The need for continuous monitoring of key clinical parameters in hospitals is well recognized. Figure 1 shows typical time constants for blood gases, ions and enzymes in response to acute ventilatory changes and interventions. Although it can be seen that relatively low rates of data collection are necessary for many medical measurements, it is also clear that intermittent measurement of P02, PCO2 and pH are not sufficient to provide safe and effective management of the patient. Very frequent or continuous monitoring is often essential. This figure also shows why the emphasis of a large number of research efforts in this country and in Europe and Japan have as their goal the development of continuous blood gas sensors, i.e., sensors that continuously monitor blood pH, partial pressure of oxygen and partial pressure of carbon dioxide. These are three (3) of the most frequent parameters measured in hospitals and the ones having the shortest time constant. Considering that in the United States alone close to 25 million blood gas samples per year are taken from patients, the potential market for continuous monitoring sensors is enormous. The emergence of microelectronics and microfabrication technologies over the past 30 years are now pointing to a possible resolution of the well recognized need for real time monitoring of critically ill patients through catheter-based sensors. Although physicians will always prefer non-invasive monitoring techniques, there are a number of parameters that presently can only be monitored by invasive method. The emerging ability to miniaturize chemical sensors using silicon microfabrication or fiber-optic techniques offer an excellent opportunity to solve this need. In fact, the development of in vivo biomedical sensors with satisfactory performance characteristics has long been considered the ultimate application of these emerging technologies.

The fabrication of microstructures for sensing and actuating has evolved in the last few years to include special technologies and materials, not commonly used in the integrated circuit industry, but which are, at least to some degree, compatible with those planar structures. Microsensors and actuators are now more frequently being constructed in the third dimension to enable mechanical motion, or to incorporate packaging components. This paper presents some of the materials and technologies for microstructure engineering which are being developed and implemented in current microsensor designs.

In this paper, we discuss the recent progress in the development at SRI International of a catheter-based macromachined electrochemical sensor - a micron-sized, silicon-based electrochemical sensor for in-vivo monitoring of pH, CO2, and O2. The manufacture of the sensor is based on a generic technology for producing electrochemical sensors in silicon, which we refer to as Room Temperature Micro-Electronic Chemical Smart Sen-sor (RT-MECSS).

The paper discusses the fabrication and the test results of a blood urea nitrogen sensor (BUN) and a chloride sensor. The BUN sensor consists of a potentiometric ammonium ion sensor covered by a polymer membrane that contains the immobilized enzyme urease. The chloride sensor is a liquid membrane type electrode. Both electrodes are batch fabricated. The sensors are part of a multispecies sensor chip. The results of the sensor in aqueous solutions and blood will be given. Good uniformity and reproducibility is obtained. The BUN sensor has a linear range of 1 to 20 mM urea and a coefficient of variation of 3% in normal blood.

A number of new low cost and high performance sensors are being made using silicon. These sensors are based on a collection of photolithographic and silicon wafer fabrication techniques borrowed from the integrated circuit industry that have been combined with new processes and structures to form a technology generally known as silicon micromachining. The result has been devices suitable for measuring such parameters as pressure, strain, acceleration or motion, and chemical concentration in markets ranging from medical through industrial, military aerospace, consumer and automotive. In many of these new markets, the precision or performance of traditional instrument grade parts is required in substantially high production volume and much lower cost. Silicon batch fabrication meets this requirement.

A new type of fiber optic pressure transducer has been developed and tested for biomedical application. It consists of two parts: a pressure-sensing membrane and a fiber optic displacement transducer which is based on the light intensity loss caused by angular misalignment between two fibers. The light intensity change is registered by a photodiode. The pressure transducer has been evaluated for static and dynamic pressure measurements. Not only does it have adequate linearity and frequency response and is safe and small in size, but it also may be an attractive device for designing catheters for multiple site static pressure detection.

Electrical impedance has been one of the many "tools of great promise" that physicians have employed in their quest to measure and/or monitor body function or physiologic events. So far, the expectations for its success have always exceeded its performance. In simplistic terms, physiologic impedance is a measure of the resistance in the volume between electrodes which changes as a function of changes in that volume, the relative impedance of that volume, or a combination of these two. The history and principles of electrical impedance are very nicely reviewed by Geddes and Baker in their textbook "Principles of Applied Biomedical Instrumentation". It is humbling, however, to note that Cremer recorded variations in electrical impedance in frog hearts as early as 1907. The list of potential applications includes the measurement of thyroid function, estrogen activity, galvanic skin reflex, respiration, blood flow by conductivity dilution, nervous activity and eye movement. Commercial devices employing impedance have been and are being used to measure respiration (pneumographs and apneamonitors), pulse volume (impedance phlebographs) and even noninvasive cardiac output.

The object of this study was to show the potential use of tissue electrical impedance as a feedback parameter for the control of therapeutic Radio Frequency (R.F.) energy. Radio Frequency energy is a form of electromagnetic energy with an usable frequency range between 200KHZ and 2MHZ. The low frequency limit is determined by the physiological stimulation of muscle, and the high frequency is limited by the physics of the delivery system.

We have developed a unique electrode catheter which combines a deflectable tip with a high torque shaft. The control handle is at the proximal end of the catheter. The control handle and the electrical connector which extends out the back of it are symmetrical about the catheter axis for optimum control. The tip deflects to the desired curve when a piston at the distal end of the handle is pushed forward by a thumb ring, and straightens again when the piston is pulled back. I will talk today about how we have developed this catheter. Although this may seem far afield from an SPIE subject, remember that any catheter used for transferring electrical or laser energy inside the body must have some rather unique properties. The tip or delivery end of the catheter must be placed and controlled correctly from outside the body two or three feet proximal from the tip. This catheter is used by cardiac electrophysiologists to study the electrical conduction system of the heart. During the past ten years the science of electrophysiology (EP) has seen major advances. But the tools of the trade, the catheters for performing the studies, have not significantly improved. Our deflectable tip catheter is indeed a significant advancement.

To monitor regional ventricular function in man following cardiac surgery we have developed an implantable ultrasonic sensor which is attached to the epicardial surface during surgery to measure wall thickening. During the last 18 months we have implanted sensors in 33 patients for up to 2 days postoperatively. Good quality signals were obtained from 25 of the sensors, and all were extracted sucessfully in the intensive care unit. Thickening fraction (TF) was highly variable among patients, dropped transiently after surgery, but remained relatively stable in most patients during the recovery period. Four patients developed sustained systolic thinning suggestive of ischemia or infarction during the recovery period. However, none of these patients were symptomatic or had ECG changes or other positive tests for ischemia. We conclude that the method is a safe, reliable, and accurate technique to monitor the recovery of ventricular function following surgery and that it provides quantitative information not obtainable by other methods.

Because coronary arteriography may underestimate the severity of coronary artery disease, other methods to assess the physiologic significance of a coronary lesion have been sought. Experimental data have confirmed that the ratio of peak flow to resting flow, coronary vasodilator reserve (CVDR), is a quantitative measure of the functional significance of a coronary a stenosis. A 20 MHz pulsed Doppler catheter with a 1 mm outer diameter and an innerlumen for guidewire placement was developed in 1985 and has been used for clinical measurement of CVDR. The technique appears safe, and reliable signals can be obtained in the vast majority of patients studied. Limitations of the technique include possible changes in vessel diameter with delivery of a vasodilator stimulus, possible elevation of baseline flow above normal resting values which would diminish the CVDR, and inability to measure absolute coronary flow. These limitations could be overcome by the development of an intravascular Echo-Doppler device in the future.

We have developed and built a miniature implantable 10 or 20 MHz pulsed ultrasonic Doppler sensor to provide continuous monitoring of blood flow in patients. The sensor is made from silicone rubber material and during surgery can be wrapped around a blood vessel and secured in place with a releasable tie. No tissue puncturing techniques are required. The wire leads and a release cable are sheathed in 2 mm diameter tubing and are exteriorized through the skin. Two to six days postoperatively when monitoring is no longer needed, the release cable is activated to release the tie, and the sensor is extracted from the patient by pulling on the tubing. We have tested 38 sensors on 6 carotid and 17 coronary arteries (2.5-4.5 mm diameter) and 15 ascending aortas (15-19 mm diameter) in 20 chronically instrumented dogs for up to 16 days. At the end of the implantation period, the sensors were extracted from the awake dogs with no visible behavioral reaction. Postmortem exams showed that none of the vessels were thrombosed or damaged, and the sensors were free of tissue. Doppler audio signals acquired with 20 MHz crystals from vessels 2.5-4.5 mm diameter had signal-to-noise ratios greater than 32 db, and the velocity signals had excellent linear correlations (r=0.99) with EMF sensors and timed blood collections. It has been found that the implantable sensor is a simple, reliable, and safe method of providing continuous monitoring of blood flow during and after surgery. The sensor has been granted F.D.A. Investigational Device Exemption (I.D.E.) approval, and we have begun clinical intraoperative evaluations on coronary artery bypass grafts (3-5 mm in diameter).

Patterns of activation (A) and perhaps repolarization (R) depend on myocardial fiber structure and intercellular resistance, parallel and perpendicular to fiber orientation. Gray scale maps of A and R were measured from Langendorff preparations of left guinea pig ventricles stained with a voltage-sensitive dye (di-4-ANEPPS). Action potentials (124) were recorded from syncytia (6x6 and 12x12 mm) under SA node control, or stimulated at the edges (4) of a patch viewed with a photodiode array. At the end of the runs, muscles were marked with ink, fixed, sectioned every 5 pm as a function of depth, up to 1 mm. Fiber axis rotated less than 15 degrees in depth for the first 0.5 mm of epicardium and was aligned with respect to optical maps of A and R. Fast and slow A pathways matched respectively the parallel and perpendicular orientations of the fiber axis. R did not follow the fast axis of fiber orientation, but appears to travel either transverse or 45 degrees to it. R typically occurred at the apex of the ventricle suggesting that these cells have intrinsically shorter action potential durations (APD's). Under SA node control, Purkinje fibers accelerated the conduction velocity of the A process 4 fold over electrical pacing. The average velocity of R, however, remained the same whether SA node or electrically paced, demonstrating that Purkinje fibers do not drive the R process. During hypoxia, A patterns and conduction velocity remained stable, but APD's decreased dramatically within 10 minutes and the pattern of R become random and its velocity decreased. Thus R is also an active process, dependent on cell-to-cell coupling and highly susceptible to hypoxia.

External gamma-camera imaging using tumor-seeking radiotracers has been shown to be insensitive in the task of identifying small tumors situated deep inside the body. Our approach is to use probes containing miniature radiation detectors which, when maneuvered close to a tumor, overcome the limitations of external imaging. We discuss our results from using such probes for bronchoscopic and surgical tumor detection. A limitation that prevents the detection of small tumors with these probes is the spatial variation of the background count rate due to distant background sources of high radiotracer uptake. We present two techniques for solving this problem geometric background subtraction and imaging.

This paper discusses the design of a holographic endoscope and high resolution contact recording holography. A holographic microscopic reconstruction technique is then proposed to allow for the visualization of the holograms using phase contrast and interference microscopy.

Balloon angioplasty is a well established non-surgical treatment of ischemic vascular disease. Balloon dilatation increases the lumen in a stenosed artery by overstretching the wall and fracturing the atherosclerotic plaque. Fluoroscopy is adequate to guide the proper placement of guide wire and balloon catheter. Fluoroscopy largely fails, however, when a different recanalization strategy is choosen to address the major problems associated with balloon dilatation. In the past few years, more than twenty different recanalization catheters have been developed that physically remove obstructing plaque. In the femoral artery, both mechanical and thermal methods appear to be quite successful in traversing total occlusions in spite of 'blind' guidance by fluoroscopy. However, subsequent balloon dilatation is often needed. The femoral artery is large and runs a fairly straight course. Perforation of the femoral artery is a minor complication. In the coronary arteries, in contrast, the novel angioplasty methods have met with variable success. These arteries are small, tortuous and move continuously. The anatomy and composition of the plaque is complex and the remainder of the diseased wall may be thin. In the coronary arteries, the margin between recanalization and perforation is small. The latter is a potentially fatal complication. Thus, there is a great need for a catheter that is capable of high resolution imaging and tissue identification in obstructed arteries of small caliber. Intra-arterial echo imaging, possibly combined with laser fluorescence spectroscopy, seems a promising approach. The design of a catheter that combines these powerful diagnostic features with steerability, flexibility and controlled ablation is now the major engineering challenge in interventional cardiology.

Endovascular imaging has undergone rapid development this decade. This paper represents a vascular surgeon's personal views of this new technology. This paper will focus primarily on vascular endoscopy and endovascular ultrasonography, the two fields that will probably have the most clinical applications. Spectra analysis, which is still in its infancy stages will not be discussed.

The application and relevance of intravascular ultrasound is reviewed. We studied 89 arterial segments with prototype intravascular ultrasound imaging systems. For data analysis we use pathology as the reference standard. Qualitative and quantitative comparative studies are presented to elucidate the ultrasonic characterization of arterial morphology and composition. The ability to measure quantitatively arterial lumen area, lumen diameter, wall thickness, plaque area as well as plaque composition make intravascular ultrasound unique.

External ultrasound has achieved an important niche in the spectrum of diagnostic imaging modalities. Its real-time capability, ease of use, and relative low cost have brought it to prominence as an important diagnostic tool. Medical ultrasound imaging, driven by advances in technology and by clinical needs, continues to improve its diagnostic capabilities. Key technologies for ultrasound development are new transducers, advances in signal processing algorithms, and increased computer power. Although external ultrasound image quality continues to steadily improve, certain clinical limitations such as organ access and tissue attenuation have spurred the development of more invasive scanning techniques. Endorectal, endovaginal, and transesophageal probes provide better access to--and provide superior images for--the prostate, uterus and heart. Intraluminal ultrasound is an emerging field of imaging, employing miniature, high-frequency probes which can be inserted into arteries to monitor interventional procedures. To put these developments into perspective, this manuscript reviews the capabilities and limitations of existing ultrasound technology and discusses the impetus for future developments.

Although one of the great merits of medical application of ultrasound is the non-invasive character of the technique, there are several examples in which an invasive or semi-invasive application can be useful. Probably due to then poor sensitivity of the piezoelectric crystals some very early invasive echo applications are based on tube-or catheter-mounted echo elements. Echo-endoscopy for the inspection of body cavities dates back to the fifties. Not much later investigators began to develop intra-esophageal probes for inspection of the heart and great vessels. Transesophageal echography has become very important in cardiology today. Intravascular echo imaging has gained much interest. Such a technique would be very helpful in the diagnosis of obstructive arterial disease and the evaluation of therapeutic intervention. This echo technique has been suggested for steering of catheter-based desobstruction methods. A few examples of combined echo and desobstruction techniques as reported in literature are mentioned. With high-frequency intravascular real-time imaging it is possible to observe detailed information on arterial wall, fibrotic proliferation and atherosclerotic plaque.

We have developed and tested a two,dimensional catheter ultrasound system operating at 20MHZ and providing a 360 degree cross-sectional image at the catheter tip. Initial clinical trials have been performed in the peripheral (leg) vessels in the context of both diagnostic angiographic studies and therapeutic procedures (balloon angioplasty and atherectamy). The catheters provided high quality images of the extent and distribution of atherama in the vessel wall and were clinically useful in gauging the adequacy of the vascular repair. There were no complications to the patients arising from use of the imaging catheters.

Previous in vitro ultrasonic imaging in our laboratory with a prototype subassembly suitable for use in an imaging catheter was performed at 20 MHz using a rather large transducer having a 2 mm aperture focussed at 10 mm [1,2,3]. More recently ultrasonic images obtained at 30 MHz with a 1 mm active aperture focussed at 3 mm, have been compared with high resolution magnetic resonance images (MRI). Axial and lateral ultrasonic resolutions have improved from 100 and 450 μm, respectively, to 100 and 150 μm using the newer, smaller transducer. Signal-to-noise ratios as visualized in the resulting images remains unchanged, and penetration of normal vessel walls is complete.

Mechanically driven ultrasound imaging catheters (UICs) using single, focused elements are examples of a class of emerging catheter tip ultrasound devices beginning to be used for intravascular imaging. To develop a practical, cost effective imager that is both a good catheter from the interventionalist's standpoint and provides high quality images, we have had to develop new materials, design methods, and fabrication techniques. Present day diagnostic catheter's now being used widely by interventional radiologists were re-examined, and modified for acoustic imaging. Design philosophies that embodied the conventional wisdom of 20 or so years of ultrasound transducer manufacture had to be dissected, and in some cases discarded, and fabrication techniques more familiar to the semiconductor industry applied to the manufacture of intricate, miniature ultrasound crystals.

Ultrasonic images of cadaver arteries are presented in this paper. The images were produced from within the arteries by a pulsed transducer capable of being mounted on a 0.8 mm catheter equipped with both a guide wire and a laser delivery system. The image data were acquired, processed, and displayed by a computer based real-time system. Example images show some of the display formats which are available for elucidating the internal structure of arteries with this imaging modality. Transverse and radial sections of the artery are shown in both A-Scan and B-Scan modes. Derived thickness images are also presented. In each mode the internal artery structure, including plaque areas, is clearly apparent. Histologic sections are shown beside ultrasonic data to verify the utility of these images.

In this study we examine the feasibility of using high resolution ultrasonic imaging to monitor and control tissue removal by laser. Design goals for the imaging system are that it be capable of determining the location and thickness of target lesions and adjacent and underlying normal structures, as well as the extent of ablative removal and thermal damage underlying the ablation zone. The prototype devices discussed here employed a single ultrasonic transducer operating in pulse-echo mode. Measurements have been performed on in vitro tissue samples. Information obtained includes vessel and lesional thickness as successive tissue layers are removed from the vessel wall by laser action. Later, observations and processing algorithms can be transferred to scanned or phased array devices. This information should enable the ablation process to be controlled so as to preferentially remove lesional material and to minimize the danger of vessel perforation and damage to normal tissue.

A display has been developed for intravascular ultrasonic imaging. Design of this display has a primary goal of providing guidance information for therapeutic interventions such as balloons, lasers, and atherectomy devices. Design considerations include catheter configuration, anatomy, acoustic properties of normal and diseased tissue, catheterization laboratory and operating room environment, acoustic and electrical safety, acoustic data sampling issues, and logistical support such as image measurement, storage and retrieval. Intravascular imaging is in an early stage of development so design flexibility and expandability are very important. The display which has been developed is capable of acquisition and display of grey scale images at rates varying from static B-scans to 30 frames per second. It stores images in a 640 X 480 X 8 bit format and is capable of black and white as well as color display in multiplevideo formats. The design is based on the industry standard PC-AT architecture and consists of two AT style circuit cards, one for high speed sampling and the other for scan conversion, graphics and video generation.

A novel approach to arterial imaging using ultrasonic data is presented. A catheter-mounted ultrasound probe is used to acquire data from within the arterial lumen. Signals reflected from the vessel wall are manipulated by specially created software in order to recreate a three dimensional model of the arterial segment. Of the available reconstruction strategies, the use of volume-elements (voxels) has been selected to create detailed and realistic images. A 2-D image is first formed from a matrix of picture elements (pixels), derived from co-ordinates obtained from the received ultrasound signal. The 3-D image is then recreated in full voxel space from a series of these 2-D pixel images using linear interpolation between the 2-D slices. The system has been used to recreate 3-D images of arterial models, post-mortem human arterial specimens and segments of femoral artery using data acquired 'in vivo'.

Normal human arteries have a well-defined structure on intravascular images. The intima appears very thin and is most likely represented by a bright reflection arising from the internal elastic lamina. The smooth muscle tunica media is echo-lucent on the ultrasound image and appears as a dark band separating the intima from the adventitia. The adventitia is a brightly reflective layer of variable thickness. The thickness of the intima, and therefore of the atherosclerotic plaque can be accurately measured from the ultrasound images and correlates well with histology. Calcification within the wall of arteries is seen as bright echo reflection with shadowing of the peripheral wall. Fibrotic regions are highly reflective but do not shadow. Necrotic liquid regions within advanced atherosclerotic plaques are seen on ultrasound images as large lucent zones surrounded by echogenic tissue. Imaging can be performed before and after interventional procedures, such as laser angioplasty, balloon angioplasty and atherectomy. Intravascular ultrasound appears to provide an imaging modality for identifying the histologic characteristics of diseased arteries and for quantifying plaque thickness. It might be possible to perform such quantification to evaluate the results of interventional procedures.

The mechanical properties of commonly used fluid transfusion or pressure monitoring catheters are similar to the properties required of catheters which include sensing devices. Consequently, bending and torsional stiffness of commercial catheters and tubes were measured at both room and body temperature. Five of these usually placed with the aid of fluoroscopy had an average Young's modulus of 5714x101 dyne/cm at 21°C which decreased 29% at body temperature; a shear modulus of 70.5x101 dyne/cm 4 at 21°C which decreased 13% at body temperature, and plastic deformation of 8% when loaded for 1 minute at 37°. Four of these were composed of a composite material. Catheters which are balloon directed during insertion had moduli values approximately 1/3 of these or less. The drag forces produced on balloons used on such catheters were measured for fluid velocities ranging from 10-50 cm/sec. Using this information the average force applied to a balloon throughout a cardiac cycle was calculated; values of 1280 dynes for a .6 ml balloon and 2490 dynes for a 1 ml balloon were found. The maximum wall thicknesses to catheter radii for single lumen catheters were determined for various material moduli which would allow the catheter tip to be directed by a balloon during its passage into the right heart.