With the current developments in optics, including laser optics, there should be more progress on the development of instrumentation for dermatological diagnostics, even clinical diagnostics. This report introduces new microscopy, much still under development, and research. A frank appeal for this program is made for multi-discipline help from optical physicists, biophysicists, biomedical engineers, even laser biomedical engineers, investigative dermatologists, and pathologists. If one is allowed to say, the most current advanced clinical diagnostic microscope is the polarizing microscope under the stimulus of Rox Anderson, Lynn Drake, Steven Jacques, and Peter Dorogi. The other microscopes for clinical dermatological research to be developed include the confocal scanning microscopy with the emphasis for living tissue, the ultrasonic biomedical microscope (UBM) for dermatology, and the holographic microscope, related to our studies on the biomedical aspects of optical phase conjugation. All these are introduced briefly for our studies and for definite pleas to help us.

Over the past ten years there has been a revolution in the PC based image processing and analysis market. Given the large number of hardware and software products that are currently available, how should one select a system that is right for them? This paper discusses, in light of common biomedical image processing problems, some of the issues that should be considered when buying such a system.

A Coulter diff3/50 research microscope under computer control was used to digitize biopsy slides at differing resolutions and in full color. Then, a broad set of candidate features were extracted using color analysis, template matching, statistical texture analysis, frequency domain techniques, and surface modeling by both cellular logic filters and relative extrema analysis. In all, over 600 candidate features were measured for selection and classifier design. Standard pattern recognition techniques for classifier design assume that objects cluster into distinct classes. For the lymphoma problem, where the classes (subtypes) form a continuum based on the percentage of large cells, this discrete class assumption does not apply. Estimation theoretic techniques were combined with pattern recognition for this project to design classifiers that exploit the continuous nature of lymphoma subtyping.

This chapter describes the principles and practice of the optical recording of cellular metabolism. Redox imaging of cells, tissues, and organs based on intrinsic fluorescent probes of cellular metabolism. Cellular metabolism may be noninvasively interrogated through the `optical method' based on the fluorescence intensity of intrinsic probe molecules (Chance, 1991). The intrinsic fluorescent probes, which report on cellular metabolism, are the reduced pyridine nucleotides, NAD(P)H, and the oxidized flavoproteins.

The ex vivo human was examined with a laser scanning confocal microscope. Back scattered light was used to image the surface epithelial cells and the deeper lenticular fibers. The membranes of the lenticular fibers are visualized within the ocular lens. This noninvasive optical examination of both the surface and the deeper structures of the ocular lens is demonstrated.

There are many different methods being developed for generating images through tissue using light as the transmitted beam. The potential advantages are enormous: at low intensities, the procedure is safe requiring no ionizing radiation, and is therefore suitable for continuous monitoring of patients (such as those with head trauma), or for use in screening operations (such as mammography). Furthermore, the apparatus will be inexpensive when compared with MRI, CT, and PET scanners, and will have the capability of providing information about the tissue oxygenation status, rather than merely providing morphological information.

A new high resolution transillumination technique has been developed to detect small lesions in teeth and in tissues. It uses collimated, narrow light beams synchronized with a highly sensitive receiver to scan the sample. Extracted teeth and tissues have been examined with this technique. The teeth included a variety of conditions, some with caries, some previously restored, and others with metal pins inserted into them. The results show that a small carious lesion as small as 1 mm² can be detected from either side of the tooth. Cysts and cancerous lesions in tissues were also detected. The new light imaging technique can find carious lesions better than dental x rays, however, the solution of mental pins inside a tooth is less than that of x rays. The technique is totally harmless to teeth and tissues and can be used as often as necessary even during any clinical procedure. The investigation shows that the resolution for carious lesions could be increased further to about 0.1 mm by incorporating presently available new semiconductor devices and that such a device could be designed for clinical oral use.

Three-dimensional surface imaging of the face using visible light can play a valuable role in orthomorphic facial surgery and orthodontic treatment. To provide such data we have constructed a clinical facial imager that is (1) completely safe for patients, (2) permits full facial data acquisition in less than two seconds, (3) provides resolution better than 0.5 mm in x, y, and z coordinates. The equipment uses a moire fringe technique developed at the National Engineering Laboratory (UK) by which data for each side of the face is acquired by two separate cameras. Once generated the two half-face images are fused using an internal calibration technique and the full 3-D facial image exported to a work station for rendering, and display. Patients representative of a broad range of skin tones and facial shape have been imaged and the results are discussed in relation to the clinical application of the equipment.

The new technology of digital imaging colposcopy is the subject of experimentation by a number of physicians, some of them working in clinical environments. One of the interesting applications of digital imaging colposcopy is the quantitative measurement of areas of dysplasia, or of anatomical features, on the cervix. It has been suggested that patient care can be managed in response to findings based on these area measurements. We call attention to several sources of error which may not be immediately apparent to users of these systems and which may have implications for the clinical utility of these measurements. We suggest guidelines for the measurement protocols which should be a minimum requirement for use of the systems, and indicate new directions for research to improve the efficacy of this important new technology.

Intravascular ultrasound imaging (IVUS) is a catheter based imaging technique in which a miniaturized crystal is mounted at the tip of a catheter to provide high resolution, two- dimensional ultrasound images of the vascular system. With ultrasound crystals of the appropriate imaging frequency, any region of the vascular system can be imaged intraluminally by this catheter system. IVUS images are generated from ultrasound waves that are reflected from the vessel wall. The reflections demonstrate the morphologic details of the vessel and the geometry of the lumen (Figure 1). A detailed analysis of the entire length of the vessel can be made by simply advancing or withdrawing the catheter. The vessel wall morphology and lumen geometry obtained with IVUS have been shown to correlate well with histologic specimens.

Annual angiographic assessment to determine the presence or progression of allograft coronary artery disease (CAD) has been unable to modify the natural history of this disease. Coronary angioscopy is a sensitive method to detect the early presence of coronary artery disease and in a retrospective analysis severity of CAD by angioscopy correlated with the time since transplantation. The purpose of this study was to prospectively evaluate progression of coronary artery disease over a one year period in 40 cardiac transplant recipients. The progression of coronary artery disease as assessed by angioscopy is directly related to time after transplantation and therefore angioscopy may be the method of choice for detection and evaluation of therapeutic regimens to control allograft coronary artery disease.

Percutaneous transluminal coronary rotational atherectomy (PTCRA) is an exciting new device to recannulate obstructed coronary arteries. This device works as a high speed `drill,' selectively cutting hard atherosclerotic plaque while preferentially sparing the softer, less diseased vascular luminal surface. At speeds as high as 200,000 rpm the plaque is pulverized into small particles easily handled by the circulatory system with no untoward clinical sequela. Balloon angioplasty does not remove atherosclerotic plaque. It dilates the vessel by mechanically stretching, compressing and splitting the plaque and vessel lining. We compare morphological and surface luminal characteristics of vessels post PTCRA to vessels post PTCA.

Coronary angioscopy was used in an attempt to visualize the internal architecture of cardiac vessels before and after deployment of Palmaz-Schatz stents in 50 patients. The vessel was successfully visualized in 48 (96%) of these patients. In 24 patients, angioscopy was performed both after preliminary balloon angioplasty and then again after stent deployment. In all 24 patients the diameter of the lumen appeared larger after stent deployment as compared to after balloon angioplasty. In 16 of these 24 patients a dissection was documented by angioscopy after balloon angioplasty. The dissection was absent after stent deployment in all 16 patients. In seven patients, thrombus that was not apparent by angiography was visualized by angioscopy. Moreover, in four patients, thrombus that was suggested by angiography could not be confirmed by angioscopy.

Current intravascular ultrasound imaging technology is able to determine the extent and distribution of pathologic processes within the vessel wall, but is not highly sensitive in discriminating between different types of pathologic tissue. `Tissue characterization' refers to a set of computer-based techniques that utilize features of the ultrasound signal beyond basic amplitude to help define the structure of the tissue of interest. Although preliminary results with this approach are encouraging, additional work is needed to define its clinical application in vascular disease.

Intraluminal vascular ultrasound (IVUS) is developing rapidly as a method to define the transmural anatomy of vascular structures for both diagnostic and therapeutic applications. Current devices use various configurations of transducers at the end of an intraluminal catheter to make real time images of cardiac structures and vessel walls. These devices may have diagnostic applications, particularly at the time of angiography or catherization for symptomatic cardiovascular disease.

We have developed and subjectively evaluated a lossy classified vector quantization (CVQ) using a subsampling and prediction decollation scheme. Both interframe and intraframe codings were evaluated using a sequence of x-ray CT images. Both 10:1 and 15:1 compression ratios were evaluated using nine head images from three patients. Thirteen radiologists evaluated the quality of images by viewing them on film, and comparing them to the original images on film. Although there are large variations in individual evaluations of image quality, there was overall agreement among all readers to a statistically significant level. With the proposed algorithm, the interframe coding approach produces better quality than the intraframe at the same level of data compression. Even though some data compressions were not statistically significant from the originals, the average responses were slightly worse than those for the original image. The effect of data compression on diagnostic accuracy was not evaluated.

Optical properties of the fiber taper used in the fiberoptically coupled CCD x-ray imaging system will affect the image quality of the overall system significantly. In this paper, three important characteristics were studied: optical efficiency and its impact on SNR, contrast degradation caused by stray light, and spatial resolution loss (resulting from misalignment in corresponding pixels in cascaded components) when the fiber taper is coupled with another `pixel array' such as a CCD imager.

The introduction of sapphire tips enabled the use of Nd:YAG lasers in contact surgery, thereby adding tactile feeling, surgical precision, and reduction of required optical power levels. However, the acceptance of sapphire tips has not been widespread. This is due to their high cost, limited reliability, cooling requirements, and somewhat inconvenient handling. The introduction of sculpted fibers has revolutionized Nd:YAG contact surgery by advancing a fully disposable, superior scalpel/coagulator at a considerably lower cost. Handheld contact fibers are excellent surgical tools for incision, excision, hemostatis, and vaporization of lesions.

Three-hundred-eight nm laser-induced autofluorescence spectra of the normal human brain, astrocytoma grade IV and glioblastoma grade IV specimens, have been recorded in vitro two hours after surgical resection. Typical fluorescence spectra for normal (N) and malignant (M) tissue show 4 maxima at about 352, 362, 383, and 460 nm. These spectra are analyzed in detail. Subtle differences in normalized spectra of N and M tissues appear to be large enough for diagnosis. Several criteria such as maxima and minima absolute intensity and intensity ratios at typical wavelengths are computed and used to classify the tissue. This preliminary study shows that fluorescence spectroscopy with 308 nm UV excitation could be a valid technique for discriminating tumor types. However, it should be noted that these measurements are made in vitro. Living tissues may have different spectral characteristics, therefore future in vivo investigations must be performed.

Endoscopic laser stereotaxis (ELS) involves integration of rigid-flexible endoscopy and Nd:YAG laser to 3-D - 2-D multiplanar image-guided stereotactic procedures. The major advantages of ELS include: minimally invasive (burrhole or small craniotomy surgery), direct intraoperative visualization, hemostasis, evacuation or resection assessment, and wide exploration of intracranial cavities or ventricles. ELS has been used in the management of 152 clinical cases including biopsy, aspiration, resection, and internal decompression of deep and subcortical intracranial lesions, and different types of fenestration procedures. Image-guidance combined with endoscopic techniques may offer a safe, accurate alternative to conventional neurosurgical procedures in treating small solid, cystic, and intraventricular lesions as well as fenestration procedures.

Percutaneous Lumbar Diskectomy, or PLDD, is the third generation in the evolution of percutaneous techniques to treat herniated lumbar discs. Until the development of the percutaneous techniques, the hallmark of the surgical therapy for a disc herniation was to remove the herniated fragment of disc, and to follow the fragment into the disc space to remove as much of the nuclear material as possible. Even in the best of surgical hands, and with the most selective of patient candidates, there is still a five to ten percent failure rate of this surgery. One of the causes of surgical failure is recurrent herniation of the same disc.

Intracranial lesions are invisible and exploration can be hazardous to the patient's postoperative quality of life. The first generation of neurosurgeons relied on their knowledge of neurology and anatomy to try to locate the patient's lesion. In general, this did not produce satisfactory results. Improvements in neuroradiology, neuroanesthesia, and neurosurgical techniques have diminished the proportion of the brain viewed as inoperable. These improvements have significantly improved patient outcomes.

In the 1960s stereotactic surgery was done merely by taking two projective images, using hand calculations to determine the numerical values, and deriving a set of specific measurements that guided a needle through an eloquent part of the brain. This type of surgery developed through the years, taking advantage of the new technologies in imaging modalities, networking, advanced video systems, high speed computer workstations, to the most sophisticated technology currently available, digitizing systems using infrared sensing technology.

In this decade the concept and development of computer assisted stereotactic neurological surgery has improved dramatically. First, the computer network replaced the tape as the data transportation media. Second, newer systems include multi-modality image correlation and frameless stereotactics as an integral part of their functionality, and offer extensive assistance to the neurosurgeon from the preplanning stages to and throughout the operation itself. These are very important changes, and have spurred the development of many interesting techniques. Successful systems include the ISG and NSPS-3.0.

The computer-assisted neurological surgery planning system (NSPS), developed by the Neurological Surgery Department, Wayne State University, Detroit, MI, is designed to offer neurosurgeons a safe and accurate method to approach intracranial lesions. Software consisting of the most advanced technologies in computer vision, computer graphics, and stereotactic numeric analysis forms the kernel of the system. Our paper discusses the functionalities and background theories used in NSPS.

Computed tomographic (CT) and magnetic resonance imaging (MRI) studies reconstructed in a stereotactic space can be used for accurate localization of intracranial lesions located in deep or eloquent regions in the brain, and for optimization of subsequent surgical removal. We describe our experience with 163 patients who underwent computer-assisted volumetric resection. The planning for the stereotactic volumetric neurosurgical methodology utilized the Zamarano-Dujovny localizing unit, the neurosurgical planning software (NSPS) system, which generates 2-dimensional and 3-dimensional views of the area of surgical interest, the arc setting parameters, and reconstructed CT images corresponding to the surgeon's eye-view perspective.

In stereotactic neurosurgery several localizing systems such as robotic or articulated arms and sonic digitizers can correlate on-site surgical points with precaptured multi-modality images. We discuss a strategy for intraoperative real-time localization of brain tumors using an opto- electronic system with light emitting diodes (LED).

Over the past 20 years the authors have been involved in the design and development of various aspects of multi-modality imaging for stereotaxy. The current system, which we call CASS (Computer Assisted Stereotactic Surgery), is a patented design which is capable of acquiring MR, CT, PET, SPECT, angiographic, and other images via tape, camera, Ethernet, or analog-to-digital conversion into a computer display system for multi-modality image viewing and stereotactic simulation.

Our original brain mapping techniques have been expanded so that MR and CT images can be displayed in a three-dimensionally simulated localization environment. Various combinations of MR images as well as CT images (or combinations of both and angiography) can be selectively displayed and viewed in three-dimensional stereotactic space. Data from the Talairach anatomical library, the architectonics of the atlases of Van Buren and Borke, Schaltenbrand and Bailey, Schaltenbrand and Wahren, and Brodmann's cortico-architectonics have been used to develop a detailed anatomical atlas library and brain mapping system based on brain reference structures common to each of these databases. The data in this mapping and imaging environment can be interrogated to create computerized anatomical displays showing any given functional anatomical region in two-dimensional displays or three-dimensional relief. This composite mapping system allows the interrogation and cross referencing of data from virtually any other brain mapping or localization system.

Histologic studies of normal and pathologic human cerebral tissue were performed following irradiation with CO₂ laser beams. The action of CO₂ laser radiation was studied on cerebral and medullar tumors and on atherosclerotic plaque. We have used CO₂ laser surgical systems with output powers from several watts to 30 W equipped with either free hand optomechanical articulated arms or laser microscope adapters. Based on histologic examination of different tissue samples under light and scanning electron microscopy we have established the main advantages and limitations of this technique in neurosurgery. These findings suggest that the CO₂ laser has advantageous clinical application for a large number of neurosurgical procedures.

Frame based stereotactic surgery allows the surgeon to precisely approach a predetermined target. Although useful for diagnostic and functional procedures, mechanical instruments fail to indicate position quickly during open craniotomy. We have developed a system employing an infrared optical digitizer to indicate position on either CT, MRI, or PET scans. The system consists of a base ring attached to the patient's head during surgery, hand held instruments of any type, a camera array, and a computer display. Light emitting diodes on the instruments and head ring are tracked by three linear CCDs suspended over the surgical field. The position of the surgical instrument relative to the patient's head is computed by a personal computer. Surgical position is indicated on an individual CT, MRI, or PET slice. A graphics workstation provides three dimensional display of position.

MR imaging was performed using a superconducting 1.5 Tesla Siemen's Magnetom SP whole body imaging system with a standard planar elliptical spine coil used as an rf receiver. The imaging sequence consisted of a gradient-echo sequence with incremented flip angle series and k-space segmentation such that 6 or 8 phase-encoding steps were acquired in rapid sequence, constituting one segment. Sixteen or 20 interleaved segments were acquired so as to complete the 120, 128, or 160 X 256 matrix. A fat saturation pulse was applied prior to each segment so as to null signal from epicardial fat and thus enhance signal from coronary flow. A repetition time (TR) of 13 msec and an echo time (TE) of 8 msec were used with a slice thickness of 4 or 5 mm and 230 mm field-of-view (in-plane resolution of 1.4 - 1.9 mm X 0.9 mm). All imaging was performed within a single breath-hold.

The stable xenon/CT method has been expanded to study in addition to the tissue flow the flow of the vessels in the brain. This method was then applied to a series of baboon experiments to measure the flow and diameters of vessels and their dependence on the partial pressure of carbon dioxide (PACO₂). The method uses the value of the flow to distinguish vessels from tissues. The results from an experimental animal study show that (1) the location of the vessels with a diameter larger than 0.3 mm can be detected, (2) the vessel flow follows a Poiseuilles type law, (3) the average of the vessel flow is much higher near the skull at low PaCO₂, whereas at high PaCO₂ the average vessel flow is nearer the center.