The present invention relates generally to a computer based medical system that can be used to position and navigate electrophysiology catheters and ablation catheters inside the heart of a patient. The catheters are used to create an image of the interior of the heart, and electrophysiology information collected from the catheters is presented to a physician user to guide a therapy or diagnosis.
Wittkampf in his U.S. Pat. Nos. 5,697,377 and 5,983,126, discloses a system for determining the position or location of a catheter in the heart. These patents are incorporated by reference in its entirety herein.
In the Wittkampf system current pulses are applied to orthogonally placed patch electrodes placed on the surface of the patient. These patches are used to create axes specific electric fields inside the patient. The patents teach the delivery of small amplitude low current pulses supplied continuously at three different frequencies, one on each axis. Any measurement electrode placed in these electric fields (for example in the heart) experiences voltages depending on its location between the various patches or surface electrodes on each axis. The voltage on the measurement electrode in the field when referred to a stable positional reference electrode indicates the position of the measurement electrode in the heart with respect to that reference. The three voltages give rise to a location of the measurement electrode in “three space”.
Although the Wittkampf system is both safe and effective there are several sources of error that result in errors in position or electrode location. Principle among these is the respiration artifact. For these reasons there is a need for improved location and navigation methods to improve the functionality of the system for the physician.
The present invention is disclosed in the context of a system for collecting and displaying “heart” data to a physician. This is an illustrative and not limiting example of the use of the methodologies. For example it should be clear that the techniques can be used guide the delivery of general surgical devices into the body or to guide and position drug delivery devices among other applications.
In the present invention, several improved location and navigation methodologies are presented and these are used to determine more accurately the location in space of a multitude of electrodes located on catheters inserted into any heart chamber. This improved location data is used to create a geometric representation of the whole heart chamber. In addition, the improved methodologies allow for the collection and presentation of electro physiologic (EP) data to the physician. This “activation mapping” aspect of the system allows the system to present EP data and catheter location data on an accurate representation of the heart. This improves the functionality of the system for both diagnosis and intervention. A typical use would include the use of ablation catheters to find the origin of a cardiac arrhythmia and then to treat it with RF energy to ablate tissue.
In the prior art and for many years physicians have measured electrical potentials on the interior walls of the heart with electrodes in contact with that tissue. In general the time course of electrical impulse at each electrode is displayed in a “timing window”. This plot of voltage against time is displayed and the physician looks at this data and visually determines the “site” or location of earliest activation. That is the physician determines which of the electrodes is closest to the ectopic focus or other origin of depolarization signals in the heart.
In the present invention multiple electrodes are swept over the surface of the heart. Each electrode is “located” in space using the location algorithm and the voltage present on that electrode is measured and stored. The timing information related to the measured voltage and a fiducial electrographic event such as the R-wave is also stored by the system.
These measurements are made continuously in a beating heart and the measurements are made without regard to the specific cardiac phase. To present this information to a physician the electrode location information is used to project a measurement point onto the geometry created by the roving catheter. The “timing” information or the voltage “magnitude” information is interpreted as a false color that is presented and projected to a point on the heart surface geometry.
It should be recognized and understood that the electrophysiology data or activation mapping data can change during the heartbeat. As a consequence a typical display would include animation of the EP data showing the passage of electrographic wave fronts over the surface of the heart as represented by the geometry.
Several related methods are taught including:
Respiration Compensation
A method of compensation for respiration is disclosed. The compensated signal allows for greater location accuracy of electrodes in the heart. In brief the respiration motion artifact is measured and subtracted from the measurement or roving electrode position data to improve accuracy.
Virtual Reference Generation
A method of generating a non physical reference position which permits the operation of the system without a fixed physical reference electrode attached to the heart. This virtual reference is a benefit to the patient and a significant improvement for the physician.
Geometry Capture
The improved location methodologies allow for the collection of location data from one or more catheter electrodes all during the heart contraction cycle and without keeping track of the phase of the heart. The data is “reduced” to create a “shell” which represents the shape of the heart chamber swept by the electrodes.
EP Mapping
One or more catheters having one or more electrodes are moved around in the heart chamber. Throughout this process the electrodes are in contact with the heart wall. The electrical potentials on these electrodes are measured and there location in space noted. The electrical data is then projected on the geometry as an activation map. The activation mapping methodology permits the display of peak to peak voltages at various locations on the geometry. The activation mapping methodology permits the display of timing relationships between the measured voltages and a fiducial timing reference at various locations on the geometry.
The embodiments of the invention shown are illustrative and various modifications may be made to the invention without departing from the scope of the invention. Throughout the figures identical reference numerals refer to equivalent structure, wherein:
System Level Overview and Basic Location Methodology
It should also be appreciated that in use the patient will have most or all of the conventional 12 leads ECG system in place as well and this ECG information is available to the system although not illustrated in the figure.
Each patch electrode is coupled to the switch 24 and pairs of electrodes are selected by software running on computer 20, which couples the patches to the signal generator 25. A pair of electrodes, for example 18 and 19, are excited by the signal generator 25 and they generate a field in the body of the patient 11 and the heart 10. During the delivery of the current pulse the remaining patch electrodes are referenced to the belly patch 21 and the voltages impressed on these remaining electrodes are measured by the A to D converter 26. Suitable lowpass filtering of the digital data is subsequently performed in software to remove electronic noise and cardiac motion artifact after suitable low pass filtering in filter 27. In this fashion the surface patch electrodes are divided into driven and non-driven electrode sets. While a pair of electrodes a driven by the current generator 25 the remaining non-driven electrodes are used a references to synthesize the orthogonal drive axes.
The belly patch electrode 21 is seen in the figure is an alternative to a fixed intra-cardiac electrode 31. In many instances a coronary sinus electrode or other fixed electrode in the heart can be used as a reference for measuring voltages and displacements. All of the raw patch voltage data is measured by the A to D converter 26 and stored in the computer under the direction of software. This electrode excitation process occurs rapidly and sequentially as alternate sets of patch electrodes are selected and the remaining members of the set are used to measure voltages. This collection of voltage measurements is referred to herein as the “patch data set”. The software has access to each individual voltage measurement made at each patch during each excitation of each pair of electrodes.
The raw patch data is used to determine the “raw” location in three space (X,Y,Z) of the electrodes in side the heart such as the roving electrode 17. The patch data is also used to create a respiration compensation value used to improve the raw location data for the electrode locations.
If the roving electrode 17 is swept around in the heart chamber while the heart is beating a large number of electrode locations are collected. These data points are taken at all stages of the heart beat and without regard to the cardiac phase. Since the heart changes shape during contraction only a small number of the point represent the maximum heart volume. By selecting the most exterior points it is possible to create a “shell” representing the shape of the heart. The location attribute of the electrodes within the heart are measured while the electric field is impressed on the heart by the surface patch electrodes.
The guide bar 64 is graduated in milliseconds and it shows the assignment of time relationship for the false color image in the geometry. The relationship between the false color on the geometry image 62 and the guide bar is defined by interaction with the user in panel 66 best seen in
Also shown on panel 66 of
For completeness in description the tracing 78 labeled “V1” in
In summary the basic software process proceeds stepwise by first selecting a set of electrodes and then driving them with current pulses. While the current pulses are being delivered the voltages on several of the other remaining surface electrodes and intracavitary electrodes are measured and stored. At this point it is preferred but not required to compensate these measured values with coefficients taken from the respiration compensation process.
It is this process that collects the various data points associated with multiple endocardial electrode locations. Each point in this set has coordinates in space. In general several dozen points are collected. A larger data set results in a more complex and higher resolution representation of the heart however, it is computationally more expensive. The raw location data is corrected for respiration and other artifacts and then a geometry process is started. In this process the exterior most location points in the data are used to create a shape. The preferred surface is a convex hull using standard algorithms such as Qhull. This surface is then resampled over a more uniform grid and interpolated to give a reasonably smooth surface stored as a “geometry” for presentation to the physician. The algorithm is used to compute the convex hull shape is well known and it is one of many potential algorithms suitable for use in implementing the invention. This shape estimates the boundary of the interior of the heart from the set of points. The process then proceeds to resampled the convex hull on a regular grid of points in physical space. By resampling the computed hull shape on the regular grid, a larger set of points is generated. Most significantly this enlarged set of points ensures that computational points are available along the length of each edge of the hull. The next process uses an algorithm for smoothing the convex hull shape. This process forms a mathematically differentiable shape approximating the physiologic shape of the heart chamber. Any of a number of interpolation processes can be adopted to implement this portion of the process. The final process causes the model to exit to a display routine or other process where the computed shape is used for further analysis.
This geometry surface is also used a display surface to present the activation maps. This is also the surface that the EP data is project on.
As described the EP catheters are moved over the surface of the heart and while in motion they detect the electrical activation of the heart or EP signals on the surface of the heart. During each measurement the real time location of the catheter electrode is noted along with the value of the EP voltage or signal. Since this data is not taken with the location data used to create the geometry a projection process is used to place the electrical information on the nearest heart surfaces represented by the geometry. The preferred implementation is to select two close points or locations in the EP data set and to “drop” a perpendicular to the “nearest” surface point on the geometric surface. This new point is used as the “location” for the presentation of EP data in the images presented to the physician.
Respiration Compensation Methodology
The basic location methodology described above provides a first order indication of the location of a roving or other catheter electrode within the heart chamber. However the primary artifact to mitigate in the basic location methodology is due to patient ventilation or breathing. For example, with reference to the belly patch 21 reference, error displacements exceeding two centimeters due to breathing have been noted in the left atrium experiments in human data as measured with a with a roving catheter electrode 17. The same data when referenced to a fixed electrode (coronary sinus electrode) the measured error displacement still exceed about one centimeter. In these experiments the data is collected with a 0.25 Hz low pass filter setting. Using higher low pass settings (necessary for geometry creation) makes the situation even worse, plus it allows a significant stroke signal component that will be due both to actual displacement during contraction, as well as impedance changes due to blood ejection. For these reasons a simple filtering algorithm is inadequate to compensate for respiration.
First, there are changes in the current paths from the axes drives when the lungs fill with air. This alteration of path changes the measurement. This changes the potential between the tracked or roving electrode 17 and the reference electrode. Respiration also moves the patches apart and this displacement of the patch drive electrodes with respect to the heart changes the measurement. Secondly, the resistivity of blood returning to the heart from the lungs is altered following inspiration. A compressive effect of filling lungs on the heart may cause some displacement the catheters themselves may actually move due to stretching or tension upon inspiration.
In addition with the belly patch 21, or another body surface patch as reference, there may be displacement of the patch sense electrode with respect to the heart.
Because the frequency components of ventilation can be so low (approximately 0.1-0.2 Hz) only very heavy filtering would adequately remove artifacts but this would result in an unacceptably slow responding system. For this reason filtering alone does not improve the system.
Fixed Electrode Reference
The large ventilatory displacements seen with a body surface reference electrodes is evident in the Medtronic Localisa™ system which is commercially available and similar to the Wittkampf system described in the incorporated reference. This system uses a fixed intra-cardiac electrode reference catheter which is usually a screw in tip pacing lead. A screw in tip is added expense and risk for the patient. An alternative coronary sinus catheter or other fixed reference may be substituted but these may move during the study, rendering a navigation map inaccurate. Even with a “fixed cardiac reference” the ventilation artifact remains problematic.
Virtual Reference
Two discrete, independent but complementary methods are discussed herein to address the problem of a surface electrode reference: the virtual reference VRV and a method of respiration compensation or RCV. While the virtual reference has less artifact than just using a single belly patch, even greater suppression of respiration artifact can be achieved using subsequent respiration compensation. Or, if an intra-cardiac reference is desired and is available, the respiration compensation method may still be applied using intra-cardiac reference.
The key objective of the virtual reference is to facilitate the basic location methodology without requiring a separate stable intra-cardiac reference electrode and catheter.
Potentials on any of the six electrodes are acquired for all samples except when an electrode is driven. Sampling while an electrode is acting as a source or sink is explicitly avoided, as the potential during this time, at the electrode will include the electrode impedance and the effects of high local current density.
Drive Axes Synthesis
Before discussing the virtual reference in detail, and with reference to
The drive configurations that comprise the data used to generate the X axis are shown in
The net X axis potential on a given sense electrode, “E” in the heart chamber, using the principle of reciprocity but essentially equivalent to Wittkampf's method, is obtained by sampling the electrode potential obtained while driving each of the 4 drive configurations, and combining them arithmetically.
Virtual Reference Computation
As previously discussed it is useful to place an electrode in the heart attached to the heart wall to operate as a location reference for measuring the location of other catheters. This can be done with a screw in type pacing lead or a pacing type lead placed in the coronary sinus. Although this is a useful expedient, the presence of additional catheters in the chamber is always unwelcome during ablation procedures and the like. One aspect of the present invention is the creation of the virtual reference, which eliminates the need for a fixed physical reference electrode. The virtual reference is a computed value that is applied to the measured values of location electrode data to simulate or emulate the position of a fixed lead. The process proceeds as follows:
The data from any four of the six patches (the system electrical reference being a seventh “belly patch”) is available whenever a pair of patches is driven. (Again, the sensed data from the driven patches is not used.)
While other choices may be contemplated, the preferred method for computing the virtual reference potential for the X axis is nominally the midpoint of the chest (YB) and back (YA) patches, since each of them is used with exactly ½ of the weight of the normal electrode Ex, described above.
The net X axis potential for electrode E is solved for by subtracting the value of the X axis virtual electrode value from the measured X axis value. A similar process can be used to find the other three space values for the roving or other electrode.
Therefore in summary, a set of patch electrodes is placed on the skin or surface of the patient. Two of these are selected to source current with respect to a ground reference, while the remaining electrodes measure voltage with respect to the ground reference. An electrode placed in the heart exposed to the field from the current pulse is also measured with respect to ground. In practice the catheters within the heart may contain multiple electrodes and each electrode ring potential is measured. Preferably at least one electrode is fixed to the interior surface of the heart and forms a fixed reference, which is also measured with respect to ground. The patch electrode data set, the internal electrode data set and the virtual electrode data set are all used to determine the location of the electrodes within the heart. After the voltage measurements are made a different pair of surface electrodes is excited by the current source and the voltage measurement process of the remaining patch electrodes and interior electrodes takes place. The sequence occurs rapidly on the order of 30 kHz. To a first approximation the voltage on the electrodes within the heart bears a linear relationship with position between the patch electrodes that establish the field within the heart. Although useful location information can be acquired with a linear model, it is preferable to compensate for respiration which varies rhythmically as the patient breaths and non-linearity within the bio-impedance of the body. In a preferred method correction factors are applied to the raw location information to improve the accuracy of the location value.
Application of Respiration Compensation
As the patient breaths the lung's volume changes and the impedance between patch electrodes varies not only as a function of electrode position but as respiration as well. Consequently it alters the perceived location of the electrode. Although the displacement or inaccuracies associated with respiration are small, it is important to remove them if possible because the ablation location positioning can be critical to achieve the therapeutic result desired by the physician.
A weighting optimization is used that requires the acquisition of data at a stable catheter electrode position for a long enough epoch to ensure that a breath was acquired (10 seconds is adequate). During this “ventilation nulling” or determination of weights, the error term (i.e. that which is to be minimized) is the perturbation of the position of “E”.
A 10-second buffer of data from a stable electrode “E” is acquired and the algebra described in the previous section is applied. This yields a time vector of data. For respiration compensation, the choice of reference is immaterial: in other words the electrode data may have the virtual reference applied (Enet), another intra-cardiac reference, or it may merely be referred to the electrical (belly patch) reference.
The gist of the respiration compensation method requires recognizing two facts: the electrode that is “apparently” moving (due to the above-cited ventilatory mechanisms) is known to be essentially stable; and a related component of the movement artifact is present on the un-driven patch electrodes.
Six sets of patch data at each time are acquired as follows. First, the patch data is high pass filtered with a corner frequency in the neighborhood of 0.01 Hz. This removes DC information from the patch data, since we're only interested in time varying (respiration) artifact above this frequency.
Again, considering the X axis only, a vector of patch data designated “R” is comprised of the following, for a given time sample. Note that the four drive dipoles (D0-D3) are used, and sampling from all six patches is obtained. The table B in
Respiration Compensation Algorithm
Consider now, a set of sampled data from an acquired 10 second epoch where the electrode “E” was essentially stationary. Since the data from intra-cardiac electrode data E(i) (“i” indicates time sample number) is by user enforcement, from an anatomically stable site, we subtract the mean of the vector from the samples, at which point we're left a noise vector, or variation about zero. Denote this Ems(i). Ideally we would like this to be zero. The respiration compensation weights, W(0-5) are determined by minimizing the expression which is a linear least squares problem, amenable to singular value decomposition or other linear solution methods. The representative expression is seen as expression “C” in
Once determined, the compensation is applied by merely subtracting in real time sample the weighted sum of the R vectors derived from the patch data, yielding a net compensated sample Eri defined in
Data Over-fitting & Mitigation
Initial results revealed that the above solution was prone to “over-fitting” of the data, meaning that relatively large weights would result from the optimization, with very small residuals. This weighting, however, would generalize poorly to other test data from the same patient. Note that after filtering, the NavX data has very low bandwidth, and after DC removal, consists primarily of the ventilation artifact, assuming a stable anatomic location. The approximating power of the respiration compensation data to match the normal electrode data with optimal weighting is very high. The over-fitting problem was solved by adding a small amount (0.012 ohms RMS) of noise to the data prior to solving for the weights.
It was also found that best results are obtained by concatenating more than one epoch of data from the patient, each individually from stable locations. Mean subtraction of the intra-cardiac electrode data applied separately to each epoch removes the “location” information, again leaving the noise components in the data. Given the approximating power of the method, the ability to deal with ventilation impact across the chamber is enhanced with a greater representation of optimizing data.
Stability of the input data is important; if the catheter is moved or changes position significantly during the epoch acquisition, the algorithm will attempt to determine weights to nullify that displacement.
Method & Algorithm Summary
The electrode on an “active” or roving intra-cardiac catheter is designated via software. Ideally this would be the ablation tip electrode 17 or the first ring electrode.
The clinician parks the catheter in an anatomically stable site as indicated by fluoroscopy. Next the -physician will interact with the display 23 by clicking a GUI button to acquire a 10-second prospective buffer. The user is asked for confirmation to accept the acquisition. The clinician is asked to move the catheter to one more stable site in the chamber, ideally far removed from the first, and a second 10-second buffer is acquired. The clinician is given the option to quit and apply, or acquire additional stable sites.
In software the following events occur:
The electrode data for each separate acquisition (epoch) is mean subtracted.
The epochs are concatenated.
12 milliohms of Gaussian noise is added to the electrode channel data and the 6 patch channels.
Equation 1 or expression “C” on
From this point forward, the basic location computation incorporates weighted subtraction per equation 2 or expression “D” on
Geometry Generation
As discussed above a specific electrode called the roving electrode 17 is placed on a roving catheter which is swept around the interior of the heart by the physician. The location of this roving electrode is monitored continuously and without regard to the heart phase. The improved location process previously described is used to collect the data points. As a result, the multitude of electrode locations taken over time form a “cloud” of points. A mathematical algorithm such as a convex hull algorithm is used to construct a surface surround this cloud of electrodes and this algorithm or process “prefers” the most exterior points so that a shell is formed around the cloud. It is generally assumed that the most exterior points occurred during distally while the more interior points most likely occurred during systole. The quality of the shell model or its accuracy depends upon both the accuracy of the location algorithm and the number of points collected. The exterior shell is referred to throughout the specification as the geometry of the chamber.
Although there are numerous ways to use the sequential data, one useful technique is to construct a normal from the surface and to note the point at which it intersects a superimposed hull of greater volume. The distance between the two surfaces is calculated along the direction of the normal and this distance measurement is used to compute velocity and acceleration for the wall at that location.
Although a representative illustration of the methodology is given various modifications can be made without departing from the scope of the invention.
The present application claims the benefit of U.S. Provisional Patent Application 60/461,004 filed Apr. 7, 2003, which is incorporated by reference herein. The present Application is also a Continuation in Part of U.S. patent application Ser. No. 09/107,371 filed Jun. 30, 1998. Each application is incorporated by reference in its entirety herein.
Number | Name | Date | Kind |
---|---|---|---|
3954098 | Dick et al. | May 1976 | A |
4173228 | Van Steenwyk et al. | Nov 1979 | A |
4304239 | Perlin | Dec 1981 | A |
4380237 | Newbower | Apr 1983 | A |
4431005 | McCormick | Feb 1984 | A |
4444195 | Gold | Apr 1984 | A |
4478223 | Allor | Oct 1984 | A |
4522212 | Gelinas et al. | Jun 1985 | A |
4559951 | Dahl et al. | Dec 1985 | A |
4572186 | Gould et al. | Feb 1986 | A |
4572206 | Geddes et al. | Feb 1986 | A |
4573473 | Hess | Mar 1986 | A |
4613866 | Blood | Sep 1986 | A |
4628937 | Hess et al. | Dec 1986 | A |
4641649 | Walinsky et al. | Feb 1987 | A |
4649924 | Taccardi | Mar 1987 | A |
4660571 | Hess et al. | Apr 1987 | A |
4674518 | Salo | Jun 1987 | A |
4697595 | Breyer et al. | Oct 1987 | A |
4699147 | Chilson et al. | Oct 1987 | A |
4706670 | Andersen et al. | Nov 1987 | A |
4721115 | Owens | Jan 1988 | A |
4777955 | Brayton et al. | Oct 1988 | A |
4821731 | Martinelli et al. | Apr 1989 | A |
4840182 | Carlson | Jun 1989 | A |
4890623 | Cook et al. | Jan 1990 | A |
4898176 | Petre | Feb 1990 | A |
4898181 | Kessler | Feb 1990 | A |
4899750 | Ekwall | Feb 1990 | A |
4911174 | Pederson et al. | Mar 1990 | A |
4922912 | Watanabe | May 1990 | A |
4940064 | Desai | Jul 1990 | A |
4945305 | Blood | Jul 1990 | A |
4945342 | Steinemann | Jul 1990 | A |
4951682 | Petre | Aug 1990 | A |
5000190 | Petre | Mar 1991 | A |
5005587 | Scott | Apr 1991 | A |
5025786 | Siegel | Jun 1991 | A |
5029588 | Yock et al. | Jul 1991 | A |
5042486 | Pfeiler et al. | Aug 1991 | A |
5054492 | Scribner et al. | Oct 1991 | A |
5054496 | Wen et al. | Oct 1991 | A |
5056517 | Fenici | Oct 1991 | A |
5058583 | Geddes et al. | Oct 1991 | A |
5081993 | Kitney et al. | Jan 1992 | A |
5090411 | Higuchi | Feb 1992 | A |
5156151 | Imran | Oct 1992 | A |
5158092 | Glace | Oct 1992 | A |
5161536 | Vilkomerson et al. | Nov 1992 | A |
5211165 | Dumoulin et al. | May 1993 | A |
5220924 | Frazin | Jun 1993 | A |
5228442 | Imran | Jul 1993 | A |
5237996 | Waldman et al. | Aug 1993 | A |
5255678 | Deslauriers et al. | Oct 1993 | A |
5273038 | Beavin | Dec 1993 | A |
5282471 | Sato | Feb 1994 | A |
5295484 | Marcus et al. | Mar 1994 | A |
5297549 | Beatty et al. | Mar 1994 | A |
5305745 | Zacouto | Apr 1994 | A |
5311866 | Kagan et al. | May 1994 | A |
5323781 | Ideker et al. | Jun 1994 | A |
5324284 | Imran | Jun 1994 | A |
5325860 | Seward et al. | Jul 1994 | A |
5341807 | Nardella | Aug 1994 | A |
5345936 | Pomeranz et al. | Sep 1994 | A |
5360006 | Geiser et al. | Nov 1994 | A |
5372138 | Crowley et al. | Dec 1994 | A |
5377678 | Dumoulin et al. | Jan 1995 | A |
5385146 | Goldreyer | Jan 1995 | A |
5391199 | Ben-Haim | Feb 1995 | A |
5409000 | Imran | Apr 1995 | A |
5411025 | Webster, Jr. | May 1995 | A |
5433198 | Desai | Jul 1995 | A |
5458126 | Cline et al. | Oct 1995 | A |
5551426 | Hummel et al. | Sep 1996 | A |
5553611 | Budd et al. | Sep 1996 | A |
5558091 | Acker et al. | Sep 1996 | A |
5588432 | Crowley | Dec 1996 | A |
5601084 | Sheehan et al. | Feb 1997 | A |
5622174 | Yamazaki | Apr 1997 | A |
5662108 | Budd et al. | Sep 1997 | A |
5669382 | Curwen et al. | Sep 1997 | A |
5687737 | Branham et al. | Nov 1997 | A |
5697377 | Wittkampf | Dec 1997 | A |
5701897 | Sano | Dec 1997 | A |
5713363 | Seward et al. | Feb 1998 | A |
5722402 | Swanson et al. | Mar 1998 | A |
5738096 | Ben-haim | Apr 1998 | A |
5797396 | Geiser et al. | Aug 1998 | A |
5824005 | Motamedi et al. | Oct 1998 | A |
5840031 | Crowley | Nov 1998 | A |
5846198 | Killmann | Dec 1998 | A |
5848972 | Triedman et al. | Dec 1998 | A |
5871019 | Belohlavek | Feb 1999 | A |
5908446 | Imran | Jun 1999 | A |
6004269 | Crowley et al. | Dec 1999 | A |
6095976 | Nachtomy et al. | Aug 2000 | A |
6603996 | Beatty et al. | Aug 2003 | B1 |
6990370 | Beatty et al. | Jan 2006 | B1 |
Number | Date | Country | |
---|---|---|---|
20040254437 A1 | Dec 2004 | US |
Number | Date | Country | |
---|---|---|---|
60461004 | Apr 2003 | US |
Number | Date | Country | |
---|---|---|---|
Parent | 09107371 | Jun 1998 | US |
Child | 10819027 | US |