Cardiac mapping systems

CARDIAC MAPPING SYSTEMS

RELATED APPLICATION

0003This application is a continuation of U.S. Provisional Application Serial No. 60/163,842, filed November 5, 1999, which is incorporated herein in its entirety, and claims priority thereto. FIELD OF INVENTION

0004The present invention 1 elates generally to catheters which conform to the shape ol a cardiac chambei or vessel (or other space) and which are useful in mapping these cardiac chambers and vessels, and more particularly to the use of conformal catheters to generate cardiac chamber maps showing the mechanical movement and electrical properties of the cardiac chamber.

0005BACKGROUND OF THE INVENTION Non-surgical cardiology mapping applications can be broadly divided into two markets with distinct, but somewhat overlapping needs interventional cardiology and

0006electrophysiology

0007Interventional cardiologists, sometimes referred to in lay terms as "plumbers", are interested in rapid evaluation of tissue health, and rapid application of appropriate treatment, if possible, the same day. They are less concerned with electrograms than they are with what low voltage readings mean to tissue viability Low voltage electrograms may indicate tissue that is dying from suboptimal blood flow (ischemic) or already dead (necrotic), or possibly "hibernating". Hibernating tissue may often be revascularized and restored to health with proper drug therapy. Interventional cardiologists are also concerned with the mechanical motion of areas of cardiac tissue that is the natural result of muscle fiber contraction during a cardiac cycle. Normal muscle tissue will cause movement of the area of its locus, and movement is decreased when the muscle is dysfunctional or starved for blood. Low voltage electrograms coupled with low local wall motion status usually indicates a restricted blood flow to the indicated area, that the interventional cardiologists may be able to correct with arterial stints, properly placed laser therapy to induce collateral circulation development, or drug therapy to enhance blood flow to the area.

0008Electrophysiologists, sometimes referred to in lay terms as "electricians", are concerned primarily with cardiac electrical anomalies known as arrhythmias. Although there are a number of common dysfunctions those patients present with, there are also many unique arrhythmias that can be treated if diagnosed properly. By studying the timing, waveshape, and slope, and amplitude information present in a group of electrograms taken over the chamber surface, the physician can locate specific sites of origins of electrical stimulation (foci) or aberrant conductive pathways in the tissue that can be interrupted.

0009The electrophysiologists therefore prefer to see a time-ordered and correlated sequence of dozens of electrograms in order to formulate a path of diagnosis and therapy. It is of primary importance in this case to be able to locate fairly precisely the position of each electrode and its corresponding electrogram. The relative timing of arrival of a pacing pulse, corresponding to the cardiac QRS (the portion of the electrogram indicting ventricular contraction) at each electrode is necessary to determine problem areas The depolaπzation/repolaπzation waveshape is analyzed to pick "activation points" on the waveform The EP often desires to stipulate amplitude and slope parameters for picking these points on each electrogram When the time durations between individual electrodes are analyzed, a picture of the propagation across the chamber walls emerges Slow rates of pacing signal propagation may indicate ischemic or necrotic tissue, since this type is less conductive than viable tissue

0010Other methods of acquiring w aveform electrical data are commonly used with other cardiac data acquisition instruments A major drawback of these instruments lies in the fact that they acquire anatomical location and electrical signal lntormation one point at a time The sensor used to acquire signal data is also used to provide discrete anatomical location information The doctor must therefore move the single point sensor throughout the cardiac chamber to slowly acquire a geometrical model of the chamber and electrical signal information at each location For these single point measurements to be interpreted as a self-consistent data set. one must assume that the electrical state of the heart persists long enough for the acquired single point data to be correlated properly This may not always be the case in an electrophysiology procedure With the present method used in conjunction with a multi-pole catheter, both anatomical location and electrical signal data are gathered simultaneously and saved for analysis as a self-consistent data set Anatomical location data for each catheter electrode is derived by image processing of the fluoroscopic images and electrical data is provided by standard analysis means though computer data reduction. Therefore, many data points are captured simultaneously, each providing location and electrical signal data. This means the data set in inherently self-consistent and is captured more rapidly than by single-point means.

SUMMARY OF THE INVENTION

0012The present invention provides a mapping system useful to both interventional and electrophysiological cardiologists. The systems of the present invention are currently designated ZIP (Zynergy Interventional Product/Program) and ZEP (Zynergy Electrophysiology Product/Program).

0013The present invention relates to a catheter system for aiding in the acquisition of electrograms from the human cardiac chambers, both atria and ventricles, as well as cardiac vessels, in real time. The catheters of the present invention advantageously conform to the unique shape of the internal walls of the cardiac chamber or vessel of the patient under examination Generally, these chambers or vessels may be viewed as a simple tube, egg-shaped, or cone-shaped spaces. The straightened catheter inserted into a major blood vessel takes on a geometπc form that is a close approximation to the chamber or (space) shape when deployed inside the target chamber, and is thus anchored in the target chamber

0014In this anchored position, the device^'s multipole electrode/marker complement is in intimate contact with the chamber walls at multiple locations, e\ enly distributed across the internal surface of the chamber, or across a portion of the internal surface of the chamber. Prototype multipole electrode/marker complements include as many as 64 electrodes and/or markers, although a greater or lesser number may be employed. In the cardiology application, this is a major advantage in obtaining multitude of electrograms quickly and consistently for later processing. When properly processed and interpreted as a group, these electrograms can be used to determine the form and site of electrical dysfunction. Competitive systems, such as those discussed in U.S. Patent No 6.066,094, use single or quad-pole catheters to laboriously obtain electrograms from a number of sites within the chamber by repetitively moving and recording individual datum

0015The mapping system of the present invention can also provide mechanical data on chamber dimensions and moφhology changes during the cardiac cycle by simply noting the size and shape of the deployed catheter fluoroscopically at multiple times during chamber movement The location of the catheter's electrodes may be traced on sequential radiographs taken over one or more cardiac cycles to determine total range of motion of smaller segments of the cardiac tissue, directly related to tissue viability. Competitive products using a single location catheter take a number of position readings and reconstruct the chamber moφhology by connecting these data with geometric planes. The more points taken, the closer the physical definition becomes. The multipole conformal catheter largely eliminates, or at least greatly reduces, the need to re-deploy the catheter. The conformal catheter nevertheless may be re- deployed more than once, and data sets combined to form an arbitrarily detailed picture of the chamber/vessel as desired.

0016The procedure of acquiring and analyzing these electrograms and mechanical motion and shape data to locate specific problem sites is called cardiac mapping. A logical extension of this capability is an instrument that uses the conformal catheter as a therapeutic catheter for ablation or drug delivery.

0017The conformal catheter, while oriented toward cardiac mapping applications, can also be used in other medical applications using the same mapping methods described herein. Alternatively, the same methods may be used to determine the three- dimensional location of any radiopaque markers on any device that may be made to conform to the internal shape of a chamber or space in the body. Gastrointestinal mapping may be performed, with or without actual electrical signal acquisition, using these methods. Mapping of cranial sinuses are another application. In general, if the internal shape of a space or cavity in the body must be determined, the methods and devices described herein may be used in conjunction with a standard fluoroscope. DETAILED DESCRIPTION OF THE EMBODIMENTS

0018The ability to acquire and display both maximum voltage and mechanical motion loops is key to providing the interventional cardiologists with the data they need for diagnosis The present invention provides the tools needed for this acquisition and display. Two types of data aie useful to interventional cardiologists ( 1 ) electrogram peak-to-peak voltages of activation corresponding to the heart's intrinsic QRS pacing impulse, and (2) time-sequential changes in each catheter electrode position over a cardiac cycle The best way to display these types of information is in the form of maps projected onto a virtual model of the cardiac chamber being diagnosed. Maximum voltage levels obtained from each electrode of the catheter are graded by color and projected onto the model The model also contains the virtual conformal catheter tor spatial reference Thus the user may rotate the model in three-dimensional space and note areas where low voltage electrograms have been obtained.

0019A second map, possibly displayed on another monitor, will concurrently display the wall motion map. This map will indicate the mechanical movement of each catheter electrode in three-dimensional space over a single cardiac cycle by drawing small, often closed-boundary, loops that trace the path of motion of each electrode on the model chamber. While this is a static image, it may also be of benefit to allow the viewing of the entire sequence of images from which the electrode position data was deπved in a film-clip fashion. This may give the physician a more realistic sense of decreased motion in problem areas of the chamber. This data is developed from a sequence of discrete radiological images captured at stable time increments over the cardiac cycle, possibly taking such a sequence from three or more fluoroscopic beam angles. This is discussed in detail below. Ancillary displays must be included to allow the viewing of standard ECG lead waveforms. The physician may want to refer to standard lead-set views to verify his diagnosis.

0020Data useful to electrocardiologists is based on the same raw data used to obtain the voltage and position data described in the previous paragraphs, i.e., the electrograms and the radiological images of the deployed catheter. This information, however, is processed and presented in a different way

0021In electrocardiological applications, the electrogram data must be available for display in the form of both isochronal and isopotential maps. The isochronal map is projected onto a model chamber, with various graded colors showing the activation times of each electrogram relative to the time zero pacing source, either a pacing electrode or the heart's natural sinoatπal node pacer. The virtual catheter may or may not be present in this view for spatial reference, since anatomical features will be present on the model chamber and have already been correlated to catheter electrode positions Rapid or slow conductance pathways will show up readily on such a display. This map may be a static display or a dynamic picture of waveform propagation across the chamber.

0022The isopotential map displays the voltage present at each electrode at any given time in the cardiac cycle, graded by color to represent voltage. Again, this may be dynamic or static, but following a potential wavefront across the chamber over the course of the cardiac cycle is valuable. The individual electrogram waveforms themselves should be available for display whenever the user moves the pointing device over a particular electrode on the model This may be used for a quick validation method when problem areas are located Alternatively, the user may use a pointing device to group electrodes together and display a group electrogram display in a strip-chart style

0023Since the electrophysiologist will eventually design an ablation procedure to cauterize specific areas of endocardium, it is essential that the location of each electrode on the catheter be identified to relatively close tolerances in the actual chamber The conformal catheter electrodes may also be used singly or in groups for locuscd ablation at specific sites The same fluoroscopic image capture and analysis techniques used for the ZIP systems are used in the ZEP system

0024Because the electrophysiologist may deliberate for several days before a therapy is initiated, the length of the data acquisition procedure is not as critical. In these applications, therefore, more emphasis will be put on obtaining additional images from additional viewing angles before the virtual catheter and its multiple electrodes ai e created and located ithin the chamber Taking multiple data sets from more than one cathetei location and deployment, and merging these data sets to reduce positional error, will further augment this procedure

0025It should be noted that electrodes are used only when electrical information is required from the catheter. In other applications, the catheter can be used only as a means of determining the three-dimensional shape of the cavity or chamber it is inserted into In this case radiopaque markers are used to determine the catheter geometry and electrodes are not used at all. The process of finding the three- dimensional position of these markers is the same as for finding electrodes. In general, any geometry can be used to support the fluoroscopic markers that are used to determine the final three-dimensional shape of the deployed device. For instance, a balloon or basket type of device fitted with radiopaque markers could be deployed m an anatomical cavity or chamber, and the two-dimensional marker information reduced to three-dimensional location information as it is done for the spiral catheter, as described below All that need be known is the manner in which the device is constructed and the physical relationship between markers that is determined by the device design This mechanical relationship information is used to determine the location of markers that cannot necessarily be seen at all, but whose location may be derived irom the locations ol related markers that are visible and whose three- dimensional locations can be located by the method described later in this document. Marker locations are chosen prudently to allow determination of all geometrical information necessary to recreate the true geometry of the deployed device For instance, if the deployed device is simply a straight linear catheter, then markers need only be placed on the distal and proximal ends, and perhaps in the middle From the two-dimensional fluoroscopic images, finding any two of these markers allows the reconstruction of the entire catheter in thiee dimensions because the spacing of the markers is already known, as is the geometry of the catheter, i.e . a straight line. Similarly, if the catheter is deployed in a "C" shape, the material properties and the spacing between the proximal, distal, and middle markers in several two-dimensional views may be used to reconstruct the three-dimensional shape of the deployed device by computing bending radπ, etc of the mateπal.

0026In the case of the deployed device being a spiral conformal catheter, the true location of the electrodes on the catheter is a priori information, and the intermediate electrode positions can be computed by knowing the spacing of the electrodes on the catheter; this spacing was determined when the device was manufactured. The need is to identify the key electrodes on the image and to compute separation distances between them as input to a function that creates a first pass approximation of the spiral model The distances computed from the image must yield enough information to determine the radius of the spiral at each coil, and the overall length of the spiral.

0027There are presently two methods implemented to find the electrodes. The most operator-transparent method is to algoπthmically search the enhanced image for particulai patterns that identify individual electrodes and directly compute the required distance information When such machine vision techniques are employed, it Irees the user from any interaction with the system except to visually v erify that the virtual spiral thus generated is tiulv a very close approximation to the fluoroscopic image of the real catheter The image processing techniques described elsewhere in this document are used in this mode.

0028The second approach is to have the user interact w ith the system to match a virtual spiral image to the actual image Letting the user move numbered target icons over whichever corresponding key electrodes that he can visually identify in the image does this The user can position other generic targets over any other intermediate electrodes that he can easily find. Given the location of at least a few key electrodes and several intermediate electrodes, the system software will then attempt to generate a spiral equation or set of equations that is used to overlay a virtual spiral on the catheter image

0029Creating and fitting a mathematical model of the spiral catheter onto the image of the actual catheter, is a mathematically complex operation A method of identifying key electrodes on the catheter is provided as follows The catheter is constructed in such a way that every fifth electrode is uniquely marked so as a show up in the fluoroscopic image easily. Other types of electrode markers may be used besides simply changing the electrode length, such as making a series of bands or radiopaque marks periodically along the catheter.

0030Since it is unlikely that this first approximation of the model will be a close fit, further means of refining the model are needed. One method implemented is that of having the key electrode markers provided on the model "grabbed and dragged" by the user to position them over their unique counteφarts on the actual catheter image. The spiral model can be re-computed quickly enough that each time such a point is moved, the model will update in real time to make a closer fit to the image. This process can continue until the model is sufficiently close to the real image. When the user verifies acceptance of the model, the data or equations for the model must be saved to disk for archiving puφoses

0031In the case of other devices and geometπes, the same method is used, but with different equations used to make a first approximation to the deployed shape of the device The true shape is then iteratively determined by user interaction with the virtual and true two-dimensional fluoroscopic ιmage(s) as described above.

0032The question of scale of the image and the resultant model must be addressed There is one school of thought that says that absolute unit scaling is not necessary That is to say, the actual size of the model in centimeters is not so important as is the relative size of the catheter features. It is clear that to derive absolute distances for the model it is necessary to analyze two orthogonal images of the same catheter deployment at the same point in the cardiac cycle. This can be accomplished, at least nominally, even with a single head fluoroscope if the ECG is used as a time zero marker for cycle and image acquisition timing. Only with the use of two such images can true length of separation lines can be measured. If, on the other hand, at least two electrodes of known separation distance can be found in the image they can be used as a "yardstick" by which all other distances are measured and scaled to.

0033A feature that has tremendous value to the user is that of including one or more anatomical landmarks that can be associated with the virtual catheter or other device to aid the human eye in determining the true anatomical location of various electrodes on the catheter, and thus its location in the cardiac chamber or other space.

0034The AV groove, coronary sinus, pulmonary vein/artery, and the aortic root are all possible candidates for these markers in the case of the cardiac catheter application.

0035These markers are tied to the virtual catheter m the same way key electrodes are and are processed with the other electrodes/markers for display with the virtual model

0036It is assumed that these ancillary markers are present in the fluoroscopic images along with the device electrodes/markers. In order to use these landmarks to locate the virtual model in three-dimensional space, at least one marker must be present to position a virtual anatomical feature over the position of the identical location in the three-dimensional model. Locations of additional features can assist the development of a more realistic model by forcing it to conform in these reference locations and then extrapolating intervening features and associating them with the catheter electrode positions In addition, it is necessary that these landmarks not move relative to the device electrodes/markers if they are to be used to derive mechanical motion information as described elsewhere in this document.

0037In the case of the conformal catheter used in the left ventricle of the heart, a

0038"wedge" type catheter with three markers/electrodes is inserted into the pulmonary artery, immediately at the top of the heart, above the ventricle. This point was chosen because it moves very little with respect to the rest of the heart, and is still within the viewing area of the fluoroscopic image containing the conformal catheter. Standard projective geometry algoπthms are used to calculate the three- dimensional coordinates of markers from several two-dimensional views of the markers taken at different angles. All that is necessary to be known is that all markers appear in all views, and the exact angle in each of two orthogonal planes that the views were taken from. For instance, if the device under study is a simple straight line, two or more co-linear points in the standard XY Cartesian plane define it. If the points have any non-zero Z coordinate value, then the XY plane is "tipped" into the XZ or YZ plane, and the angle of tip must also be known for the true three- dimensional model to be located in space. In the case of the conformal catheter, the angle ol the fluoroscope head in both XY lateral plane as well as the angle, if any, in the Z coronal/caudal plane must be known when the view is captured as two- dimensional imagery

0039Up to this point in the disclosure, systems have been described in which real catheter images were used to derive a virtual model. We will now describe systems and methods by which the user will interact with the catheter only as a virtual entity. Once the conformal device is in a virtual form, it may be "wrapped" with another \ irtual anatomical surface that represents the chamber or space around the device. However, unless a common reference point links the two structures, they cannot be rotated as one structure, thereby losing much of the potential benefit. This is discussed further later in this document. The major advantage is that this image can be rotated and translated mathematically for display so that the user can view it, and its associated anatomical landmarks from any perspective, thus gaining a greatly enhanced perception of how the deployed catheter is located within the cardiac chamber. Graphics programming exists which is capable of displaying a realistic version of the catheter, complete with electrodes, and an iconic representation of the landmarks A means of rotating this image is included in such software. Once the three-dimensional coordinates of each electrode have been calculated, various standard means are available for constructing a virtual model of the device. Graphing computer applications or virtual reality modeling languages such as MathCAD

0040(MathsoftV.5 0), IDL V 4.0 (Research Assoc), and VRML (Platinum Technology), although there are many other applications that may be used. Each application provides means of rotating the resultant image so that it can be displayed from any desired angle This provides a means of "flying around or over" the image to view its true three-dimensional shape

0041In order to present the most useful model possible, a fictitious "ghost" chamber image is constructed around the deployed catheter model to aid the user in perspective and location of the device within the chamber or space in which it is deployed

0042This image need not be extremely detailed, but recognizable as the cardiac chamber or other space under study and yet be transparent enough to easily see the virtual catheter inside it We assume of course that the conformal catheter or other device has indeed expanded to fit the actual physical chamber or space snuggly. Assuming this is the case, the outside dimensions of the model, along with the relative locations of the anatomical landmarks/markers, are the basis for the size of the ghost chamber. The outside dimensions of the ghost chamber are determined from other data or simply chosen by the computer, since only the inside dimensions are determined by the device. A preferred system uses a method that derives physical information from actual fluoroscopic studies to create the virtual chambers and their external dimensions. Although there are a vaπety of software graphics applications available to create this virtual chamber, a method commonly used is to create a "mesh or screen" model that is mathematically fitted around the catheter and scaled properly to show other anatomical features, such as chamber wall thickness, correctly. This mesh model, if the mesh size is chosen properly, has the advantage of displaying very realistic changes in chamber size and moφhology over the cardiac cycle. This approach gives the impression of viewing an actual beating heart when displaying the local wall motion, isochronal, isopotential maps in the cardiac application, or other information displays in other applications. The ghost chamber or space can also be graphically "sectioned" to display the chamber as a cutaway view or "unrolled" to form a flat plane in the same manner that world maps represent the truly spherical Earth. This plane can be used as a physical background for the isopotential or isochronal map displays in the cardiac application or for other information display in other applications. Electrograms (EGM). the cardiac signals from each catheter electrode, are acquired and displayed for the user, along with standard electrocardiograms (ECG). The preferred system contains analog-to-digital converter boards installed in the computer to capture this electrical data.

0043The preferred systems of the invention use multi-channel A/D boards that are equipped with bandpass filters for each input channel. These filters are configured under computer control, and are currently set for a DC to lKHz range. In general, the computer also configures gain and offset, as well as sample rate, but these abilities depend upon the ND board chosen. A sampling rate of 2000 samples per second is sufficient for EGM/ECG signals, but may differ for other applications. The conventional manner of displaying such signals as time graphs is followed. The maximum number of electrodes on the catheter determines how many of these signals there are to display Vaπous user controls customize the display as to the number of signals to display on a given page, the vertical screen area assigned to each signal, and whether there is horizontal and/or vertical scrolling allowed.

0044Although the vertical axis scale of all the signals is the same, the number of heartbeats acquired for display determines the horizontal time axis length. A typical number is probably about 2-3 beats, representing approximately 2-3 seconds at 60

0045BPM If a pacing electrode is used to set the heart rate, the pacing rate will determine the numbci of beats captured

0046Immediately after the EGM data is acquired and displayed, it is quickly scanned by the computer system to determine if valid signals have been acquired for each electrode (channel). This is done by various standard methods such as frequency domain analysis, wavelet analysis, or other methods. Generally, noisy or non-existent capture results in only noise of greatly attenuated peak-to-peak values While not nccessaiy, checking the validity of the captured data, both image and electrical signal inputs, is desirable to minimize asted time and poor reconstruction of the virtual model(s) later in the process

0047The instrument marks any bad channels so the user has the option of attempting to re-acquire the data to improve it. Means is also provided to pπnt the signals themselves and store the raw data for future display or analysis. If only certain captured information is bad, the user marks it so it is not used in the subsequent analysis or three-dimensional virtual model construction, but it is still acquired along with all other signals and stored. To summaπze the operation of the preferred system, a sequence of operation is described in the following examples This may be viewed as the general manner in which the svstem will be used to obtain and study the data from the catheter or conformal device EXAMPLES

0048Example 1 - Instrument/System Setup

0049The user will first use several menus in the application to choose how he wishes to acquire image and signal data, and to document the doctor, patient, and procedure These menus allow the following information to be inputted ( 1 ) patient history, personal information, attending/referπng phvsician. and prognosis, and procedure notes, (2) for image acquisition, the number and angle of the fluoroscopic views to acquire, (3) the number of cardiac cycles over which to obtain image and signal data, (4) the video frame rate at which image data is acquired (generally this will be the maximum rate at which the fluoroscopic camera will operate), and (5) sampling rate and filter characteristics with which signal data is acquired

0050Example 2 - Image/Signal Acquisition

0051After the svstem setup is accomplished as described in Example 1. the user is prompted to start the procedure by positioning the fluoroscope head at the first angle chosen for image acquisition The doctor then activates the fluoroscope and simultaneously triggers the image/signal acquisition The system automatically acquires the data for a time sequence over the number of cardiac cycles chosen duπng system setup Each sequence begins at a common time point as determined by external synchronizing pulses such as the "R-wave" trigger from the ECG machine

0052The doctor deactivates the fluoroscope The system presents image data m a "cme" moving picture style or one frame at a time as desired for review Additionally, the doctor may also display the EGM and ECG data captured for judgment of validity.

0053Bad data is marked or the acquisition at this angle is repeated until acceptable data is obtained The acquisition procedure is repeated at a number of fluoroscopic angles as determined by the setup. All data may be reviewed and acquisition repeated as necessary until a complete image/signal data set is obtained.

0054Example 3 - Image/Signal Processing Following the procedures in Example 2, image data is now stored as a number of time-sequence images, one for each fluoroscopic viewing angle acquired. In order to calculate the three-dimensional coordinates of each radiopaque electrode/marker, the two-dimensional images must be processed one at a time, to locate the markers. This is accomplished as follows- Starting with the first frame of the first time sequence of images, the system automatically processes the image and marks any electrodes found. The doctor confirms the success or failure of the marker locations. If the process was not at all, or only partially successful at locating all electrodes/markers, the doctor uses the mouse to aid the system in identifying key markers The system uses its own internally generated marker location data or the doctor^'s determination of marker location in the first fi ame of the sequence as a starting point in searching for markers in each successive frame of the sequence. In this manner, all visible electrodes/markers are located in each frame of the sequence. This procedure is repeated for all the image time sequences captured over all of the fluoroscopic head angles. The system processes this multitude of image XY marker location information to calculate the three-dimensional XYZ locations of all markers. The location of those markers/electrodes that are hidden in all images may be inteφolated from the conformal device geometry. Standard projective geometry techniques are used for this step All the signal data captured is processed to determine time and voltage relationships as determined by the procedure.

0055Example 4 - Data Display

0056After image and signal data has been processed as described in Example 3, the data may be displayed in various ways, depending on whether interventional or electrophysiology related data is desired. In one embodiment, a virtual model of the conformal device is generated and displayed using standard graph/curve-plottmg computer programs and software The user may view this three-dimensional model from any angle using a pointing device or other methods provided by the ZIP/ZEP system A fictitious graphical ghost model of the chamber or space is built around the conformal device using the dimensions of the conformal device as a scaling guide Standard graphical construction computer applications software and techniques is

0057used to accomplish this step such as Corel Draw¹ , AutoCAD ¹ , or the IDL^{1 M}

0058graphical modeling language by Research Associates The resultant three- dimensional image of the ghost chamber with the conformal device inside it is displayed and may be viewed from various angles The virtual ghost chamber/cavity model is used as a surface onto which signal data or text is superimposed. In interventional cardiology procedures this may be electrode voltage or electrogram relative timing information The vicinity of each marker/electrode may be color- graded to represent this information in a color-coded manner. In an electrophysiology procedure, electrode information pertaining to relative timing of the EGM signals and their propagation over the chamber surface duπng a cardiac cycle may be displayed. The mechanical motion or movement of individual electrode/markers over the time sequence of a cardiac cycle, for instance, can be visualized using the ghost chamber surface. The location of each marker in each frame of the time sequence can be emphasized or highlighted in such a way as to show its apparent motion over time by playing a composite time sequence of image frames in a rapid fashion. Taking the three-dimensional marker location data from the series of fluoroscopic angle time sequences forms this composite time sequence. Since each time sequence at each fluoroscopic head angle starts at the same point m time relative to the cardiac cycle, they are themselves synchronized. In this way, the movement of each marker/electrode can be traced over a cardiac cycle and played as a movie to show that motion in a realistic manner. Additionally, the ghost chamber foπned around the conformal device may be regenerated for each frame of the time sequences over a card'ac cycle, thus also displaying the changes in chamber size and shape. This lends a more realistic view of the three-dimensional virtual chamber deformations as a function of time. After all the data is reviewed, the user may make a decision to redeploy the conformal device to acquire and append new two-dimensional image/signal data to the existing database to further improve, widen, or focus the area being studied.

0059Example 5 - Data Archiving The user is prompted by the system to save all acquired and processed data to a suitable archiving means such as a CD-ROM or magnetic tape or disk Since all the data taken is sa ed, the entire procedure may be reviewed by using the system to re- play and review the data at a later time.

0060The ability to review a procedure and all its associated data is extremely valuable in planning subsequent procedures on the same patient, as well as a teaching tool. Since each patient and associated procedure may be saved permanently on a standard CD ROM, media storage bulk and cost is kept low. Example 6 - Platform

0061An Intel-based personal computer is adequate to accommodate all necessary control I/O, image capture hardware, and video terminal devices. It is immediately assumed that the machine will be dedicated to ZIP/ZEP system use, and all system l esources are available for software operation A 900 MHz Pentium III processor with a PCI bus architecture and 512 MB of system RAM allows fast DMA storage of radiographic image sequences without processor intervention.

0062Care must be taken in the selection of the system graphics card, but an AGP (Advanced Graphics Port) card with 8/16 MB of graphics memory is recommended as a starting point The necessity of housing one or two image capture boards (frame grabbers) makes it mandatory that a desk or tower style case be used It also is prudent and cost effective to include a UPS (Uninterruptable Power Supply) and line noise filter to eliminate data loss by power outage during a procedure

0063Example 7 - Mass Storage Because large numbers of radiographic images from the fluoroscope will be captured and stored, it is necessary to be capable oi archiving these images on both hard disk and CD-ROM media At current hardwaie rates it is probably not possible to dump image sequences at 30 frames per second to the CD-ROM, so disk buffering is necessary The images aie all 8 bit monochrome VGA quality with 640 x 480 pel resolution, making a ~308kB file for each frame. A standard 10-20 GB hard disk is adequate to accumulate images from several procedures before it is necessary to spool it off to the CD-ROM Example 8 - Operating System

0064Because of the relatively low cost and reasonably robust environment that the

0065Windows NT Terminal Server operating system offers, this is our first choice, with a

0066UNIX (Linux) based system also available. Example 9 - Video Displays

0067Networking the display terminal(s) as clients is a major advantage gained at a relatively modest cost of the network adapter and server software. We can now choose either minimum video terminal or a full-blown PC (even a laptop computer) as the client. The terminals may be added or deleted as the user desires, and other devices such as touch screens can be incoφorated. This versatility and ability to upgrade gradually may be greatly appreciated in today's crowded catheter labs.

0068The displays must of course have VGA (640x480) resolution at a minimum, but 600x800 or better can be had at a small incremental cost. Display size is also variable, but at least one of the working terminals should be 21 inches. The advent of reasonably priced LCD displays, with their good viewing angles, flat screens, and small footprints are attractive.

0069A technician can be located in the control/observation room several feet from the physician while monitoring and helping guide the data acquisition from his terminal. As the physician is placing the catheter and attending to the procedure, the technician is acquiring and processing the data for display on the doctor's terminal.

0070Moreover, the doctor can also have two or more displays available to him for viewing two or more maps or data displays simultaneously.

0071Example 10 - Camera Control & Image Capture

0072All modern fluoroscopes present standard NTSC video output spigots, although they may also be capable of outputting video signals at higher resolutions and formats. We are forced to accommodate the lowest common denominator here, which is NTSC 480x640 video, as our radiographic frame capture source.

0073Similarly, we assume a single head fluoroscope as the standard, although biplane machines are becoming more common. A biplane machine would speed up acquisition of frame sequences by acquiring two frame angles nearly simultaneously.

0074The image capture board is capable of streaming video to memory at the maximum 30 frames/second rate to avoid being a bottleneck, and to provide real-time view of fluoroscopic images on the computer terminal as an alternative to the doctor watching the fluoroscope monitor It may be desirable to add footswitch input capability that allow s the physician to trigger a single- or multiple- frame sequence capture Some type of pacing or R- wave trigger circuits is necessary, and is provided by the hospital pacing and ECG systems

0075Example 1 1 Software Fluoroscopic Image Capture and Storage The invention provides means for interacting with the user to capture and store monochrome grayscale images from a fluoroscope In general, this discussion will l efer to these images and the capture process as if only one fluoioscopic view is being captured and analyzed, but two or more, views are actually required Typically these will be at least one or more right and left anterior oblique views, left or right lateral views, or anterior/posterior views

0076The images to be captured are available as a standard RS-170 analog video output from the fluoroscopic imaging instrument. This standard specifies an image that is 640 pixels wide and 480 pixels high, at a frame rate of 30/second. Each frame then requires a bit more than 308 Kbytes of storage when digitized at a 256 level (8 bit) grayscale value. Catheter labs that have either a single or dual head can be accommodated. In the case of a single head fluoroscope, it is obvious that images from one view will be captured, the head angle changed by some amount, and the images from another view captured. In this way the user quickly builds a library of images that are processed to obtain true three-dimensional information on each electrode position and movement over an entire cardiac cycle. It is also necessary to capture electrograms that record electrical activity during cardiac cycle corresponding to the radiographic sequence captured. So each viewing angle frame sequence will be stored along with EGM information from each catheter electrode. Such a database, once captured for a patient, is processed to provide all lnfoπuation discussed above

0077In reality, at each fluoroscopic head angle, we are actually capturing sequences of single frames evenly spaced over the cardiac cycle, usually at a rate of about 20 frames/cycle. Because of this necessity, the capture of a sequence of frames must be triggered by and synchronized with, some cardiac event, such as the ECG R- wave signal, so that they are correlated correctly in time with each other.

0078In the case of dual head machines, the same trigger can be used, but the single frame grabber input multiplexer is used to switch between the two heads, since the heads are never active at the same time, but are triggered with a one-frame delay between them. In this case, frames captured in this staggered manner are useable. The 30/second frame rate means a complete image every 33 milliseconds, or even every

007966 milliseconds is adequate to fully document a rapid cardiac cycle of 400 msec (150

0080BPM).

0081It is desirable to allow a series of these images to be viewed as an active film clip too, to validate good image capture before the head is moved. In any case, the system must display the individual images as thumbnail views and have the ability to show one or more images in expanded form.

0082Facilities for auto-numbering images for disk storage and retrieval are also implemented. Finally, the user has the ability to select one of the images for further processing into the virtual spiral in the next module. The experienced user will select the image based on this judgment of how well certain electrodes show up on the image.

0083Once a set of images is selected to operate on, the software performs various types of image processing to enhance the image of the deployed catheter and its electrodes. Region-of-interest (ROI) selection is applied first that encloses the portion containing the catheter image and the reference catheter to narrow the image area operated on. Histogram equalization and particle (blob) analysis are then used, and the found blobs are sorted by size and perimeter to select only those blobs that are within the size range of one or two electrodes. This eliminates spurious blobs that may be misinteφreted as electrodes.

0084It should be noted that drivers for the image capture board must be written or integrated into the software and board programming done before the images are captured. This allows the specific proprietary hardware from various image capture manufacturers to be used to its fullest extent in preprocessing the images.