Medical Devices : Diagnostics


Algorithms to Classify T Cell Activation by Autofluorescence Imaging

Building on award-winning work, UW–Madison researchers have discovered that autofluorescence intensity images of NAD(P)H can accurately classify T cells as activated or not activated (‘naïve’ or ‘quiescent’), and have developed algorithms to classify T cell activation based on the images. Specifically, adapting pre-trained convolutional neural networks (CNNs) for the T cell activity classification task, T cells can be classified with 92 percent accuracy. These pre-trained CNNs perform better than classification based on summary statistics (e.g., cell size) or CNNs trained on the autofluorescence images alone.

This invention provides a way to non-invasively detect T cell activation by imaging NAD(P)H intensity. These algorithms can be applied to NAD(P)H images taken with commercial imaging flow cytometers / sorters, and fluorescence microscopes. If increased accuracy of T cell activation is needed for a specific application, additional measurements of the other NAD(P)H and FAD fluorescence endpoints can be obtained and used for classification.

Method and Device to Screen and Sort Cancer Immunotherapy Cells

UW–Madison researchers have developed a highly accurate label-free method to non-invasively detect T cell activation by detection of free-NAD(P)H fraction, NAD(P)H α1. NAD(P)H α1 can be measured by fluorescence lifetime imaging or spectroscopy systems. The device could also sort T cells based on NAD(P)H α1. If increased accuracy of T cell activation is needed for a specific application, additional measurements of the other NAD(P)H and FAD autofluorescence endpoints can be obtained and used for classification.

Methods for Detection, Staging and Surveillance of Colorectal Adenomas and Carcinomas

UW–Madison researchers have identified a panel of protein biomarkers that can be used to noninvasively “stage” cancer (including lymph node status), identify high-risk precancerous polyps, and longitudinally monitor polyp presence before and after polypectomy. These protein markers reliably distinguish between stage 1, 2 and 3 cancers, as well as low- and high-risk adenomas. Further, select biomarkers revert toward normal levels with the disappearance of such adenomas, suggesting that these markers can be used to monitor for tumor resurgence and long-term care. Overall, this diagnostic advancement should improve colorectal cancer detection, patient treatment and outcomes while also reducing the associated costs.

Digital Otoscope for Optimal Access, Visualization

UW–Madison researchers have designed an otoscope featuring a small camera that is mounted on a narrow tip and able to ‘look around’ obstructions such as earwax. The narrow tip also permits other medical instruments to be inserted into the ear while the otoscope is being used (e.g., a curette for removing earwax or foreign objects). A remarkable view of the tympanic membrane is achieved, facilitating proper diagnosis.

Notable features include a disposable, light-conducting speculum sleeve with distal tip smaller than 2 mm. In addition, images may be captured directly from the device and stored in the patient record in compliance with Federal law.

Implantable Cancer Drug Delivery Device Signals the Future of Personalized Medicine

UW–Madison researchers have developed a new microfluidic device that allows efficient, minimally invasive delivery of drugs within a tumor, sparing patients from the unnecessary drug toxicity of full and indeterminate chemotherapy regimens.

With nothing more than a hypodermic needle, researchers and clinicians are able to administer small implantable devices containing concentrations of chemotherapeutic compounds to the primary tumor. Each device remains anchored and stable by deploying small barbs upon implantation. Specific drugs or drug combinations can be delivered to different areas of the tumor. Surgical removal of the tumor with the devices in place enables assessment of drug efficacy on affected cells.

Biomarkers for Detecting Prostate Cancer

UW–Madison researchers have identified eight genetic markers, or biomarkers, for prostate cancer. They can be detected in histologically normal prostate samples and/or the bodily fluids of men with no history of prostate cancer.

The biomarkers act as red flags, exhibiting abnormal methylation levels when cancer is present in peripheral prostate tissue (this is called cancer ‘field defect’). These changes are believed to represent early stages of the cancer process.

The biomarkers are associated with the genes CAV1, EVX1, MCF2L, FGF1, WNT2, NCR2, EXT1 and SPAG4.

Bedside Diagnosis of Swallowing Disorders

UW–Madison researchers have developed software that helps clinicians more easily analyze HRM data. Using a specially adapted manometer inserted through the nasal tract, a series of pressure measurements can be made at different points along the pharynx and esophagus. A computer program uses pattern recognition software to identify changes in pressure when the patient swallows. This data is output as diagnostic values indicating swallowing function.

Enhancing Light-Based Tissue Diagnostics by Dimpled Waveguide

UW–Madison researchers have developed a method of controllably and efficiently distributing the intense light available from solid-state light sources (LEDs, laser diodes, "white-light" GaN diodes, superluminescent light diodes) using a planar plate or waveguide film containing a two-dimensional array of conical indentations to distribute and/or concentrate near-point sources of light distributed at the periphery of the array.

The plate, or waveguide, is made of thin glass or other optically transparent material. Impressed into this layer are inverted cone-shaped indentations that are filled or coated by highly reflective silver. When photons from one or more light diodes strike the more numerous cones (the ratio of light sources to dimples can exceed 1:10), the light deflects laterally and radiates in proportion to the density and distribution of the conical indentations. Distributions can range from uniform to highly pixelated spots of high intensity light.

For clinical spectroscopic probes, an array of inverted reflective cones buried in a transparent planar waveguide deflects LED light from the periphery of the guide, focusing it through an array of apertures and onto the tissue to be diagnosed. The light reflected from the tissue bears a signature characteristic of either healthy or cancerous (breast) tissue. This reflected light is detected by an array of annular photodetectors, each surrounding one of the exit apertures.

Detection of the Catabolic State by Determining the Rate of Systemic Oscillations in Carbon Isotope Ratios

UW–Madison researchers have developed a method to determine if an individual is transitioning from a healthy state to an unhealthy state by monitoring breath and measuring the oscillation pattern in the relative amount of carbon isotopes. Breath taken from the individual is measured for a relative amount of a first isotope to a second isotope over a total time interval. Changes in the functional oscillation pattern of the isotope ratio can be correlated with and provide information about the health of the individual. Diet is not a confounding factor in the functional oscillation pattern.

To determine whether an individual is transitioning between a healthy and unhealthy state, a healthy functional oscillation pattern in the relative amount of the isotopes is identified during a time interval when the individual is healthy. Then, it can be determined that the individual is transitioning from a healthy state to an unhealthy state when the healthy oscillation pattern and a test oscillation pattern are distinct in period of oscillation, oscillations per unit time and/or variability in oscillation period. To determine the severity of an infection in an individual, the relative amount of a first isotope to a second isotope over a time period is measured. The degree of difference between the functional oscillation pattern for the individual compared to a reference pattern can be determined to indicate the severity of the infection.

Algorithm Improves Resolution of Time-Frequency Analysis for Medical Diagnostics, Telecommunications

UW-Madison researchers have developed a pseudo-wavelet algorithm known as the “damped-oscillator oscillator detector” (DOOD). This algorithm is unique among all wavelet and pseudo-wavelet algorithms in that it is the only algorithm that is explicitly based on modeling data as a “driving force” that interacts with a hypothetical set of mathematical oscillators. In the DOOD algorithm, an entirely new spectral density can be defined as the time rate of change in the energy specifically due to interaction with the data driving force, referred to as the data power. The data power measure is more sensitive to the presence or absence of data oscillators than traditional energy measures.

The DOOD algorithm allows an enormous frequency range to be spanned over as many orders of magnitude as desired. The instantaneous phase of oscillation and correlation functions can be calculated easily. The inverse of the DOOD transform is accomplished readily, which means that the DOOD algorithm also can be used to compress data. Any time-frequency or correlation analysis that can be accomplished by conventional means also can be accomplished using the DOOD algorithm, with the advantages of greater flexibility in defining the frequency range and better time resolution.

Liquid Crystal-Based Assay for Rapid and Precise Detection of Endotoxin in the Presence of Masking Agents

UW–Madison researchers have developed methods and devices for the simple and low cost, yet rapid, sensitive and selective detection of endotoxin in the presence of masking agents. Their assay uses micrometer-sized droplets of liquid crystal dispersed in aqueous solution. When a sensor containing the liquid crystal droplets is exposed to a solution containing endotoxin, the alignment of the liquid crystals quickly changes. This change in alignment is unaffected by common cations, surfactants, buffers or chelating agents and can be detected easily using polarized light or other means.

Integrated, Miniaturized Fiber Optic Probe for Light-Based Diagnostics

UW–Madison and Duke University researchers have developed an integrated fiber optic probe that allows for in vivo sensing of biochemical and morphological changes in local tissue. The probe replaces a spectrometer by bonding thin, flexible photodetector elements directly to the fiber probe tip, which makes local detection of light feasible.  The fiber is processed further to incorporate a mutual ground plane, an insulator and metal lines for transmitting the detected signal. The structure may be constructed to be compact enough to fit within the shaft of a needle, allowing probing of tissue without the need for biopsy. By directly integrating photodetectors with an optical fiber, the probe provides a compact structure that can be placed in close proximity to a sample to increase throughput and decrease cost, making it practical for clinical use.

Computational Algorithms for Identifying, Suppressing and Reversing Epilepsy

UW-Madison researchers have developed a protocol that accounts for each of the conditions required for the development of epileptogenesis and determines a treatment to reverse, or “unlearn,” epilepsy.  Because this protocol addresses factors in addition to neuronal hyperexcitability, it may prove more effective than current methods. 

The new technique involves acquiring and analyzing neural activity data from a subject to determine epileptic patterns based on neuronal hyperexcitability, spatial connectivity and temporal connectivity.  Treatment using an electrical stimulus then is focused based on the determined patterns and administered to the subject.

Imaging Spectrometer for Early Detection of Skin Cancer

A UW-Madison researcher has developed a portable imaging spectrometer for the early detection of skin cancer. A handheld scanner uses light emitting diodes to illuminate a region of skin and the reflected light is collected by an objective lens. A micro-lens array then divides the region into smaller images that are processed to reveal their spectral content.

Because spectral and image data are acquired in one step, this new device provides two effective indicators to detect skin cancer. Physicians can evaluate the image data while the spectral data is compared to spectra of known cancerous or healthy regions.

Using Endogenous Fluorescence to Identify Cancerous Cells

UW-Madison researchers have developed methods to identify, detect and characterize diseases, such as cancer, using non-linear infrared imaging. Changes in the fluorescent properties of tissue indicate changes in cellular metabolism that may signify the presence of disease. Specifically, the fluorescent properties of flavin adenine dinucleotide (FAD), a redox cofactor involved in several important metabolic reactions, can indicate the presence of cancer, particularly epithelial tumors such as breast tumor cells.

To detect cancer, tissue is exposed to near infrared radiation, which excites endogenous FAD fluorophors. The FAD fluorophors then emit measurable fluorescent signals that vary with different tissue properties. A partially or fully automated system analyzes the signals and compares them to previously acquired reference data. The findings can be used to identify, locate and characterize the presence and stage of carcinomas.

Improved Method of Fluorescence Spectroscopy using Monte Carlo Simulation for Medical Diagnostics

UW-Madison researchers have developed an improved method for fluorescence spectroscopy using a Monte Carlo based simulation to extract intrinsic fluorescence spectra and detect abnormal cells.  The improved method utilizes certain aspects of previous optical diagnostic technologies such as fiber-optic probe geometry optimization and diffuse reflectance spectroscopy.  However, unlike previous methods of extracting optical data exclusively from diffuse reflectance spectra, the new fluorescence spectroscopy technique extracts intrinsic fluorescence from the raw tissue fluorescence and does not require empirical correction factors for the specific probe geometry and/or instrument configuration. 

The optical properties of the tissue are first acquired through an established method of Monte Carlo based modeling of diffuse reflectance.  Simulations are run to generate a three-dimensional grid of the amount of photon energy deposited per unit volume.  Then the location and intensity of fluorescence can be calculated as a function of emission wavelength based on the probability of incident light causing fluorescence and the generated grid of deposited energy.  Finally, the fluorescence location and intensity and probability of a photon escaping the medium surface are used to determine the concentration of the fluorophore and other intrinsic fluorescence properties that are independent of absorption and scattering.

The fluorophore concentration is the diagnostic metric used to determine the condition of a tissue.  In such a case the bio-optical diagnostic would be compared to an optical property database to determine the types of cells in the region of interest.  For example, pre-cancerous and cancerous tissues contain a different concentration of fluorophores due to abnormal growth of the cells.  Concentration of fluorescently labeled drugs or reporter genes can also be quantified in vivo. The new method of fluorescence spectroscopy provides the means to detect abnormalities in the tissue for the early detection of cancer and other diseases, which will increase the probability of successful treatment and full patient recovery.

Depth-Resolved Reflectance Instrument

UW-Madison researchers have developed an improved reflectance instrument and method to collect and analyze optical information from a pre-cancerous or cancerous target in a turbid medium, such as epithelial tissue. The instrument uses a smart fiber-optic probe to deliver a selected wavelength of light to tissue and sense the reflected light from specific layers. Altering the angles of illumination and detection relative to the tissue surface and the source-detector separation allows the clinician to probe at various depths beneath the surface of the tissue. Specially designed modeling for two-layered tissue enables the user to extract information from each individual layer for diagnosis.

Optimizing Probes to Improve Spectroscopic Measurement in Turbid Media

UW-Madison researchers have developed a method, apparatus and corresponding computer program to determine the optimal probe geometry for use in a particular tissue or other turbid medium. Light transport through the medium is modeled to simulate the diffuse reflectance properties that would be measured by a specific probe geometry. An inversion algorithm then converts the diffuse reflectance properties into optical properties. Those optical properties are compared to the known optical properties of the tissue to determine how well they match. When this process is repeated for additional probe geometries, the program indicates which geometry gives the most accurate measurement of the optical properties of the medium.

Method for Extraction of Optical Properties from Diffuse Reflectance Spectra

UW-Madison researchers have developed a flexible and efficient Monte Carlo-based method for analyzing the results of diffuse reflectance spectroscopy and extracting absorption and scattering measurements from tissue. The method is valid for a wide range of optical properties, yet remains computationally feasible. It uses an iterative process that repeatedly models reflectance values from a set of estimated tissue optical parameters using a Monte Carlo model. It then calculates the error between the measured diffuse reflectance and the modeled reflectance values. Next, the tissue’s scattering and absorption characteristics are calculated from the estimated optical parameters that result in minimum error. The concentrations of the absorbers in the tissue, including concentrations of oxygenated and deoxygenated hemoglobin, are easily extracted from the absorption coefficient and can be readily converted into two important physiological parameters: hemoglobin saturation and total hemoglobin content.

Ultrasound Determination of Vascular Age

A UW-Madison researcher has combined direct measurements of atherosclerotic burden with existing risk paradigms to determine an individual’s “vascular age.” Atherosclerotic burden is determined from measurements of carotid artery intimal-media thickness (CIMT) acquired by high-resolution ultrasound -- a non-invasive, highly reproducible technique for detecting and quantifying atherosclerosis. CIMT was combined with population-based nomograms from the Atherosclerosis Risk in Communities Study (1993. Stroke 24:1297-1304) to create mathematical algorithms for determining vascular age, which in turn is used in conjunction with traditional risk assessment models to improve evaluation of individual coronary heart disease risk.