System and method for lung volume gated X-ray imaging

Abstract

A method of imaging a patient's lungs includes the steps of positioning the patient in a first orientation relative to an x-ray imaging device and obtaining a volume measurement of the patient's respiration. The method includes determining a respiratory phase of the patient based on the volume measurement of the patient's respiration while the patient is positioned in the first orientation relative to the x-ray imaging device and while obtaining the volume measurement of the patient's respiration. The method includes, based on determining that the respiratory phase of the patient matches a predefined respiratory phase, gating the x-ray imaging device to generate x-ray projections of the patient's lungs.

Description

Technical Field

[0001] Some embodiments of the present disclosure relate to medical imaging, and more particularly, to systems and methods for performing geometrically resolved radiographic x-ray imaging.Some embodiments of the present disclosure relate to radiation therapy. Background Art

[0002] Computational modeling of human anatomy helps understand the behavior of the anatomy under different physiological conditions. Although state-of-the-art imaging techniques allow physicians to visualize the behavior of the anatomy, state-of-the-art techniques that accurately image the complex movements of the heart and lungs are generally too expensive to be widely adopted. In addition, cardiac motion that is intricate and unrelated to the respiratory cycle, such as heart deformation, may appear as noise in CT- or radiographic-based respiratory motion measurements. Therefore, the accuracy of mathematical models describing respiratory motion will also be reduced by seemingly random cardiac motion. One way to address this issue is to eliminate lung motion in the image by having the patient hold their breath. While this method prevents the patient's respiratory motion, it can be difficult to identify important information about lung health in these breath-hold images. Summary of the invention

[0003] Embodiments of the present disclosure attempt to address at least one of the problems of the prior art by providing a geometrically resolved radiographic x-ray imaging system (GREX), at least to some extent.

[0004] According to some embodiments, the GREX imaging system acquires images of the patient's chest by computationally targeting specific respiratory phases and specific cardiac phases. By taking snapshots of the chest for the targeted respiratory phases in a series of imaging planes, the GREX imaging system obtains time-independent anatomical geometry. Based on the acquired respiratory and cardiac signals, the time-independent anatomical geometry is co-registered and interpolated to produce an interactive image of the chest volume. The interactive image has a key advantage over conventional static imaging. Using interactive images, GREX creates a biomechanical model that quantitatively describes how the chest geometry changes during the patient's breathing. Based on basic physical laws, the biomechanical model provides estimates of important quantities, such as lung tissue elasticity, stress, strain, and respiratory compliance. These properties and other properties provide diagnostic capabilities that are currently unavailable to doctors and medical practitioners. Doctors can use such information provided by the GREX imaging system to inform patients of the causes of lung disease.

[0005] In some embodiments, the GREX imaging system comprises a unique hardware and software add-on package for an existing digital diagnostic x-ray unit. Together, the hardware and software add-on package improves the diagnostic quality of a conventional digital diagnostic x-ray unit by providing new enhanced imaging capabilities using the patient's biometric signals to inform a customized imaging program that is more diagnostically informative. The hardware and software add-on package is a universal upgrade for any commercially available digital diagnostic x-ray unit (e.g., a conventional x-ray unit).

[0006] According to some embodiments, a method of imaging a lung of a patient is provided. The method includes positioning a patient in a first orientation relative to an x-ray imaging device (e.g., an x-ray unit) and obtaining a volumetric measurement of the patient's respiration. The method includes, while the patient is positioned in the first orientation relative to the x-ray imaging device, and while obtaining the volumetric measurement of the patient's respiration, determining a respiratory phase of the patient based on the volumetric measurement of the patient's respiration. The method includes, based on determining that the respiratory phase of the patient matches a predefined respiratory phase, gating the x-ray imaging device to generate x-ray projections of the patient's lungs.

[0007] In some embodiments, the predetermined breathing phase is a first predefined breathing phase among a plurality of predetermined breathing phases, and the method further comprises, when obtaining a volume measurement of the patient's breathing, gating an x-ray imaging device to generate an x-ray projection of the patient's lungs based on determining that the patient's breathing phase matches any one of the plurality of predefined breathing phases.

[0008] In some embodiments, x-ray measurements of the patient's lungs are obtained only when the patient's breathing phase determined by the volumetric measurement of the patient's breathing matches one of the plurality of predefined breathing phases.

[0009] In some embodiments, the plurality of predefined breathing phases comprises an early expiratory phase, a late expiratory phase, a maximal expiratory phase, an early inspiratory phase, a late inspiratory phase, and a maximal inspiratory phase of a complete breathing cycle of the patient.

[0010] In some embodiments, the x-ray projection is a first x-ray projection, and the method further includes repositioning the patient to a second orientation relative to the x-ray imaging device. In some embodiments, the method further includes: while the patient is positioned at the second orientation relative to the x-ray imaging device, and while continuing to obtain a volumetric measurement of the patient's respiration, continuing to determine a respiratory phase of the patient based on the volumetric measurement of the patient's respiration, and based on determining that the respiratory phase of the patient matches a predefined respiratory phase, gating the x-ray imaging device to produce a second x-ray projection of the patient's lung. In some embodiments, the method includes generating a static image cube corresponding to the predefined respiratory phase using the first x-ray projection and the second x-ray projection.

[0011] In some embodiments, a static image cube corresponding to the predefined breathing phase is generated using less than ten x-ray projections acquired from various angles during the predefined breathing phase.

[0012] In some embodiments, the volume measurement of the patient's breathing includes a measurement of the patient's chest rise.

[0013] In some embodiments, volumetric measurements of the patient's breathing are obtained using one or more volumetric breathing phase sensors from the group consisting of a three-dimensional (3D) scanner, a spirometer, and a belly belt.

[0014] In some embodiments, the method further comprises creating a point cloud of the patient's chest surface. A volume measurement of the patient's respiration is determined based on the point cloud of the patient's chest surface.

[0015] In some embodiments, one or more locations of the patient's chest are measured using 3D imaging techniques to obtain a point cloud of the patient's chest surface.

[0016] In some embodiments, the method further comprises identifying one or more anatomical landmarks on the surface of the patient's chest using the point cloud of the surface of the patient's chest, and inferring the positions of one or more internal anatomical landmarks within the patient's chest based on the point cloud of the surface of the patient's chest.

[0017] In some embodiments, the patient's breathing phase is a future breathing phase, and determining the patient's breathing phase based on a volumetric measurement of the patient's respiration includes predicting the future breathing phase based on one or more current and / or past breathing phases.

[0018] In some embodiments, a method is provided. The method includes positioning a patient at a first orientation relative to a radiation source, obtaining a measurement of the patient's respiration, and obtaining a measurement of the patient's cardiac function. The method also includes, while the patient is positioned at the first orientation relative to the radiation source and while obtaining the measurement of the patient's respiration, determining a respiratory stage of the patient based on the measurement of the patient's respiration and determining a cardiac stage of the patient based on the measurement of the patient's cardiac function. The method also includes, based on determining that the patient's respiratory stage matches a predefined respiratory stage and determining that the patient's cardiac stage matches a predefined window of a cardiac cycle, gating the radiation source to expose the patient to radiation.

[0019] In some embodiments, the radiation source is an x-ray imaging device, and gating the radiation source to expose the patient to radiation includes gating the x-ray imaging device to produce x-ray projections of the patient's lungs.

[0020] In some embodiments, the radiation source is a radiotherapy source, and gating the radiotherapy source to expose the patient to radiation includes gating the radiotherapy source to irradiate a region of a lung of the patient at a therapeutic dose.

[0021] In some embodiments, the method further comprises obtaining a measurement of cardiac function of the patient from a plurality of cardiac cycles of the patient prior to gating the radiation source to expose the patient to radiation, and determining an average interval between a predefined cardiac phase and the beginning of a predefined window of the cardiac cycle using the measurement of cardiac function of the patient from the plurality of cardiac cycles. In some embodiments, determining that the cardiac phase of the patient matches the predefined window of the cardiac cycle comprises predicting the predefined window of the cardiac cycle by detecting the predefined cardiac phase in real time and waiting for a certain period of time corresponding to the average interval between the predefined cardiac phase and the beginning of the predefined window of the cardiac cycle.

[0022] In some embodiments, the measurements from a plurality of cardiac cycles of the patient are waveform measurements of the plurality of cardiac cycles, and the method further comprises validating that the waveform measurements of the plurality of cardiac cycles are statistically stable.

[0023] In some embodiments, the predefined window of the cardiac cycle is a quiescent window of the cardiac cycle.

[0024] In some embodiments, a method for determining a biophysical model of a lung based on a plurality of x-ray measurements corresponding to different respiratory phases of a patient's lung is provided. The method includes extracting a plurality of displacement fields of lung tissue from the plurality of x-ray measurements corresponding to different respiratory phases of the lung. Each displacement field represents movement of lung tissue from a first respiratory phase to a second respiratory phase, and each respiratory phase has a corresponding set of biometric parameters. The method also includes calculating one or more biophysical parameters of the biophysical model of the lung using the plurality of displacement fields of lung tissue between different respiratory phases of the lung and the corresponding set of biometric parameters.

[0025] In some embodiments, the one or more biophysical parameters define a physical relationship between a biometric parameter associated with different breathing phases of the lung and a plurality of displacement fields of lung tissue.

[0026] In some embodiments, the physical relationship between the biometric parameters associated with different breathing phases of the lung and the multiple displacement fields of lung tissue is defined as follows:

[0027]

[0028] vector describes the normal stress caused by tidal volume, describes the normal stress caused by airflow, describes the shear stress caused by the airflow, and Describes the tissue movement caused by cardiac motion, and the displacement of the tissue at any point in the closed-loop trajectory Expressed as tidal volume (T v )、airflow(A f ) and cardiac stage (H c ) is the sum of scaled stress, strain, and perturbation cardiac motion vectors.

[0029] In some embodiments, the method further includes generating a plurality of medical image cubes corresponding to different respiratory stages of the lung based on a plurality of x-ray measurements corresponding to different respiratory stages of the lung. A plurality of displacement fields of the lung tissue are extracted from the plurality of medical image cubes corresponding to different respiratory stages of the lung by delineating the lung tissue from the rest of the first medical image cube using image segmentation. The method includes, for corresponding voxels in the first medical image cube, determining a displacement vector between voxels in the first medical image cube and the second medical image cube using an intensity-based structure mapping between the first medical image cube and the second medical image cube, and iteratively refining the displacement vectors of different voxels in the first medical image cube and their corresponding parts in the second medical image cube.

[0030] In some embodiments, the set of biometric parameters associated with a respective breathing stage comprises tidal volume and airflow of the lungs in the respective breathing stage and a cardiac stage corresponding to the respective breathing stage of the lungs.

[0031] In some embodiments, the method further includes generating a plurality of medical image cubes corresponding to different breathing stages of the lung based on a plurality of x-ray measurements corresponding to different breathing stages of the lung. In some embodiments, the method includes: selecting one or more of the plurality of medical image cubes as reference medical image cubes, determining a biometric parameter set associated with each reference medical image cube, and selecting a biometric parameter set based on biometric measurements of the lung between two biometric parameter sets associated with the two reference medical image cubes. In some embodiments, the method further includes simulating a medical image cube between the two reference medical image cubes by applying a biometric parameter set based on the biometric measurements of the lung to a biophysical model.

[0032] In some embodiments, the different respiratory phases of the lungs include early exhalation, late exhalation, maximal exhalation, early inspiration, late inspiration, and maximal inspiration of a complete respiratory cycle of the patient.

[0033] In some embodiments, one or more sensors are used to measure a patient's biometric signals as one or more sequences of time series, the one or more sensors including one or more of a 3D spatial position locator, a respiratory phase sensor, and a cardiac phase sensor.

[0034] In some embodiments, the 3D spatial position localizer is configured to measure the real-time body movements of the patient caused by breathing and heartbeat and output them as a time series.

[0035] In some embodiments, the breathing phase sensor is configured to measure one or more physiological metrics related to the patient's breathing, including tidal volume and its first time derivative.

[0036] In some embodiments, the heart phase sensor is configured to measure periodic and stationary electrical signals generated by the patient's heart.

[0037] In some embodiments, biometric signals of the patient measured by one or more sensors are used to trigger an x-ray unit to acquire x-ray images of the patient at specific respiratory and cardiac stages.

[0038] In some embodiments, the x-ray unit includes a clock, and the biometric signals of the patient measured by the one or more sensors are synchronized with the clock of the x-ray unit. In some embodiments, the corresponding values ​​of the biometric signals are recorded as being associated with the acquired x-ray images.

[0039] In some embodiments, the patient's biometric signals measured during the training window are used to construct an optimized breathing prediction model for predicting the desired breathing phase at which the x-ray unit is triggered to capture x-ray images of the patient.

[0040] In some embodiments, a method for generating a 3D x-ray image cube image from a 2D x-ray image of a patient is provided. The method includes converting a first plurality of x-ray images of a lung captured at different projection angles into a second plurality of x-ray images of the lung corresponding to different respiratory stages. The method also includes generating a static image cube from each of the second plurality of x-ray images of the corresponding respiratory stage using back projection, and combining the static image cubes corresponding to the different respiratory stages of the lung into a 3D x-ray image cube image by temporal interpolation.

[0041] In some embodiments, converting the first plurality of sets of x-ray images of the lung captured at different projection angles into the second plurality of sets of x-ray images of the lung corresponding to different respiratory stages further comprises capturing the first plurality of sets of x-ray images of the lung at different projection angles. Each set of the first plurality of sets of x-ray images corresponds to a different respiratory stage of the lung at a particular projection angle. The converting further comprises reorganizing the first plurality of sets of x-ray images of the lung into the second plurality of sets of x-ray images of the lung by the associated respiratory stage of the first plurality of sets of x-ray images of the lung. Each set of the second plurality of sets of x-ray images corresponds to a respective respiratory stage of the lung.

[0042] In some embodiments, the x-ray images within any particular group are geometrically resolved and temporally independent.

[0043] In some embodiments, different breathing phases of the lung correspond to different tidal volume percentiles of movement of the lung.

[0044] In some embodiments, the different respiratory phases of the lungs include early exhalation, late exhalation, maximal exhalation, early inspiration, late inspiration, and maximal inspiration of a complete respiratory cycle of the patient.

[0045] In some embodiments, multiple x-ray images of the lungs captured at different projection angles all correspond to the same respiratory phase.

[0046] In some embodiments, one or more sensors are used to measure a patient's biometric signals as one or more sequences of time series, the one or more sensors including one or more of a 3D spatial position locator, a respiratory phase sensor, and a cardiac phase sensor.

[0047] In some embodiments, the method further comprises identifying a cardiac stage gating window using one or more cardiac stage sensor measurements, predicting a respiratory stage using one or more respiratory stage sensor measurements, identifying consistency between the cardiac stage gating window and the predicted respiratory stage for generating x-ray imaging pulses, and marking an x-ray image corresponding to the x-ray imaging pulse using the respiratory stage, cardiac stage, and 3D spatial orientation locator measurements.

[0048] In some embodiments, the 3D spatial position localizer is configured to measure the real-time body movements of the patient caused by breathing and heartbeat and output them as a time series.

[0049] In some embodiments, the breathing phase sensor is configured to measure one or more physiological metrics related to the patient's breathing, including tidal volume and its first time derivative.

[0050] In some embodiments, the heart phase sensor is configured to measure periodic and stationary electrical signals generated by the patient's heart having characteristic features corresponding to the heart phase.

[0051] In some embodiments, two different filters are used to remove signal drift and noise from the patient's biometric signal after synchronization with the x-ray unit's clock.

[0052] In some embodiments, biometric signals of the patient measured by one or more sensors are used to trigger an x-ray unit to acquire x-ray images of the patient at specific respiratory and cardiac stages.

[0053] In some embodiments, the x-ray unit includes a clock, and the biometric signals of the patient measured by the one or more sensors are synchronized with the clock of the x-ray unit. In some embodiments, the corresponding values ​​of the biometric signals are recorded as being associated with the acquired x-ray images.

[0054] In some embodiments, a biometric signal of the patient is measured during a training window prior to capturing any x-ray images of the patient, and the biometric signal of the patient measured during the training window includes multiple complete breathing cycles of the patient.

[0055] In some embodiments, a biometric signal of the patient measured during a training window is used to identify a plurality of tidal volume percentiles within a complete respiratory cycle, each tidal volume percentile corresponding to one of the different respiratory phases.

[0056] In some embodiments, the patient's biometric signals measured during the training window are used to construct an optimized breathing prediction model for predicting the desired breathing phase at which the x-ray unit is triggered to capture x-ray images of the patient.

[0057] In some embodiments, the optimized breathing prediction model is based on an autoregressive integrated moving average (ARIMA) model.

[0058] In some embodiments, the desired respiratory phase for capturing x-ray images of the patient is configured to coincide with a cardiac gating window during which heart-induced lung motion varies slowly.

[0059] In some embodiments, the cardiac gating window is selected based on the location of T and P waves in an electrocardiogram (ECG) signal such that the heart-induced lung motion varies slowly.

[0060] In some embodiments, different breathing phases of the lungs at a specific projection angle are acquired from at least two breathing cycles.

[0061] Additional aspects and advantages of embodiments of the present disclosure will be set forth in part in the following description and in part will be obvious from the following description, or may be learned from practice of embodiments of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly describe the embodiments of the present disclosure or the technical solutions in the prior art, the drawings necessary for the description of the embodiments or the prior art will be briefly explained. Obviously, the drawings in the following description are only some embodiments of the present disclosure. For those skilled in the art, other drawings can also be obtained according to the structures shown in these drawings without creative work.

[0063] Figure 1 is a schematic block diagram of a GREX imaging system including a hardware box, acquisition software, and post-processing software according to some embodiments of the present disclosure.

[0064] Figure 2 is a schematic flowchart of a GREX image acquisition process according to some embodiments of the present disclosure.

[0065] Figure 3 is a schematic block diagram depicting a top view of a 3D spatial position locator of a GREX imaging system according to some embodiments of the present disclosure.

[0066] Figure 4 Depicted is an example of two breaths that share the same maximum inspiratory phase in time but have completely different tidal volumes, in accordance with some embodiments of the present disclosure.

[0067] Figure 5Depicted are examples of tidal volume percentiles (two bottom sub-plots) for regular breathing patients (upper left sub-plot) and irregular breathing patients (upper right sub-plot), respectively, according to some embodiments of the present disclosure.

[0068] Figure 6 is a schematic flowchart of a synchronization process according to some embodiments of the present disclosure.

[0069] Figure 7 Depicted are synchronized heart electrocardiogram (ECG) signals and lung respiration signals according to some embodiments of the present disclosure.

[0070] Figure 8 is a schematic flow chart of a drift and signal noise removal process according to some embodiments of the present disclosure.

[0071] Fig. 9 Depicted are exemplary breaths plotted as tidal volume versus time in accordance with some embodiments of the present disclosure.

[0072] Fig.10 Depicting a plot of airflow versus tidal volume according to some embodiments of the present disclosure Fig. 9 The same breath shown in .

[0073] Fig.11 is a schematic flow chart of a respiratory phase prediction process according to some embodiments of the present disclosure.

[0074] Fig.12 Depicted is an exemplary gating window in time selected based on the location of T and P waves in an ECG signal to allow for slow variations in cardiac induced lung motion, in accordance with some embodiments of the present disclosure.

[0075] Fig.13 is a schematic flow chart of another GREX procedure using cardiac stage prediction rather than real-time identification of cardiac gating windows according to some embodiments of the present disclosure.

[0076] Fig.14 Depicted from left to right are examples of triggering images at early exhalation, late exhalation, maximal exhalation, early inspiration, late inspiration, maximal inspiration of the breathing phases according to some embodiments of the present disclosure.

[0077] Fig.15 is a schematic flowchart of an image acquisition triggering process according to some embodiments of the present disclosure.

[0078] Fig.16 is a schematic block diagram of variables used in an exemplary GREX image reconstruction algorithm according to some embodiments of the present disclosure.

[0079] Fig.17is a schematic block diagram of operating components of a GREX imaging system during an imaging procedure according to some embodiments of the present disclosure.

[0080] Fig.18 is an exemplary depiction of tissue depth traversed by 0° (A) and 90° (B) x-ray projections according to some embodiments of the present disclosure.

[0081] Fig.19 is an exemplary depiction of depth resolution according to a narrow projection angle range (1) and a wide projection angle range (2) according to some embodiments of the present disclosure.

[0082] Fig. 20 is an exemplary illustration of the beta imaging angle orientation relative to three principal imaging planes (left) and relative to the imaging isocenter of a GREX imaging system in a top-down viewing direction (right) according to some embodiments of the present disclosure.

[0083] Fig.21 is an exemplary block diagram showing how a GREX imaging system according to some embodiments of the present disclosure transfers 2D x-ray projections to a static image cube and ultimately to an image cube image.

[0084] Fig. 22 is created according to some embodiments of the present disclosure Fig.21 A graphical depiction of the process is shown in the static image cube.

[0085] Fig.23 is a schematic flowchart of an image filtering process according to some embodiments of the present disclosure.

[0086] Fig.24 Depicted is an exemplary closed-loop lung tissue trajectory of a patch of tissue located in the left lung near the heart resulting from heart-lung interaction in accordance with some embodiments of the present disclosure.

[0087] Fig.25 is a schematic flow chart of the operation of components of a biomechanical model according to some embodiments of the present disclosure.

[0088] Fig.26A is a graphical depiction of the motion trajectory of lung tissue elements during a respiratory cycle according to some embodiments of the present disclosure.

[0089] Fig.26B is a graphical depiction of displacement vectors between different breathing phases according to some embodiments of the present disclosure.

[0090] Fig. 27 is a schematic flowchart of a multi-resolution 3D optical flow algorithm according to some embodiments of the present disclosure.

[0091] Fig.28Ais a schematic flow chart of generating images by biometric interpolation according to some embodiments of the present disclosure.

[0092] Fig.28B is a block diagram illustrating the creation of an intermediate image cube using a biometric data matrix according to some embodiments of the present disclosure.

[0093] Fig.29 Depicting examples of a standard radiograph of a healthy patient and a standard radiograph of a diseased patient with a stage 1b left upper lung tumor (indicated by arrow).

[0094] Fig.30 Depicting examples of GREX parameter graphs showing indicators of a healthy patient's condition, and GREX parameter graphs showing indicators of a health condition of a diseased patient, in accordance with some embodiments of the present disclosure.

[0095] Figures 31A to 31B is a flow chart of a method of imaging a patient's lungs according to some embodiments of the present disclosure.

[0096] Figures 32A to 32B is a flow chart of a method of gating a radiation source according to some embodiments of the present disclosure.

[0097] Figures 33A to 33C is a flow chart of a method of determining a biophysical model of a patient's lungs according to some embodiments of the present disclosure.

[0098] Figures 34A to 34C is a flow chart of a method for generating a 3D x-ray image cube image according to some embodiments of the present disclosure.

[0099] Fig.35 An exemplary patient positioning fixture (PPF) (eg, a swivel chair) for supporting a patient is depicted in accordance with some embodiments.

[0100] Fig.36 Depicted is an exemplary biological event monitoring process (BEMP) card in accordance with some embodiments. DETAILED DESCRIPTION

[0101] Reference will now be made in detail to the embodiments of the present disclosure. Throughout the specification, identical or similar elements and elements having identical or similar functions are denoted by identical reference numerals. The embodiments described herein with reference to the accompanying drawings are explanatory, illustrative, and are used to generally understand the present disclosure. The embodiments should not be construed as limiting the present disclosure.

[0102] In the specification, unless otherwise specified or limited, relative terms such as "center", "longitudinal", "lateral", "front", "back", "right", "left", "inside", "outside", "lower", "upper", "horizontal", "vertical", "above", "below", "upper", "top", "bottom" and their derivatives (e.g., "horizontally", "downward", "upward", etc.) should be interpreted as referring to directions described subsequently or as shown in the drawings discussed. These relative terms are for convenience of description and do not require that the present disclosure be constructed or operated in a specific orientation.

[0103] In the present invention, unless otherwise specified or limited, the terms "install", "connect", "couple", "fix" and the like are used in a broad sense, and may be, for example, a fixed connection, a detachable connection or an integral connection; it may also be a mechanical or electrical connection; it may also be a direct connection or an indirect connection through an insertion structure; it may also be an internal connection between two elements, which can be understood by technicians in the relevant field according to the specific circumstances.

[0104] In the present invention, unless otherwise specified or limited, a structure in which a first feature is "on" or "under" a second feature may include an embodiment in which the first feature is in direct contact with the second feature, and may also include an embodiment in which the first feature and the second feature are not in direct contact with each other but are in contact through an additional feature formed therebetween. In addition, a first feature being "on", "above" or "on top" a second feature may include an embodiment in which the first feature is exactly or obliquely "on", "above" or "on top" the second feature, or simply means that the height of the first feature is higher than the height of the second feature; and a first feature being "under", "below" or "bottom" a second feature may include an embodiment in which the first feature is exactly or obliquely "under", "below" or "bottom" the second feature, or simply means that the height of the first feature is lower than the height of the second feature.

[0105] Figure 1 is a schematic block diagram of a GREX imaging system 100 including a hardware box 102 , acquisition software 104 (eg, stored in non-transitory memory), and post-processing software 106 (eg, stored in non-transitory memory), according to some embodiments of the present disclosure.

[0106] There are three measurement sensors (e.g., 3D spatial position locator 300, respiratory phase sensor 110, and cardiac phase sensor 112) in hardware box 102, which independently collect the patient's biometric signals as time series, and the collected time series serve as input to acquisition software 104. Acquisition software 104 processes and filters the biometric time series measurements to generate imaging trigger signals (e.g., to x-ray unit 108). The imaging trigger signals are targeted to a specific respiratory phase, and optionally to the patient's cardiac phase. The respiratory phase and cardiac phase are each defined by their corresponding biometric time series measurements. A connector cable transmits the imaging trigger signals from acquisition software 104 to x-ray unit 108, which acquires respiratory and cardiac phase radiographic images at the targeted phase. Once acquired, a series of images for respiratory and cardiac phases defining a complete respiratory cycle are input into post-processing software 106. The post-processing software constructs a biomechanical model of lung motion based on the images for respiratory and cardiac phases. The biomechanical model is then used to generate other diagnostic results in the post-processing software. Although the present application uses x-ray images as an example, it is obvious to those skilled in the art that the methods disclosed in the present application can be applied to other types of medical images without difficulty. For example, the process of constructing a biomechanical model is not limited to the use of x-ray images, and other types of medical images (e.g., CT scans, MRI, etc.) can be used.

[0107] Figure 2 is a schematic flow chart of a GREX image acquisition process 200 according to some embodiments of the present disclosure (eg, Figure 1 104 executes the acquisition software 104). Specifically, the acquisition software 104 synchronizes and processes the time series measurements from the hardware box 102 to remove potential signal drift and noise. Next, the acquisition software 104 implements a stage prediction algorithm that predicts the respiratory and cardiac stages based on the synchronized drift-free and noise-free time series input. Based on the respiratory and cardiac stage prediction results, the acquisition software 104 uses a logic algorithm 118 to search for targeted respiratory and cardiac stage consistency. The targeted respiratory and cardiac stage consistency defines the conditions that prompt the image trigger generation and the acquisition of the resulting image by the x-ray unit 108. Once the GREX image is acquired, the post-processing software uses the GREX image that was previously unavailable to medical professionals to perform new lung disease diagnoses. For example, the acquired GREX image is used to construct a biomechanical model that defines the thoracic geometry as varying with respiratory and cardiac stages, but does not explicitly include time parameters.

[0108] In this document, the GREX-based imaging system is divided into three sections. Section 1 describes an embodiment of the hardware box 102. Section 2 describes an embodiment of the acquisition software 104. Section 3 describes an embodiment of the post-processing software 106. Each section describes in more detail the components of the hardware box 102. Figure 1 Components and functions of the hardware box 102, acquisition software 104 and post-processing software 106 shown in FIG.

[0109] Chapter 1. Hardware Box 102

[0110] In some embodiments, the hardware 102 has at least two responsibilities: one is to acquire biometric signals that define the anatomical geometry of the thorax, and the other is to communicate with a digital diagnostic x-ray unit (eg, x-ray unit 108).

[0111] The biometric signal inputs defining the anatomical geometry of the thorax include: thorax dimensions (measured by the 3D spatial localizer 300, Figure 3 ), respiratory phase (measured by respiratory phase sensor 110), and cardiac phase (measured by ECG monitor). Within hardware box 102, the biometric signals are sampled in real time at 100 Hz to create a time series curve for each signal input. The output of hardware box 102, i.e., the time series curve, is then passed to the acquisition software.

[0112] Figure 3 is a schematic block diagram depicting a top view of a 3D spatial position locator 300 of a GREX imaging system 100 according to some embodiments of the present disclosure.

[0113] Chapter 1.1-3D spatial orientation locator 300

[0114] The 3D spatial position locator 300 measures the real-time body movements of the patient caused by breathing and heartbeat, and outputs them as a time series in a (e.g., Cartesian, polar, superordinate, etc.) coordinate space. Figure 3 As shown in , the 3D spatial position locator 300 includes three separate 3D cameras (e.g., cameras 302a to 302c) fixed to a gimbal on the ceiling of a room. There are 120° angular increments separating the three cameras 302a of the system from each other. The patient is located between the x-ray unit 108 and the detector panel 304 so that the patient is located at the center of the ceiling-mounted 3D spatial position locator 300. The system measures the chest expansion (e.g., rise and fall) of the patient in a predefined coordinate space. Each camera 302 has an unobstructed view of the patient's torso and one of the x-ray unit 108 and the detector panel 304.

[0115] Using the real-time depth map, each 3D camera 302 creates a surface reconstruction of the patient. Simultaneous information from all three 3D cameras 302 is combined to form a volumetric skin surface position measurement that changes in real time based on the patient's respiratory and cardiac phase (e.g., using a ray casting technique). The 3D spatial position locator 300 uses the volumetric skin surface position measurement in at least two ways: (i) to define the patient's spatial boundaries and (ii) to determine the patient's tissue position. Based on assumptions such as skin thickness, rib thickness, muscle thickness, and bone position derived from the normative medical internal radiation dose (MIRD) anatomical database, the 3D spatial position locator 300 roughly estimates the real-time spatial position of the lungs in the patient. The 3D spatial position locator 300 uses the MIRD data to calculate the spatial boundary conditions of the lungs, which are then made available to the post-processing software 106. For example, the 3D spatial position locator 300's estimate of the spatial boundary conditions of the lungs produces an initial chest geometry, which the post-processing software 106 uses to simulate the cumulative tissue density along the rays originating from the x-ray unit 108.

[0116] Section 1.2 - Breathing Phase Sensor 110

[0117] The respiratory phase sensor 110 in the hardware box 102 measures key physiological metrics related to breathing, namely tidal volume and its first time derivative (e.g., the rate of change of tidal volume over time or airflow). There are two methods to measure tidal volume: direct tidal volume measurement and indirect tidal volume measurement. Direct tidal volume measurement is performed using a mouth spirometer, which consists of a turbine in a tube that rotates at a rate proportional to the volume of air inhaled or exhaled by the patient. Indirect tidal volume measurement is performed using an abdominal belt (or any other geometric measurement of the patient's chest, as described herein), which measures the change in the patient's abdominal circumference during breathing (e.g., the change in the patient's abdominal circumference during breathing). Figure 1 3D spatial position locator 300).

[0118] As used herein, the term tidal volume refers to the difference between the current lung volume and a predefined benchmark (e.g., the volume during maximum exhalation of normal breathing without additional effort, the volume during maximum inspiration of normal breathing without additional effort, or any other suitable benchmark volume). Based on the ideal gas law, the air density difference between room air and body air causes the lung air to increase the tidal volume by 11%. In order to maintain mass, the body increases in volume by 11% over the amount of air inhaled. Therefore, the tidal volume of the lungs can be calculated by using external measurements of the body calibrated to the body's air content. In addition, the 3D spatial orientation locator 300 provides an auxiliary check of the accuracy of air content and tidal volume measurements by identifying the patient's volume expansion during breathing. The volume expansion of the body is compared with estimates of the amount of air in the trachea, lungs, and bronchi from x-ray images.

[0119] Section 1.3 - Heart Phase Sensor 112

[0120] like Figure 1 As shown, the heart phase is measured using an electrocardiogram (ECG) monitor or a blood volume pressure device (e.g., via a heart phase sensor 112). For example, when using an ECG monitor, the clinician places a lead on each arm of the patient and a ground lead on the lower left side of the patient's abdomen (away from the diaphragm and abdominal band). The human heart generates a periodic and stationary electrical signal that has characteristic characteristics corresponding to the phase of the heartbeat. A stationary signal is a signal that is a random process and whose joint probability distribution does not change over time. The blood volume pressure device uses a light source and a light sensor to measure the light attenuation in the patient's finger. The circulating blood volume driven by the heartbeat causes the light attenuation magnitude in the patient's finger to vary. The light attenuation magnitude is proportional to the heart phase.

[0121] Typically, a digital diagnostic x-ray unit (e.g., x-ray unit 108) is turned on by an analog plunger attached to a plug port. The plug port is uniquely configured to accept a plunger with a specific pin configuration. In each pin configuration, there is a "get pin" that accepts a voltage signal that turns on (e.g., gates) the x-ray unit. When the end user presses the plunger, the plunger sends a voltage pulse to the digital diagnostic x-ray unit, which activates an imaging beam that passes through the patient's body.

[0122] like Figure 1As shown, the hardware box 102 uses a connector cable to communicate the same voltage pulse (e.g., gating signal) from the acquisition software to the digital diagnostic x-ray unit 108. The voltage pulse signal turns on the digital diagnostic x-ray beam (e.g., gated x-ray beam) when the voltage pulse signal exceeds a predefined voltage threshold, and turns off the x-ray unit 108 when the voltage pulse signal on the pin becomes less than the predefined voltage threshold. In some embodiments, the hardware box 102 generates a square wave signal with a pulse height greater than the predetermined voltage threshold and maintains the pulse height for the duration of the image exposure. That is, a rectangular pulse with a voltage Y lasting X seconds is generated for the x-ray unit 108 to take an x-ray image of the patient. Subthreshold voltages less than Y will not trigger the scanner, so that the value of Y must be greater than the predefined voltage threshold to initiate an x-ray image. The pulse duration X is the amount of exposure time of the image, starting at the moment when the voltage Y exceeds the predefined voltage threshold and ending when the voltage drops below the predetermined voltage threshold. The voltage dropping below the predefined voltage threshold turns off the x-ray unit 108. The pulse duration is defined by the manufacturer's specifications, but is typically on the order of a few milliseconds.

[0123] Chapter 2. Obtaining Software 104

[0124] The acquisition software 104 is designed to acquire the measured spatial, cardiac, and pulmonary time series of the hardware box 102 and determine when to trigger the x-ray unit 108 to acquire x-ray images at specific respiratory and cardiac phases. The acquisition software 104 then accurately overlays (e.g., synchronizes) the measured cardiac and pulmonary phases, processes (e.g., filters) the acquired biometric time series, identifies the appropriate imaging time, and generates an electronic trigger signal for the digital diagnostic x-ray unit 108. The electronic trigger signal (e.g., a gate signal) turns on the x-ray unit 108, thereby acquiring a snapshot image of the chest geometry. The spatial, cardiac, and pulmonary values ​​associated with the snapshot are recorded to define the chest surface geometry when the image was taken. The entire process is automated and independent of user input. Figure 1 and 2 As shown, the process implemented by the acquisition software 104 includes four sub-sections: Section 2.1 (synchronization module 114), Section 2.2 (signal processing module 116), Section 2.3 (logic algorithm 118), and Section 2.4 (trigger generation module 120).

[0125] Chapter 2.1 - Synchronization Module 114

[0126] The input of the synchronization module 114 is synchronized with the clock of the x-ray unit 108 by the synchronization module 114, and the input includes the signal of the 3D spatial position locator 300, the signal of the respiratory phase sensor 110 and the signal of the cardiac phase sensor 112. It should be noted that the physiological biometric signals are collected in an asynchronous manner and therefore need to be synchronized. One source of asynchrony is that the respiratory cycle is slower than the cardiac cycle and is completely independent of the cardiac cycle. As previously explained, the respiratory cycle and the cardiac cycle are measured using different sensors, respectively. The synchronization module 114 is configured to synchronize the respiratory and cardiac phase sensors with the acquired image. When the image is taken, the image will show the anatomical geometry of the chest at a certain moment. The moment is recorded using the local timing system of the x-ray unit 108, and the local timing system does not have to be synchronized with the respiratory and cardiac sensor biometric time series.

[0127] It should be noted that time alone does not distinguish between periods of irregular breathing and periods of regular breathing. In other words, if time is the only defining dimension of breathing phase, images taken during normal breathing and images taken during abnormal breathing (such as coughing) are computationally indistinguishable from each other. Figure 4 Depicted is an example of two breaths 400 (e.g., 400a and 400b) from the same individual that are temporally identical inspiratory phases but have completely different tidal volumes, according to some embodiments of the present disclosure. When stacked on top of each other, the two breaths 400 are actually distinct because they have different tidal volume magnitudes despite having similar maximal inspiratory phases. In some embodiments, tidal volume refers to a capacity value (e.g., in ml) measured relative to a reference volume. For example, a reference volume represents the minimum volume of the lungs (e.g., at maximum exhalation) during normal breathing of a patient (e.g., without extra effort or forceful exhalation). In some embodiments, the reference volume is different for each patient. In some embodiments, the reference volume is represented as 0. In some embodiments, tidal volume is measured at one point in time. Figure 4 A representation of the changes in tidal volume over a period of time is shown in .

[0128] To overcome the problem of the time dimension, the GREX imaging system 100 defines the respiratory phase by physiological values ​​obtained from various physiological sensors in the hardware box 102, which is a more informative respiratory phase dimension than time. The synchronization module 114 is mainly used to allow a seamless transition between the x-ray unit 108 and the GREX imaging system 100. In some embodiments, the acquisition software 104 uses a 30-second long training window, which is discussed in detail in Section 2.2 below, during which the collected tidal volume time series observations are used to calculate tidal volume percentiles. The acquisition software 104 uses tidal volume percentiles to define the respiratory phase rather than the peak-to-peak time interval of the periodic cosine wave. Due to the tidal volume variation between breaths, the tidal volume percentiles of the acquisition software 104 are a more informative method of defining the geometry of the lungs than the peak-to-peak periodic cosine curve.

[0129] Figure 5 Depicts examples of tidal volume percentiles (two bottom sub-plots) for regular breathing patients (upper left sub-plot) and irregular breathing patients (upper right sub-plot), respectively, according to some embodiments of the present disclosure. To quantitatively assess the tidal volume histogram for irregular breathing, the ratio between normal inspiratory tidal volume and extreme inspiratory tidal volume is used as a measure of the breathing phase. The ratio has a threshold that defines the probability that the patient is breathing irregularly. Figure 5 As shown in the two bottom subplots of , the vertical lines show where the 85th, 90th, 95th, and 98th percentile tidal volumes are located in the tidal volume histogram. Figure 5 Compared with the lower right sub-graph of Figure 5 ), the more normal tidal volume percentiles (85th and 90th) are located closer to the extreme tidal volume percentiles (95th and 98th).

[0130] Figure 66 is a schematic flow chart of a synchronization process 600 between different sensors in the hardware box 102 and the diagnostic x-ray unit 108 according to some embodiments of the present disclosure. It should be noted that the hardware box 102 continuously acquires the coordinates of the respiratory phase defined by the tidal volume percentile, the cardiac phase defined by the ECG, and the thoracic geometry defined by the 3D spatial position locator 300. Before synchronization with the x-ray unit 108, the sensor signals supporting these measurements need to be synchronized with each other. To this end, the measurement channel differences and cable resistivity differences between the different sensors are also corrected by matching the trace lengths using impedance matching in the digital / analog converter. Because the clock (e.g., timing system) of the x-ray unit 108 is usually not synchronized with the clock of the hardware box 102, the connector cable is configured to connect to the data acquisition board and interface it with the clock of the x-ray unit 108. The clock of the x-ray unit 108 passes through the distribution buffer and performs trace matching for each channel in the analog / digital converter. The converted digital signal is passed to the field programmable gate array, where synchronization of all signals is ensured. When the acquisition software 104 properly synchronizes the respiratory and cardiac phase sensor signals, the result should be similar to Figure 7 The example shown in .

[0131] In order to avoid unnecessary radiation exposure to the patient, the acquisition software 104 does not send any trigger signal to activate the x-ray unit 108 through the connector cable without proper synchronization between the different components of the GREX imaging system 100. In some embodiments, the 30-second training window is used to verify the synchronization between the clock of the hardware box 102 and the clock of the x-ray unit 108. Therefore, the 30-second training window should contain samples worth 30 seconds. If each sensor in the hardware box 102 and the clock of the x-ray unit 108 do not show samples worth exactly 30 seconds, synchronization has not occurred. For this reason, if this above-mentioned check procedure contains inconsistencies, the synchronization system will restart to correct the inconsistencies. It should be noted that the 30-second training window is for illustrative purposes, and those skilled in the art should understand that the length of the training window will vary as long as there is enough data to perform the synchronization process.

[0132] Chapter 2.2 - Signal Processing Module 116

[0133] After the spatial orientation, respiratory phase, and cardiac phase signals are synchronized, the acquisition software 104 processes the sensor signals to remove noise and ultimately predict the patient's accurate tidal volume. The noise in the measured lung and cardiac time series comes from the sensor electronics, electrodes, and from background electrical signals. A set of filters specifically designed to maintain time accuracy removes noise from the measured lung and cardiac time series so that the biometric time series maintains time accuracy after being filtered by the set of filters.

[0134] In some embodiments, two different filters (e.g., wavelet filters) are used to remove signal drift and noise from the biometric time series. Signal drift can skew measurements taken over time, making measurements taken at the beginning of data collection inconsistent with measurements taken at the end of data collection. Signal noise is not physiological in nature and can cause serious problems when calculating a patient's airflow from tidal volume measurements. Figure 8 is a schematic flow chart of a drift and signal noise removal process 800 according to some embodiments of the present disclosure.

[0135] The acquisition software 104 requires a smooth tidal volume time series to calculate the first time derivative of tidal volume, such as airflow. If the tidal volume time series is not smooth, the first time derivative of tidal volume will not produce a smooth curve; instead, the curve will contain discontinuities that violate the biophysical reality of breathing. Fig. 9 Depicted are exemplary breaths plotted as tidal volume versus time in accordance with some embodiments of the present disclosure. Fig.10 Depicting a plot of airflow versus tidal volume according to some embodiments of the present disclosure Fig. 9 The same breathing shown in , which is a continuous closed loop used by the post-processing software 106 for biomechanical modeling (described in Section 3.1).

[0136] The acquisition software 104 uses the filtered and time-accurate time series curves to perform two different functions. The first function of the acquisition software 104 is to generate a breathing prediction using a short prediction horizon. Fig.11 1 is a schematic flow diagram of a respiratory phase prediction process 1100 according to some embodiments of the present disclosure. The short prediction horizon is the "look ahead time" of the prediction algorithm. The prediction algorithm predicts future moments when a desired tidal volume and airflow (e.g., respiratory phase) will occur. The "desired" respiratory phase is targeted so that the diagnostic x-ray unit 108 can be triggered to acquire the desired chest geometry.

[0137] A short prediction horizon also reduces another source of inaccuracy in the breathing prediction, such as breath-to-breath variations in breathing amplitude and breathing period. As the limits of the prediction horizon approach zero, the variation in lung geometry approaches zero (e.g., the lung geometry is assumed to be nearly constant). In other words, within a short prediction horizon, the lung geometry is unlikely to change significantly. Thus, a short prediction horizon reduces the impact of human breathing variations on the prediction accuracy of the respiratory motion model.

[0138] The time-accurate filtered tidal volume time series is used as the input value of the respiratory prediction algorithm. The respiratory prediction algorithm provides a fast, real-time and accurate forecast of the respiratory stage. For example, the respiratory prediction algorithm is based on autoregressive integrated moving average (ARIMA). ARIMA is suitable for respiratory prediction because the ARIMA model does not assume that the input value is stable and is composed of polynomials. The polynomial coefficients of the ARIMA model are estimated during the 30-second training window acquired at the beginning of the imaging study. The number of polynomial coefficients of the ARIMA model, such as the model order, is checked by nonlinear optimization, which attempts to minimize the information criterion search function to reduce or eliminate overfitting. If the model order is optimal for the collected training data, a tidal volume histogram is constructed (discussed in Section 2.1) and a probability density function is calculated using the log-likelihood objective function. The tidal volume distribution is used to check irregular breathing, as discussed in Section 2.1. If irregular breathing is detected, the training data is discarded and re-acquired. If irregular breathing is not detected, the ARIMA model coefficients are estimated by the maximum likelihood method using the training data and the probability density function. The 30-second training window can also be used as a device check prior to imaging. Fig.11 A flow chart for predicting breathing phase is shown.

[0139] The second function of the acquisition software 104 is to identify the cardiac phase so that the heart is in the same phase in each desired chest geometry. The duration of the prediction horizon is a key parameter of the acquisition software 104 in its efforts to accurately predict human breathing, because human breathing is a quasi-random function (because each breath has some unique aspects of its own). In some embodiments, the duration of the prediction horizon is longer than the sum of the delay time of the digital diagnostic x-ray unit 108 and the exposure time of the x-ray imaging. The sum of the delay time of the digital diagnostic x-ray unit 108 and the exposure time of the x-ray imaging is extremely short, on the order of 10 milliseconds. Therefore, the duration of the prediction horizon is also short (on the order of 1 to 2 sensor measurement samples at an operating frequency of 100 to 1000 Hz).

[0140] When the acquisition software 104 looks for coincidence between cardiac and respiratory phases, it is unlikely that the phases represented by a single point in each time series will align. Therefore, imaging methods that look for single point coincidences take longer to complete because the acquisition software 104 must wait for a low probability coincidence to occur. In contrast, cardiac gating windows extend the size of the coincidence window, making imaging studies less time consuming.

[0141] To further reduce computation time in the signal processing software, the signal processing software does not predict the cardiac phase. Instead, the signal processing software targets a specific gating window where the heart does not rapidly displace the lungs. Fig.12Depicts an exemplary gating window in time selected based on the position of the T and P waves in the ECG signal to allow for slow changes in heart-induced lung motion according to some embodiments of the present disclosure. The dashed line is the lung motion induced by the heartbeat. The rate of change of lung motion (e.g., velocity) is the slope of the dashed line. When the slope of the dashed line is small, the rate of change is also small, making the corresponding cardiac phase an ideal gating window. Fig.12 It is shown that a gating window occurs between the T wave and the P wave (consistently), favoring the P wave. In the gating window, the lung motion caused by the heartbeat is minimal.

[0142] The previous paragraph discussed how to identify an ideal cardiac gating window that will minimize the physical impact of the heart on the lungs while still maintaining a window of opportunity for the targeted respiratory phase to coincide with the desired cardiac phase. In some other embodiments, the GREX imaging system 100 predicts rather than gates the cardiac phase based on signal processing differences that distinguish cardiac phase from respiratory phase (e.g., cardiac phase is periodic and stable). Because cardiac phase is periodic and stable, an unsupervised multilayer perceptron using a back-propagation method can be used to predict the next heartbeat based on pattern extraction rather than a time series prediction process.

[0143] Fig.13 is a schematic flow chart of another GREX process 1300 using real-time identification of cardiac stage prediction rather than cardiac gating windows according to some embodiments of the present disclosure. In this case, twenty seconds (20 to 22 heartbeats) of the training window are used to train the algorithm, while the remaining ten seconds (10 to 11 heartbeats) are used to validate the multi-layer perceptron node weights. The node weights are iteratively determined by gradient descent optimization until the model error in the training set is minimized. If the multi-layer perceptron provides poor prediction of cardiac stage in the validation data, the trained model is applied to the 10-second validation data and the node weights are recalculated.

[0144] Chapter 2.3 - Logical Algorithms 118

[0145] There are known medical algorithms and systems for identifying the T and P waves of the ECG within the cardiac cycle. Because the cardiac phase sensor 112 continuously measures the cardiac cycle, the time interval between the T wave and the subsequent P wave (which is equal to a constant fraction of the cardiac cycle and is therefore proportional to the heart rate) is also known. ECG features within the time interval can be used by the logic algorithm 118 to introduce a short lag before initiating a gating window, so that the gating window can start, for example, halfway between the T wave and the P wave, and close after the logic algorithm 118 identifies the P wave.

[0146] Fig.14An example of triggered x-ray image capture windows corresponding to different respiratory phases of a respiratory cycle according to some embodiments of the present disclosure is depicted, from left to right, early exhalation, late exhalation, maximum exhalation, early inspiration, late inspiration, maximum inspiration of the respiratory cycle. In this example, the acquisition software 104 identifies a minimum of six respiratory phases representing a single respiratory cycle. During a training period, the acquisition software 104 generates a sample distribution, and the logic algorithm 118 calculates the tidal volume percentile for the respiratory phase that will define the logic algorithm 118. Fig.14 The arrows shown in represent six cardiac gating windows that coincide with the targeted respiratory phase for which the logic algorithm 118 will generate an imaging trigger signal. Once the acquisition software 104 acquires the respiratory phase, it creates an automated check to prevent redundant imaging of the same respiratory phase in the future.

[0147] Fig.15 is a schematic flow chart of an image acquisition triggering process 1500 according to some embodiments of the present disclosure. The cardiac phase sensor measurements are used to identify the cardiac gating window, as previously described in conjunction with Fig.12 The measured value of the breathing phase sensor 110 is used to predict the breathing phase, as previously described in conjunction with Fig.11 As described. The logic algorithm 118 identifies the consistency between the cardiac phase gating window and the predicted respiratory phase. When a consistency is found, the respiratory phase list is checked to determine whether the respiratory phase has been acquired. If the respiratory phase has been previously acquired, an imaging trigger pulse is not generated. If the respiratory phase has not been previously acquired, an imaging trigger pulse is generated to take a snapshot of the patient's anatomy. The respiratory phase, cardiac phase, and 3D spatial orientation locator 300 measurements are recorded and labeled with images. If all respiratory phases have been acquired, the respiratory phase list is updated to prevent redundant images from being taken.

[0148] Section 2.4 - Trigger Generation (Gating)

[0149] The x-ray unit 108 has a port with a series of electrical pins. One of those pins accepts an electrical pulse that defines the timing and duration of radiation exposure. Based on the respiratory phase identified within the cardiac gating window by the logic algorithm 118, the trigger generator generates a square wave trigger event as an electrical pulse. An optical cable with a vendor-specific plug attachment carries the generated trigger signal to the x-ray unit 108.

[0150] Chapter 3. Post-processing software

[0151] Biometrically informed imaging trigger events recognized by the hardware box 102 (Section 1) and acquisition software 104 (Section 2) provide better quality inputs to the image reconstruction algorithms (and remove poor quality inputs). Specifically, the quality enhancement stems from the fact that image reconstruction and image post-processing techniques are enhanced by the process of acquiring biometrically targeted images during normal respiration. The act of biometrically targeting images during normal respiration allows multiple images of the anatomical geometry of the same patient acquired from different angles and at different times (e.g., different respirations) to be more accurately correlated because basic assumptions of the underlying radiological mathematics assume anatomical equivalence across the various imaging angles detected. The enhanced images are used as observations, and the biometric signals are used as inputs to a complex biomechanical model of the thorax geometry.

[0152] Section 3.1- GREX image acquisition based on digital tomosynthesis

[0153] Multiple imaging angles are required to reconstruct a 3D volume. In the context of GREX imaging, each angle needs to be acquired for each respiratory phase. The acquisition software 104 (Chapter 2) generates a trigger signal that allows the x-ray unit 108 to repeatedly image the chest cavity with a specific geometric structure. The same geometric structure imaged at different imaging angles and different respirations constitutes a set of 2D projection images, which are used to reconstruct the 3D volume. There are many known methods for reconstructing a 3D volume based on multiple 2D projection images. One such exemplary method is a convolution-back projection algorithm that directly reconstructs a 3D density function using a set of 2D projections called the "FDK image reconstruction algorithm", which is disclosed in Feldkamp, ​​LA, Davis, LC, Kress, JW's "Practical cone beam algorithm" (J Opt Soc Am1, 612-619 (1984)).

[0154] Fig.16 is a schematic block diagram of variables used in an image reconstruction algorithm according to some embodiments of the present disclosure. Once a single projection is acquired, the x-ray unit 108 is moved by an angle β, and the detector plane is moved to remain perpendicular to the x-ray unit 108. In some embodiments, the patient is moved, and the x-ray unit 108 remains in the same orientation between acquired projections. Fig.17 Schematically shows Figure 317. An example of how the x-ray unit 108 and the detector panel 304 in FIG. 17 are moved to obtain multiple imaging angles. For example, the x-ray unit 108 moves from position 1700a to position 1700b to position 1700c. The detector panel 304 moves from position 1702a to position 1702b to position 1702c. The detector plane rotates about an axis parallel to the detector plane, and the imaging plane rotates about its parallel axis z. The positions of pixels in the detector plane and corresponding pixels in the imaging plane are separated by a distance s. The anatomical information (f(x, z / y)) in the image plane ((x, z) plane) at any depth y is calculated by equation (1).

[0155]

[0156] In equation (1), N0 is the total number of projections, β is the angle of each projection, d is the distance from the source to the image plane, s is the distance from the pixel to the detector, p is the detector axis perpendicular to the rotation axis, ξ is the detector axis parallel to the rotation axis, R(β, p, ξ) corresponds to the cone beam projection data (e.g., the function R is the detector readout for a given angle, p coordinate, and ξ coordinate), h is the convolution filter, and W(p) is the weighting function. Basically, equation (1) represents a combination of convolution, backprojection, and weighting steps.

[0157] Information at a point in the mid-plane (y=0) is calculated from projection data along the intersection of the detector plane and the mid-plane. Projections that intersect the detector plane along lines parallel to the mid-plane but not in itself (constant non-zero y) define a plane. This plane is considered to be the mid-plane of another tilted arrangement. If a complete set of projections is acquired (note that "complete" means the case where all rotation angles about the normal are acquired), the density of the tilted plane is reconstructed using the Radon transform. Acquiring a complete set of projections requires a 360° rotation of the source around the imaging object along a circle in the tilted plane; for example, in CT imaging, a full 360° rotation occurs around the imaging object. Note that the boldfaced word "imaging object" means - more precisely, the boldfaced word explicitly expresses and defines - the basic assumptions of the Radon transform (incorporated into the R(β, p, ξ) term of equation (1)), which, if violated, will hinder the reconstruction from being representative of the underlying true anatomy being imaged: "imaging object" means the different angles of the probe that assume the Radon transform input to be a single fixed, spatially non-moving and unchanging stationary anatomical object. GREX imaging's biometric targeting (effectively "pre-selection") of only accurate (geometrically identical breathing phases that are identical even when occurring in different breaths) Radon transform inputs is by definition and in practice follows from the basic assumptions of the transform, since the breathing phase (i.e., varying in the time domain) is used by GREX imaging for its unique short prediction horizon and forward-looking biometric targeting, both geometrically and anatomically and physically (e.g., definitionally accurate formulation) and also in practice strictly limited GREX.

[0158] It should be noted that 360° rotation is not practical for GREX imaging because the large number of projections required to reconstruct the 3D volume of the torso increases clinical procedure time and increases radiation dose to the patient. In practice, GREX imaging can use projection angles ranging up to 90° between -45°≤β≤45° or 0°≤β<90°. Those skilled in the art will appreciate that experimental testing can identify a more optimal projection angle range, but the theoretical projection angle range does not exceed 90°. In some cases, -45°≤β≤45° may be preferred over 0°≤β<90° because -45°≤β≤45° keeps the radiation dose in the imaging procedure at a reasonably practicable level. Compared to the 0°≤β≤90° projection angle range, x-ray photons traverse less human tissue in the -45°≤β≤45° projection angle range, and therefore can be lower energy photons, thereby depositing less dose, such as in Fig.18 A and 18B.

[0159] In order to produce high quality 2D images without delivering excessive radiation dose, the photon energy must be high enough to partially penetrate the patient's body but not so high that the photons completely penetrate the patient. Thicker patients require higher photon energy than thinner patients. When |β|>45°, the human body is much thicker than when |β|<45°. In general, when β→0°, the photon energy decreases. In some embodiments, the GREX imaging system 100 acquires a total of 30 projections for six respiratory phases at five different projection angles, but other different numbers of projections are possible depending on the specific application of the GREX imaging technology. For example, in breast tomosynthesis, the symmetrical curvature of the breast means that the breast surface is essentially equidistant from the source at all projection angles, meaning that tomosynthesis is well suited for breasts. In addition, the breast does not move when placed in the cradle, which is a typical clinical tomosynthesis method that can be similar to breath-hold lung imaging.

[0160] The GREX imaging system 100 can utilize biometric surface information collected by the 3D spatial localization localizer (Section 1.1) to handle varying trunk curvature, while addressing the challenges of lung and heart motion through GREX-based respiratory phase geometry constraints coupled with highly accurate and fast GREX-based forward targeting (Sections 2.1 and 2.2, respectively). The biometric surface information also facilitates image post-processing to account for attenuated tissue density in the imaging field, thereby quantifying previously neglected attenuation sources and ultimately resulting in higher fidelity image reconstruction. In summary, GREX imaging technology makes digital tomosynthesis possible for non-breath-hold ("dynamic") lung and heart imaging.

[0161] The GREX-based 3D volume reconstruction method can be patient-specific (personalized medicine, with a personalized number of discrete angles and arcs), or used as a "universal minimum procedure time and universal minimum delivered dose" (about 5 discrete angles, plus or minus 3 angles, depending on the statistical reconstruction method used and the number of previous GREX-based datasets available for this person).

[0162] like Fig.19 As shown, when the x-ray unit 108 rotates over a small range of angles, the depth resolution is low compared to when the x-ray unit 108 rotates over a wider range of angles. Any motion, no matter how small, will cause image artifacts in the reconstructed image, which often leads to false positive cancer detections. However, the unique hardware box 102 (Section 1) of the GREX imaging system 100 acquires biometric signals that inform the acquisition software 104 (Section 2) when to take images so that the thorax geometry is clinically the same, thereby enabling intelligent thorax digital tomosynthesis.

[0163] In some embodiments, the GREX imaging technique allows the projection angle to vary during one respiratory phase, while multiple projection angles captured at different time points (defined by the quantitative definition of "respiratory phase" of GREX imaging) still all correspond to the same respiratory phase because they are considered to all capture one thorax geometry. In addition, the GREX-based tomosynthesis method allows for the acquisition of depth information because the final number of photon counts of the detector across all pixels and the distribution of photon counts in space at the detector surface reflect a single thorax geometry that has been probed at multiple angles.

[0164] It should be noted that the GREX imaging technique allows the x-ray unit 108 and detector panel 304 to be mounted on a non-motorized arm or stand. The role of the manual arm in an exemplary GREX imaging procedure is described below and is described in Fig. 20 It shows:

[0165] 1. In the β1 orientation, all six necessary thoracic geometries (also known as the "quantitatively defined respiratory phases of GREX") are imaged.

[0166] 2. Next, the clinician reorients the x-ray unit 108 and detector panel 304 to image the patient in orientation β2.

[0167] 3. The orientation of the x-ray unit 108 and detector panel 304 relative to the patient is verified by the 3D space localizer, which allows imaging at β2.

[0168] 4. Now, at β2: If only 4 of the 6 required chest geometry images are acquired during the first breath, the patient simply continues breathing normally in the subsequent breaths at position β until the remaining 2 chest geometries are acquired.

[0169] 5. Now that all 6 breathing geometries have been acquired at β1 and at β2, the x-ray arm can be repositioned to acquire each imaging angle in sequence from (β2→β3; β3→β4; β4→β5).

[0170] 6. At the end of the program, the x-ray arm has moved only 4 times in total, as indicated by the → arrows in the workflow (β1→(β2; β2→β3; β3→β4; β4→β5).

[0171] By moving the arm only four times during the procedure, GREX imaging technology minimizes procedure length, clinician-equipment interaction during the procedure, and wear and tear on the x-ray arm because the clinician does not interact with the x-ray arm as much as moving the arm from β1 to β5 for all six respiratory phases (e.g., performing 30 x-ray device reorientations).

[0172] In some embodiments, the 3D spatial positioning positioner allows for separate a priori position checking and safety interlocks of the three individual elements - namely, patient posture and orientation, x-ray unit 108 orientation, and detector spatial orientation, as well as collective consistency of the orientation / alignment of each element on axis and relative to the other elements. Such a priori position checking and safety interlocks result from the geometrical definition of respiratory phases by GREX imaging. Thus, the GREX imaging system 100 (through a prospective respiratory phase prediction algorithm) is inherently adapted to give the user software-based safety and quality assurance control capabilities that (in the case of a safety interlock) prevent the trigger algorithm from initiating a "beam on" if either (or both) of the x-ray arm or detector or patient posture is improperly positioned (or inconsistent) in space at a particular angle β.

[0173] In some embodiments, since the 3D spatial positioning locator records the coordinates of all devices during the procedure, the image reconstruction technique can benefit from the (a posteriori) quantification of each beta angle and its associated uncertainty.

[0174] Fig.21 is an exemplary block diagram showing how the GREX imaging system 100 according to some embodiments of the present disclosure reconstructs a static image cube from 2D projection data taken from each imaging angular position. Specifically, Fig.21 An exemplary GREX imaging case is shown, where the coronal and sagittal views themselves form the outer limits / boundaries of the imaging angle positions (e.g., projection angles). For each of the 6 breathing geometries, at a single imaging angle position (β i ). The sum of the i=1 to i=n planar projections then informs the reconstruction of the depth information in (x, y, z) of the (v1, f1) geometry probed using the i=1 to i=n projections (Equation 1), each with a focal plane at a unique (compared to the other projections) depth.

[0175] 2D projection data are acquired six times (for early inspiration (EI), late inspiration (LI), maximum inspiration (MI), early expiration (EE), late expiration (LE), and maximum expiration (ME)) at each projection angle (β). For simplicity, the projection angles range from 0° to 90°. The x-ray unit 108 moves to the next projection angle only when all respiratory phases have been acquired at the previous projection angle. The 2D projections are ordered according to the respiratory phase in order to reconstruct a static image cube (e.g., from the β projections that detect the thorax geometry at (v1, f1)). i =1 to β i =n projection angles) to represent the thoracic volume at each respiratory stage. Next, the static image cube is temporally interpolated using the method discussed in Section 3.3 below.

[0176] The x-ray projections acquired at each angle β are obtained at the targeted respiratory phase identified using the ARIMA model (Section 2.2). After acquiring the targeted respiratory phase, the x-ray unit 108 moves to the next imaging angle position. Fig. 22 An example of acquiring only maximum inspiration and maximum expiration by the acquisition software 104 is shown. In the example, the maximum inspiration and maximum expiration are imaged at β1, and then the x-ray unit 108 moves to β2 so that the maximum inspiration and maximum expiration phases can be acquired at β2. When the targeted respiratory phase is acquired at all imaging angle positions, the images are sorted according to the corresponding respiratory phase of the image. Sorting the x-ray images according to the respiratory phase groups the projections according to the six respiratory phases defined by biometry. Although the respiratory phases are acquired during different respirations, the accurate ARIMA model ensures that the tidal volume and airflow parameters are the same between the x-ray projections captured at different projection angles. It should be noted that GREX imaging defines the "same chest geometry" in a biometric manner, so that the lungs are in the "same chest geometry" at multiple time points. GREX imaging detects the same biometric respiratory phase at different angles because the ARIMA model (Section 2.2) is a fast prediction method. The short prediction range of the ARIMA model is used to minimize the chest geometry prediction error. The relative equivalence of the depth information at each individual angle (together, used to generate real structure depth information) depends on the equivalence (consistency, within approximate tolerance) of the thorax geometry across different detection angles. Thus, respiratory phase prediction accuracy ensures successful image reconstruction.

[0177] The respiratory phase ordered projections are used to create a static image cube by the canonical FDK image reconstruction algorithm given in equation (1) discussed previously (or a similar cone beam geometry image reconstruction algorithm). The image reconstruction algorithm uses the respiratory phase ordered projections and creates a respiratory phase ordered static image cube. Each respiratory phase will have a separate image cube. The static image cubes are called static because they represent the anatomical structure in only one respiratory phase. The static cubes representing all targeted respiratory phases are combined and temporally interpolated (described in Section 3.3) to create a 3D image cube image from the 3D static image cube.

[0178] The GREX imaging system 100 maintains radiation dose to the lowest level that can be reasonably achieved through statistical image reconstruction. Each acquired image increases the dose of the entire imaging process (a clinically undesirable outcome), but provides additional information for image reconstruction (a clinically desired outcome). Traditional forms of image reconstruction based on Fourier transforms or filtered back projection tend to display image artifacts due to their inability to handle missing information (e.g., missing projection angle β). For example, if projections are made every 10° instead of 5°, there will be half as much information available to create a static image cube, but the former will only deliver half the dose compared to the latter. Statistical iterative image reconstruction addresses the missing information caused by having an incomplete image data set.

[0179] There are many existing statistical image reconstruction algorithms that can be used by the GREX imaging system 100 to accomplish the task of image reconstruction (e.g., constructing a static image cube). However, the GREX imaging system 100 improves upon conventional statistical image reconstruction algorithms by implementing a unique feedback step and adhering to boundary conditions based on the law of mass conservation.

[0180] Basic principles of physics can be applied to the statistical image reconstruction of GREX, because GREX images are biometrically defined by biophysical quantities governed by the laws of physics. By biometrically labeling each image and the resulting image cube, and by acquiring a continuous stream of biometric data during the procedure, even when the patient is not being imaged, the respiratory dynamics of mass exchange (inhalation and exhalation) and volume changes can also be known (making it possible to resolve the constant lung tissue mass that is consistent during the scan). The law of conservation of mass can be applied because the static image cube can be composed of a moving organ such as the lung for the first time. This is attributed to the speed and accuracy of the prospective prediction / triggering algorithm that accurately labels and acquires the same chest geometry at different times (Sections 2.2 and 2.3). In other words, from one static image cube to the next, the tissue mass in the static image cube should not change (e.g., it will not change due to the law of conservation of mass). According to the ideal gas law, the ratio of room temperature air to air in the lungs is 1.11. Given the tidal volume of the image cube from the sensor data, combined with the 1.11 ratio at room temperature and the mass / volume air deviation curve, the mass of inspired air can be determined (in absolute terms and also in relative terms as the ratio between the image cubes of two different breathing phases).

[0181] GREX imaging is very useful based on the boundary condition of mass conservation because, for example, the presence of air can artificially darken voxels, thereby adversely affecting the ability of statistical image registration algorithms to accurately determine object density. GREX imaging produces more accurate image reconstructions by correcting the acquired projections for air volume differences and isolating quantities that should be constant throughout the scan (e.g., lung tissue mass) to generate a static image cube.

[0182] Consider two GREX projections acquired at different imaging angles but with the same biometrically defined respiratory phase, the volume of air in the lung is the same, but the way the air displaces tissue between the two projections may be different and darken the brightness of the nodule in the second projection, which is affected and erroneous compared to the first projection. The result of this error is that a piece of tissue (the "nodule") visible in the first projection is not visible in the second projection, which will ultimately dull the intensity of the "nodule" in the resulting image cube (or cause it to be misidentified as background). The mass conservation boundary condition of GREX imaging is implemented as a feedback step that checks to confirm lung mass conservation between the above image cube with the erroneously darkened nodule and the different (correct and anatomically representative "nodule" brightness) image cube from the later respiratory phase. The feedback step of GREX imaging corrects the level of the second projection of the erroneous image cube by updating the expected geometry during reconstruction based on simulations using the first projection as the highest criterion. In this way, GREX statistical image reconstruction will produce a more accurate static image cube.

[0183] In addition to pre-existing statistical image reconstruction algorithms, GREX's post-processing software 106 also incorporates edge-defining filters (discussed in Section 3.2), spatial boundary conditions (discussed in Section 3.2), and smooth transitions between respiratory phases (discussed in Section 3.3) into digital diagnostic x-ray images to improve anatomical imaging.

[0184] Chapter 3.2 - Image Filters

[0185] The quality of a digital diagnostic x-ray image depends on the x-ray unit 108 settings and the anatomical study site of the image. Each patient and anatomical site has a different electron density through which x-rays pass to generate an image. For example, a higher x-ray energy is required to image the femur than the x-ray energy required to image the thorax because the lungs are mainly composed of air, while the femur is composed of bone. Given that higher energy x-rays can penetrate the body to a greater extent than lower energy x-rays, the number of x-rays that come out of the body and reach the flat panel detector is different for high energy x-rays and low energy x-rays that image the same anatomical geometry. Too many x-rays coming out of the body can cause the flat panel detector to be overexposed in a manner similar to overexposure in optical photography. If the x-ray unit 108 settings are not optimal for the anatomical image study site, the image quality will be greatly reduced. In clinical practice, commercial vendors have designed imaging protocols for their digital diagnostic x-ray units that roughly estimate the optimal x-ray unit settings for a selected anatomical site. But these rough estimates of optimal tube settings are not tailored to account for potentially significant anatomical differences between anatomical sites in different patients (e.g., the stomach of an overweight male versus the stomach of an average weight male). In fact, existing imaging protocol settings by vendor groups, as rough estimates only, rarely produce optimal quality images.

[0186] If the optimal x-ray unit settings or other imaging parameters are not known prior to imaging (as is currently the case within the medical community), strategic deployment of digital image filters can improve image quality for non-optimal x-ray unit settings. The improved image enhances visibility of anatomical features that are not readily visible to the human eye. For example, in the coronal plane, each rib may not be visible in a digital diagnostic radiographic image. The post-processing software 106 filters the coronal plane image using an edge enhancement filter (such as a Laplacian filter) to display the boundaries of all ribs on the resulting image even when the rib boundaries on the original image are too subtle for the human eye (e.g., the eye of a radiologist) to detect. The post-processing software 106 overlays the filtered image with the original image, which highlights the enhanced (e.g., post-filtered) and previously invisible rib edges on the original image. The image filters available for users to apply include Laplace filter, Hanning filter, Butterworth filter, Parzen filter, Wiener filter, Metz filter, Ramp filter, nonlinear spatial mean filter and hybrid filter.

[0187] In some embodiments, the post-processing software 106 uses the skin surface measurements of the 3D spatial position locator 300 to calculate the optimal imaging parameters for images taken at each respiratory stage. As the patient breathes, the electron density of the body changes as more air is inhaled and the chest circumference increases. Increasing the patient's diameter, reducing the distance between the patient and the x-ray unit 108, and reducing the distance between the patient and the detector panel 304 will produce additional image noise in the resulting x-ray image. The 3D spatial position locator 300 tracks the patient's skin surface position relative to the x-ray unit 108 and the detector panel 304 for each image. This skin surface position tracking provides unique measurements for digital diagnostic x-ray studies.

[0188] The field of digital diagnostic radiology currently relies on scaling the dose index readings measured by ion chambers to the patient's approximate body mass index. Radiologic technicians currently take only two measurements: the first measurement is on a solid water cylinder phantom of equivalent tissue density, 16 cm in diameter, and the second measurement is on a phantom of the same material, 32 cm in diameter. The x-ray unit 108 has built-in protocols that are "one size fits all" that are vendor-defined and targeted to specific anatomical locations. For example, regardless of the patient's chest diameter, the vendor only provides a single protocol with built-in imaging settings for the technician to choose. In other words, a man with a large chest gets the same imaging settings as a man with a thin chest.

[0189] The 3D spatial positioner 300 of the GREX imaging system 100 generates a real-time and personalized measurement of the patient's chest diameter. The measurement will inform the technician to select the x-ray unit settings personalized for the patient. When the patient breathes, the chest diameter changes. The changing patient chest diameter prevents the technician from setting the imaging parameters in an optimal manner to match the patient's chest diameter. In addition, the real-time and personalized measurement of the patient's chest diameter can be used to remove image noise in post-processing procedures, as well as simulate x-ray images obtained with optimal imaging parameters.

[0190] Fig.23is a schematic flow chart of an image filtering process 2300 according to some embodiments of the present disclosure, which calculates a noise-free lung image and simulates an image obtained with an optimal x-ray unit setting. After acquiring the image as described in Section 2, the trachea is identified using a line profile of a line segment that passes transversely (from left to right) through the neck region. The neck is composed of muscle, bone, and arteries, but the trachea stands out from all other tissues because it contains only air with a density significantly lower than that of the tissue. The line segments will show where the air is located, and small areas containing pixels designated as air will be identified. Image noise in an x-ray image is calculated by subdividing the entire image into smaller blocks. Gaussian noise is estimated independently for each block, and the block with the smallest amount of noise is used for texture mapping. The texture mapping technique uses a gradient covariance matrix to estimate the initial texture level in each block. The block with the smallest initial noise level re-estimates the noise level through an iterative process, which continues until the noise estimate of the block converges through additional iterations of the gradient covariance matrix. It is assumed that the weakly textured block is located in air, where the air is located away from the patient, such as the upper corner of the x-ray image. The estimation of the noise level of the block gives a baseline noise level for the entire image. The baseline noise level identified by the iterative gradient covariance matrix is ​​then subtracted from the entire image to obtain a noise-free air density estimate in the trachea.

[0191] The post-processing software 106 overlays the human skeleton model (which is scaled individually for each patient) onto the surface position estimate provided by the 3D spatial position locator 300 to estimate the initial position of the lung. The individually scaled patient skeleton is strictly registered to the x-ray image using visible landmarks on the skin surface (e.g., clavicle, rotator cuff, scapula, vertebrae, etc.). After the skeleton is strictly registered to the x-ray image, the skeleton provides the position of the rib cage. The rib cage itself provides boundary conditions for the pixel values ​​of the noiseless air density estimate of the edge of the lung and the trachea. The edge of the lung (through the rib cage position) and the pixel values ​​of the noiseless air density estimate of the trachea are automatically identified as seed positions by a region growing algorithm (the region growing algorithm will start from the position and then grow radially outward). Region growing is a segmentation method based on a region. This segmentation method first identifies a group of initial seed points in the image, and then checks the adjacent pixels of the initial seed point and determines whether the adjacent pixels should be added to the region. Repeat the process in the same way as the general data clustering algorithm. In other words, the region growing algorithm uses the initial placement of seed pixels to expand outward using a statistical process of swallowing up "similar" pixels. The region growing algorithm will continue (swallowing similar pixels) until the identified pixels are statistically dissimilar to the swallowed clusters.

[0192] In practice within the GREX imaging system, the region growing algorithm "stops" (e.g., detects pixel dissimilarity) at important anatomical landmark interfaces (e.g., lungs defined by dense pixels of the rib cage). Pixels that the region growing algorithm does not identify as belonging to lung tissue will be masked (the masked image is defined as an image that enhances the structure once subtracted from the original image) to form two separate images. The two separate resulting images are (i) the segmented lungs and (ii) the remaining body tissue. In order to provide an accurate and noise-free segmented lung volume (e.g., a lung volume that is not visually obscured by non-lung tissue) that gives radiologists improved diagnostic visibility, tissue associated with body masking (e.g., non-lung, and therefore visually uninformative) is removed from the lung image. For example, pixels belonging to the intercostal muscles within each simulated imaging ray projection are completely subtracted from the segmented lung image. In addition to the above-described subtraction of body masking from the lung image to produce improved lung tissue visualization, body masking can also be used to provide a second check of the patient surface orientation calculated by the 3D spatial orientation locator 300. For example, the post-processing software 106 counts the number of pixels identified as body occlusion by the region growing algorithm, and then calculates the body diameter at various locations along the height of the torso. This body diameter calculation should be very consistent with the 3D spatial position locator 300's estimate of the patient's body diameter. If not, it may indicate that the 3D spatial position locator 300 needs to be recalibrated to increase its accuracy.

[0193] If a clinical user requires a more accurate body occlusion view (e.g., for clinical or educational reasons), the body occlusion image is simulated using the optimal x-ray unit settings, thereby removing noise forms and potential artifact sources from the body occlusion. The body occlusion and segmented lungs can then be recombined to form an artifact noise-free x-ray image with global enhancement for clinical applications (e.g., structural topography).

[0194] Section 3.3 - Biomechanical Modeling

[0195] The biomechanical model used in the post-processing software 106 is created based on the first principles of physics, namely the law of conservation of mass and the ideal gas law. The goal of the biomechanical model of the post-processing software 106 is to determine biophysical quantities that enhance the ability of clinicians to diagnose disease. The relevant biophysical quantities include, but are not limited to, stress and strain of lung tissue elements.

[0196] Mechanical systems that are loaded with force will generate stresses. In the case of the lung, the elements of the mechanical system are represented by the lung tissue. The lung tissue visible and distinguishable in medical imaging consists of parenchyma containing alveolar sacs, alveolar walls, bronchi and blood vessels. The parenchyma is directly responsible for lung function. Suitable tissue elements for biomechanical modeling should be small enough to be internally homogeneous, but should be statistically stable in response to respiratory stimuli. Typical voxel sizes in medical imaging of the lung range from 1 mm 3 Up to 3mm 3 , equivalent to 125 to 375 alveoli. The voxels are assumed to be of nearly uniform density and contain enough alveoli to provide a stable response to respiratory stimulation. The alveoli are arranged in a hexagonal array, which expands due to the expansion normal stress from each shared alveolar wall. The sum of all expansion normal stresses within the lung tissue element provides an estimate of the pressure experienced by the alveoli and caused by the respiratory stimulus. The expansion stress is balanced by the elastic response stress on the alveolar wall, and there is a balance between the two stresses when the airflow through the tissue element is zero. The spatial translation of the orientation of the lung tissue element in any direction due to the changing lung tidal volume can be modeled by the material response of the tissue element to the normal stress on the corresponding surface of the lung tissue element. In other words, the biomechanical model contains vectorized terms that describe the response of the lung tissue elements to increased tidal volume, wherein the vectorized terms involve normal stresses.

[0197] Strain is defined as the response of a mechanical system to stress. From the perspective of a material element (e.g., tissue), stress is the deforming force and strain is the restoring force. The stress vector on the surface of a lung tissue element has two components: (i) normal stress (associated with outward or inward directional motion of the lung tissue element causing expansion or contraction, respectively), and (ii) shear stress, which is perpendicular to the normal stress and results from the pressure imbalance induced by airflow. It should be noted that, by definition, the perpendicular shear stress component does not contribute to lung volume changes. A graphical illustration of the relationship between normal stress / shear stress and tidal volume / airflow can be found at Fig.10 . Exhalation is at the lowest tidal volume value where the airflow is 0 (the leftmost point on the curve). During inspiration, the airflow is positive and the tidal volume increases (the upper part of the curve). When the airflow slows, the tidal volume reaches its maximum value. As the tidal volume approaches the maximum value (the rightmost point on the curve), the airflow decreases dramatically and the lungs gradually begin to expel air. As the lungs expel more air, the tidal volume begins to decrease. The tidal volume decreases when the airflow is negative, and continues to decrease until the tidal volume reaches the residual volume and the airflow is zero (e.g., exhalation). In summary, the movement of lung tissue elements is defined by the tidal volume and airflow respiratory stimuli. The movement of lung tissue elements is caused by the normal and shear stresses acting on the elements. Changes in the shape of lung tissue elements (e.g., compression, elongation) are not modeled by stress; instead, changes in the shape of the elements are modeled by strain.

[0198] The model creates a new diagnostic perspective in the medical field because medical analysis using the model is controlled by quantitative force analysis. The balance of stress and strain forces at each voxel in the lung fundamentally assesses lung health because the functional behavior of the lung (e.g., lung tissue motion) can now be visualized and analyzed from a clinical perspective.

[0199] The first step in constructing the model is to identify which tissues are part of the lungs and which are not. To accomplish this task, the GREX imaging technology acquires multiple snapshots of the thoracic geometry in multiple configurations (6 different breathing phases). The lungs are segmented based on a region growing segmentation algorithm using the air density of the trachea as the initial starting point for the region growing algorithm. The GREX imaging system 100 uses two different deformable image registrations to construct the biomechanical model: one for lung tissue and the other for non-lung tissue (e.g., chest wall, ribs, liver, heart, trachea, esophagus, etc.). The results of the region growing algorithm distinguish between lung tissue and non-lung tissue before the deformable image registration step.

[0200] The reason the GREX imaging system uses two different deformable registrations (one for the lungs and another for the non-lungs) is that the motion and material properties of the lungs are different from those of the non-lungs. If only one deformable image registration were used for the entirety of the lungs and non-lungs, the registration would assign a greater weight to the non-lung tissue than to the lung tissue. This would result in the lung tissue being assigned an unrealistically low weight, and therefore an unrealistically limited degree of motion. The GREX imaging technique is intended to intuitively illustrate the subtleties of lung motion, including subtleties of motion at the lung surface which are computationally complex.

[0201] One deformable image registration algorithm that can track tissue movement between images is the optical flow algorithm. Fig. 27 is a schematic flow chart of a multi-resolution 3D optical flow algorithm 2700 according to some embodiments of the present disclosure. Specifically, the 3D multi-resolution optical flow algorithm performs deformable image registration, which identifies structures in two different images based on the brightness or contrast of each structure, or both. To this end, given that non-lung tissue contains characteristically brighter (e.g., at least 10 times denser) anatomical structures (e.g., ribs, pectoral muscles, sternum, etc.) than lung tissue, using a single 3D multi-resolution optical flow algorithm for non-lung and lung regions will cause the lower density lung tissue to be artificially de-prioritized in the algorithm. The algorithm's computational resources will be preferentially allocated to brighter contrast structures (e.g., ribs, etc.) present in non-lung tissue. This result in which computational resource allocation preferentially ignores lung tissue dynamics is a nuisance to the purpose of GREX imaging to intuitively illustrate the subtleties of lung tissue movement.

[0202] Since the preferential resource allocation to high contrast structures is inherent to the 3D multi-resolution optical flow algorithm, the task of deformable image registration in the chest is divided into two separate subsidiary tasks (e.g., two more uniform regions inside): (i) image registration of the lungs, and image registration of the non-lungs. In some embodiments, in order to resolve the two separate task areas from the chest as a whole, it is necessary to identify the lung surface (e.g., the boundary between the lungs and the non-lungs) before the 3D multi-resolution optical flow deformable image registration. It is basically possible to identify the lung surface using a region growing algorithm, which starts in an air-filled (e.g., visually dark) region inside the lungs, grows outward toward the lung surface boundary, and reaches high pixel contrast at the lung surface boundary. Executing the region growing algorithm is the first step.

[0203] The lungs are not attached to the chest wall. Therefore, the movement of the lungs is relatively independent of the movement of the chest cavity. In other words, instead of a predictable push-pull response at the lung surface boundary, there are other types of tissue dynamics that come into play. For example, horizontally adjacent chest voxels move vertically downward, so the lung voxels move horizontally to the space where the chest voxels previously stayed.

[0204] In order to accurately model complex motion dynamics, the GREX imaging system 100 quantifies the shear force experienced by the surfactant layer at the boundary of the lung surface. The use of two separate segments for lung and non-lung tissue provides the basis for force estimation. The force estimation procedure (performed for each reconstructed respiratory phase image) is performed by subtracting the segmented lung pixels from the x-ray image (e.g., removing their assigned values). The previously segmented lungs will be masked from the original image, which will provide an image containing all the remaining tissues. The segmented chest geometry of each individual image must be registered to each chest geometry of the other images in order to know the orientation of each lung tissue element on all images. The multi-resolution optical flow algorithm performs image registration by calculating a displacement vector field, which shows the displacement of each pixel between two images taken at different respiratory stages. In fact, knowing the displacement vector field allows accurate spatial calculation of all lung tissue elements in the chest geometry. The difference between the displacement vector field of the segmented lung registration and the displacement vector field of the non-lung tissue provides the magnitude and direction of the shear force between the lung and the chest wall.

[0205] Based on observations of over 150 unique patients, the relationship between displacement and tidal volume is linear. The relationship between displacement and airflow is also linear. The output of the multiresolution optical flow algorithm is a displacement vector in coordinate space for each measured tidal volume and airflow value. When the displacement vectors are calculated for all respiratory phases, the result is a closed-loop trajectory (e.g., Fig.10As shown), which is used as an observation of the biomechanical model. As described in more detail below, the biomechanical model parameters - which may include parameters representing the normal stress associated with tidal volume, the normal stress associated with airflow, and the shear stress associated with airflow - are solved using QR decomposition for each lung tissue element, respectively. The parameters are specific to each lung tissue element (e.g., each lung tissue element has a unique solution) and together describe the response of the lung tissue element to respiratory stimulation. The biomechanical model parameters are vectors globally scaled by the measured tidal volume and airflow (e.g., tidal volume and airflow are scalar values). The relationship between the biomechanical model parameters, such as the angle between two (vectorized) parameters, can assist in diagnosing potential morbidity in the lungs. Based on the unique biomechanical model vector parameters of the lung tissue elements (e.g., each piece of tissue has different vector parameters), the displacement of the lung tissue elements is scaled to the thoracic geometry, where the thoracic geometry is defined by the tidal volume and airflow measurements. The biomechanical model can be calculated for each patient, or shared by multiple patients.

[0206] In some embodiments, the biomechanical model roughly estimates the motion of lung tissue as a function of multiple factors, including the tidal volume (T) of the lung. v )、airflow(A f ) and cardiac stage (H c ). These values ​​are global values, for example, cardiac phase is the same for all tissue elements in the chest. Global values ​​are treated as scalar numbers and are measured using the hardware discussed in Section 1. Note that tidal volume, airflow, and cardiac phase are all time-varying measurements. Using vector and The unique stress and strain values ​​for each tissue element are expressed mathematically by the following equation:

[0207]

[0208] in The vector describes the normal stress caused by the tidal volume, describes the normal stress caused by airflow, describes the shear stress caused by the airflow, and Describes the tissue motion caused by the perturbed cardiac motion. In general, the tissue displacement at any point in the closed-loop trajectory It is expressed as the sum of stress, strain, and perturbation cardiac motion vectors scaled by tidal volume, airflow, and cardiac phase, respectively.

[0209] Fig.24 Depicts an exemplary closed-loop lung tissue motion trajectory of a piece of tissue located in the left lung near the heart caused in part by the interaction of the heart with the lungs according to some embodiments of the present disclosure. It should be noted that Fig.24The wave behavior depicted in Figure 1 results from the interaction between the heart and lungs. Fig.24 It is a schematic block diagram illustrating how to determine the position of a piece of lung tissue moving in a closed-loop lung tissue motion trajectory based on the above-mentioned biomechanical model according to some embodiments of the present disclosure. Fig.24 It is shown how the three vectors of the biomechanical model described in the above equations can be added together to calculate the displacement of a single tissue element from the origin to any position on the closed-loop trajectory.

[0210] The main advantage of using a physiologically based biomechanical model to interpolate images between acquired respiratory phases is the ability to use quantitative physical quantities to check the accuracy of the output of the biomechanical model. According to the ideal gas law, at room temperature, the ratio of the change in lung volume to the tidal volume is 1.11. In other words, the ratio of the density of room air to the density of air in the lungs is 1.11. Therefore, the volume integral of the divergence of the normal stress vector should be 1.11 (e.g., Where V is the total body volume.) The ideal gas law provides a "sanity check" to produce applicable quality assurance information about the 3D spatial position locator 300, the respiratory phase sensor 110, and each interpolated image.

[0211] One example of how the biomechanical modeling of the GREX imaging technology can enhance diagnosis is an early lung tumor that is invisible to the radiologist's eyes when the image is examined. The radiologist cannot see the tumor because it is too small for the imaging sensitivity. Although the radiologist cannot see the tumor, the presence of the tumor still affects the balance of forces inside the lung because the electron density of the tumor is greater than that of healthy lung tissue. The higher electron density of the tumor means that the tumor has different material and mechanical properties (e.g., different characteristic stress and strain parameters) that affect the movement of the tumor and the movement of the local area around the tumor (e.g., healthy tissue close to the tumor site). The effect of the tumor on the local healthy lung tissue can be roughly analogized to the effect of mass on the space-time continuum under general relativity: when there is a large mass object, the space-time around the object is curved, causing light to behave differently near the object compared to the behavior of light when there is no mass. The same analogy applies to the lungs, causing the tumor to deform the trajectory of motion of the adjacent healthy tissue to move differently compared to the trajectory of motion of the healthy lung. The displacement vector mapping of the biomechanical model makes changes in lung tissue composition and biomechanical properties apparent to the medical practitioner: when a tumor is present, the displacement vector field displays an unnatural degree of vector curling and / or other altered properties. The creation of parametric maps by the post-processing software 106 of the GREX imaging technology discussed in Section 3.4 visually displays such diagnostically important information that was previously invisible to the end user. The parametric maps of the GREX imaging technology are an example of the new diagnostic perspective that the GREX platform brings to medicine.

[0212] Fig.25 is a schematic flow chart of operations 2500 of components of a biomechanical model according to some embodiments of the present disclosure.

[0213] The process of the biomechanical model provides a quantitative means for biometric interpolation between two images (2D or 3D) acquired at different respiratory phases. Equation (2) shows that the solution of the biomechanical model is the displacement between two respiratory phases As described above, the displacement between two breathing phases is found by performing deformable image registration to index the two breathing phases to each other. It usually includes the following three steps:

[0214] • Step 1: Perform region-based segmentation to delineate the structural boundaries between lung and non-lung.

[0215] Step 2: Use 3D multi-resolution optical flow to perform intensity-based structure mapping to match the “same” structure in two separate image cubes.

[0216] Step 3: In some embodiments, iteratively refine the displacement between two breathing phases until it is optimized.

[0217] Fig.26A is a graphical depiction of the motion trajectory of lung tissue elements during a respiratory cycle according to some embodiments of the present disclosure. Fig.26A As shown, each imaging angle has 6 images, corresponding to the six corresponding respiratory stages of a complete respiratory cycle, from left to right: early inspiration (EI), late inspiration (LI), maximum inspiration (MI), early expiration (EE), late expiration (LE) and maximum expiration (ME), as shown in FIG. Fig.14 . According to finite strain theory, the vector joining the orientation of the particle in the undeformed configuration and the deformed configuration is called the displacement vector of the particle. Using a specific voxel in the EI image cube as a reference, the six images show the deformation of the reference voxel from its original position, size and shape during the respiratory cycle. As the lung fills with more and more air, the reference voxel begins to "bubble". In other words, as the voxel "bubbles", the lung problem corresponding to the reference voxel deforms. The displacement between the two respiratory phases The degree of voxel bubbling, or the deformation of the corresponding lung tissue, is quantified. Deformable image registration assumes that a reference voxel in the EI image cube moves to its new position in the LI image cube, deforming at the same time as more air is drawn into the lung tissue. As the respiratory cycle moves forward, the reference voxel maintains its motion trajectory, as shown in the MI, EE, LE, and ME image cubes, respectively. In other words, for each voxel in the reference image cube, a set of displacement vectors is calculated across all six image cubes.

[0218] Fig.26B It is shown that there are five deformable registrations (2→1, 3→1, 4→1, 5→1, 6→1) between the EI image cube and the other five image cubes. These six image cubes represent the thoracic anatomy of the patient at six corresponding predefined respiratory phases. For each voxel, there is a vector at the image cube corresponding to each of the six respiratory phases. Assuming that the vector corresponding to the EI image cube is zero, the six displacement vectors for a given voxel can be expressed as:

[0219] Or expressed as

[0220] The parameter "n" is a time-dependent parameter corresponding to the corresponding breathing phase.

[0221] Similarly, for a particular respiratory phase “n”, the tidal volume (T v )、airflow(A f ) and cardiac stage (H c ) can be expressed as:

[0222] Or expressed as [B n ].

[0223] For each voxel in the image cube, the parameter matrix The biomechanical model can use six displacement vectors and the corresponding biometric data matrix described above Solve. The four vectors describe the tissue properties of the voxel, which control the displacement vector and deformation of the voxel. For each voxel in the reference image cube, we have:

[0224]

[0225] As noted above, there are many deformable image registration algorithms that can perform image registration for GREX imaging, including 3D multi-resolution optical flow algorithms. The 3D multi-resolution optical flow algorithm computes smooth (e.g., fluid-like) transitions between images acquired at different observed tidal volumes. The calculated displacement between two respiratory phases supply Fig.24 Once the parameters of the biomechanical model are By solving, for example, the least squares regression, changing the tidal volume (T v )、airflow(A f ) and cardiac stage (H c ) will produce the entire closed-loop trajectory. Using the closed-loop trajectory for all tissue elements in the thorax gives a new image cube that can be considered as a biometrically interpolated image cube. In summary, Table 1 shows each component of the GREX biomechanical model and how to find the component.

[0226]

[0227] Post-processing software 106 fills in all potential breathing phases by biometric interpolation between acquired image cubes, thereby creating a complete chest image. Generally, an image requires at least 30 simulated images to smoothly transition between frames. Fig.28A FIG. 28 is a schematic flow chart 2800 for creating an image by biometric interpolation according to some embodiments of the present disclosure. Fig.28B As shown in , it is assumed that the biometric data matrix corresponding to the EI image cube is:

[0228] T v =20ml, (A f )=20ml / s, and H c =0.10.

[0229] For each voxel in the intermediate simulated image cube at a particular moment in the respiratory cycle, e.g., EI+Δt, EI+2Δt, EI+3Δt, etc., the parameter matrix The biomechanical model and the corresponding biometric data matrix at a specific moment are calculated.

[0230] Section 3.4 - Parameter Diagram

[0231] One clinical benefit of GREX imaging technology is the unique parametric maps. Using 2D maps, 2D colorwashes, 3D maps, and 3D vector field mapping, GREX imaging technology presents previously unavailable information to the end user about the health of the patient's chest.

[0232] Fig.29Depicting an example of a standard radiograph of a healthy patient (left) and an example of a standard radiograph of a patient with a stage 1b left upper lung tumor (as indicated by the arrow). Both figures show standard radiographs currently used in the field of radiography. Fig.29 The lesion on the right side of the left lung is an early stage lung tumor, but it is not easily seen on standard radiographs, which are difficult to interpret and only show anatomical information. Fig.30 Depicted are accompanying GREX parameter maps showing health status indicators for the same two patients in accordance with some embodiments of the present disclosure. The parameter map on the right clearly indicates a lesion in the patient's left upper lung, whereas a standard radiograph is blurry. Changing the window level and image contrast ultimately reveals a poorly ventilated area in the left upper lung. However, if the end user does not take these steps, the chances of capturing this early lung cancer tumor are low. In other words, the parameter maps generated by GREX can greatly reduce the variation in the end user's risk of disease underreporting.

[0233] First consider Fig.30 In A 2D color map of , which describes the normal stress with respect to tidal volume. By definition, the end user will expect that the lungs will be closer to the diaphragm than to the apex. The magnitude of will be larger (e.g., the diaphragm has a larger tissue displacement than the apex of the lung when tidal volume increases). The value of varies smoothly throughout the healthy lung. The gradient will be a smooth function. Fig.30 The example described in shows that for both patients, although The parameters were distributed in a similar manner, but the tissue displacement in healthy patients was twice that in diseased patients.

[0234] Next, consider the sum of the normal and shear stresses associated with the airflow 2D coloring of . In general, higher parameter values ​​occur near areas where the bronchial tree brings air into the lungs at a faster rate (mid-lung). However, in diseased lungs, the presence of a tumor can substantially alter the elastic behavior of the lung, making the tumor visually distinguishable from the distribution of healthy lung tissue. Fig.30 In the example shown, the presence of a lung tumor is clearly visible in the left lung due to the significant difference in magnitude. When disease affects an area of ​​the lung, other areas of lung tissue “pick up the mess” by ventilating healthy areas instead of the diseased areas. Fig.29As shown in the left figure of the figure, the healthy lung is more elastic than the diseased lung. In other words, the airflow resistance is greatly increased in the location of the tumor, thereby highlighting the presence of the tumor and enabling quantitative analysis of the tumor's impact on the patient's ability to properly ventilate during breathing. The ratio between the parameters associated with tidal volume and the parameters associated with airflow dramatically shows that the left upper lung has a pattern that is significantly different from that of a healthy patient. Essentially, the ratio of the parameters associated with tidal volume to the parameters associated with airflow is different from that of a healthy patient. Fig.24 Compared to the example tissue trajectory shown in , a higher percentage is interpreted as tissue moving in a more circular pattern. The tidal volume attributed to the 2D brush is and airflow The histogram of the observed total motion component also shows clear indicators of left lung disease. A healthy lung has The GREX imaging technique can involve a Bayesian process to better classify the etiology of disease using patient interviews (information voluntarily provided by the patient before the examination) combined with the distribution of the parameter histograms of the biomechanical model. In summary, lung disease (e.g., tumor presence, e.g., Fig.30 Consistent indicators of chromatin (as shown) greatly aid in early detection and disease diagnosis.

[0235] Section 3.5 - Diagnostic Disease Indicators

[0236] Lung cancer, chronic obstructive pulmonary disease (COPD), lower respiratory tract infections, and tuberculosis all have disease markers that are visible in existing digital diagnostic x-ray imaging. However, diagnostic disease markers may not always be visible in the early stages of the disease. Fig.29 and 30 As shown, GREX parameters have the potential to be used as diagnostic disease indicators in the early stages of the disease. If the medical team only considers standard radiographs, the presence of lung cancer may not be detected because lung cancer tumors are not large enough to appear clearly in standard radiographs. GREX imaging uses biometric signals combined with multiple biometrically targeted radiographs to diagnose the disease using previously unavailable information about how the lungs move. The GREX process directly addresses the problem of early lung cancer, that is, lung cancer often has poor specificity in digital diagnostic x-ray images. Existing diagnostic sensitivity limitations based on digital diagnostic x-rays are solved by using imaging that is biometrically informed by the patient's unique and disease-indicating respiratory information. Clinicians can observe diagnostic deep aspects as early as the onset of the disease through lung tissue motion visualization and parameter maps, and continue to monitor patients as the disease progresses. Machine learning and finite element analysis techniques will be used to discover subtle patterns or baseline changes in lung function to assist end users in their ability to detect the disease at the onset of the disease, thereby providing patients with a higher chance of survival and more treatment options.

[0237] Another diagnostic example of GREX’s impact on the diagnostic community is in classifying the causes of lung disease. In a biopsy, lung disease caused by asbestos and lung disease caused by cigarettes appear differently. Asbestos is a naturally occurring silicate mineral that consists of long, thin, fibrous crystals made up of millions of microscopic fibers. Inhaled asbestos fibers penetrate the alveoli and eventually form a dense web that impairs alveolar function, which in turn reduces lung function. The dense web collects cancerous tissue and is called mesothelioma. Standard radiographs can only identify mesothelioma by plaque accumulation, which appears as dense consolidation (blurring) in the lung image. GREX can detect traces of asbestos through parametric maps before the end user can visually identify the presence of asbestos. As the asbestos web forms, the elasticity of the lung is locally reduced. Small localized elasticity reductions can be seen in the parametric maps (2D brushing, parameter ratios, and histograms). For example, in tissue that has lost elasticity, the trajectory of the tissue during breathing will be more round than elliptical. This means that, and The ratio between the magnitudes will be higher than in healthy tissue. Review of the color map will clearly show areas of the lung that display patterns indicative of disease.

[0238] Inhaled particles (non-fibers) will present in a completely different way than fibrous asbestos particles. Inhaled particles are deposited in the lungs, thereby reducing lung function by "clogging" the airways, forming scar tissue and forming tumors rather than nets. GREX will show the beginning of particle deposition by detecting small local changes in lung motion dynamics that are inconsistent with healthy characteristics. Cigarettes contain tar and radon, which adhere to the alveoli and completely change lung function. GREX can track reduced lung function and provide end users with a unique tool to better show patients the degree of damage caused by smoking habits. COPD is another disease that can be explained more effectively by parametric maps than current clinical methods. COPD is currently detected by spirometry tests and standard radiography. None of these methods are particularly sensitive and cannot detect COPD at the initial onset. Detecting the disease at the earliest possible time allows patients to have more opportunities to use preventive medicines and change bad habits before it is too late.

[0239] Chapter 3.6 - Image Presentation

[0240] The focus of the image presentation graphical user interface (GUI) is to provide a clear and intuitive canvas for displaying planar images and 3D image results. All functions such as rotating views, pausing images and rulers are placed in a bin represented by a tab at the top of the screen. Selecting a function will change the tool tip to show the user the function being used. The GUI is designed to be lightweight so that even with an older computer system, the GUI can operate normally. Finally, the image presentation GUI will work together in conjunction with annotation / configuration, possible disease pointers and cloud computing platforms.

[0241] Medical end users prefer to highlight and place notes directly on medical images rather than attaching additional files to the image. Medical end users strongly request the ability to include image annotations (“eye-catching” arrows and treatment notes) and anatomical contours (drawing virtual lines around anatomical structures) on imaging studies. Medical end users are often members of large multidisciplinary care teams that collaborate to treat patients. The following examples describe the current workflow and technology capabilities of clinical care teams:

[0242] Once an imaging study is performed on a lung cancer patient, the images are read by a radiologist, radiation therapy is ordered by a radiation oncologist, and used as the basis for surgical treatment by the surgeon. Specifically, the radiologist draws a contour around the tumor site and transmits the image of the configuration to the radiation oncologist.

[0243] The radiation oncologist then writes orders for radiation therapy and / or tumor removal surgery based on the images.

[0244] Using radiographic contouring and radiation oncology tumor resection instructions, the surgeon removes the tumor.

[0245] The post-processing software 106 provides end users with tools to annotate and contour directly on the moving image, which simplifies healthcare workflow. This user-friendly capability enhances healthcare workflow and reduces the probability of medical errors. The interactive nature of our user interface means that treatment instructions and / or medical questions are now clearly visible to all users, and the relevant clinical notes are correctly displayed in the corresponding anatomical area of ​​interest.

[0246] Figures 31A to 31B 3100 is a flow chart illustrating a method 3100 for imaging a patient's lungs. In some embodiments, any or all of the operations described below may be performed without human intervention (e.g., without intervention by a technician). In some embodiments, the method 3100 is performed by or using any of the devices described herein (e.g., Figure 1 3100). Some operations of method 3100 are performed by a computer system comprising one or more processors and a memory storing instructions, which, when executed by the one or more processors, cause the one or more processors to perform the operations of method 3100. Some operations in method 3100 are optionally combined, and / or the order of some operations is optionally changed.

[0247] The method includes positioning (3102) a patient relative to an x-ray imaging device (e.g., a GREX imaging system, Figure 1). In some embodiments, method 3100 is performed at a conventional medical imaging system that is modified to perform certain operations described below. In some embodiments, positioning the patient in the first orientation includes moving (e.g., rotating) the patient to the first orientation (e.g., as described in reference Fig.35 ), while the x-ray imaging system (e.g., the x-ray unit and the detector panel) remains in a fixed position. For example, in some embodiments, the patient sits or stands on a patient positioning fixture (PPF), such as reference Fig.35 PFF 3501 described. In some embodiments, positioning the patient in a first orientation relative to the x-ray imaging device includes rotating a patient positioning fixture. For example, in some embodiments, at the beginning of method 3100, the patient positioning fixture is rotated so that the sagittal plane of the patient is positioned at a predefined angle relative to an optical axis of the x-ray imaging device (e.g., an axis along which the x-ray imaging device emits x-rays). In some embodiments, the predefined angle is selected from the group consisting of -45 degrees, -22.5 degrees, 0 degrees, 22.5 degrees, and 45 degrees.

[0248] In some embodiments, positioning the patient in the first orientation includes moving (e.g., rotating) the x-ray imaging system while the patient remains in a fixed orientation (e.g., as shown in FIG. Fig.17 discussed).

[0249] In some embodiments, when the patient is positioned in a first orientation relative to the x-ray imaging device, the patient positioning fixture maintains the patient in a fixed orientation (e.g., stabilizes the patient's orientation) such that a three-dimensional image of the patient's lungs can be reconstructed based on a "stationary object" assumption (e.g., as described throughout this disclosure).

[0250] The method includes obtaining (3104) a volume measurement of the patient's breathing (e.g., while the patient is breathing normally; the patient need not maintain their width). In some embodiments, the volume measurement of the patient's breathing is a measurement (e.g., a direct measurement) of the patient's lung volume (e.g., instantaneous lung volume) or a derivative of the patient's lung volume (e.g., flow rate). In some embodiments, as explained in more detail below, the volume measurement of the patient's breathing is a measurement that can be converted into the patient's tidal volume (e.g., by measuring chest cavity rise and fall). As used herein, the term tidal volume refers to the difference between the current lung volume and a predefined benchmark (e.g., the volume during maximum exhalation of normal breathing without additional effort, the volume during maximum inspiration of normal breathing without additional effort, or any other suitable benchmark volume).

[0251] In some embodiments, the volumetric measurement of the patient's respiration is a geometric (spatial or directional) measurement of the patient's respiration.

[0252] In some embodiments, the volumetric measurement of the patient's breathing includes (3106) a measurement of the patient's chest rise (and / or chest fall). In some embodiments, the patient's breathing volumetric measurement includes (3106) a measurement of the patient's chest rise (and / or chest fall). In some embodiments, the patient's breathing volumetric measurement includes (3106) a measurement of the patient's chest rise (and / or chest fall). In some embodiments, the patient's breathing volumetric measurement includes (3106) a measurement of the patient's chest rise (and / or chest fall). In some embodiments, the patient's breathing volumetric measurement includes (3108 ... Figure 1 ) to obtain a volume measurement of the patient's breathing. In some embodiments, the method further includes creating (3110) a point cloud of the patient's chest surface. The volume measurement of the patient's breathing is determined based on the point cloud of the patient's chest surface.

[0253] In some embodiments, the 3D point cloud is used to determine a volume measurement of the patient's respiration without generating a mesh reconstruction of the patient's chest (e.g., the raw output of the 3D point cloud is used to generate a volume measurement of the patient's respiration without first generating a mesh). In some embodiments, one or more locations of the patient's chest are measured using (3112) 3D imaging technology to obtain a point cloud of the patient's chest surface. For example, the 3D imaging technology includes laser scanning technology such as light detection and ranging (LIDAR). Such laser scanning technology is beneficial because some lasers can accurately measure the orientation of the patient's chest even when the patient is wearing a modesty garment (e.g., the patient wears a LiDAR-permeable garment when performing method 3100).

[0254] In some embodiments, method 3100 further includes identifying (3114) one or more anatomical landmarks on the surface of the patient's chest using the point cloud of the surface of the patient's chest (e.g., externally visible landmarks such as the clavicle, trachea, supraorbital bone, sternum, xiphoid process, spine, or humor). The method further includes inferring the location of one or more internal anatomical landmarks within the patient's chest (e.g., the location of the patient's lungs) based on the point cloud of the surface of the patient's chest. In some embodiments, the method includes generating a reconstruction of the patient (e.g., generating a computer model of the patient, also referred to as a "virtual patient"). In some embodiments, generating a reconstruction of the patient includes generating a reconstruction of the patient's internal anatomical structure. In some embodiments, the reconstruction of the patient's internal anatomical structure includes a computer model of tissue density (e.g., a 3D model). In some embodiments, the reconstruction includes or is used to determine absorption cross-sections (e.g., x-ray absorption cross-sections) at multiple locations within the body. In some embodiments, the reconstruction of the patient is used to determine an x-ray dose to be delivered by an x-ray imaging device (e.g., for each image).

[0255] The method includes determining (3116) a respiratory phase of the patient based on a volumetric measurement of the patient's breathing (e.g., in real time while the patient is breathing normally) while the patient is positioned in a first orientation relative to the x-ray imaging device and while a volumetric measurement of the patient's breathing is obtained. In some embodiments, the respiratory phase of the patient is defined by a volume of the lungs. Thus, in some embodiments, determining the respiratory phase of the lungs includes determining the volume of the lungs.

[0256] In some embodiments, operation 3102, etc. is performed as part of an imaging session of method 3100. Method 3100 also includes undergoing a training session prior to the imaging session, during which information about the patient's normal breathing is obtained. For example, during the training session, volume measurements of the patient's breathing are obtained at regular intervals over multiple cycles of the patient's breathing (e.g., 15, 20, 50 cycles of the patient's breathing, each cycle corresponding to one breath). Next, specific volume measurements are associated with specific breathing stages using the volume measurements from the training session. For example, a volume measurement corresponding to a tidal volume of 400 ml may be associated with maximum inspiration, a volume measurement corresponding to a tidal volume of 0 ml may be associated with maximum exhalation, and so on. In addition, statistical data (e.g., a histogram) about the patient's breathing may be obtained during the training session and used to verify that the breathing obtained during imaging is "normal" breathing (e.g., not excessive deep breathing or otherwise abnormal breathing).

[0257] During the imaging period, in some embodiments, the determined respiratory phase of the patient is (3118) a future (e.g., predicted) respiratory phase. That is, in some embodiments, determining the respiratory phase of the patient based on volume measurements of the patient's breathing includes predicting the future respiratory phase based on one or more current and / or past respiratory phases. For example, the prediction is based on a time series of respiratory phases. In some embodiments, predicting the future respiratory phase based on one or more current and / or past respiratory phases includes generating an autoregressive integrated moving average (ARIMA) model. In some embodiments, the prediction uses data from a training period.

[0258] The method includes, based on determining that the patient's breathing stage matches a predefined breathing stage (e.g., the volume of the lungs matches a predefined lung volume), gating (3120) an x-ray imaging device to generate x-ray projections (sometimes referred to as projection images) of the patient's lungs. In some embodiments, the x-ray projections are images taken at a particular angle (e.g., determined by the orientation of the patient relative to the x-ray imaging device). In some embodiments, the x-ray projections are obtained using a single x-ray exposure. In some embodiments, the method includes, based on determining that the patient's breathing stage does not match a predefined breathing stage, stopping gating the x-ray imaging device (e.g., stopping exposing the patient to x-ray radiation).

[0259] In some embodiments, the method includes determining whether the current breathing is irregular breathing based on the volume measurement of the patient's breathing. The method also includes, based on determining that the current breathing is irregular breathing, stopping gating (e.g., maintaining) the x-ray imaging device (e.g., continuing to wait for suitable breathing to obtain x-ray projections of the breathing phase).

[0260] In some embodiments, the predefined breathing phase is (3122) a first predefined breathing phase of a plurality of predefined breathing phases. In some embodiments, the method further comprises, while obtaining a volumetric measurement of the patient's respiration, gating an x-ray imaging device to generate corresponding x-ray projections of the patient's lungs based on determining that the patient's respiration phase matches any one of the plurality of predefined respiration phases. In some embodiments, an x-ray projection (e.g., x-ray measurement) of the patient's lungs is obtained (3124) only when the patient's respiration phase determined by the volumetric measurement of the patient's respiration matches one of the plurality of predefined respiration phases, thereby reducing the patient's total x-ray exposure.

[0261] In some embodiments, the plurality of predefined respiratory phases includes (3126) an early expiratory phase, a late expiratory phase, a maximum expiratory phase, an early inspiratory phase, a late inspiratory phase, and a maximum inspiratory phase of a complete respiratory cycle of the patient (e.g., as described in reference to Fig.14 In some embodiments, x-ray projections are obtained for each of a plurality of predefined breathing phases while the patient is positioned in a first orientation relative to the x-ray imaging device.

[0262] In some embodiments, the x-ray projection is (3128) a first x-ray projection, and the method further includes repositioning the patient to a second orientation relative to the x-ray imaging device (e.g., by rotating the patient or rotating the x-ray imaging system). In some embodiments, the method includes, while the patient is positioned at the second orientation relative to the x-ray imaging device, and while continuing to obtain a volumetric measurement of the patient's respiration, continuing to determine a respiratory phase of the patient based on the volumetric measurement of the patient's respiration. In some embodiments, the method includes, based on determining that the patient's respiration phase matches a predefined respiratory phase, gating the x-ray imaging device to generate a second x-ray projection of the patient's lungs.

[0263] In some embodiments, the method further comprises using the first x-ray projection and the second x-ray projection (e.g., as described in reference Fig.21In some embodiments, the static image cube is a three-dimensional reconstruction of the patient's lung volume. In some embodiments, x-ray projections are obtained for each of a plurality of predefined respiratory stages while the patient is positioned in each of a plurality of directions (including a first direction and a second direction) relative to the x-ray imaging device. In some embodiments, the plurality of directions includes at least 5 directions (e.g., -45 degrees, -22.5 degrees, 0 degrees, -22.5 degrees, and 45 degrees). In some embodiments, the plurality of directions includes more than 5 directions (e.g., 6, 7, 8, or more directions). In some embodiments, when the patient is positioned in each of a plurality of directions (including a first direction and a second direction) relative to the x-ray imaging device, no x-ray projections are obtained except for each of a plurality of predefined respiratory stages. Thus, when x-ray projections are obtained for six different respiratory stages in five directions, a total of thirty x-ray projections are obtained (e.g., as described above, the x-ray imaging device is gated to obtain only these images).

[0264] As described elsewhere in this document, these thirty x-ray projections can be used to reconstruct a biomechanical model of how the lungs move in 3D. In some embodiments, less than ten x-ray projections acquired from various angles during a predefined breathing phase are used (3130) to generate a static image cube corresponding to the predefined breathing phase (e.g., as described in reference Fig.21 described).

[0265] Those skilled in the art will recognize that method 3100 may also be applied to other types of movement in addition to lung movement due to breathing. For example, in some embodiments, a method includes positioning a patient in a first orientation relative to an x-ray imaging device. The method also includes obtaining a 3D measurement of a portion of the patient's body (e.g., a 3D measurement of the position of the portion of the patient's body). The method also includes, when the patient is positioned in the first orientation relative to the x-ray imaging device, and when obtaining the 3D measurement of the portion of the patient's body: determining, based on the 3D measurement of the portion of the patient's body, that a trigger criterion for triggering exposure to radiation performed by the x-ray imaging device is met; and based on the determination that the trigger criterion is met, gating the x-ray imaging device to generate an x-ray image of the patient. In some embodiments, the x-ray image is an image of a portion of the patient's body (e.g., a leg, abdomen, skull, etc. of the patient). In some embodiments, the trigger criterion includes a criterion that is met when the 3D measurement of the portion of the patient's body indicates that the portion of the patient's body is in a predefined position (e.g., relative to the imaging device). In some embodiments, the method includes, based on the determination that the trigger criterion is not met, stopping gating the x-ray imaging device (e.g., stopping exposing the patient to x-ray radiation). Furthermore, in some embodiments, method 3100 is applicable to other types of imaging that are not strictly x-ray based, such as positron emission tomography (PET) imaging or MRI imaging.

[0266] In addition, it should be understood that method 3100 can be applied to radiation therapy as well as radiation imaging. For example, in some embodiments, a method includes positioning a patient in a first orientation relative to a radiation therapy source. The method also includes obtaining a 3D measurement of a portion of the patient's body (e.g., a 3D measurement of the position of the portion of the patient's body). The method also includes, when the patient is positioned in the first orientation relative to the radiation therapy source, and when the 3D measurement of the portion of the patient's body is obtained: based on the 3D measurement of the portion of the patient's body, determining that a trigger criterion for triggering exposure to radiation performed by the radiation therapy source is met; and based on determining that the trigger criterion is met, gating the radiation therapy source to expose the patient to radiation (e.g., exposing the portion of the patient's body to radiation). In some embodiments, the trigger criterion includes a criterion that is met when the 3D measurement of the portion of the patient's body indicates that the portion of the patient's body is in a predefined position (e.g., relative to an imaging device). In some embodiments, the method includes, based on determining that the trigger criterion is not met, stopping exposing the patient to radiation.

[0267] It should be understood that the Figures 31A to 31BThe specific order of operations in the description is merely an example and is not intended to indicate that the described order is the only order in which the operations may be performed. Those skilled in the art will recognize various ways to reorder the operations described herein. In addition, it should be noted that the details of other processes described herein with respect to other methods described herein also apply in a similar manner to the above description of Figures 31A to 31B The method 3100 described. For example, reference Figure 2 , Figure 6 , Figure 8 , Fig.11 , Fig.13 , Fig.15 , Fig.21 , Fig.23 , Fig.25 , Fig. 27 , Figures 28A to 28B , Figures 32A to 32B , Figures 33A to 33C and Figures 34A to 34C For the sake of brevity, these details are not repeated here.

[0268] Figures 32A to 32B 3200 is a flow chart illustrating a method 3200 for gating a radiation source when the patient's respiratory phase and cardiac phase coincide. In some embodiments, any or all of the operations described below can be performed without human intervention (e.g., without intervention by a technician). In some embodiments, the method 3200 is performed by or using any of the devices described herein (e.g., Figure 1 3200). Some operations of method 3200 are performed by a computer system comprising one or more processors and a memory storing instructions, which, when executed by the one or more processors, cause the one or more processors to perform the operations of method 3200. Some operations in method 3200 are optionally combined, and / or the order of some operations is optionally changed.

[0269] The method includes positioning (3202) a patient in a first orientation relative to a radiation source. In some embodiments, the radiation source is (3204) an x-ray imaging device. In some embodiments, the radiation source is (3206) a radiotherapy source. For example, Figure 3 and Fig.35 As shown, a patient (e.g., patient 3502) is positioned in a first orientation relative to a radiation source 3504 (e.g., x-ray unit 108). In some embodiments, as shown in reference operation 3102 ( Fig.31A ) described, positioning the patient includes moving (e.g., rotating) the patient, and in some embodiments, positioning the patient includes moving (e.g., rotating) an x-ray device (e.g., an x-ray source and a detector).

[0270] The method includes obtaining (3208) a measurement of the patient's breathing (e.g., using a breathing phase sensor). In some embodiments, the measurement of the patient's breathing is the same as that described above with reference to operation 3104 ( Fig.31A ) is a volumetric measurement of a patient's respiration as described herein. In some embodiments, the measurement of the patient's respiration is a non-volumetric measurement of the patient's respiration (e.g., a time-based measurement of the patient's respiration).

[0271] The method includes obtaining (3210) a measurement of a patient's cardiac function. In some embodiments, one or more sensors (3304) are used to measure the patient's cardiac function. In some embodiments, an electrocardiogram (ECG) is used to measure the patient's cardiac function (e.g., a 3-lead or 12-lead ECG). In some embodiments, the method includes obtaining a plurality of measurements of the patient's cardiac function that provide a time series of electrical signals that control movement of the patient's heart.

[0272] Returning to the imaging phase, the method includes, while the patient is positioned at a first orientation relative to the radiation source and while obtaining a measurement of the patient's respiration (3212), determining (3214) the patient's respiratory stage based on the measurement of the patient's respiration and determining (3216) the patient's cardiac stage based on the measurement of the patient's cardiac function (e.g., in real time). For example, in some embodiments, the patient's cardiac stage is determined using a zero instruction set computer (ZISC) processor, as described in reference to Fig.36 The ZISC processor is capable of identifying, in a single cardiac cycle or a portion of a cardiac cycle, that a predefined marker in the cardiac cycle has occurred (eg, an S wave or a T wave).

[0273] In some embodiments, the method further includes obtaining (3218) measurements of the patient's cardiac function from a plurality of cardiac cycles of the patient prior to gating the radiation source to expose the patient to radiation (operation 3222 below). In some embodiments, the method includes determining an average interval between a predefined cardiac phase and the beginning of a predefined window of the cardiac cycle using the measurements of the patient's cardiac function from the plurality of cardiac cycles. For example, the predefined window of the cardiac cycle represents the interval between the top of the R wave and the beginning of the gating window. For example, in some embodiments, operation 3202 is performed as part of an imaging session of method 3200, etc. Method 3200 further includes undergoing a training period prior to the imaging session, during which information about the patient's cardiac function is obtained. For example, during the training period, ECG measurements of the patient's cardiac function are obtained at regular intervals over a plurality of cardiac cycles (e.g., 15, 20, 50 cardiac cycles, where one cycle corresponds to a complete period of cardiac motion, e.g., from one T wave to the next T wave). The ECG measurements from the training period are then used to predict periods of quiescence of cardiac motion during the imaging phase, as described below.

[0274] In some embodiments, the measurements of multiple cardiac cycles from the patient are (3220) waveform measurements of the multiple cardiac cycles (e.g., ECG measurements), and the method includes validating that the waveform measurements of the multiple cardiac cycles are statistically stable.

[0275] The method also includes gating (3222) the radiation source to expose the patient to radiation based on determining that the patient's respiratory phase matches a predefined respiratory phase and determining that the patient's cardiac phase matches a predefined window of the cardiac cycle. In some embodiments, when the radiation source is gated, the patient's lungs are exposed to radiation. In some embodiments, the predefined cardiac window corresponds to a quiescent period of cardiac movement (e.g., a period of time during a cardiac cycle when cardiac movement is minimal, as described in reference to FIG. 32). Fig.12 In some embodiments, the radiation source is gated within the same cardiac cycle as the determined cardiac phase. In this way, by gating the exposure of the patient's lungs based on the coincidence of the respiratory phase and the predefined cardiac phase window, a precise area of ​​the lungs (e.g., a precise area of ​​lung tissue) is exposed to radiation without disturbance or motion caused by cardiac movement.

[0276] In some embodiments, the radiation source is an x-ray imaging device. Gating the radiation source to expose the patient to radiation includes (3224) gating the x-ray imaging device to produce x-ray projections of the patient's lungs. In some embodiments, as described in reference Figures 31A to 31B As described in method 3100 of , x-ray projections may be obtained in this manner for multiple orientations of the patient relative to an x-ray imaging source and multiple breathing phases. These x-ray projections may then be used to generate images of lung motion and / or a biophysical model of the lung (e.g., by correlating the motion of lung tissue with biophysical parameters such as stress, strain elasticity, etc.). In some embodiments, according to method 3200, these images are acquired within a predefined cardiac window to minimize disturbance or motion of the lung due to cardiac movement.

[0277] In some embodiments, the radiation source is a radiotherapy source. Gating the radiation source to expose the patient to radiation includes (3226) gating the radiotherapy source to irradiate a region of the patient's lung with a therapeutic dose. In these situations, it is important to provide as much radiation dose as possible to diseased tissue (e.g., cancerous tissue) and as little radiation dose as possible to healthy tissue. Method 3200 improves radiotherapy devices by delivering a more accurate dose to diseased tissue while minimizing radiation provided to healthy tissue.

[0278] In some embodiments, determining that the patient's cardiac phase matches a predefined window of the cardiac cycle comprises detecting in real time (e.g., using Fig.36) and waits for a certain period of time to predict (3228) a predefined window of the cardiac cycle, the period corresponding to the average interval between the predefined cardiac phase and the beginning of the predefined window of the cardiac cycle. For example, in some embodiments, the peak of the T wave is detected, and the average interval from the T wave to the ideal gating window (when the radiation source is gated) is waited.

[0279] It should be understood that the Figures 32A to 32B The specific order of operations in the description is merely an example and is not intended to indicate that the described order is the only order in which the operations may be performed. Those skilled in the art will recognize various ways to reorder the operations described herein. In addition, it should be noted that the details of other processes described herein with respect to other methods described herein also apply in a similar manner to the above description of Figures 32A to 32B The method 3100 described. For example, reference Figure 2 , Figure 6 , Figure 8 , Fig.11 , Fig.13 , Fig.15 , Fig.21 , Fig.23 , Fig.25 , Fig. 27 , Figures 28A to 28B , Figures 31A to 31B , Figures 33A to 33C and Figures 34A to 34C For the sake of brevity, these details are not repeated here.

[0280] Figures 33A to 33C is a flow chart of a method 3300 for generating a model of mechanical properties of a lung by fitting data from registered images, according to some embodiments. In some embodiments, any or all of the operations described below may be performed without human intervention (e.g., without intervention by a technician). In some embodiments, method 3300 is performed by or using any of the devices described herein (e.g., Figure 1 3300). Some operations of method 3300 are performed by a computer system comprising one or more processors and a memory storing instructions, which, when executed by the one or more processors, cause the one or more processors to perform the operations of method 3300. Some operations in method 3300 are optionally combined, and / or the order of some operations is optionally changed.

[0281] The method includes extracting (3302) a plurality of displacement fields of lung tissue from a plurality of x-ray measurements (e.g., x-ray images, also referred to as x-ray projections) corresponding to different respiratory phases of the lung. Each displacement field represents movement of lung tissue from a first respiratory phase to a second respiratory phase, and each respiratory phase has a corresponding set of biometric parameters. In some embodiments, the x-ray measurements are x-ray projections (also referred to as x-ray projection images) obtained according to method 3100 and / or method 3200. In some embodiments, extracting the displacement field from the lung tissue includes identifying a portion of the lung tissue in one or more first x-ray projections from the first respiratory phase that corresponds to the same portion of the lung tissue in one or more second x-ray projections from the second respiratory phase, and determining the displacement of the portion of the lung tissue from the first respiratory phase to the second respiratory phase (e.g., with reference to FIG. 3104 ). Fig.28B In some embodiments, the identifying is performed using a deformable image registration algorithm, as discussed above.

[0282] In some embodiments, the plurality of x-ray measurements includes a plurality of x-ray images, including an x-ray image acquired for a first respiratory phase at each of a plurality of directions of an x-ray imaging device relative to the patient, thereby forming a plurality of x-ray images corresponding to the first respiratory phase and a plurality of x-ray images corresponding to a second respiratory phase. In some cases, at least one of the x-ray images corresponding to the first respiratory phase is acquired during a different respiratory cycle of the patient than a different x-ray image corresponding to the first respiratory phase (e.g., the image is acquired during the same phase but in a different breath). In some embodiments, the method includes grouping the plurality of x-ray images by respiratory phase.

[0283] In some embodiments, the method includes extracting a plurality of vector fields, of which a displacement field is one example.

[0284] In some embodiments, one or more sensors are used (3304) to measure a biometric signal of the patient as one or more sequences of time series, wherein the one or more sensors include a 3D spatial position locator (e.g., 3D spatial positioner 300, Figure 3), a respiratory phase sensor, and a cardiac phase sensor. In some embodiments, the 3D spatial position locator is configured (3306) to measure the real-time body movements of the patient caused by breathing and heartbeat and output them as a time series. In some embodiments, the respiratory phase sensor is configured (3308) to measure one or more physiological measurements related to the patient's breathing, including tidal volume and its first-order time derivative. In some embodiments, the cardiac phase sensor is configured (3310) to measure periodic and stationary electrical signals (e.g., ECG signals) generated by the patient's heart. For example, the cardiac phase sensor measures periodic and stationary electrical signals having characteristic features corresponding to the phase of the heartbeat.

[0285] In some embodiments, biometric signals of the patient measured by one or more sensors are used to trigger (3312) an x-ray unit to acquire x-ray images of the patient at specific respiratory and cardiac stages (eg, as described with reference to method 3200).

[0286] In some embodiments, the x-ray unit includes (3314) a clock, and the biometric signals of the patient measured by the one or more sensors are synchronized with the clock of the x-ray unit. In some embodiments, corresponding values ​​of the biometric signals are recorded as associated with the acquired x-ray images.

[0287] In some embodiments, the patient's biometric signals measured during the training window are used (3316) to construct an optimized respiratory prediction model for predicting a desired respiratory phase at which the x-ray unit is triggered to capture x-ray images of the patient.

[0288] The method includes calculating (3318) one or more biophysical parameters of a biophysical model of the lung using a plurality of displacement fields of lung tissue between different respiratory phases of the lung and a corresponding set of biometric parameters. In some embodiments, calculating the one or more biophysical parameters includes calculating one or more derivatives of the displacement field (e.g., curl, gradient, etc.). In some embodiments, the biophysical parameters are biomechanical parameters (e.g., stress, strain, elastic modulus, elastic limit, etc.). In some embodiments, the one or more biophysical parameters define (3320) a physical relationship between the biometric parameters associated with different respiratory phases of the lung and the plurality of displacement fields of the lung tissue. In some embodiments, the set of biometric parameters associated with the respective respiratory phases includes (3322) tidal volume and airflow of the lung in the respective respiratory phase and a cardiac phase corresponding to the respective respiratory phase of the lung. In some embodiments, the physical relationship between the biometric parameters associated with the respective respiratory phases of the lung and the plurality of displacement fields of the lung tissue is defined as follows:

[0289]

[0290] vector describes the normal stress caused by tidal volume, describes the normal stress caused by airflow, describes the shear stress caused by the airflow, and Describes the tissue movement caused by cardiac motion, the displacement of tissue at any point in the closed-loop trajectory Expressed as tidal volume (T v )、airflow(A f ) and cardiac stage (H c ) is the sum of scaled stress, strain, and perturbation cardiac motion vectors.

[0291] In some embodiments, the different respiratory phases of the lungs include (3324) early exhalation, late exhalation, maximal exhalation, early inspiration, late inspiration, and maximal inspiration of a complete respiratory cycle of the patient.

[0292] In some embodiments, the method further comprises displaying a visual representation of the biophysical parameter. Fig.30 Provides biophysical parameters where and In some embodiments, displaying the visual representation includes displaying an image of the lungs, wherein the color of locations within the image of the lungs corresponds to the biophysical parameter (e.g., displaying the image of the lungs using a color map of the biophysical parameter). Displaying such a visual representation improves the x-ray imaging device itself by increasing the accuracy of the diagnosis. For example, compared to Fig.29 The conventional radiograph shown in Fig.30 The patient's left lung lesion is more easily seen in the image.

[0293] In some embodiments, the method further comprises determining the number of x-ray measurements corresponding to different respiratory phases of the lung (e.g., reference Fig.21 In some embodiments, multiple displacement fields of the lung tissue are extracted (3328) from the multiple medical image cubes corresponding to different respiratory stages of the lung by outlining the lung tissue from the rest of the first medical image cube using image segmentation, and, for corresponding voxels in the first medical image cube, determining a displacement vector between voxels in the first medical image cube and the second medical image using an intensity-based structure mapping between the first medical image cube and the second medical image. Multiple displacement fields of the lung tissue are also extracted from the multiple medical image cubes corresponding to different respiratory stages of the lung by iteratively refining the displacement vectors of different voxels in the first medical image cube and their corresponding parts in the second medical image cube.

[0294] In some embodiments, the method further includes: selecting (3330) one or more of the plurality of medical image cubes as a reference medical image cube, determining a biometric parameter set associated with each reference medical image cube, selecting a biometric parameter set based on a biometric measurement of the lungs between two biometric parameter sets associated with the two reference medical image cubes, and simulating a medical image cube between the two reference medical image cubes by applying the biometric parameter set based on the biometric measurement of the lungs to a biophysical model.

[0295] It should be understood that the Figures 33A to 33C The specific order of operations in the description is merely an example and is not intended to indicate that the described order is the only order in which the operations may be performed. Those skilled in the art will recognize various ways to reorder the operations described herein. In addition, it should be noted that the details of other processes described herein with respect to other methods described herein also apply in a similar manner to the above description of Figures 32A to 32B The method 3100 described. For example, reference Figure 2 , Figure 6 , Figure 8 , Fig.11 , Fig.13 , Fig.15 , Fig.21 , Fig.23 , Fig.25 , Fig. 27 , Figures 28A to 28B , Figures 31A to 31B , Figures 32A to 32B and Figures 34A to 34C For the sake of brevity, these details are not repeated here.

[0296] Figures 34A to 34C 3D x-ray image cube image from a 2D x-ray image of a patient. In some embodiments, any or all of the operations described below can be performed without human intervention (e.g., without intervention by a technician). In some embodiments, method 3300 is performed by or using any of the devices described herein (e.g., Figure 1 3400). Some operations of method 3400 are performed by a computer system comprising one or more processors and a memory storing instructions, which, when executed by the one or more processors, cause the one or more processors to perform the operations of method 3300. Some operations in method 3400 are optionally combined, and / or the order of some operations is optionally changed.

[0297] In some embodiments, one or more sensors are used (3402) to measure a patient's biometric signals as one or more sequences of time series, wherein the one or more sensors include one or more of a 3D spatial orientation locator, a respiratory phase sensor, and a cardiac phase sensor (e.g., as described above with reference to method 3100 and method 3200).

[0298] In some embodiments, the method further includes identifying (3406) a cardiac stage gating window using one or more cardiac stage sensor measurements, predicting a respiratory stage using one or more respiratory stage sensor measurements, identifying consistency between the cardiac stage gating window and the predicted respiratory stage for generating x-ray imaging pulses, and marking an x-ray image corresponding to the x-ray imaging pulse using the respiratory stage, cardiac stage, and 3D spatial orientation locator measurements (e.g., as described above with reference to method 3200).

[0299] In some embodiments, the 3D spatial position locator is configured (3408) to measure real-time body movements of the patient caused by breathing and heartbeat and output them as a time series (e.g., as described above with reference to method 3100 and method 3200).

[0300] In some embodiments, the respiratory phase sensor is configured (3410) to measure one or more physiological metrics related to the patient's respiration, including tidal volume and its first-order time derivative, such as the rate at which tidal volume changes with time or airflow.

[0301] In some embodiments, the heart phase sensor is configured (3412) to measure periodic and stationary electrical signals generated by the patient's heart having characteristic features corresponding to the heart phase.

[0302] In some embodiments, after synchronization with the x-ray unit's clock, two different filters are used (3414) to remove signal drift and noise from the patient's biometric signal.

[0303] In some embodiments, biometric signals of the patient measured by one or more sensors are used (3416) to trigger the x-ray unit to acquire x-ray images of the patient at specific respiratory and cardiac stages.

[0304] In some embodiments, the x-ray unit includes (3418) a clock. Biometric signals of the patient measured by the one or more sensors are synchronized with the clock of the x-ray unit, and corresponding values ​​of the biometric signals are recorded in association with the acquired x-ray images.

[0305] In some embodiments, a biometric signal of the patient is measured (3422) during a training window (e.g., a training period) prior to capturing any x-ray images of the patient, and the biometric signal of the patient measured during the training window includes multiple complete breathing cycles of the patient (e.g., as described above with reference to method 3100 and method 3200).

[0306] In some embodiments, a plurality of tidal volume percentiles within a complete respiratory cycle are identified (3424) using a biometric signal of the patient measured during a training window, each tidal volume percentile corresponding to one of the different respiratory phases.

[0307] In some embodiments, the patient's biometric signals measured during the training window are used (3426) to construct an optimized respiratory prediction model for predicting a desired respiratory phase at which the x-ray unit is triggered to capture x-ray images of the patient.

[0308] In some embodiments, the optimized respiration prediction model (3428) is based on an autoregressive integrated moving average (ARIMA) model.

[0309] In some embodiments, a desired respiratory phase for capturing x-ray images of the patient is configured (3430) to coincide with a cardiac gating window during which cardiac-induced lung motion varies slowly.

[0310] In some embodiments, a cardiac gating window is selected (3432) based on the location of T waves and P waves in an electrocardiogram (ECG) signal so that cardiac-induced lung motion varies slowly.

[0311] The method includes converting (3434) first sets of x-ray images of the lung captured at different projection angles into second sets of x-ray images of the lung corresponding to different respiratory stages.

[0312] In some embodiments, converting the first plurality of sets of x-ray images of the lung captured at different projection angles into the second plurality of sets of x-ray images of the lung corresponding to different respiratory stages further comprises (3436) capturing the first plurality of sets of x-ray images of the lung at different projection angles. Each set of the first plurality of sets of x-ray images corresponds to a different respiratory stage of the lung at a particular projection angle. In some embodiments, converting the first plurality of sets of x-ray images of the lung captured at different projection angles into the second plurality of sets of x-ray images of the lung corresponding to different respiratory stages further comprises reorganizing the first plurality of sets of x-ray images of the lung into the second plurality of sets of x-ray images of the lung according to the associated respiratory stages of the first plurality of sets of x-ray images of the lung. Each set of the second plurality of sets of x-ray images corresponds to a respective respiratory stage of the lung.

[0313] In some embodiments, the x-ray images within any particular group are (3438) geometrically resolved and temporally independent.

[0314] In some embodiments, different respiratory phases of the lungs correspond (3440) to different tidal volume percentiles of movement of the lungs.

[0315] In some embodiments, the different respiratory phases of the lungs include (3442) early exhalation, late exhalation, maximal exhalation, early inspiration, late inspiration, and maximal inspiration of a complete respiratory cycle of the patient.

[0316] In some embodiments, multiple x-ray images of the lungs captured at different projection angles all correspond to (3444) the same respiratory phase.

[0317] In some embodiments, different breathing phases of the lungs are acquired (3446) from at least two breathing cycles at a specific projection angle.

[0318] The method includes generating (3448) a static image cube from each set of the second plurality of sets of x-ray images at a corresponding respiratory phase using back projection.

[0319] The method includes combining (3450) static image cubes corresponding to different respiratory phases of the lung into a 3D x-ray image cube image by temporal interpolation.

[0320] It should be understood that the Figures 34A to 34C The specific order of operations in the description is merely an example and is not intended to indicate that the described order is the only order in which the operations may be performed. Those skilled in the art will recognize various ways to reorder the operations described herein. In addition, it should be noted that the details of other processes described herein with respect to other methods described herein also apply in a similar manner to the above description of Figures 34A to 34C The method 3100 described. For example, reference Figure 2 , Figure 6 , Figure 8 , Fig.11 , Fig.13 , Fig.15 , Fig.21 , Fig.23 , Fig.25 , Fig. 27 , Figures 28A to 28B , Figures 31A to 31B , Figures 32A to 32B and Figures 33A to 33C For the sake of brevity, these details are not repeated here.

[0321] Fig.35An exemplary patient positioning fixture (PPF) 3501 (e.g., a swivel chair) for supporting a patient 3502 is depicted in accordance with some embodiments. In some embodiments, the PPF 3501 rotates (e.g., along rotation 3503) to position the patient 2502 at multiple angles (e.g., directions) relative to a radiation source 3504 (e.g., an x-ray imaging system or a radiotherapy system). For example, the PPF moves in a manner that moves the patient to a desired position (e.g., the patient does not need to move independently) so that the radiation device 3504 can capture x-ray images of the patient at various angles. In some embodiments, the patient 3502 rotates (e.g., along rotation 3502) to achieve the multiple angles without rotating the PPF 3501. In some embodiments, the PPF 3501 and / or the patient 3505 are automatically rotated and / or moved (e.g., using a motor) to a desired position. In some embodiments, a technician rotates and / or moves the patient 3502. In some embodiments, the flat panel detector unit 3505 is positioned behind the patient relative to the radiation device 3504.

[0322] In some embodiments, one or more cameras 3506-1 to 3506-m are used to detect objects within a predefined area (e.g., a room) surrounding the PPF 3501. For example, the camera 3506 captures whether an object will collide with the PPF 3501 as the PPF 3501 moves (and rotates within) the predefined area. In some embodiments, a collision avoidance method for the PPF 3501 (and the patient 3502) is provided.

[0323] In some embodiments, one or more 3D imaging sensors (e.g., a LIDAR sensor or a structured light sensor) are used to geometrically monitor the breathing of patient 3502 , as described with reference to method 3100 .

[0324] Fig.36An exemplary bio-event monitoring process (BEMP) card 3600 is depicted. The BEMP card 3600 includes a programmable analog filter; a programmable analog / digital converter (ADC) / digital signal processor (DSP) 3605; and a zero instruction set computer (ZISC) processor 3606. The BEMP card 3600 receives an analog input signal 3602 comprising a biometric signal of a patient. In some embodiments, the analog input 3602 comprises an ECG signal. In some embodiments, the analog input is a 3-lead ECG signal or a 12-lead ECG signal. The ZISC processor 3606 detects predefined patterns in the analog input in real time. For example, the ZISC processor 3606 detects an R wave in the ECG signal to predict when will be the next stationary period of the patient's cardiac cycle (e.g., in a TP interval). In some embodiments, the ZISC processor 3606 outputs a ZISC 4-bit weighted output that can be used to trigger a radiation source (e.g., an x-ray imaging system or a radiation therapy system). In some embodiments, a clock signal 3603 for the BEMP card 3600 is provided from an external source (eg, a radiation source) so that the BEMP card 3600 can be synchronized with the radiation source.

[0325] The computational techniques discussed throughout the Post-Processing Software 106 section are computationally resource intensive. Many community hospitals and small clinics do not have access to the computer hardware necessary to create images, compute biomechanical models, and provide the results to the end user in a fast, seamless experience. The functionality of the Post-Processing Software 106 is performed in the cloud, enabling the end user to access the powerful visualization tools of the Post-Processing Software 106 on a processing console, office desktop, or work laptop.

[0326] References throughout this specification to “an embodiment,” “some embodiments,” “one embodiment,” “another example,” “an example,” “a specific example,” or “some examples” mean that a particular feature, structure, material, or characteristic described in conjunction with the embodiment or example is included in at least one embodiment or example of the present disclosure. Thus, the appearance of phrases such as “in some embodiments,” “in one embodiment,” “in an embodiment,” “in another example,” “in an example,” “in a specific example,” or “in some examples” in various places throughout the specification does not necessarily refer to the same embodiment or example of the present disclosure. Furthermore, particular features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples.

[0327] Although illustrative embodiments have been shown and described, those skilled in the art should understand that the above embodiments should not be construed as limiting the present disclosure, and that changes, substitutions, and modifications may be made in the embodiments without departing from the spirit, principle, and scope of the present disclosure.

Claims

1. A method for imaging a patient's lungs, comprising: positioning the patient in a first orientation relative to an x-ray imaging device; obtaining a volume measurement of the patient's breathing; In obtaining the volume measurement of the patient's respiration: determining a breathing phase of the patient based on the volumetric measurement of the patient's respiration; as well as gating the x-ray imaging device to acquire a plurality of x-ray measurements corresponding to different respiratory phases of a lung based on the respiratory phase of the patient determined from the volume measurements; extracting a plurality of displacement fields of lung tissue from a plurality of x-ray measurements corresponding to different breathing phases of the lung, wherein each displacement field represents movement of the lung tissue from a first breathing phase to a second breathing phase, and One or more biophysical parameters of a biophysical model of the lung are calculated using the multiple displacement fields of the lung tissue between different breathing phases of the lung and a corresponding set of biometric parameters associated with each breathing phase, wherein the set of biometric parameters includes tidal volume and airflow of the lung in the corresponding breathing phase and a cardiac phase corresponding to the corresponding breathing phase of the lung.

2. The method according to claim 1, wherein: The x-ray imaging device is gated based on determining that the breathing phase of the patient matches one of a plurality of predefined breathing phases.

3. The method of claim 2, wherein x-ray projections of the patient's lungs are obtained only when the breathing phase of the patient determined by the volume measurement of the patient's breathing matches one of the plurality of predefined breathing phases.

4. The method according to any one of claims 2 to 3, wherein the plurality of predefined breathing phases comprises an early expiratory phase and a late expiratory phase of a complete breathing cycle of the patient.

5. The method according to claim 2, wherein: The plurality of x-ray measurements is a first plurality of x-ray measurements; and The method further comprises: repositioning the patient to a second orientation relative to the x-ray imaging device; acquiring a second plurality of x-ray measurements corresponding to different breathing phases of a lung while obtaining the volumetric measurement of the patient's breathing while the patient is positioned in the second orientation relative to the x-ray imaging device: continuing to determine a breathing phase of the patient based on the volumetric measurement of the patient's respiration; as well as gating the x-ray imaging device to generate a second x-ray projection of the patient's lungs based on determining that the patient's respiratory phase matches each respiratory phase; as well as A static image cube corresponding to each respiratory phase is generated using the first plurality of x-ray measurements and the second plurality of x-ray measurements.

6. The method of claim 5, wherein the static image cube corresponding to each breathing phase is generated using less than ten x-ray projections obtained from various angles during the each breathing phase.

7. The method of any one of claims 1 to 3 and 5-6, wherein the volume measurement of the patient's breathing comprises a measurement of the patient's thoracic rise.

8. The method of any one of claims 1 to 3 and 5-6, wherein the volumetric measurements of the patient's breathing are obtained using one or more volumetric breathing phase sensors from the group consisting of a three-dimensional (3D) scanner, a spirometer, and a belly belt.

9. The method of any one of claims 1 to 3 and 5-6, further comprising creating a point cloud of a surface of the patient's thorax, wherein the volume measurement of the patient's respiration is determined based on the point cloud of the surface of the patient's thorax.

10. The method of claim 9, wherein one or more locations of the patient's thorax are measured using 3D imaging techniques to obtain the point cloud of the surface of the patient's thorax.

11. The method according to claim 10, further comprising: identifying one or more anatomical landmarks on the surface of the patient's thorax using the point cloud of the surface of the patient's thorax; and The locations of one or more internal anatomical landmarks within the patient's thorax are inferred from the point cloud of the surface of the patient's thorax.

12. The method according to any one of claims 1 to 3 and 5-6, wherein: The respiratory phase of the patient is a future respiratory phase; Determining the patient's breathing phase based on the volumetric measurement of the patient's breathing includes predicting the future breathing phase based on one or more current and / or past breathing phases.

13. The method of claim 1, wherein the physical relationship between the biometric parameters associated with the different breathing phases of the lung and the plurality of displacement fields of the lung tissue is defined as follows: where the vector describes the normal stress caused by tidal volume, describes the normal stress caused by airflow, describes the shear stress caused by the airflow, and Describes the tissue movement caused by cardiac motion, the displacement of the tissue at any point in the closed-loop trajectory ( ) are expressed as tidal volume T v ,airflow A f and heart stage H c The scaled sum of stress, strain, and perturbation cardiac motion vectors.

14. A system for imaging a patient's lungs, comprising: X-ray imaging equipment; one or more processors; as well as A memory storing instructions that, when executed by the one or more processors, cause the one or more processors to perform a set of operations comprising: positioning a patient in a first orientation relative to the x-ray imaging device; obtaining a volume measurement of the patient's breathing; In obtaining the volume measurement of the patient's respiration: determining a breathing phase of the patient based on the volumetric measurement of the patient's respiration; gating the x-ray imaging device to acquire a plurality of x-ray measurements corresponding to different respiratory phases of a lung based on the respiratory phase of the patient determined from the volume measurements; extracting a plurality of displacement fields of lung tissue from a plurality of x-ray measurements corresponding to different breathing phases of the lung, wherein each displacement field represents movement of the lung tissue from a first breathing phase to a second breathing phase, and One or more biophysical parameters of a biophysical model of the lung are calculated using the multiple displacement fields of the lung tissue between different breathing phases of the lung and a corresponding set of biometric parameters associated with each breathing phase, wherein the set of biometric parameters includes tidal volume and airflow of the lung in the corresponding breathing phase and a cardiac phase corresponding to the corresponding breathing phase of the lung.

15. A system for imaging a patient's lungs, comprising: X-ray imaging equipment; one or more processors; as well as A memory storing instructions which, when executed by the one or more processors, cause the one or more processors to perform the method of any one of claims 1 to 13.

16. A non-transitory computer-readable storage medium storing instructions that, when executed by a system comprising an x-ray imaging device and one or more processors, cause the one or more processors to perform a set of operations comprising: positioning a patient in a first orientation relative to the x-ray imaging device; obtaining a volume measurement of the patient's breathing; In obtaining the volume measurement of the patient's respiration: determining a breathing phase of the patient based on the volumetric measurement of the patient's respiration; gating the x-ray imaging device to acquire a plurality of x-ray measurements corresponding to different respiratory phases of a lung based on the respiratory phase of the patient determined from the volume measurements; extracting a plurality of displacement fields of lung tissue from a plurality of x-ray measurements corresponding to different breathing phases of the lung, wherein each displacement field represents movement of the lung tissue from a first breathing phase to a second breathing phase, and One or more biophysical parameters of a biophysical model of the lung are calculated using the multiple displacement fields of the lung tissue between different breathing phases of the lung and a corresponding set of biometric parameters associated with each breathing phase, wherein the set of biometric parameters includes tidal volume and airflow of the lung in the corresponding breathing phase and a cardiac phase corresponding to the corresponding breathing phase of the lung.

17. A non-transitory computer-readable storage medium storing instructions that, when executed by a system comprising an x-ray imaging device and one or more processors, cause the one or more processors to perform the method of any one of claims 1 to 12.

Images