A computer readable storage medium, as used herein, is not to be construed as being transitory signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through a waveguide or other transmission media (e.g., light pulses passing through a fiber-optic cable), or electrical signals transmitted through a wire.

[0216] Computer readable program instructions described herein can be downloaded to respective computing / processing devices from a computer readable storage medium or to an external computer or external storage device via a network, for example, theISUS Ref: 53516-001.2PCTInternet, a local area network, a wide area network and / or a wireless network. The network may comprise copper transmission cables, optical transmission fibers, wireless transmission, routers, firewalls, switches, gateway computers and / or edge servers. A network adapter card or network interface in each computing / processing device receives computer readable program instructions from the network and forwards the computer readable program instructions for storage in a computer readable storage medium within the respective computing / processing device.

[0217] Computer readable program instructions for carrying out operations of the present invention may be assembler instructions, instruction-set-architecture (ISA) instructions, machine instructions, machine dependent instructions, microcode, firmware instructions, state-setting data, or either source code or object code written in any combination of one or more programming languages, including an object oriented programming language such as Smalltalk, C++ or the like, and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The computer readable program instructions may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection may be made to an external computer (for example, through the Internet using an Internet Service Provider). In some embodiments, electronic circuitry including, for example, programmable logic circuitry, field-programmable gate arrays (FPGA), or programmable logic arrays (PLA) may execute the computer readable program instructions by utilizing state information of the computer readable program instructions to personalize the electronic circuitry, in order to perform aspects of the present invention.

[0218] Aspects of the present invention are described herein with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer readable program instructions.ISUS Ref: 53516-001.2PCT

[0219] These computer readable program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks. These computer readable program instructions may also be stored in a computer readable storage medium that can direct a computer, a programmable data processing apparatus, and / or other devices to function in a particular manner, such that the computer readable storage medium having instructions stored therein comprises an article of manufacture including instructions which implement aspects of the function / act specified in the flowchart and / or block diagram block or blocks.

[0220] The computer readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other device to produce a computer implemented process, such that the instructions which execute on the computer, other programmable apparatus, or other device implement the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0221] The flowchart and block diagrams in the Figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of instructions, which comprises one or more executable instructions for implementing the specified logical function(s). In some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and / or flowchart illustration, and combinations of blocks in the block diagrams and / or flowchart illustration, can be implemented by special purpose hardware-basedISUS Ref: 53516-001.2PCTsystems that perform the specified functions or acts or carry out combinations of special purpose hardware and computer instructions.

[0222] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and / or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0223] The corresponding structures, materials, acts, and equivalents of all means or step plus function elements in the claims below are intended to include any structure, material, or act for performing the function in combination with other claimed elements as specifically claimed. The description of the present invention has been presented for purposes of illustration and description, but is not intended to be exhaustive or limited to the invention in the form disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the invention. The embodiment was chosen and described in order to best explain the principles of the invention and the practical application, and to enable others of ordinary skill in the art to understand the invention for various embodiments with various modifications as are suited to the particular use contemplated.

[0224] A number of implementations have been described. Nevertheless, it will be understood that various modifications may be made without departing from the spirit and scope of the disclosure. Accordingly, other implementations are within the scope of the following claims.

Claims

ISUS Ref: 53516-001.2PCTCLAIMSWhat is claimed is:

1. A computer- implemented method for non-invasive detection and evaluation of coronary circulatory dysfunction, comprising:obtaining spectral Doppler ultrasound data representing coronary blood flow at one or more anatomical locations, the spectral Doppler ultrasound data being acquired during a resting state and during a stress state;extracting, by at least one processor, a plurality of flow-related features from the spectral Doppler ultrasound data;processing the extracted flow-related features using an Artificial Intelligence (Al)-based model to evaluate coronary flow values corresponding to the resting state and the stress state;computing a non-invasive coronary flow reserve (CFR) value based on a ratio between the coronary flow value during the stress state and the coronary flow value during the resting state; andgenerating, based on the processing and the computed non-invasive CFR value, a physiological assessment output indicative of the presence and / or severity of coronary circulatory dysfunction.

2. The method of Claim 1, wherein the plurality of flow-related features comprises at least one of: flow velocity parameters, flow intensity parameters, diastolic flow characteristics, and flow direction components.

3. The method of Claim 1, wherein processing the extracted features using the AI- based model comprises:applying a dynamic model configured to derive time-varying characteristics of coronary blood flow from the extracted flow-related features; andapplying a static model configured to determine non-linear relationships between the derived characteristics and coronary circulatory dysfunction indicators.ISUS Ref: 53516-001.2PCT4. The method of Claim 1, wherein extracting the plurality of flow-related features comprises computing one or more of:diastolic acceleration time;diastolic deceleration time;systolic time;diastolic time;end-diastolic velocity;end-systolic velocity;mean diastolic velocity;mean systolic velocity;peak diastolic velocity;peak systolic velocity;time-averaged velocity;diastolic time- velocity integral;systolic time-velocity integral;a resistive index; anda pulsatility index.

5. The method of Claim 1, wherein the Al-based model is configured to process spectral Doppler ultrasound data synchronized with cardiac cycle phases based on electrocardiogram (ECG) signals.

6. The method of Claim 1, wherein the Al-based model is further configured, during a training or a validation phase, to be refined based ona comparison between physiological assessment outputs and invasive coronary physiological measurements including at least one of: invasive CFR and index of microcirculatory resistance (IMR).

7. The method of Claim 1, wherein computing the non-invasive CFR value comprises converting Doppler-derived velocity measurements during the resting state and the stress state into estimated volumetric coronary flow values.ISUS Ref: 53516-001.2PCT8. The method of Claim 7, wherein computing the non-invasive CFR value comprises analyzing the Doppler-derived velocity measurements to determine at least one of:net positive flow;net negative flow; anda ratio of net positive flow to net negative flow.

9. The method of Claim 1, further comprises:synchronizing extraction of the plurality of flow-related features with cardiac cycle phases based on electrocardiogram signals.

10. The method of Claim 1, wherein the spectral Doppler ultrasound data comprises Pulsed Wave Doppler (PWD) data acquired from at least one of a left anterior descending artery, a right coronary artery, and a left circumflex artery.

11. The method of Claim 1, wherein the coronary circulatory dysfunction is selected from a group consisting of: Coronary Artery Disease (CAD) and Coronary Microvascular Dysfunction (CMD).

12. The method of Claim 11, wherein identifying CMD comprises:determining that the computed non-invasive CFR value is below a pathological threshold; anddetermining an absence of epicardial CAD indicators based on additional diagnostic input data;wherein the coronary circulatory dysfunction is classified as CMD when the pathological CFR value is present in the absence of epicardial coronary artery disease indicators.

13. The method of Claim 12, wherein the additional diagnostic input data comprise at least one of coronary angiography or Coronary Angiography Computed Tomography (CTCA).

14. The method of Claim 12, wherein identifying CMD comprises detecting retrograde coronary blood-flow components below a Doppler baseline and / or quantifying reverse perfusion patterns in distal segments of primary coronary arteries.ISUS Ref: 53516-001.2PCT15. A system for non-invasive detection and evaluation of coronary circulatory dysfunction, comprising:an ultrasound data acquisition interface configured to receive spectral Doppler ultrasound data representing coronary blood flow acquired at one or more anatomical locations during a resting state and during a stress state; at least one processor; anda memory storing instructions which, when executed by the at least one processor, cause the system to:extract a plurality of flow-related features from the spectral Doppler ultrasound data;process the extracted features using an Artificial Intelligence (AI)- based model to determine coronary flow values corresponding to the resting state and the stress state;compute a non-invasive Coronary Flow Reserve (CFR) value based on a ratio between the coronary flow value during the stress state and the coronary flow value during the resting state; andgenerate a physiological assessment output indicative of a presence or severity of coronary circulatory dysfunction, based on the computed non-invasive CFR.

16. The system of Claim 15, wherein the Al-based model comprises:a dynamic model configured to derive time-varying characteristics of coronary blood flow from the extracted features; anda static model configured to determine non-linear relationships between the derived characteristics and coronary circulatory dysfunction indicators.

17. The system of Claim 15, wherein identifying coronary circulatory dysfunction comprises jointly evaluating the computed non-invasive CFR value and retrograde coronary flow magnitude.

18. The system of Claim 15, wherein the instructions further cause the system to synchronize extraction of the plurality of flow-related features with cardiac cycle phases based on electrocardiogram signals.ISUS Ref: 53516-001.2PCT19. The system of Claim 15, wherein the spectral Doppler ultrasound data comprises Pulsed Wave Doppler data acquired from at least one of a left anterior descending artery, a right coronary artery, and a left circumflex artery.

20. The system of Claim 15, wherein the coronary circulatory dysfunction comprises at least one of Coronary Artery Disease (CAD) and Coronary Microvascular Dysfunction (CMD).

21. A non-transitory computer-readable medium storing instructions which, when executed by at least one processor, cause the processor to perform the method of Claim 1.

22. The non-transitory computer-readable medium of Claim 21, wherein the operations further comprise integrating patient- specific data with the spectral Doppler ultrasound data, the patient-specific data comprises at least one of physiological measurements, demographic information, medical history, and laboratory test results.

23. The non-transitory computer-readable medium of Claim 21, wherein the physiological assessment output comprises a virtual coronary evaluation including at least one of an anatomical location of vascular impairment and a severity classification of dysfunction.

24. The non-transitory computer-readable medium of Claim 21, wherein the operations further comprise synchronizing extraction of the plurality of flow-related features with cardiac cycle phases based on electrocardiogram signals.

25. The non-transitory computer-readable medium of Claim 21, wherein the spectral Doppler ultrasound data comprises Pulsed Wave Doppler data reconstructed from Digital Imaging and Communications in Medicine (DICOM) images using image processing to extract spectral and flow-related information.

26. A computer-implemented method for training an Artificial Intelligence (Al)-based model for evaluating coronary circulatory dysfunction, comprising:ISUS Ref: 53516-001.2PCTreceiving training data comprising spectral Doppler ultrasound data representing coronary blood flow acquired at one or more anatomical locations;extracting, by at least one processor, a plurality of flow-related features from the spectral Doppler ultrasound data;training the Al-based model to map the extracted flow-related features to at least one of: coronary flow values, Coronary Flow Reserve (CFR), and physiological assessment outputs indicative of coronary circulatory dysfunction; andupdating one or more parameters of the Al-based model based on comparison between model outputs and reference physiological measurements.

27. The method of Claim 26, wherein training the Al-based model comprises:training a dynamic model configured to derive time-varying characteristics of coronary blood flow from the extracted flow-related features; andtraining a static model configured to determine non-linear relationships between the derived characteristics and coronary circulatory dysfunction indicators.

28. The method of Claim 26, wherein the plurality of flow-related features comprises at least one of: diastolic flow parameters, systolic flow parameters, velocity-based parameters, intensity -based parameters, and directional flow components including antegrade and retrograde flow.

29. The method of Claim 26, wherein training the Al-based model comprises associating retrograde coronary blood-flow patterns with microvascular dysfunction labels.

30. A system for training an Artificial Intelligence (Al)-based model for evaluating coronary circulatory dysfunction, comprising:ISUS Ref: 53516-001.2PCTat least one processor; and a memory storing instructions which, when executed by the at least one processor, cause the system to perform the method of Claim 26.

31. A non-transitory computer-readable medium storing instructions which, when executed by at least one processor, cause the processor to perform the method of Claim 26.

Images