Patent Grant
12026832
This application is also related to U.S. patent application Ser. No. 16/420,030, entitled “System and Method for Geometrically-Resolved Radiographic X-Ray Imaging”; U.S. patent application Ser. No. 16/420,011, entitled “System and Method for Biophysical Lung Modeling” and PCT Application PCT/US2019/033751, entitled “System and Method for Lung-Volume-Gated X-ray Imaging.” Each of these applications is incorporated by reference in their entirety.
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.
Computational modeling of human anatomy facilitates an understanding of the anatomical behavior that typifies different physiological conditions. While state-of-the-art imaging techniques can allow a physician to visualize anatomic behavior, the state-of-the-art technology that accurately images the complex movement of the heart and lung is often too expensive to be widely adopted. Moreover, cardiac motion, e.g., the deformation of the heart, which is an intricate process and unrelated to the breathing cycle, may appear as noise in CT-based or radiographic breathing motion measurements. Consequently, the accuracy of mathematical models that describe breathing motion will also be degraded by seemingly random heart motion. One way of addressing this issue is to remove lung motion from images by making patients hold their breath. While this approach stops the patient's breathing motion, crucial information on lung health may be hard to identify in these breath-hold images.
Embodiments of the present disclosure seek to solve at least one of the problems existing in the related art to at least some extent by providing a geometrically-resolved radiographic x-ray imaging system (GREX).
According to some embodiments, a GREX imaging system acquires images of a patient's chest by computationally targeting specific breathing phases and a specific cardiac phase. By taking snapshots of the chest at targeted breathing phases in a series of imaging planes, the GREX imaging system obtains anatomic geometries that are independent of time. The time-independent anatomic geometries are co-registered and interpolated, based on the collected breathing and cardiac signals, to create a volumetric interactive movie of the chest. The interactive movie has key advantages over conventional static imaging. Using the interactive movie, GREX creates biomechanical models that quantitatively describe how the chest geometry changes during patient respiration. The biomechanical models, based on fundamental physical laws, provide estimates of important quantities such as lung tissue elasticity, stress, strain, and respiratory compliance. These properties and others provide diagnostic capabilities that are currently unavailable to physicians and practitioners in medicine. Physicians can leverage such information provided by the GREX imaging system to inform on the etiology of the patient's lung disease.
In some embodiments, the GREX imaging system includes a unique hardware and software add-on package for existing digital diagnostic x-ray units. Together, the hardware and software add-on package increase the diagnostic quality of a conventional digital diagnostic x-ray unit by providing new and enhanced imaging capabilities through using the patient's biometric signals to inform a custom imaging procedure that gives more diagnostically informative information. The hardware and software add-on package is a universal upgrade for any commercially available digital diagnostic x-ray unit (e.g., a legacy x-ray unit).
In accordance with some embodiments, a method of imaging a patient's lung is provided. The method comprises positioning the patient at a first orientation relative to an x-ray imaging apparatus (e.g., an x-ray unit) and obtaining a volumetric measurement of the patient's breathing. The method includes, while the patient is positioned at the first orientation relative to the x-ray imaging apparatus, and while obtaining the volumetric measurement of the patient's breathing, determining, based on the volumetric measurement of the patient's breathing, a breathing phase of the patient. The method includes, in accordance with a determination that the breathing phase of the patient matches a predefined breathing phase, gating the x-ray imaging apparatus to produce an x-ray projection of the patient's lung.
In some embodiments, the predefined breathing phase is a first predefined breathing phase of a plurality of predefined breathing phases and the method further includes, while obtaining the volumetric measurement of the patient's breathing, in accordance with a determination that the breathing phase of the patient matches any of the plurality of predefined breathing phases, gating the x-ray imaging apparatus to produce an x-ray projection of the patient's lung.
In some embodiments, x-ray measurements of the patient's lung are not obtained except when the breathing phase of the patient, as determined by the volumetric measurement of the patient's breathing, matches one of the plurality of predefined breathing phases.
In some embodiments, the plurality of predefined breathing phases includes an early exhalation phase, a late exhalation phase, a maximum exhalation phase, an early inhalation phase, a late inhalation phase, and a maximum inhalation phase of a complete breathing cycle of the patient.
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 apparatus. In some embodiments, the method further includes, while the patient is positioned at the second orientation relative to the x-ray imaging apparatus, and while continuing to obtain the volumetric measurement of the patient's breathing, continuing to determine, based on the volumetric measurement of the patient's breathing, a breathing phase of the patient and, in accordance with a determination that the breathing phase of the patient matches the predefined breathing phase, gating the x-ray imaging apparatus 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 breathing phase, using the first x-ray projection and the second x-ray projection.
In some embodiments, the static image cube, corresponding to the predefined breathing phase, is generated using less than ten x-ray projections obtained from various angles at the predefined breathing phase.
In some embodiments, the volumetric measurement of the patient's breathing includes a measurement of the patient's chest rise.
In some embodiments, the volumetric measurement of the patient's breathing is obtained using one or more volumetric breathing phase sensors of the group consisting of: a three-dimensional (3D) scanner, a spirometer, and an abdominal belt.
In some embodiments, the method further includes creating a point cloud of a surface of the patient's chest. The volumetric measurement of the patient's breathing is determined from the point cloud of the surface of the patient's chest.
In some embodiments, the point cloud of the surface of the patient's chest is obtained using a 3D imaging technique to measure one or more positions of the patient's chest.
In some embodiments, the method further includes 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 a location of one or more internal anatomical landmarks within the patient's chest from the point cloud of the surface of the patient's chest.
In some embodiments, the breathing phase of the patient is a future breathing phase and determining, based on the volumetric measurement of the patient's breathing, a breathing phase of the patient includes forecasting the future breathing phase from one or more current and/or past breathing phases.
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 breathing and obtaining a measurement of the patient's cardiac function. The method further includes, while the patient is positioned at the first orientation relative to the radiation source, and while obtaining the measurement of the patient's breathing, determining, from the measurement of the patient's breathing, a breathing phase of the patient and determining, from the measurement of the patient's cardiac function, a cardiac phase of the patient. The method further includes gating the radiation source to expose the patient to radiation based on a determination that the breathing phase of the patient matches a predefined breathing phase and a determination that the cardiac phase of the patient matches a predefined window of the cardiac cycle.
In some embodiments, the radiation source is an x-ray imaging apparatus and gating the radiation source to expose the patient to radiation comprises gating the x-ray imaging apparatus to produce an x-ray projection of the patient's lung.
In some embodiments, the radiation source is a radiation therapy source and gating the radiation therapy source to expose the patient to radiation comprises gating the radiation therapy source to irradiate a region of the patient's lung at a therapeutic dose.
In some embodiments, the method further includes, before gating the radiation source to expose the patient to radiation, obtaining measurements of the patient's cardiac function from a plurality of cardiac cycles of the patient and, using the measurements of the patient's cardiac function from the plurality of cardiac cycles, determining an average interval between a predefined cardiac phase and a beginning of the predefined window of the cardiac cycle. In some embodiments, determining that the cardiac phase of the patient matches the predefined window of the cardiac cycle includes predicting the predefined window of the cardiac cycle by, in real-time, detecting the predefined cardiac phase and waiting a length of time corresponding to the average interval between the predefined cardiac phase and the beginning of the predefined window of the cardiac cycle.
In some embodiments, the measurements from the plurality of cardiac cycles of the patient are waveform measurements of the plurality of cardiac cycles and the method further includes validating, as statistically stable, the waveform measurements of the plurality of cardiac cycles.
In some embodiments, the predefined window of the cardiac cycle is a quiescent window of the cardiac cycle.
In some embodiments, a method of determining a biophysical model for a lung of a patient from multiple x-ray measurements corresponding to different breathing phases of the lung is provided. The method includes extracting multiple displacement fields of lung tissue from the multiple x-ray measurements corresponding to different breathing phases of the lung. Each displacement field represents movement of the lung tissue from a first breathing phase to a second breathing phase and each breathing phase has a corresponding set of biometric parameters. The method further includes calculating one or more biophysical parameters of a biophysical model of the lung using the multiple displacement fields of the lung tissue between different breathing phases of the lung and the corresponding sets of biometric parameters.
In some embodiments, the one or more biophysical parameters define a physical relationship between the biometric parameters associated with the different breathing phases of the lung and the multiple displacement fields of the lung tissue.
In some embodiments, the physical relationship between the biometric parameters associated with the different breathing phases of the lung and the multiple displacement fields of the lung tissue is defined as follows:−
=TV
1+Af(
2+
3)+Hc
4
The 1 vector describes normal stress caused by tidal volume,
2 describes normal stress caused by airflow,
3 describes shear stress caused by airflow, and
4 describes tissue motion introduced by heart motion, and the displacement (
−
0) of tissue at any point in a closed loop trajectory is expressed as a summation of the stress, strain, and perturbing heart motion vectors scaled by the tidal volume (Tv), airflow (Af), and cardiac phase (Hc) respectively.
In some embodiments, the method further includes generating multiple medical image cubes corresponding to the different breathing phases of the lung from the multiple x-ray measurements corresponding to the different breathing phases of the lung. The multiple displacement fields of lung tissue are extracted from the multiple medical image cubes corresponding to different breathing phases of the lung further by delineating the lung tissue from a remaining portion of a first medical image cube through image segmentation. The method includes, for a respective voxel in the first medical image cube, determining a displacement vector between the voxel in the first medical image cube and a second medical image cube using 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 counterparts in the second medical image cube.
In some embodiments, the set of biometric parameters associated with a respective breathing phase includes a tidal volume and an airflow of the lung at the respective breathing phase and a cardiac phase corresponding to the respective breathing phase of the lung.
In some embodiments, the method further includes generating multiple medical image cubes corresponding to different breathing phases of the lung from the multiple x-ray measurements corresponding to different breathing phases of the lung. In some embodiments, the method includes choosing one or more of the multiple medical image cubes as reference medical image cubes, determining a set of biometric parameters associated with each reference medical image cube and selecting a set of biometric parameters based on biometric measurements of the lung between two sets of biometric parameters associated with 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 the set of biometric parameters based on biometric measurements of the lung to the biophysical model.
In some embodiments, the different breathing phases of the lung include early exhalation, late exhalation, maximum exhalation, early inhalation, late inhalation, and maximum inhalation of a complete breathing cycle of the patient.
In some embodiments, one or more sensors are used for measuring biometric signals of the patient as one or more sequences of time series, including one or more of a 3D spatial position localizer, a breathing phase sensor, and a cardiac phase sensor.
In some embodiments, the 3D spatial position localizer is configured for measuring the patient's real-time body movement caused by respiration and heartbeats and outputting them as time series.
In some embodiments, the breathing phase sensor is configured for measuring one or more physiologic metrics related to the patient's breathing, including a tidal volume and its first-order time derivative.
In some embodiments, the cardiac phase sensor is configured for measuring periodic and stationary electrical signals generated by the patient's heart.
In some embodiments, the biometric signals of the patient measured by the one or more sensors are used for triggering an x-ray unit to acquire an x-ray image of the patient at a specific breathing and cardiac phase.
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 x-ray unit's clock. In some embodiments, respective values of the biometric signals are recorded to be associated with the acquired x-ray image.
In some embodiments, the biometric signals of the patient measured during a training window are used for building an optimized breathing prediction model for predicting a desired breathing phase at which an x-ray unit is triggered to capture an x-ray image of the patient.
In some embodiments, a method of generating a 3D x-ray image cube movie from 2D x-ray images of a patient is provided. The method includes converting first multiple sets of x-ray images of a lung captured at different projection angles into second multiple sets of x-ray images of the lung corresponding to different breathing phases. The method further includes generating a static image cube from each set of the second multiple sets of x-ray images at a respective breathing phase using back projection and combining the static image cubes corresponding to the different breathing phases of the lung into a 3D x-ray image cube movie through temporal interpolation.
In some embodiments, the converting first multiple sets of x-ray images of a lung captured at different projection angles into second multiple sets of x-ray images of the lung corresponding to different breathing phases further comprises capturing the first multiple sets of x-ray images of the lung at different projection angles. Each set of the first multiple sets of x-ray images corresponds to the different breathing phases of the lung at a particular projection angle. The converting further comprises re-organizing the first multiple sets of x-ray images of the lung by their associated breathing phases into the second multiple sets of x-ray images of the lung. Each set of the second multiple sets of x-ray images corresponds to a respective breathing phase of the lung.
In some embodiments, the x-ray images within any particular set are geometrically resolved and temporally independent.
In some embodiments, the different breathing phases of the lung correspond to different tidal volume percentiles of the lung's movement.
In some embodiments, the different breathing phases of the lung include early exhalation, late exhalation, maximum exhalation, early inhalation, late inhalation, and maximum inhalation of a complete breathing cycle of the patient.
In some embodiments, the multiple x-ray images of the lung captured at different projection angles all correspond to the same breathing phase.
In some embodiments, one or more sensors are used for measuring biometric signals of the patient as one or more sequences of time series, including one or more of a 3D spatial position localizer, a breathing phase sensor, and a cardiac phase sensor.
In some embodiments, the method further comprises identifying a cardiac phase gating window using one or more cardiac phase sensor measurements, predicting a breathing phase using one or more breathing phase sensor measurements, identifying a coincidence between the cardiac phase gating window and the predicted breathing phase for generating an x-ray imaging pulse, and tagging an x-ray image corresponding to the x-ray imaging pulse with the breathing phase, the cardiac phase, and 3D spatial position localizer measurements.
In some embodiments, the 3D spatial position localizer is configured for measuring the patient's real-time body movement caused by respiration and heartbeats and outputting them as time series.
In some embodiments, the breathing phase sensor is configured for measuring one or more physiologic metrics related to the patient's breathing, including a tidal volume and its first-order time derivative.
In some embodiments, the cardiac phase sensor is configured for measuring periodic and stationary electrical signal generated by the patient's heart, with characteristic features that correspond to the cardiac phase.
In some embodiments, two distinct filters are used to remove signal drift and noise from biometric signals of the patient after being synchronized with an x-ray unit's clock.
In some embodiments, the biometric signals of the patient measured by the one or more sensors are used for triggering an x-ray unit to acquire an x-ray image of the patient at a specific breathing and cardiac phase.
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 x-ray unit's clock. In some embodiments, the respective values of the biometric signals are recorded to be associated with the acquired x-ray image.
In some embodiments, the biometric signals of the patient are measured during a training window before capturing any x-ray image of the patient and the biometric signals of the patient measured during the training window include multiple complete breathing cycles of the patient.
In some embodiments, multiple tidal volume percentiles within a complete breathing cycle are identified using the biometric signals of the patient measured during the training window, each tidal volume percentile corresponding to one of the different breathing phases.
In some embodiments, the biometric signals of the patient measured during the training window are used for building an optimized breathing prediction model for predicting a desired breathing phase at which an x-ray unit is triggered to capture an x-ray image of the patient.
In some embodiments, the optimized breathing prediction model is based on an autoregressive integrated moving average (ARIMA) model.
In some embodiments, the desired breathing phase for capturing the x-ray image of the patient is configured to coincide with a cardiac gating window during which heart induced lung motion is changing slowly.
In some embodiments, the cardiac gating window is chosen based on locations of T wave and P wave in an electrocardiogram (ECG) signal such that the heart induced lung motion is changing slowly.
In some embodiments, the different breathing phases of the lung at a particular projection angle are collected from at least two breathing cycles.
Additional aspects and advantages of embodiments of present disclosure will be given in part in the following descriptions, become apparent in part from the following descriptions, or be learned from the practice of the embodiments of the present disclosure.
In order to describe embodiments of the present disclosure or technical solutions in the prior art more clearly, the drawings necessary in the descriptions of the embodiments or the prior art will be briefly explained. Obviously, the drawings in the following description are just some embodiments of the present disclosure. To those skilled in the art, in the premise of no creative labor, other drawings can also be obtained according to structures illustrated in these drawings.
Reference will be made in detail to embodiments of the present disclosure. The same or similar elements and the elements having same or similar functions are denoted by like reference numerals throughout the descriptions. The embodiments described herein with reference to drawings are explanatory, illustrative, and used to generally understand the present disclosure. The embodiments shall not be construed to limit the present disclosure.
In the specification, unless specified or limited otherwise, relative terms such as “central”, “longitudinal”, “lateral”, “front”, “rear”, “right”, “left”, “inner”, “outer”, “lower”, “upper”, “horizontal”, “vertical”, “above”, “below”, “up”, “top”, “bottom” as well as derivative thereof (e.g., “horizontally”, “downwardly”, “upwardly”, etc.) should be construed to refer to the orientation as then described or as shown in the drawings under discussion. These relative terms are for convenience of description and do not require that the present disclosure be constructed or operated in a particular orientation.
In the present invention, unless specified or limited otherwise, the terms “mounted,” “connected,” “coupled,” “fixed” and the like are used broadly, and may be, for example, fixed connections, detachable connections, or integral connections; may also be mechanical or electrical connections; may also be direct connections or indirect connections via intervening structures; may also be inner communications of two elements, which can be understood by those skilled in the art according to specific situations.
In the present invention, unless specified or limited otherwise, a structure in which a first feature is “on” or “below” 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 contacted via an additional feature formed therebetween. Furthermore, a first feature “on,” “above,” or “on top of” a second feature may include an embodiment in which the first feature is right or obliquely “on,” “above,” or “on top of” the second feature, or just means that the first feature is at a height higher than that of the second feature; while a first feature “below,” “under,” or “on bottom of” a second feature may include an embodiment in which the first feature is right or obliquely “below,” “under,” or “on bottom of” the second feature, or just means that the first feature is at a height lower than that of the second feature.
There are three measurement sensors in the hardware box 102 (e.g., 3D spatial position localizer 300, breathing phase sensor 110, and heart phase sensor 112) that independently collect biometric signals of a patient as time series, and the collected time series serve as the inputs into the acquisition software 104. The acquisition software 104 processes and filters the biometric time series measurements to generate an imaging trigger signal (e.g., that gets x-ray unit 108). The imaging trigger signal targets a specific breathing phase and, optionally, a cardiac phase of the patient. The breathing phase and cardiac phase are each defined by their respective biometric time series measurements. A connector cable transmits the imaging trigger signal from the acquisition software 104 to the x-ray unit 108, which acquires breathing and cardiac phase radiographic images at the targeted phases. Once acquired, a series of breathing and cardiac phase targeted images, which define a complete breathing cycle, are input into the post-processing software 106. The post-processing software builds a biomechanical model of lung motion from the breathing and cardiac phase targeted images. The biomechanical model is then used to generate other diagnosis results in the post-processing software. Although this application uses x-ray image as an example, it would be apparent to those skilled in the art that the approaches disclosed in this application can be adapted to be applied to other types of medical images with little (if any) effort. For example, the process of building the biomechanical model is not limited to the use of x-ray images and can use other types of medical images (e.g., CT-Scan, MRI, etc.).
In this document, a GREX-based imaging system is divided into three sections. Section 1 describes embodiments of the hardware box 102. Section 2 describes embodiments of the acquisition software 104. Section 3 describes embodiments of the post-processing software 106. Each section describes, in greater detail, the components and functions that constitute the hardware box 102, the acquisition software 104, and the post-processing software 106 as shown in
Section 1. Hardware Box 102
In some embodiments, the hardware box 102 has at least two responsibilities. The first is to collect biometric signals that define the chest's anatomic geometry. The second is to communicate with the digital diagnostic x-ray unit (e.g., x-ray unit 108).
The biometric signal inputs that define the chest's anatomic geometry include: the chest dimensions (measured via the 3D Spatial Localizer 300,
Section 1.1—3D Spatial Position Localizer 300
The 3D spatial position localizer 300 measures the patient's real-time body movement caused by respiration and heartbeats and outputs them as a time series in a coordinate space (e.g., Cartesian, polar, hyper, etc.). As shown in
Using real-time depth maps, each 3D camera 302 creates a surface rendering of the patient. Simultaneous information from all three 3D cameras 302 is combined to form a volumetric skin surface position measurement (e.g., using a ray-casting technique) that changes, in real-time, based on the patient's breathing and cardiac phase. The 3D spatial position localizer 300 uses the volumetric skin surface position measurements in at least two ways: (i) defining the spatial boundary of the patient and (ii) determining the tissue location of the patient. Based on the assumptions such as skin thickness, rib thickness, muscular thickness, and skeletal position, which are derived from the canonical Medical Internal Radiation Dose (MIRD) Anatomic Database, the 3D spatial position localizer 300 approximates the lung's real-time, spatial position inside the patient. The 3D spatial position localizer 300 uses the MIRD data to calculate the lung's spatial boundary conditions, which are then made available to the post-processing software 106. For example, the 3D spatial position localizer 300's estimation of the lung's spatial boundary condition creates an initial chest geometry that the post-processing software 106 uses to simulate the cumulative tissue density along a ray that originates at the x-ray unit 108.
Section 1.2—Breathing Phase Sensor 110
The breathing phase sensor 110 in the hardware box 102 measures key physiologic metrics related to breathing, namely the tidal volume and its first-order time derivative (e.g., the rate of tidal volume changes 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 with a mouth spirometer, which is comprised of a turbine within a tube that spins at a rate proportional to the volume of air that the patient inhales or exhales. Indirect tidal volume measurement is performed with an abdominal belt (or any other geometric measurement of the patient's chest, as described herein) that measures the patient's abdominal circumference changes during breathing (as shown in
The term tidal volume, as used herein, means a difference between a current lung volume and a predefined baseline (e.g., a volume during maximum exhalation of a normal breath without extra effort applied, a volume during maximum inhalation of a normal breath without extra effort applied, or any other suitable fiducial volume). Based on the ideal gas law, differences in air density between room air and internal air lead to air in the lung expanding 11% more than the tidal volume. To conserve mass, the body expands 11% more volumetrically than the volume of air inhaled. Therefore, the lung's tidal volume can thus be calculated by using an external measurement of the body that is calibrated to internal air content. Moreover, the 3D spatial position localizer 300 provides a secondary check of both the air content and the accuracy of the tidal volume measurement by identifying the patient's volumetric expansion during breathing. The volumetric expansion of the body is compared with estimates for the volume of air in the trachea, lungs, and bronchi from the x-ray images.
Section 1.3—Heart Phase Sensor 112
As shown in
Typically, the digital diagnostic x-ray unit (e.g., x-ray unit 108) is turned on with an analog plunger that attaches to a plug port. The plug port is uniquely configured to accept a plunger with a specific pin configuration. In every pin configuration, there is an “acquisition pin,” which accepts the voltage signal that turns on (e.g., gates) the x-ray unit. When an end-user depresses a plunger, the plunger sends a voltage pulse to the digital diagnostic x-ray unit, which activates the imaging beam through the patient's body.
As shown in
Section 2. Acquisition Software 104
The acquisition software 104 is designed to collect the hardware box 102's measured spatial, heart and lung time series and determine when to trigger the x-ray unit 108 to acquire an x-ray image at a specific breathing and cardiac phase. The acquisition software 104 then accurately overlays (e.g., synchronizes) the measured heart and lung phases, processes (e.g., filters) the collected biometric time series, identifies appropriate imaging times, and creates an electronic trigger signal for the digital diagnostic x-ray unit 108. The electronic trigger signal (e.g., gate signal) turns the x-ray unit 108 on, acquiring a snapshot image of the chest geometry. The spatial, heart, and lung values that are associated with the snapshot are recorded to define the chest's surface geometry when the image was taken. The entire process is automated and is user-input-independent. As shown in
Section 2.1—Synchronization Module 114
Inputs to the synchronization module 114, including 3D spatial position localizer 300's signal, breathing phase sensor 110's signal, and heart phase sensor 112's signal are synchronized with the x-ray unit 108's clock by the synchronization module 114. Note that the physiological biometric signals are asynchronously collected, and therefore require synchronization. One source of asynchronicity is that the breathing cycle is slower than, and completely independent from, the cardiac cycle. As explained previously, the breathing cycle and the cardiac cycle are measured separately with different sensors. The synchronization module 114 is configured to synchronize the breathing and heart phase sensors with the acquired images. When an image is taken, the image displays the chest's anatomic geometry at a moment in time. That moment in time is recorded using the x-ray unit 108's native timing system, which is not necessarily synchronized with the breathing and heart sensor biometric time series.
It should be noted that time alone does not differentiate between time periods of irregular breathing versus time periods of regular breathing. In other words, if time is the sole defining dimension of breathing phase, an image taken during a normal breath and an image taken during an abnormal breath (like a cough) are computationally indistinguishable from one another.
To overcome the issue with the time dimension, the GREX imaging system 100 defines the breathing phase by the physiologic values that are acquired from the various physiologic sensors in the hardware box 102, which are more informative breathing phase dimensions compared to time. The synchronization module 114 serves primarily 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, discussed in detail below within Section 2.2, during which the collected tidal volume time series observations are used to calculate tidal volume percentiles. The acquisition software 104 uses the tidal volume percentiles to define the breathing phase rather than the peak-to-peak time interval of a periodic cosine wave. The acquisition software 104's tidal volume percentiles are a more informative method of defining the lung geometry compared to the peak-to-peak periodic cosine curves because of breath-to-breath tidal volume variations.
To avoid the unwanted radiation exposure for the patient, the acquisition software 104 does not send any triggering signals to activate the x-ray unit 108 via the connector cable without appropriate synchronization between different components of the GREX imaging system 100. In some embodiments, the 30-second training window is used for verifying the synchronization between the hardware box 102's clock and the x-ray unit 108's clock. Therefore, the 30-second training window should contain 30 seconds worth of samples. If every sensor in the hardware box 102 as well as the x-ray unit 108's clock do not show exactly 30 seconds worth of samples, then synchronicity has failed to occur. To that end, if this aforementioned checking procedure contains a discrepancy, the synchronization system will restart to correct the discrepancy. It should be noted that the 30-second training window is for illustrative purpose and one skilled in the art would understand that the length of the training window varies as long as there is sufficient data for performing the synchronization process.
Section 2.2—Signal Processing Module 116
After the spatial position, breathing phase, and cardiac phase signals are synchronized, the acquisition software 104 processes the sensor signals to remove noise and ultimately predicts accurate tidal volume of the patient. Noise in the measured lung and heart time series comes from the sensor electronics, the electrodes, as well as from background electrical signals. A suite of filters that specifically maintains temporal accuracy removes the noise from the measured lung and heart time series such that the biometric time series maintain temporal accuracy after being filtered by the suite of filters.
In some embodiments, two distinct filters (e.g., wavelet filters) are used to remove signal drift and noise from the biometric time series. Signal drift skews measurements taken over time so that measurements at the beginning of the data collection are not consistent with measurements taken at the end of the data collection. Signal noise is not physiologic in nature and causes serious problems when calculating the patient's airflow from tidal volume measurements.
The acquisition software 104 needs smooth tidal volume time series to compute the first order time derivative of the tidal volume, e.g., airflow. If the tidal volume time series were not smooth, the first order time derivative of the tidal volume would not yield a smooth curve; rather, the curve would contain discontinuities that violate the biophysical realities of breathing.
The acquisition software 104 performs two distinct functions using the filtered and temporally accurate time series curves. The acquisition software 104's first function is to generate a breathing phase prediction using a short prediction horizon.
A short prediction horizon also mitigates another source of inaccuracy in breathing prediction, e.g., breath-to-breath variations in breathing amplitude and breathing period. As the limit of prediction horizon approaches zero, the change in the lung geometry approaches zero (e.g., the lung geometry is deemed to be virtually constant). In other words, the lung geometry is very unlikely to change dramatically during a short prediction horizon. The short prediction horizon therefore reduces the impact of human breathing variation on the breathing motion model's predictive accuracy.
Temporally accurate, filtered tidal volume time series serve as input values for the breathing prediction algorithm. The breathing prediction algorithm provides a fast, real-time, and accurate forecast of the breathing phase. For example, the breathing prediction algorithm is based on an autoregressive integrated moving average (ARIMA). ARIMA is appropriate for breathing prediction because ARIMA models do not assume that input values are stationary, and is comprised of polynomials. The polynomial coefficients of the ARIMA model are estimated during the 30-second training window taken at the start of the imaging study. The number of polynomial coefficients of the ARIMA model, e.g., model order, is checked with nonlinear optimization that seeks to minimize an information criteria search function to reduce or eliminate overfitting. If the model order is optimal for the collected training data, then builds a tidal volume histogram (discussed in Section 2.1) and calculates the probability density function using a log-likelihood objective function. The tidal volume distribution is to check for irregular breathing as discussed in Section 2.1. If irregular breathing was detected, the training data is discarded and reacquired. If irregular breathing was not detected, the training data and probability density function is used to estimate the ARIMA model coefficients with a maximum likelihood approach. The 30-second training window also serves as an equipment check prior to imaging.
The acquisition software 104's second function is to identify the cardiac phase so that the heart is in the same phase in each desired chest geometry. The prediction horizon's duration is a crucial parameter in the acquisition software 104's endeavor to accurately predict human breathing because human breathing is a quasi-random function (because each breath has some unique aspect of its own). In some embodiments, the prediction horizon's duration is longer than the sum of the digital diagnostic x-ray unit 108's latency time and the x-ray imaging's exposure time. The sum of the digital diagnostic x-ray unit 108's latency time and the x-ray imaging's exposure time is very short, on the order of 10 milliseconds. As a result, the prediction horizon's duration is also short (on the order of 1-2 sensor measurement samples at an operating frequency of 100-1000 Hz).
When the acquisition software 104 looks for coincidence between the cardiac and breathing phases, the probability of having phases represented by a single point in each time series align is low. As a result, an imaging method that looks for single-point coincidences takes a longer time to complete since the acquisition software 104 has to wait for low probability coincidences to occur. In contrast, the cardiac gating window extends the size of the coincidence window so that the imaging study takes less time.
To further reduce computation time in the signal processing software, the signal processing software does not predict cardiac phase. Instead, the signal processing software targets a specific gating window in which the heart is not rapidly displacing the lung.
The preceding paragraph discusses how to identify the ideal cardiac gating window that would minimize the heart's physical effect on the lungs while still maintaining a window of opportunity for the targeted breathing 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 a signal processing difference that distinguishes the cardiac phase from the breathing phase, e.g., the cardiac phase is periodic and stable. Because the cardiac phase is periodic and stable, an unsupervised multi-layer perceptron using a backpropagation approach can be used to predict the next heart beat based on the pattern extraction rather than a time series prediction process.
Section 2.3—Logic Algorithm 118
There are known medical algorithms and systems for identifying an ECG's T wave and P wave within a cardiac cycle. Because the heart phase sensor 112 takes continuous measurements of 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 thus proportional to the heart rate) is also known. The ECG features within the time interval can be used by the logic algorithm 118 to introduce a short time lag before initiating the gating window such that the gating window may start, e.g., halfway between the T and P waves and close after the logic algorithm 118 identifies the P wave.
Section 2.4—Trigger Generation (Gating)
X-ray unit 108 has a port that contains a series of electrical pins. One of those pins accepts an electrical impulse that defines the radiation exposure timing and duration. Based on the logic algorithm 118's identified breathing phases within a cardiac gating window, the trigger generator generates a square wave trigger as an electrical impulse. A fiber optic cable with a vendor-specific plug attachment carries the generated trigger signal to the x-ray unit 108.
Section 3. Post-Processing Software
The biometrically-informed imaging trigger identified by the hardware box 102 (Section 1) and acquisition software 104 (Section 2) provides better quality inputs (and remove poor quality inputs) to the image reconstruction algorithm. Specifically, the quality enhancement arises from the fact that image reconstruction and image post-processing techniques are enhanced through the process of acquiring biometrically targeted images during normal breathing. The act of biometrically targeting images during the normal breathing allows for more accurate association of multiple images of the same patient's anatomic geometry taken from different angles and at different times (e.g., different breaths) because the fundamental assumptions of the underlying radiology mathematics assume anatomical equivalence across the various probed imaging angles. The enhanced images serve as observations and the biometric signals serve as inputs for a complex biomechanical model of the chest geometry.
Section 3.1—Digital Tomosynthesis-Based GREX Image Acquisition
Multiple imaging angles are needed to reconstruct a 3D volume. In the context of the GREX imaging, each angle needs to be acquired for each breathing phase. The acquisition software 104 (Section 2) creates a trigger signal that allows the x-ray unit 108 to repeatedly image the chest in a specific geometry. Identical geometries imaged at different imaging angles and different breaths constitute a set of 2D projection images that are used to reconstruct the 3D volume. There are many known methods of reconstructing a 3D volume from the multiple 2D projection images. One such exemplary method is a convolution-back-projection algorithm for direct reconstruction of a 3D density function using a set of 2D projections called “FDK image reconstruction algorithm” disclosed in Feldkamp, L. A., Davis, L. C., Kress, J. W. “Practical cone beam algorithm”. J Opt Soc Am 1, 612-619 (1984).
In Equation (1), N0 is the total number of projections, β is the angle of each projection, d is the source-to-image plane distance, s is the pixel-to-detector distance, 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 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 a weighting function. Basically, Equation (1) represents a combination of convolution, back-projection, and weighting steps.
Information at points that lie in the mid-plane is calculated from the projection data along the intersection of the detector plane and the mid-plane (y=0). Projections that intersect the detector plane along a line parallel to the mid-plane, but not in it (constant, nonzero y), themselves define a plane. This plane is treated as if it were the mid-plane of another, tilted arrangement. If a complete set of projections are acquired (note that “complete” denotes the case where all rotation angles about the normal are acquired) the tilted plane's density 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; In CT imaging, for example, a full 360° rotation around the imaged object occurs. Note that the bolded words “the imaging object” imply—more accurately, the bolded words explicitly communicate and define—the fundamental assumption of the Radon transform (folded into Equation (1)'s R(β, p, ξ) term) that, if violated, stymies the reconstruction's representativeness of the underlying ground truth anatomy that was imaged: “The imaging object” implies that Radon transform inputs are assumed to be different angular probes of one single, fixed, non-moving and non-changing in time and space, stationary object of anatomy. The GREX imaging's biometric targeting (indeed “pre-selection”) of only the accurate (the geometrically identical breathing phases that are identical despite occurring in different breaths) Radon transform inputs adheres to the transform's fundamental assumption definitionally and practically, since breathing phase (that is: changing in the time domain) is for GREX defined strictly geometrically and anatomically and physically (e.g., a definitionally correct formulation) and also practically with GREX imaging's uniquely short prediction horizon and prospective biometric targeting.
Note that 360° rotation is not practical for GREX imaging because the high number of projections needed to reconstruct the torso's 3D volume increase clinical procedure time and increase patient dose of radiation. Instead, GREX imaging may use projection angles ranging up to 90° between −45°≤β≤45° or 0°≤β<90°. One skilled in the art would understand that experimental testing may identify more optimal projection angle ranges, but the theoretical projection angle ranges does not exceed 90°. In some cases, the −45°≤β≤45° is likely preferable to the 0°≤β<90° because the −45°≤β≤45° keeps the radiation dose from the imaging procedure as low as reasonably achievable. The x-ray photons traverse less human tissue in the −45°≤β≤45° projection angle range compared to the 0°≤β≤90° projection angle range, and therefore can be lower energy photons thereby depositing less dose, as shown in
To produce a high quality 2D image without delivering excessive radiation dose, photon energy must be high enough to partially penetrate the patient's body but not too high so that the photons totally penetrate the patient. Thicker patients require higher photon energy than thinner patients. When |ß|>45°, the human body is substantially thicker than when |ß|<45°. In general, as β→0° the photon energy decreases. In some embodiments, the GREX imaging system 100 acquires six breathing phases at five distinct projection angles for a total of 30 projections although other different numbers of projections may be possible depending on the specific application of the GREX imaging technology. For example, in breast tomosynthesis, the symmetric 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 breast. Furthermore, the breasts do not move when placed in a cradle, a typical clinical tomosynthesis approach that could be analogized to breath-hold lung imaging.
With the lung and heart motion challenge resolved by the GREX-based geometric definition of breathing phase plus GREX-based highly accurate and fast prospective targeting prediction algorithm (Section 2.1 and 2.2, respectively), the GREX imaging system 100 can handle the varying torso curvature with the biometric surface information gathered by 3D Spatial Positioning Localizer (Section 1.1). Biometric surface information also assists in the image post-processing to account for attenuating tissue density in the imaging field, thus quantifying previously neglected sources of attenuation and ultimately resulting in higher fidelity image reconstruction. In sum, the GREX imaging technology makes it possible for the digital tomosynthesis to work for non-breath-hold (“dynamic”) lung and heart imaging.
GREX-based approach to 3D volume reconstruction can be either patient-specific (personalized medicine, with a personalized number of discrete angles and arc subtended) or used as a “universal minimum procedure time and universal minimum delivered dose” (in the neighborhood of 5 discrete angles, plus or minus 3 angles depending on the statistical reconstruction methods used and number of prior GREX-based datasets available for the person).
As shown in
In some embodiments, the GREX imaging technology allows the projection angle to vary during one breathing phase whereas multiple projection angles captured at different time points (as defined by GREX imaging's quantitative definition of “breathing phase”) still all correspond to the same breathing phase because they are deemed to be all capturing one chest geometry. Moreover, the GREX-based tomosynthesis approach allows the acquisition of depth information because the detector's final photon count numbers across all pixels and the distribution of photon counts in space at the detector surface are reflective of a single chest geometry that has been probed at multiple angles.
Note that the GREX imaging technology allows the x-ray unit 108 and detector panel 304 to be mounted on unmotorized arms or stands. The manual arm's role in an exemplary GREX imaging procedure is described and illustrated in
By moving the arm only 4 times during the procedure, the GREX imaging technology minimizes the length of the procedure, the extent of clinician-equipment interaction during the procedure, and the extent of wear-and-tear to the x-ray arm because the clinician is not interacting with the x-ray arm as much as moving the arm from ß1 through 135 for all 6 breathing phases (e.g., performing 30 x-ray equipment reorientations).
In some embodiments, the 3D Spatial Positioning Localizer allows for a-priori position checks and safety interlocks of three individual elements separately—namely the patient's posture and position, the x-ray unit 108's position, and detector's position in space—as well as the collective consistency of each element's position/alignment upon an axis & relative to the position of the other elements. This a-priori position check and safety interlocks result from GREX imaging's defining the breathing phase geometrically. Therefore, the GREX imaging system 100 (via the prospective breathing phase prediction algorithm) is inherently suitable for giving users software-based safety and quality assurance control capabilities that (in the case of a safety interlock) prevent the triggering algorithm from initiating “beam on” if either (or both) the x-ray arm or detector or patient posture were incorrectly positioned in space (or inconsistent) for a specific angle, ß.
In some embodiments, since the 3D Spatial Positioning Localizer records the coordinates of all equipment during a procedure, image reconstruction techniques can benefit from the (a-posteriori) quantification of each ß angle and its associated uncertainty.
2D projection data is acquired six times (for early inhale (EI), late inhale (LI), max inhale (MI), early exhale (EE), late exhale (LE), and max exhale (ME) at each projection angle (ß). For simplicity, the projection angles range from 0° to 90°. The x-ray unit 108 will move to the next projection angles only when all breathing phases are acquired at the prior projection angle. The 2D projections are sorted according to breathing phase so that a static image cube is reconstructed (e.g., from the βi=1 to the βi=n projection angles that probed the chest geometry at (v1,f1)) to represent the chest volume at each breathing phase. The static image cubes are then temporally interpolated using the methods discussed in Section 3.3 below.
The x-ray projections acquired at each angle ß are taken at the targeted breathing phases identified using the ARIMA model (Section 2.2). After the targeted breathing phases are acquired, the x-ray unit 108 moves to the next imaging angle position.
The breathing phase-sorted projections are used to create static image cubes through the previously discussed canonical FDK image reconstruction algorithm given in Equation (1) (or a similar cone-beam geometry image reconstruction algorithm). The image reconstruction algorithm uses the breathing phase-sorted projections and creates breathing phase sorted, static image cubes. Each breathing phase will have a separate image cube. The static image cubes are called static because they represent anatomy in only one breathing phase. Static cubes that represent all of the targeted breathing phases are combined and temporally interpolated (described in Section 3.3) to create a 3D image cube movie from 3D static image cubes.
GREX imaging system 100 keeps the radiation dose to the lowest levels reasonably achievable through statistical image reconstruction. Each acquired image increases the dose of entire imaging process (a clinically undesirable consequence) but provides additional information for image reconstruction (a clinically desirable result). Traditional forms of image reconstruction, based on the Fourier transform or filtered back projection, have the tendency to display image artifacts due to an inability to handle missing information, e.g., missing projection angles, ß. For example, if a projection is taken every 10° instead of 5°, there will be half as much information available for creating a static image cube, but only half the dose was delivered with the former compared to the latter. Statistical iterative image reconstruction addresses the missing information caused by having an incomplete image dataset.
There are numerous statistical image reconstruction algorithms that already exist which GREX imaging system 100 can use to complete the task of image reconstruction (e.g., to construct the static image cubes). However, GREX imaging system 100 improves on traditional statistical image reconstruction algorithms by implementing a unique feedback step and complying with the law of conservation of mass-based boundary condition.
Basic principles of physics can be applied to GREX's statistical image reconstruction because GREX images are biometrically defined by biophysical quantities that are governed by physical law. By biometrically tagging each image and the resulting image cube, as well as by collecting the continuous biometric data stream during the procedure even when not imaging the patient, the breathing dynamics of mass exchange (inhalation and exhalation) and volume change are known (making it possible to solve for the unchanging lung tissue mass that is consistent over the course of the scan). The law of conservation of mass can be applied because static image cubes can be constructed from a moving organ such as lung for the first time. This is due to the prospective prediction/triggering algorithm's speed and accuracy (Section 2.2 and 2.3) which accurately labels and acquires the same chest geometry at different moments in time. Stated differently, the tissue mass in a static image cube should not change (e.g., does not change due to the law of Conservation of Mass) from one static image cube to the next. Based on the ideal gas law, the ratio of room temperature air to air inside lung is 1.11. Given the tidal volume of the image cube from sensor data, and with the 1.11 ratio and the mass/volume air curves for deviations from room temperature, the mass of the inhaled air (in absolute terms and also in relative terms as a ratio between two different breathing phase image cubes) can be determined.
GREX imaging's conservation of mass-based boundary condition is highly useful because, for instance, the presence of air may artificially darken a voxel, thereby adversely affecting the ability of the statistical image registration algorithm to accurately determine object density. By correcting acquired projections for air volume differences and consistently isolating a quantity that should be constant (e.g., lung tissue mass) over the course of the scan, GREX imaging produces more accurate image reconstruction for generating the static image cubes.
Considering two GREX projections taken at different imaging angles but with the same biometrically defined breathing phase, the volume of air in the lung is identical, but the way the air displaces tissue might be different between the two projections and dim the brightness of a nodule in the second projection, which was affected and is erroneous, compared to the first projection. The consequence of this error is that a piece of tissue (“the nodule”) that was visible in the first projection is not visible in the second projection, which will ultimately dull the intensity of “the nodule” (or cause it to be mistaken for background) in the resulting image cube. GREX imaging's law of conservation of mass boundary condition is implemented as a feedback step that would check that lung mass was conserved between the aforementioned image cube with the erroneously dim nodule and a different (correct and anatomically representative of “the nodule's” brightness) image cube from a later breathing phase. GREX imaging's feedback step corrects at the level of the erroneous image cube's second projection during reconstruction by updating the expected geometry based on a simulation using the first projection as the gold standard. In this way, GREX statistical image reconstruction would produce more accurate static image cubes.
In addition to pre-existing statistical image reconstruction algorithms, GREX's post-processing software 106 incorporates edge defining filters (discussed in Section 3.2), spatial boundary conditions (discussed in Section 3.2), and smooth transitions between breathing phases (discussed in Section 3.3) into the digital diagnostic x-ray images to improve anatomic imaging.
Section 3.2—Image Filters
A digital diagnostic x-ray image's quality depends on the x-ray unit 108 settings and the image's anatomic study site. Each patient and anatomic site has different electron densities, through which x-rays traverse to generate an image. For example, imaging the femur requires a higher x-ray energy than the x-ray energy required to image the chest because the lung is mostly comprised of air, and the femur is comprised of bone. Given that higher energy x-rays penetrate the body at to a greater extent than lower energy x-rays penetrate the body, the number of x-rays that emerge from the body to reach the flat-panel detector are different for high energy x-rays and low energy x-rays imaging the same anatomic geometry. Too many x-rays emerging from the body results in the flat panel detector's overexposure, in a manner similar to overexposure in optical photography. If the x-ray unit 108 setting is not optimal for the anatomic image study site, the image quality will be greatly reduced. In clinical practice, commercial vendors have devised imaging protocols for their digital diagnostic x-ray units that roughly estimate optimal x-ray unit settings for a selected anatomic site. But these rough estimates of optimal tube settings are not customized to address the potentially significant anatomic variations between anatomic sites within different patients (e.g. the stomach of an overweight man versus the stomach of an average weight man). Indeed, the vendor community's existing imaging protocol settings, as just rough estimates, seldom result in an optimal quality image.
If optimal x-ray unit settings or other imaging parameters are unknown prior to imaging (as currently the case within the medical community), the strategic deployment of digital image filters can improve the image quality of non-optimal x-ray unit settings. An improved image enhances the visibility of anatomic features that are poorly visible to the human eye. For example, in the coronal plane, every rib might not be visible in a digital diagnostic radiographic image. The post-processing software 106 filters the coronal plane images with an edge enhancing filter, like a Laplacian filter, to display the boundary of all ribs on the resulting image even when the rib boundary on the original image is too subtle for the human eye (e.g. the radiologist's eye) to detect. The post-processing software 106 overlays the filtered image with the original image, which highlights the enhanced (e.g., post-filtration) and previously invisible rib edges onto the original image. Image filters that will be made available for the user to apply include the Laplacian filter, Hanning filter, Butterworth filter, Parzen filter, Wiener filter, Metz filter, Ramp filter, nonlinear spatial mean filter, and hybrid filters.
In some embodiments, the post-processing software 106 uses the 3D spatial position localizer 300's skin surface measurements to calculate optimal imaging parameters for images taken at each breathing phase. As a patient breathes, the electron density of the body changes as more air is inhaled and the chest circumference increases. Increasing patient diameter, decreasing distance between the patient and the x-ray unit 108, and decreasing distance between the patient and the detector panel 304 produce additional image noise in the resulting x-ray images. The 3D spatial position localizer 300 tracks the patient's skin surface position for each image in relation to the x-ray unit 108 and detector panel 304. This skin surface position tracking provides a unique measurement for digital diagnostic x-ray studies.
The digital diagnostic radiology field currently relies on scaling ion chamber-measured dose index readings to the patient's approximated body-mass index. A radiology technician currently only takes two measurements: the first measurement with a tissue density equivalent solid water cylindrical phantom with a 16 cm diameter, and the second measurement with a 32 cm diameter materially-identical phantom. The x-ray unit 108 has built-in protocols that are vendor-defined and “one size fits all (patients)” for a particular anatomic location. For example, regardless of a patient's chest diameter, the vendor provides only a single protocol with built-in imaging settings for the technologist to select. In other words, a man with a fat chest gets the same imaging settings as a man with a skinny chest.
The 3D spatial position localizer 300 of the GREX imaging system 100 produces a real-time and personalized measurement of a patient's chest diameter. The measurement will inform the technologist in selecting x-ray unit settings that are personalized for the patient. When the patient is breathing, the chest diameter is changing. A changing patient chest diameter prevents the technologist from optimally setting the imaging parameters to match the patient's chest diameter. Moreover, the real-time and personalized measurement of a patient's chest diameter can be used to remove image noise in post-processing procedures, and to simulate an x-ray image taken with optimal imaging parameters.
The post-processing software 106 overlays a human skeleton model (which is individually scaled for each patient) onto the surface position estimate provided by the 3D spatial position localizer 300 in order to estimate the lung's initial location. The individually scaled patient skeleton is rigidly registered to the x-ray image using visible landmarks on the surface of the skin (e.g., clavicle, rotator cuff, scapula, vertebrae, etc.). After the skeleton has been rigidly registered to the x-ray image, the skeleton provides the position of the rib cage. The rib cage itself provides a boundary condition for the edge of the lungs and for the pixel values close to the trachea's noise-free air density estimate. Both the edge of the lungs (via the rib cage location) and the pixel values close to the trachea's noise-free air density estimate are automatically identified by the region growing algorithm as seed locations (from which the region growing algorithm will initiate and subsequently grow radially outward from). Region growing is a region-based segmentation method. This approach to segmentation first identifies a set of initial seed points within an image and then examines neighboring pixels of the initial seed points and determines whether the neighboring pixels should be added to the region. The process is iterated on, in the same manner as general data clustering algorithms. In other words, region growing algorithms use initial placement of seed pixels to expand outward using a statistical process that annexes “similar” pixels. The region growing algorithm will continue (to annex similar pixels) until the identified pixels are statistically dissimilar from the annexed cluster.
In practice within the GREX imaging system, the region growing algorithm “stops” (e.g., detects pixel dissimilarity) at important anatomic landmark interfaces (such as the lungs, which are bounded by the intense pixels of the rib cage). Pixels that the region growing algorithm does not identify as belonging to lung tissue will be masked (a mask image is defined as an image that enhances structures once subtracted from an original image) to form two separate images. Those two separate resulting images are (i) the segmented lungs and (ii) the remaining body tissues. In order to provide an accurate and noise-free segmented lung volume that gives the radiologist increased diagnostic visibility (e.g., a lung volume that is visually unobscured by non-lung tissue), body mask-related (e.g., non-lung and therefore visually uninformative) tissues are removed from the lung image. For example, the pixels belonging to the intercostal muscles within each simulated imaging ray projection are altogether subtracted from the segmented lung image. In addition to the aforementioned subtraction of the body mask from the lung image to yield improved lung tissue visualization, the body mask can also be used to provide a second check of the 3D spatial position localizer 300's calculated patient surface position. For example, the post-processing software 106 calculates the number of pixels that the region growing algorithm identifies as the body mask, and then calculates the body diameter at various locations along the height of the torso. This body diameter calculation should closely agree with the 3D spatial position localizer 300's estimate of the patient's body diameter. If not, it might indicate that the 3D spatial position localizer 300 needs to be re-calibrated to increase its accuracy.
If a clinical user desires a more accurate view of the body mask (e.g. for clinical or educational reasons), the body mask images are simulated with the optimal x-ray unit settings, thereby removing forms of noise and potential sources of artifacts from the body mask. The body mask and the segmented lungs could then recombine to form artifact- and noise-free x-ray images with global enhancement for clinical applications like structure contouring.
Section 3.3—Biomechanical Modeling
The biomechanical models used in the post-processing software 106 are created from the first principles of physics, namely the law of conservation of mass and the ideal gas law. The goal of the post-processing software 106's biomechanical models is to determine biophysical quantities that enhance a clinician's ability to diagnose disease. Relevant biophysical quantities include, but are not limited to, stress and strain of lung tissue elements.
A mechanical system loaded with a force produces stress. In the context of the lung, an element of a mechanical system is represented by lung tissue. Lung tissue that is visible and distinguishable in medical imaging consists of parenchyma (including alveoli sacs, alveoli walls, bronchial tubes, and blood vessels). Parenchyma is directly responsible for lung function. Suitable tissue elements for biomechanical modeling should be small enough to be homogeneous inside, but statistically stable in response to respiratory stimuli. Typical voxel sizes in lung medical imaging range from 1 mm3 to 3 mm3, which equates to 125 to 375 alveoli. The voxels are deemed to be nearly homogeneous in density and contain enough alveoli to provide a stable response to respiratory stimuli. The alveoli are arranged in hexagonal arrays that inflate due to distending normal stress from each shared alveolar wall. The sum of all distending normal stresses within a lung tissue element provides an estimate of the pressure experienced by the alveoli and caused by respiratory stimuli. The distending stress is counterbalanced by the recoil stress on the alveoli wall, and an equilibrium exists between the two stresses when the airflow through the tissue element is 0. Spatial translation of a lung tissue element's position, in any one direction, due to changing lung tidal volume can be modeled by the tissue element's material response to normal stress on the lung tissue element's corresponding face. In other words, the biomechanical model includes vectorized terms that describe the lung tissue element's response to increasing tidal volume, wherein the vectorized terms relate to normal stress.
Strain is defined as the response of a mechanical system to stress. From the perspective of an element of material (e.g., tissue), stress is a deforming force, and strain is a restoring force. Stress vectors on the lung tissue element's face contains two components: (i) normal stress (related to the lung tissue element's outward or inward directional motion which results in either expansion or contraction, respectively), and (ii) shear stress which is perpendicular to the normal stress and is caused by airflow-induced pressure imbalances. Note that, by definition, the perpendicular component of shear stress does not contribute to lung volume changes. An illustration of the relationship between normal stress/shear stress and tidal volume/airflow, can be found in the “airflow vs tidal volume” plot depicted in
The model creates a new diagnostic perspective within the field of medicine because medical analysis, using the model, is governed by quantitative force analysis. The balance of stress and strain forces at each voxel in the lung fundamentally assesses lung health because the lung's functional behavior (e.g., lung tissue motion) can now be both visualized and analyzed from a clinical perspective.
The first step in building the model is to identify what tissues are part of the lung and what tissues are not part of the lung. To accomplish this task, GREX imaging technology acquires multiple snapshots of the chest geometry in multiple configurations (6 distinct breathing phases). The lung is segmented based on region growing segmentation algorithm using the trachea's air density as an initial starting point for the region growing algorithm. GREX imaging system 100 uses two different deformable image registrations for building the biomechanical model: one acts on lung tissue and the other acts on non-lung tissue (e.g. chest wall, ribs, liver, heart, trachea, esophagus, etc.). The results of a region growing algorithm distinguish between lung tissue and non-lung tissue prior to the deformable image registration step.
The justification for GREX imaging system's using two different deformable registrations, one for lung and the other for non-lung, is that the motion and material properties of lung are different from those of non-lung. If just one deformable image registration is used for the totality of lung and non-lung, the registration would assign a greater weight to the non-lung tissue compared to the lung tissue. This would result in the lung tissue being assigned an unrealistically low weight, and therefore unrealistically limited extent of motion. GREX imaging technology is built to visually elucidate the subtleties of lung motion, including the subtleties of motion at the computationally complex lung surface.
One deformable image registration algorithm that can track tissue movement between images is the optical flow algorithm.
Since 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 internally more homogeneous regions): (i) the image registration of the lung, and (ii) the image registration of the non-lung. In some embodiments, to parse the two separate task regions from the chest as a whole, identification of the lung surface (e.g., the boundary between the lung and non-lung) is required prior to the 3D multi-resolution optical flow deformable image registration. Identification of the lung surface is fundamentally possible using the region growing algorithm, which begins inside the lung in air-filled (e.g., visually dark) regions, grows outward towards the lung surface boundary, and encounters a high pixel contrast at the lung surface boundary. Performing the region growing algorithm is the first step.
The lung is not attached to the chest wall. As a result, the lung's motion is relatively independent from the chest's motion. In other words, instead of a predictable push-pull response at the lung surface boundary, there are other types of tissue dynamics at work. For example, a horizontally adjacent chest voxel moves downward vertically, so the lung voxel moves horizontally into the space where the chest voxel previously resided.
In order to accurately model the complex motion dynamics, GREX imaging system 100 quantifies the shear force that the surfactant layer experiences at the lung surface boundary. Using two separate segmentations for lung and non-lung tissue provides the basis of force estimation. The procedure of force estimation is performed by subtracting the segmented lung pixels (e.g., remove their assigned value) from an x-ray image (performed for each reconstructed breathing phase image). The previously segmented lung will be masked from the original image which provides an image containing all remaining tissues. Each individual image's segmented chest geometry must be registered to each of the chest geometries of the other images so that the position of each lung tissue element is known across all images. The multi-resolution optical flow algorithm performs the image registration by calculating a displacement vector field that shows the displacement of every pixel between two images taken at different breathing phases. Indeed, knowing the displacement vector field allows accurate spatial accounting of all lung tissue elements in the chest geometry. Differences between the segmented lung registration's displacement vector field and the non-lung tissue's displacement vector field provide the magnitude and direction of the shear force that exists between the lung and the chest wall.
Based on observation of over 150 unique patients, the relationship between displacement and tidal volume is linear. The relationship between displacement and airflow is also linear. The multi-resolution optical flow algorithm's output is displacement vectors in a coordinate space for each of the measured tidal volume and airflow magnitudes. When the displacement vectors are calculated for all breathing phases, the result is a closed loop trajectory (as shown in
In some embodiments, the biomechanical model approximates the motion of lung tissue to be a function of multiple factors, including the lung's tidal volume (Tv), airflow (Af), and cardiac phase (Hc). These values are global values, e.g., the cardiac phase is the same for all tissue elements in the chest. Global values are treated as scalar numbers and are measured by the hardware discussed in Section 1. Note that the tidal volume, airflow, and cardiac phase are all time varying measurements. The unique stress and strain values for each tissue element is mathematically expressed by the equation below using the vectors 1,
2,
3, and
4:
−
0=TV
1+Af(
2+
3)+Hc
4 (2)
wherein the 1 vector describes normal stress caused by tidal volume,
2 describes normal stress caused by airflow,
3 describes shear stress caused by airflow, and
4 describes tissue motion introduced by the perturbing heart motion. Altogether, the displacement (
−
U0) of tissue at any point in a closed loop trajectory is expressed as a summation of the stress, strain, and perturbing heart motion vectors scaled by the tidal volume, airflow, and cardiac phase respectively.
A major advantage of using a physiologically-based, biomechanical model to interpolate images between acquired breathing phases is the ability to use quantitative physical quantities to check the accuracy of the biomechanical model's output results. According to the ideal gas law, the ratio of lung volume change to tidal volume is 1.11 at room temperature. In other words, the ratio of room air density to lung air density is 1.11. Accordingly, the volume integral of the divergence of the normal stress vector should also be 1.11 (e.g., ∫V (·
2) dV, where V is the total body volume). The ideal gas law provides a “sanity check”, yielding useful quality assurance information on the 3D spatial position localizer 300, breathing phase sensor 110, and each interpolated image.
An example of how the GREX imaging technology's biomechanical modeling enhances diagnosis is an early stage lung tumor that is not visible to a radiologist's eye at the time of image inspection. The tumor is not visible to the radiologist because it is too small for the imaging sensitivity. Even though the tumor is not visible to the radiologist, its presence still affects the balance of forces inside the lung since a tumor's electron density is greater than healthy lung tissue's electron density. The tumor's higher electron density means that the tumor has different material and mechanical properties (e.g., different characteristic stress and strain parameters) that affect both the tumor's motion and the motion of the local area around the tumor (e.g., healthy tissue close to the tumor site). The effect of a tumor on the local healthy lung tissue can be loosely analogized to the effect mass has on the space-time continuum under the general relativity theory: when a massive object is present, the space-time around the object bends so that light behaves differently in the object's vicinity compared to the light's behavior when the mass is absent. The same analogized principle applies in the lung, such that a tumor warps the motion trajectories of adjacent healthy tissues into moving differently compared to a healthy lung's motion trajectories. The biomechanical model's displacement vector map makes changes in lung tissue composition and biomechanical properties readily apparent to medical practitioners: when a tumor is preset, the displacement vector field shows an unnatural level of vector curl and/or other altered properties. GREX imaging technology's post-processing software 106's creation of parameter maps, discussed in Section 3.4, visually displays such previously invisible, diagnostically important information for the end-user. GREX imaging technology's parameter maps are examples of the new diagnostic perspective that the GREX platform brings to medicine.
The biomechanical model's process provides a quantitative means for biometrically interpolating between two images (2D or 3D) acquired at different breathing phases. Equation (2) shows that the solution for the biomechanical model is the displacement between two breathing phases (−
0). As noted above, the displacement between two breathing phases (
−
0) is found by performing deformable image registration to index the two breathing phases to one another, which is typically comprised of the following three steps:
−
0) quantifies the extent of voxel bubbling or the deformation of the corresponding lung tissue. Deformable image registration assumes that the reference voxel in the EI image cube moves to its new position in the LI image cube while deforming at the same time due to the inhalation of more air into the lung tissue. As the breathing cycle moves forward, the reference voxel keeps its motion trajectory as depicted by the MI, EE, LE, and ME image cubes, respectively. In other words, for every voxel in the reference image cube, a set of displacement vectors are computed across all the six image cubes.
) at an image cube corresponding to each of the six breathing 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:
[1=0,
2,
3,
4,
5,
6] or denoted as [
n],
wherein the parameter “n” is a time-dependent parameter corresponding to a respective breathing stage.
Similarly, for a particular breathing stage “n”, the biometric data matrix including the lung's tidal volume (Tv), airflow (Af), and cardiac phase (Hc) can be expressed as:
[Tv
For each voxel in the image cube, the biomechanical model for the parameter matrix, []=[
1,
2+
3,
4], can be solved for using the six displacement vectors ([
1=0,
2,
3,
4,
5,
6]) and the corresponding biometric data matrixes ([Tv
[]=[Bn]=[
n].
As noted above, there are many deformable image registration algorithms capable of performing image registration for GREX imaging, including the 3D multi-resolution optical flow algorithm. The 3D multi-resolution optical flow algorithm calculates smooth (e.g., fluid-like) transitions between images that are taken at different observed tidal volumes. The calculated displacement between two breathing phases (−
0) provides observed points in the tissue trajectory shown in
0
1,
2,
3,
4
The post-processing software 106's biometric interpolation between acquired image cubes populates all potential breathing phases so that a complete movie of the chest is created. Generally speaking, a movie needs at least 30 simulated images to smoothly transition between frames.
Tv=20 ml, (Af)=20 ml/s, and Hc=0.10.
For each voxel in an interim simulated image cube at a particular moment of the breathing cycle, e.g., EI+Δt, EI+2Δt, EI+3Δt, etc., can be calculated using the biomechanical model for the parameter matrix, [], and a corresponding biometric data matrix for the particular moment.
Section 3.4—Parameter Maps
One of the clinical benefits of GREX imaging technology is the unique parameter maps. Using 2D plots, 2D color washes, 3D plots, and 3D vector field maps, the GREX imaging technology presents the end-user with previously unavailable information on the patient's chest health.
First, let's consider the 2D color map of 1 in
i, the end user would expect that the magnitude of
1 will be greater near the diaphragm compared to the apex of the lung (e.g., the diaphragm has greater tissue displacement than the apex of the lung when tidal volume is increased). Generally speaking, the magnitude of
1 varies smoothly throughout a healthy lung. If the gradient of
1 was calculated across the lung, the result would be a smooth function. The examples depicted in
1 parameter is similarly distributed, the healthy patient has twice the tissue displacement as the sick patient.
Next consider the 2D color wash of (2+
3), which represents the sum of normal and shear stress associated with airflow. Generally speaking, higher parameter magnitudes occur close to regions where the bronchial tree brings air into the lung at a faster rate (mid-lung). But in a diseased lung, the presence of a tumor may substantially change the behavior of the lung's elasticity substantially such that the tumor may be visually distinguishable from the functionally healthy lung tissue distribution. In the examples shown in
1) and airflow (Af (
2+
3)) located below the 2D color washes also displays a clear indicator of disease in the left lung. Healthy lungs have a bimodal distribution of Af (
2+
3) but when disease is present in one lung, the overall distribution pattern is different in both lungs. GREX imaging technology may involve Bayes processes to leverage patient interviews (information voluntarily provided by the patient before the exam) with the distribution of the biomechanical model's parameter histograms to better classify disease etiology. In sum, consistent indicators of lung disease (e.g., of a tumor's presence, as shown in
Section 3.5—Diagnostic Disease Pointers
Lung cancer, chronic obstructive pulmonary disease (COPD), lower respiratory tract infection, and tuberculosis all have disease pointers that may be visible in existing digital diagnostic x-ray imaging. However, diagnostic disease pointers are not always visible at early disease stages. But as shown in
Another diagnostic example of where GREX would disrupt the diagnostic community is in classifying lung disease etiology. The presentation of asbestos caused lung disease and cigarette caused lung disease are different in biopsies. Asbestos is a naturally silicate minerals that consists of long, thin fibrous crystals, composed of millions of microscopic fibers. Inhaled asbestos fibers penetrate the alveoli and eventually forms a dense web that compromises the alveolar function which decreases lung function. The dense webs collect cancerous tissues and is named mesothelioma. Standard radiographs can only identify mesothelioma by the plaque buildup, which presents as dense consolidation (fuzziness) in the lung image. GREX may detect the subtle presence of asbestos through the parameter maps before an end-user could visually identify the presence of asbestos. As the web if asbestos is forming, the elasticity of the lung is reduced locally. The small local reduction in elasticity will be visible in the parameter maps (2D color washes, parameter ratios, and histograms). For example, in tissue that has lost elasticity, the tissue's trajectory during breathing will be more circular than elliptical. This means that the ratio between the and (
2+
3) magnitudes would be higher than healthy tissue. Consulting the color map would clearly show a region of the lung that is displaying a pattern indicative of disease.
Inhaled particles (non-fiber) would display completely differently than the fibrous asbestos particles. Inhaled particles deposit within the lungs, reducing lung function by “clogging” airways, forming scar tissue, and forming tumors but not webs. GREX will show the beginning of particulate deposition by detecting minute local changes in lung motion dynamics that are inconsistent with healthy characteristics. Cigarettes contain tar and radon, which adhere to the alveoli and drastically changes lung function. GREX can track the reducing lung function and give end-users a unique tool to better demonstrate to a patient how destructive the smoking habit is. COPD is another disease that can be elucidated with parameter maps more effectively than the current clinical methods. COPD is currently detected with spirometry testing and standard radiography. Neither of these methods are particularly sensitive and do not detect COPD at the earliest onset. Detecting disease at the earliest possible time can give patients more opportunity to take advantage of preventative medicine and change bad habits before the it is too late.
Section 3.6—Movie Presentation
The point of the movie presentation graphic user interface (GUI) is to provide a clean and intuitive canvas for displaying the planar movie and 3D movie results. All functions like rotate view, pause movie, and ruler are placed into bins that are represented with tabs across the top of the screen. Selecting a function changes the tooltip to show the user what function is in use. The GUI is designed to be light weight so that the GUI will function even with older computer systems. Finally, the movie presentation GUI will work in conjunction with the annotation/contouring, possible disease pointers, and cloud computing platform.
Medical personnel end-users prefer to directly highlight and place notes on medical images rather than attaching additional documents to the images. Medical end-users clamor for the ability to include image annotation (“attention-grabbing” arrows and treatment notes) and anatomic contouring (drawing a virtual line around an anatomic structure) to imaging studies. Medical end-users are often members of large, multi-disciplinary care teams that collaboratively treat a patient. The following example describes the current workflow and technological capabilities of a clinical care team:
Once an imaging study is performed on a lung cancer patient, the image is read by a radiologist, prescribed radiation treatment orders by a radiation oncologist, and used as a basis for surgical intervention by an operating surgeon. Specifically, the radiologist draws a contour around the tumor site and passes the contoured image to a radiation oncologist.
The radiation oncologist then writes instructions for the radiation treatment and/or tumor resection surgery based on the image.
Using the radiology contour and the radiation oncology tumor resection instructions, the surgeon resects the tumor.
The post-processing software 106 provides tools for end-users to annotate and contour directly on the moving images, which streamlines healthcare workflow. This user-friendly capability enhances healthcare workflow and reduces the probability of medical mistakes. The interactive nature of our user interface means that treatment instructions and/or medical concerns are now plainly visible to all users, and the relevant clinical notes are displayed correctly in the corresponding anatomical region of interest.
The method includes positioning (3102) the patient at a first orientation relative to an x-ray imaging apparatus (e.g., GREX imaging system,
In some embodiments, positioning the patient at the first orientation includes moving (e.g., rotating) the x-ray imaging system, while the patient remains at a fixed position (e.g., as discussed with reference to
In some embodiments, while the patient is positioned at the first orientation relative to the x-ray imaging apparatus, the patient positioning fixture maintains the patient in a fixed position (e.g., stabilizes the position of the patient) such that three-dimensional images of the patient's lung can be reconstructed (e.g., as described throughout this disclosure) based on a “stationary object” assumption.
The method includes obtaining (3104) a volumetric measurement of a patient's breathing (e.g., as the patient is breathing normally; the patient need not hold his or her breadth). In some embodiments, the volumetric 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., a flow-rate). In some embodiments, as explained in greater detail below, the volumetric measurement of the patient's breathing is a measurement that can be converted to the patient's tidal volume (e.g., by measuring chest rise and fall). The term tidal volume, as used herein, means a difference between a current lung volume and a predefined baseline (e.g., a volume during maximum exhalation of a normal breath without extra effort applied, a volume during maximum inhalation of a normal breath without extra effort applied, or any other suitable fiducial volume).
In some embodiments, the volumetric measurement of the patient's breathing is a geometrical (spatial or positional) measurement of the patient's breathing.
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 volumetric measurement of the patient's breathing is obtained using (3108) one or more volumetric breathing phase sensors (e.g., breathing phase sensor 110,
In some embodiments, the 3D point cloud is used to determine the volumetric measurement of the patient's breathing without generating a mesh reconstruction of the patient's chest (e.g., the raw output of the 3D point cloud is used to generate the volumetric measurement of the patient's breathing without first generating a mesh). In some embodiments, the point cloud of the surface of the patient's chest is obtained using (3112) a 3D imaging technique to measure one or more positions of the patient's chest. For example, the 3D imaging technique comprises a laser scanning technique such as light detection and ranging (LIDAR). Such laser scanning techniques are beneficial because some lasers can accurately measure positions of the patient's chest even while the patient is wearing a modesty garment (e.g., the patient dons a LIDAR-transparent garment while undergoing method 3100).
In some embodiments, the method 3100 further includes identifying (3114) one or more anatomical landmarks on the surface of the patient's chest (e.g., externally-visible landmarks, such as a clavicle, trachea, sternal notch, sternum, xiphoid process, spine, or humorous) using the point cloud of the surface of the patient's chest. The method further includes inferring a location of one or more internal anatomical landmarks within the patient's chest from the point cloud of the surface of the patient's chest (e.g., a location of the patient's lung). In some embodiments, the method includes generating a reconstruction of the patient (e.g., a generating a computer model for the patient, also referred to as a “virtual patient”). In some embodiments, generating the reconstruction of the patient includes generating a reconstruction of the internal anatomy of the patient. In some embodiments, the reconstruction of the internal anatomy of the patient includes a computer model (e.g., a 3D model) of tissue densities. In some embodiments, the reconstruction includes or is used to determine absorption cross-sections (e.g., x-ray absorption cross-sections) at a plurality of locations within the body. In some embodiments, the reconstruction of the patient is used to determine an x-ray dose to be delivered (e.g., for each image) by the x-ray imaging apparatus.
The method includes, while the patient is positioned at the first orientation relative to the x-ray imaging apparatus, and while obtaining the volumetric measurement of the patient's breathing, determining (3116), based on the volumetric measurement of the patient's breathing, a breathing phase of the patient (e.g., in real-time, as the patient is breathing normally). In some embodiments, the breathing phase of the patient is defined by the volume of the lung. Thus, in some embodiments, determining a breathing phase of the lung includes determining a volume of the lung.
In some embodiments, operations 3102 et seq. are performed as part of an imaging period of method 3100. Method 3100 further includes, prior to the imaging period, undergoing a training period during which information about the patient's normal breathing is obtained. For example, during the training period, volumetric measurements of the patient's breathing are obtained at regular intervals over a plurality of cycles of the patient's breathing (e.g., 15, 20, 50 cycles of the patient's breathing, each corresponding to one breath). The volumetric measurements from the training period are then used to associate specific volumetric measurements with specific breathing phases. For example, a volumetric measurement corresponding to a tidal volume of 400 ml may be associated with maximum inhalation, a volumetric measurement corresponding to a tidal volume of 0 ml may be associated with maximum exhalation, and so on. In addition, statistics about the patient's breathing (e.g., a histogram) can be obtained during the training period, and used to verify that breaths taken during imaging are “normal” breaths (e.g., not overly deep breaths or otherwise anomalous breaths).
During the imaging period, in some embodiments, the breathing phase of the patient that is determined is (3118) a future (e.g., predicted) breathing phase. That is, in some embodiments, determining, based on the volumetric measurement of the patient's breathing, the breathing phase of the patient includes forecasting the future breathing phase from one or more current and/or past breathing phases. For example, the forecasting is based on a time-series of breathing phases. In some embodiments, forecasting the future breathing phase from the one or more current and/or past breathing phases includes generating an autoregressive integrated moving average (ARIMA) model. In some embodiments, the forecasting uses data from the training period.
The method includes, in accordance with a determination that the breathing phase of the patient matches a predefined breathing phase (e.g., that the volume of the lung matches a predefined lung volume), gating (3120) the x-ray imaging apparatus to produce an x-ray projection (sometimes called a projection image) of the patient's lung. In some embodiments, an x-ray projection is an image taken at a particular angle (e.g., determined by the orientation of the patient with respect to the x-ray imaging apparatus). In some embodiments, an x-ray projection is obtained using a single x-ray exposure. In some embodiments, the method includes, in accordance with a determination that the breathing phase of the patient does not match a predefined breathing phase, forgoing gating the x-ray imaging apparatus (e.g., forgoing exposing the patient to x-ray radiation).
In some embodiments, the method includes determining, from the volumetric measurement of the patient's breathing, whether a current breath is an irregular breath. The method further includes, in accordance with a determination that the current breath is an irregular breath, forgoing gating (e.g., obtaining) the x-ray imaging apparatus (e.g., continuing to wait for a suitable breath at which to obtain the x-ray projection for the breathing phase).
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 includes, while obtaining the volumetric measurement of the patient's breathing, in accordance with a determination that the breathing phase of the patient matches any of the plurality of predefined breathing phases, gating the x-ray imaging apparatus to produce a respective x-ray projection of the patient's lung. In some embodiments, x-ray projections (e.g., x-ray measurements) of the patient's lung are not obtained (3124) except when the breathing phase of the patient, as determined by the volumetric measurement of the patient's breathing, matches one of the plurality of predefined breathing phases, thus lowering the total amount of x-ray exposure of the patient.
In some embodiments, the plurality of predefined breathing phases includes (3126) an early exhalation phase, a late exhalation phase, a maximum exhalation phase, an early inhalation phase, a late inhalation phase, and a maximum inhalation phase of a complete breathing cycle of the patient (e.g., as shown and described with reference to
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 apparatus (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 apparatus, and while continuing to obtain the volumetric measurement of the patient's breathing, continuing to determine, based on the volumetric measurement of the patient's breathing, a breathing phase of the patient. In some embodiments, the method includes, in accordance with a determination that the breathing phase of the patient matches the predefined breathing phase, gating the x-ray imaging apparatus to produce a second x-ray projection of the patient's lung.
In some embodiments, the method further includes generating a static image cube, corresponding to the predefined breathing phase, using the first x-ray projection and the second x-ray projection (e.g., as described with reference to
As described elsewhere in this document, these thirty x-ray projections can be used to reconstruct a biomechanical model of how the lung moves in 3D. In some embodiments, the static image cube (e.g., as described with reference to
One of skill in the art will recognize that method 3100 can be applied to other types of movement besides lung motion due to breathing. For example, in some embodiments, a method includes positioning a patient at a first orientation relative to an x-ray imaging apparatus. The method further includes obtaining a 3D measurements of a portion of the patient's body (e.g., a 3D measurement of a location of the portion of the patient's body). The method further includes, while the patient is positioned at the first orientation relative to the x-ray imaging apparatus, and while obtaining the 3D measurements of the portion of the patient's body: determining, based on the 3D measurements of the portion of the patient's body, that triggering criteria are met for triggering exposure to radiation by the x-ray imaging apparatus; and in accordance with a determination that the triggering criteria are met, gating the x-ray imaging apparatus to produce an x-ray image of the patient. In some embodiments, the x-ray image is an image of the portion of the patient's body (e.g., the patient's leg, abdomen, skull, etc.). In some embodiments, the triggering criteria include 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 at a predefined location (e.g., with respect to the imaging apparatus). In some embodiments, the method includes, in accordance with a determination that triggering criteria are not met, forgoing gating the x-ray imaging apparatus (e.g., forgoing exposing the patient to x-ray radiation). Furthermore, in some embodiments, method 3100 is applicable to other types of imaging that are not strictly based on x-rays, for example positron emission tomography (PET) imaging or MRI imaging.
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 at a first orientation relative to a radiation therapy source. The method further includes obtaining a 3D measurements of a portion of the patient's body (e.g., a 3D measurement of a location of the portion of the patient's body). The method further includes, while the patient is positioned at the first orientation relative to the radiation therapy source, and while obtaining the 3D measurements of the portion of the patient's body: determining, based on the 3D measurements of the portion of the patient's body, that triggering criteria are met for triggering exposure to radiation by the radiation therapy source; and in accordance with a determination that the triggering criteria are met, gating the radiation therapy source to expose the patient to radiation (e.g., expose the portion of the patient's body to radiation). In some embodiments, the triggering criteria include 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 at a predefined location (e.g., with respect to the imaging apparatus). In some embodiments, the method includes, in accordance with a determination that triggering criteria are not met, forgoing exposing the patient to radiation.
It should be understood that the particular order in which the operations in
The method includes positioning (3202) the patient at a first orientation relative to a radiation source. In some embodiments, the radiation source is (3204) an x-ray imaging apparatus. In some embodiments, the radiation source is (3206) a radiation therapy source. For example, as shown in
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 a volumetric measurement of the patient's breathing, described above with reference to operation 3104 (
The method includes obtaining (3210) a measurement of the patient's cardiac function. In some embodiments, one or more sensors are used (3304) for measuring the patient's cardiac function. In some embodiments, an electrocardiogram (ECG) is used for measuring the patient's cardiac function (e.g., a 3-lead or a 12-lead ECG). In some embodiments, the method includes obtaining a plurality of measurements of the patient's cardiac function that provides a time-series of electrical signals that control movement of the patient's heart.
Returning to the imaging phase, the method includes, while the patient is positioned at the first orientation relative to the radiation source and while obtaining the measurement of the patient's breathing (3212), determining (3214), from the measurement of the patient's breathing, a breathing phase of the patient and determining (3216), from the measurement of the patient's cardiac function, a cardiac phase of the patient (e.g., in real-time). For example, in some embodiments, the cardiac phase of the patient is determined using a zero instruction set computer (ZISC) processor, as described with reference to
In some embodiments, the method further includes, before gating the radiation source to expose the patient to radiation (operation 3222, below), obtaining (3218) measurements of the patient's cardiac function from a plurality of cardiac cycles of the patient. In some embodiments, the method includes, using the measurements of the patient's cardiac function from the plurality of cardiac cycles, determining an average interval between a predefined cardiac phase and a beginning of the predefined window of the cardiac cycle. 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, operations 3202 et seq. are performed as part of an imaging period of method 3200. Method 3200 further includes, prior to the imaging period, undergoing a training period 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 one complete period of heart motion, such as from one T-wave to the next T-wave). The ECG measurements from the training period are then used to predict quiescent periods of heart movement during the imaging phase, as described below.
In some embodiments, the measurements from the plurality of cardiac cycles of the patient are (3220) waveform measurements (e.g., ECG measurements) of the plurality of cardiac cycles and the method includes validating, as statistically stable, the waveform measurements of the plurality of cardiac cycles.
The method further includes gating (3222) the radiation source to expose the patient to radiation based on a determination that the breathing phase of the patient matches a predefined breathing phase and a determination that the cardiac phase of the patient matches a predefined window of the cardiac cycle. In some embodiments, when the radiation source is gated, the patient's lung is exposed to radiation. In some embodiments, the predefined cardiac window corresponds to a quiescent period of heart movement (e.g., a period of time during the cardiac cycle during which the heart movement is minimal, as described with reference to
In some embodiments, the radiation source is an x-ray imaging apparatus. Gating the radiation source to expose the patient to radiation comprises (3224) gating the x-ray imaging apparatus to produce an x-ray projection of the patient's lung. In some embodiments, as described with reference to method 3100,
In some embodiments, the radiation source is a radiation therapy source. Gating the radiation source to expose the patient to radiation comprises (3226) gating the radiation therapy source to irradiate a region of the patient's lung at a therapeutic dose. In these circumstances, it is important to provide as much of the radiation dose as possible to diseased tissue (e.g., cancerous tissue), and as little as possible to healthy tissue. Method 3200 improves radiation therapy devices by delivering more accurate doses to diseased tissue while minimizing radiation provided to healthy tissue.
In some embodiments, determining that the cardiac phase of the patient matches the predefined window of the cardiac cycle includes predicting (3228) the predefined window of the cardiac cycle by, in real-time, detecting (e.g., using BEMP card shown in
It should be understood that the particular order in which the operations in
The method includes extracting (3302) multiple displacement fields of lung tissue from the multiple x-ray measurements (e.g., x-ray images, also referred to as x-ray projections) corresponding to different breathing phases of the lung. Each displacement field represents movement of the lung tissue from a first breathing phase to a second breathing phase and each breathing phase has a corresponding set of biometric parameters. In some embodiments, the x-ray measurements are x-ray projections (also called x-ray projection images) obtained in accordance with method 3100 and/or method 3200. In some embodiments, extracting displacement fields from the lung tissue includes identifying a portion of lung tissue in one or more first x-ray projections from a first breathing phase as corresponding to the same portion of lung tissue in one or more second x-ray projections from to a second breathing phase, and determining the displacement of the portion of lung tissue from the first breathing phase to the second breathing phase (e.g., as described with reference to
In some embodiments, the multiple x-ray measurements comprise multiple x-ray images, including an x-ray image obtained for the first breathing phase for each of a plurality of orientations of an x-ray imaging apparatus with respect to the patient, thereby forming a plurality of x-ray images corresponding to the first breathing phase and a plurality of x-ray images corresponding to the second breathing phase. In some circumstances, at least one of the x-ray images corresponding to the first breathing phase was obtained during from a different breathing cycle of the patient than a different x-ray image corresponding to the first breathing phase (e.g., the images were obtained during the same phase, but different breaths). In some embodiments, the method includes grouping the multiple x-ray images by breathing phase.
In some embodiments, the method includes extracting multiple vector fields, for which displacement fields are one example.
In some embodiments, one or more sensors are used (3304) for measuring biometric signals of the patient as one or more sequences of time series, including one or more of a 3D spatial position localizer (e.g., 3D spatial localizer 300,
In some embodiments, the biometric signals of the patient measured by the one or more sensors are used for triggering (3312) an x-ray unit to acquire an x-ray image of the patient at a specific breathing and cardiac phase (e.g., as described with reference to method 3200).
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 x-ray unit's clock. In some embodiments, respective values of the biometric signals are recorded to be associated with the acquired x-ray image.
In some embodiments, the biometric signals of the patient measured during a training window are used (3316) for building an optimized breathing prediction model for predicting a desired breathing phase at which an x-ray unit is triggered to capture an x-ray image of the patient.
The method includes calculating (3318) one or more biophysical parameters of a biophysical model of the lung using the multiple displacement fields of the lung tissue between different breathing phases of the lung and the corresponding sets 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., a 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 the different breathing phases of the lung and the multiple displacement fields of the lung tissue. In some embodiments, the set of biometric parameters associated with a respective breathing phase includes (3322) a tidal volume and an airflow of the lung at the respective breathing phase and a cardiac phase corresponding to the respective breathing phase of the lung. In some embodiments, the physical relationship between the biometric parameters associated with the different breathing phases of the lung and the multiple displacement fields of the lung tissue is defined as follows:−
0=TV
1+Af(
2+
3)+Hc
4
The 1 vector describes normal stress caused by tidal volume,
2 describes normal stress caused by airflow,
3 describes shear stress caused by airflow, and
4 describes tissue motion introduced by heart motion, the displacement (
−
0) of tissue at any point in a closed loop trajectory is expressed as a summation of the stress, strain, and perturbing heart motion vectors scaled by the tidal volume (Tv), airflow (Af), and cardiac phase (Hc) respectively.
In some embodiments, the different breathing phases of the lung include (3324) early exhalation, late exhalation, maximum exhalation, early inhalation, late inhalation, and maximum inhalation of a complete breathing cycle of the patient.
In some embodiments, the method further includes displaying a visual representation of the biophysical parameters. For example, 1 to (
2+
3). In some embodiments, displaying the visual representation includes displaying an image of the lung, wherein a color of a location within the image of the lung corresponds to the biophysical parameter (e.g., an image of the lung is displayed using a color map of the biophysical parameter). Displaying such visual representation improves the x-ray imaging apparatus itself by increasing the accuracy of diagnoses. For example, it is much easier to see that the patient's left lung is diseased in
In some embodiments, the method further includes generating (3326) multiple medical image cubes corresponding to the different breathing phases of the lung from the multiple x-ray measurements corresponding to the different breathing phases of the lung (e.g., as described with reference to
In some embodiments, the method further comprises choosing (3330) one or more of the multiple medical image cubes as reference medical image cubes, determining a set of biometric parameters associated with each reference medical image cube, selecting a set of biometric parameters based on biometric measurements of the lung between two sets of biometric parameters associated with two reference medical image cubes, and simulating a medical image cube between the two reference medical image cubes by applying the set of biometric parameters based on biometric measurements of the lung to the biophysical model.
It should be understood that the particular order in which the operations in
In some embodiments, one or more sensors are used (3402) for measuring biometric signals of the patient as one or more sequences of time series, including one or more of a 3D spatial position localizer, a breathing phase sensor, and a cardiac phase sensor (e.g., as described above with reference to method 3100 and method 3200).
In some embodiments, the method further comprises identifying (3406) a cardiac phase gating window using one or more cardiac phase sensor measurements, predicting a breathing phase using one or more breathing phase sensor measurements, identifying a coincidence between the cardiac phase gating window and the predicted breathing phase for generating an x-ray imaging pulse, and tagging an x-ray image corresponding to the x-ray imaging pulse with the breathing phase, the cardiac phase, and 3D spatial position localizer measurements (e.g., as described above with reference to method 3200).
In some embodiments, the 3D spatial position localizer is configured (3408) for measuring the patient's real-time body movement caused by respiration and heartbeats and outputting them as time series (e.g., as described above with reference to method 3100 and method 3200).
In some embodiments, the breathing phase sensor is configured (3410) for measuring one or more physiologic metrics related to the patient's breathing, including a tidal volume and its first-order time derivative. For example, the rate of tidal volume changes over time or airflow.
In some embodiments, the cardiac phase sensor is configured (3412) for measuring periodic and stationary electrical signal generated by the patient's heart, with characteristic features that correspond to the cardiac phase.
In some embodiments, two distinct filters are used (3414) to remove signal drift and noise from biometric signals of the patient after being synchronized with an x-ray unit's clock.
In some embodiments, the biometric signals of the patient measured by the one or more sensors are used (3416) for triggering an x-ray unit to acquire an x-ray image of the patient at a specific breathing and cardiac phase.
In some embodiments, the x-ray unit includes (3418) a clock. The biometric signals of the patient measured by the one or more sensors are synchronized with the x-ray unit's clock and the respective values of the biometric signals are recorded to be associated with the acquired x-ray image.
In some embodiments, the biometric signals of the patient are measured (3422) during a training window (e.g., a training period) before capturing any x-ray image of the patient and the biometric signals of the patient measured during the training window include multiple complete breathing cycles of the patient (e.g., as described above with reference to method 3100 and method 3200).
In some embodiments, multiple tidal volume percentiles within a complete breathing cycle are identified (3424) using the biometric signals of the patient measured during the training window, each tidal volume percentile corresponding to one of the different breathing phases.
In some embodiments, the biometric signals of the patient measured during the training window are used (3426) for building an optimized breathing prediction model for predicting a desired breathing phase at which an x-ray unit is triggered to capture an x-ray image of the patient.
In some embodiments, the optimized breathing prediction model is (3428) based on an autoregressive integrated moving average (ARIMA) model.
In some embodiments, the desired breathing phase for capturing the x-ray image of the patient is configured (3430) to coincide with a cardiac gating window during which heart induced lung motion is changing slowly.
In some embodiments, the cardiac gating window is chosen (3432) based on locations of T wave and P wave in an electrocardiogram (ECG) signal such that the heart induced lung motion is changing slowly.
The method includes converting (3434) first multiple sets of x-ray images of a lung captured at different projection angles into second multiple sets of x-ray images of the lung corresponding to different breathing phases.
In some embodiments, converting first multiple sets of x-ray images of a lung captured at different projection angles into second multiple sets of x-ray images of the lung corresponding to different breathing phases further comprises (3436) capturing the first multiple sets of x-ray images of the lung at different projection angles. Each set of the first multiple sets of x-ray images corresponds to the different breathing phases of the lung at a particular projection angle. In some embodiments, converting first multiple sets of x-ray images of a lung captured at different projection angles into second multiple sets of x-ray images of the lung corresponding to different breathing phases further comprises re-organizing the first multiple sets of x-ray images of the lung by their associated breathing phases into the second multiple sets of x-ray images of the lung. Each set of the second multiple sets of x-ray images corresponds to a respective breathing phase of the lung.
In some embodiments, the x-ray images within any particular set are (3438) geometrically resolved and temporally independent.
In some embodiments, the different breathing phases of the lung correspond (3440) to different tidal volume percentiles of the lung's movement.
In some embodiments, the different breathing phases of the lung include (3442) early exhalation, late exhalation, maximum exhalation, early inhalation, late inhalation, and maximum inhalation of a complete breathing cycle of the patient.
In some embodiments, n the multiple x-ray images of the lung captured at different projection angles all correspond (3444) to the same breathing phase.
In some embodiments, the different breathing phases of the lung at a particular projection angle are collected (3446) from at least two breathing cycles.
The method includes generating (3448) a static image cube from each set of the second multiple sets of x-ray images at a respective breathing phase using back projection.
The method includes combining (3450) the static image cubes corresponding to the different breathing phases of the lung into a 3D x-ray image cube movie through temporal interpolation.
It should be understood that the particular order in which the operations in
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 PPF 3501. For example, the camera(s) 3506 capture if an object will collide with the PPF 3501 as the PPF 3501 is moved around (and rotated within) the predefined area. In some implementations, a method for collision avoidance between PPF 3501 (and patient 3502) is provided.
In some embodiments, one or more 3D imaging sensors (e.g., LIDAR sensors or structured light sensors) are used to geometrically monitor the breathing of the patient 3502, as described with reference to method 3100.
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 the movies, calculate the biomechanical models, and present results with a fast, smooth experience for the end-user. The post-processing software 106's functions are performed in the cloud so that the end-user can have access to the post-processing software 106's powerful visualization tools on a treatment console, office desktop, or work laptop.
Reference throughout this specification to “an embodiment,” “some embodiments,” “one embodiment”, “another example,” “an example,” “a specific example,” or “some examples,” means that a particular feature, structure, material, or characteristic described in connection to the embodiment or example is included in at least one embodiment or example of the present disclosure. Thus, the appearances of the 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 this specification are not necessarily referring to the same embodiment or example of the present disclosure. Furthermore, the particular features, structures, materials, or characteristics may be combined in any suitable manner in one or more embodiments or examples.
Although explanatory embodiments have been shown and described, it would be appreciated by those skilled in the art that the above embodiments cannot be construed to limit the present disclosure, and changes, alternatives, and modifications can be made in the embodiments without departing from spirit, principles and scope of the present disclosure.
This application claims priority to Provisional Application No. 62/682,720, entitled “Geometrically-Resolved Radiographic X-Ray Imaging,” filed Jun. 8, 2018. This application also claims priority to Provisional Application No. 62/812,818, entitled, “Geometrically-Resolved Radiographic X-Ray Imaging,” filed Mar. 1, 2019. Each of these applications are incorporated by reference in their entirety.
| Number | Name | Date | Kind |
|---|---|---|---|
| 3871360 | Van Horn et al. | Mar 1975 | A |
| 5764723 | Weinberger | Jun 1998 | A |
| 6501981 | Schweikard et al. | Dec 2002 | B1 |
| 6650931 | McClure | Nov 2003 | B1 |
| 7103144 | Wong et al. | Sep 2006 | B2 |
| 7570738 | Khamene et al. | Aug 2009 | B2 |
| 7620229 | Oosawa | Nov 2009 | B2 |
| 8009795 | Arakita et al. | Aug 2011 | B2 |
| 8242465 | Iwata | Aug 2012 | B2 |
| 8295435 | Wang | Oct 2012 | B2 |
| 8553955 | Arakita et al. | Oct 2013 | B2 |
| 8693620 | Suzuki | Apr 2014 | B2 |
| 8761864 | Sabol et al. | Jun 2014 | B2 |
| 9076201 | Negahdar et al. | Jul 2015 | B1 |
| 9576354 | Fouras et al. | Feb 2017 | B2 |
| 9639952 | Shimamura | May 2017 | B2 |
| 9925392 | Vilsmeier | Mar 2018 | B2 |
| 9993304 | Yan et al. | Jun 2018 | B2 |
| 10219787 | McConnell et al. | Mar 2019 | B2 |
| 10342996 | Jordan et al. | Jul 2019 | B2 |
| 10376179 | Parthasarathy et al. | Aug 2019 | B2 |
| 10434332 | Vilsmeier | Oct 2019 | B2 |
| 10493298 | Hampton et al. | Dec 2019 | B2 |
| 10650585 | Kiely | May 2020 | B2 |
| 10702713 | Mori et al. | Jul 2020 | B2 |
| 20030036693 | Avinash | Feb 2003 | A1 |
| 20040254492 | Zhang et al. | Dec 2004 | A1 |
| 20050113670 | Salla | May 2005 | A1 |
| 20080080788 | Nord et al. | Apr 2008 | A1 |
| 20080123812 | Sabol et al. | May 2008 | A1 |
| 20080243018 | Zuhars | Oct 2008 | A1 |
| 20090180589 | Wang et al. | Jul 2009 | A1 |
| 20090262893 | Stewart et al. | Oct 2009 | A1 |
| 20100020917 | Gagliano | Jan 2010 | A1 |
| 20100074490 | Arakita et al. | Mar 2010 | A1 |
| 20120087464 | McCroskey et al. | Apr 2012 | A1 |
| 20120253178 | Mostafavi | Oct 2012 | A1 |
| 20130018232 | D'Souza et al. | Jan 2013 | A1 |
| 20130083889 | Stancanello | Apr 2013 | A1 |
| 20130281818 | Vija et al. | Oct 2013 | A1 |
| 20130345543 | Steibel, Jr. et al. | Dec 2013 | A1 |
| 20140107390 | Brown et al. | Apr 2014 | A1 |
| 20140140592 | Lasenby et al. | May 2014 | A1 |
| 20140286556 | Fouras et al. | Sep 2014 | A1 |
| 20140313193 | Helm et al. | Oct 2014 | A1 |
| 20150005659 | Masumoto | Jan 2015 | A1 |
| 20150020313 | Shibata et al. | Jan 2015 | A1 |
| 20160074674 | Kohli | Mar 2016 | A1 |
| 20160106382 | Lu et al. | Apr 2016 | A1 |
| 20160113614 | Cetingul | Apr 2016 | A1 |
| 20180070905 | El-Baz et al. | Mar 2018 | A1 |
| 20190083004 | Meir et al. | Mar 2019 | A1 |
| 20190175276 | Krimsky | Jun 2019 | A1 |
| 20190374189 | Kiely | Dec 2019 | A1 |
| 20190378329 | Kiely | Dec 2019 | A1 |
| 20200047000 | Jordan et al. | Feb 2020 | A1 |
| 20200202620 | Kiely | Jun 2020 | A1 |
| Number | Date | Country |
|---|---|---|
| 2441550 | Mar 2008 | GB |
| WO2010083415 | Jul 2010 | WO |
| WO2012058615 | May 2012 | WO |
| WO2015055485 | Apr 2015 | WO |
| Entry |
|---|
| Data Integrity Advisors, LLC, Invitation to Pay Additional Fees and, Where Applicable, Protest Fee, PCT/US2019/033751, Jul. 31, 2019, 13 pgs. |
| Brost et al., Constrained registration for motion compensation in atrial fibrillation ablation procedures, IEEE transactions on medical imaging, Dec. 22, 2011, 31(4): 870-81. |
| Fischer et al., An MR-based Model for Cardio-Respiratory Motion Compensation of Overlays in X-Ray Fluoroscopy, IEEE transactions on medical imaging, Jan. 2018, 37(1):47. |
| Kiely, Office Action, U.S. Appl. No. 16/420,030, Aug. 16, 2019, 13 pgs. |
| Kiely, Notice of Allowance, U.S. Appl. No. 16/420,030, Jan. 8, 2020, 7 pgs. |
| Kiely, Office Action, U.S. Appl. No. 16/805,479, Apr. 9, 2020, 11 pgs. |
| Meinel FG, Schwab F, Schleede S, Bech M, Herzen J, Achterhold K, Auweter S, Bamberg F, Yildirim AO, Bahia A, Eickelberg 0. Diagnosing and mapping pulmonary emphysema on x-ray projection images: incremental value of grating-based x-ray dark-field imaging. PloS one. 2013;8(3). |
| Panayiotou M, King AP, Housden RJ, Ma Y, Cooklin M, O'Neill M, Gill J, Rinaldi CA, Rhode KS. A statistical method for retrospective cardiac and respiratory motion gating of interventional cardiac x-ray images. Medical physics. Jul. 2014;41(7):071901. |
| Siemens, At ACC 2015, Siemens Offers Solutions for More Cardiology, Less Heartache, Siemens Healthineers USA, Mar. 12, 2015, downloaded from https://www.siemens-healthineers.com/en-US/news/acc-2015-cardiology-solution-03-12-2015.html, 7 pgs. |
| Shechter G, Shechter B, Resar JR, Beyar R. Prospective motion correction of X-ray images for coronary interventions. IEEE Transactions on Medical Imaging. Apr. 4, 2005;24(4):441-50. |
| Shutterstock, Chest x-ray, Animation of breathing chest and beating heart, Feb. 28, 2014, from https://www.shutterstock.com/video/clip-5778215-chest-x-ray-animatin-breathng-beating-heart. |
| Aulakh GK et al., Multiple image x-radiography for functional lung imaging, Physics in Medicine & Biology, Dec. 14, 2017, 63(1):015009. |
| Data Integrity Advisors, LLC, International Search Report and Written Opinion, PCT/US2019/033751, Sep. 24, 2019, 16 pgs. |
| Data Integrity Advisors, LLC, International Search Report and Written Opinion, PCT/US2019/064846, Sep. 1, 2020, 15 pgs. |
| Kiely, Office Action, U.S. Appl. No. 16/805,479, Jul. 16, 2020, 13 pgs. |
| Kiely, Notice of Allowance, U.S. Appl. No. 16/805,479, Nov. 20, 2020, 7 pgs. |
| Kiely, Office Action, U.S. Appl. No. 16/420,011, Oct. 20, 2020, 12 pgs. |
| Kiely, Final Office Action, U.S. Appl. No. 16/420,011, Feb. 2, 2021, 14 pgs. |
| Leong AF et al., Real-time measurement of alveolar size and population using phase contrast x-ray imaging. Biomedical optics express, Nov. 1, 2014. 5(11):4024-38. |
| Tanaka R. et al., Computerized methods for determining respiratory phase on dynamic chest radiographs obtained by a dynamic flat-panel detector (FPD) system, Journal of Digital Imaging, Jan. 1, 2006. 19(1):41-51. |
| Xu Y., A method for volumetric imaging in radiotherapy using single x-ray projection, Medical Physics, May 2015, 42(5):2498-509. |
| Number | Date | Country | |
|---|---|---|---|
| 20190374189 A1 | Dec 2019 | US |
| Number | Date | Country | |
|---|---|---|---|
| 62812818 | Mar 2019 | US | |
| 62682720 | Jun 2018 | US |
