Improving Surgery with Transient Lighting
Fluorescence image-guided surgery (FIGS) has the potential to transform surgical medicine, improving disease outcomes and minimizing damage to healthy tissue. FIGS, during real-time surgery, can “light up” and differentiate otherwise invisible features like small clusters of cancer cells and sub-surface blood vessels. However, FIGS requires very low light or even darkness in the operating suite—a significant drawback. We have developed a transient lighting system which allows FIGS and other fluorescent imaging applications to be performed in a well lit room with no loss of image quality. Our patent-pending system uses rapidly switched, ultra-bright LEDs that provide normal room lighting to the eye, but are in fact off over 90 percent of the time. Fluorescence images can be collected during these imperceptible periods of darkness, while providing the medical team a safe, well-lit work environment. We have implemented a working prototype of this system on a clinical fluorescence microscope, and are currently incorporating transient lighting into a small animal imaging and the surgical room.
State-of-the-Art Equipment for Animal Studies
Preclinical studies are pivotal to the development of new medical technologies, and they may require specialized equipment to generate useful data. We are developing a small animal micro-radiotherapy (RT) system, based on open-source technology, that will provide an inexpensive yet powerful tool for widespread use in animal radiobiology research. This system incorporates a custom-designed binary micro-multileaf collimator (MLC) for intensity modulated RT (IMRT). It also includes WiscPlan-kV, an open-source, in-house treatment planning system for low and high energy x-ray beams that can accurately calculate dose and optimize dose distribution, which is valuable for animal research. We designed and built the MLC in our Fab Lab with sub-millimeter precision to accurately position and actuate the collimator leaf, and used our 3D printers to make small custom parts. We are proceeding to validate the assembled system.
Optical Tomography: Improving 3D Imaging
3D imaging has tremendous potential to display the interior realities of living organisms. Our optical tomography research is designed to find better techniques for gathering high-resolution and large-scale 3D datasets of tissues, organs, embryos and complete organisms. One effort involves Optical Projection Tomography (OPT), where light is projected through a sample and a camera records an image of the projection. Rotating the sample in the light path enables a 3D reconstruction of the 3D volume using the projections from multiple angles. In addition to refining the instrumentation, we are exploring ways to optically clear tissue to improve light penetration and image production. We are also using a related technique, Single Plane Illumination Microscopy (SPIM), where a specimen is moved through a focused sheet of laser light to obtain 3D fluorescent images at cellular resolution – very much like a CT scanner. SPIM and OPT are capable of generating striking pictures of entire model organisms. Both of these techniques will constitute valuable imaging tools for both researchers and medical diagnosticians.
A Visionary New Design: Multi-Source Computed Tomography
Computed Tomography (CT) is a widely used tool for the non-invasive imaging and diagnosis of disease, but industry standards are constraining its technological advancement. Adherence to an outdated data acquisition geometry—where a single x-ray tube and detector rotate around a patient—is impeding progress towards better temporal resolution (shutter speed) in CT and superior image quality in cone beam CT. We are designing and constructing a modular multi-source x-ray tube for CT that should enable faster imaging, dose reduction, and improved image quality. In our design, multiple modules can be combined to form a stationary CT scanner that increases temporal resolution beyond that possible with a standard rotating gantry. The new tube can generate the required x-ray flux for an emerging data acquisition geometry that promises to improve image quality in volumetric CT, and it can also be used to lower patient dose through a novel scanning technique which optimizes the radiation needed to produce the image.
New Production Technology for Medical Isotopes
Medical isotopes, such as Moldybdenum-99, are radioactive substances injected into the patient for safe, cost-effective medical imaging and treatment. The supply of Mo-99, used extensively in diagnosing heart conditions and certain cancers, is currently in jeopardy, as the aging nuclear reactors which generate it are shutting down. We are designing a new reactor capable of producing 100 percent of the U.S. demand for Mo-99 and operating without use of highly enriched, weapons-grade uranium. This facility will be fueled with crystalline uranyl nitrate to simplify back-end chemistry and allow for a flexible fuel recovery process. We have partnered with the U.S. Department of Energy National Laboratories (responsible for key chemistry work) and UW–Madison (responsible for required thermal, neutronics and criticality calculations). Combined findings have enabled us to generate a preliminary reactor design used as a starting point for a preliminary process flow scheme for the facility. We will continue design revisions based on additional neutronics analyses and the resolution of unknown thermal properties of the proposed fuel.
Deformable Registration: New Tools for Breast Cancer Medicine
Is there a better way to monitor the progression of cancer to improve outcomes? We are leveraging a unique longitudinal study of changes in breast density—a known prognostic factor in breast cancer—to create tools to do just that. Our starting point is a set of image datasets acquired over multiple years from screening women at high risk for breast cancer. For reliable comparisons, these images must undergo a sophisticated macro-scale manipulation using computer algorithms. Images acquired from each woman at different times must be segmented to create a region of interest of the breast tissue, and then deformably registered onto each other. A successful manipulation enables the use of image processing tools to reliably quantify breast density changes year to year. The longitudinal study will correlate this data with known clinical outcomes for the women to create a medical reference tool. Our initial efforts have focused on optimizing the image registration process by developing the segmentation algorithm for the deformable registration. We also think that this technique will have wide utility for other longitudinal image studies for medicine and microscopy.
CAMM: New Instrumentation to Track Breast Cancer
Our Compact Automated Multi-Photon Microscope (CAMM) is an imaging system designed to better track the progress of breast cancer, based on the alignment of surrounding collagen fibers. Collagen has recently been identified as a biomarker for breast cancer metastasis. The perpendicular alignment of collagen around ductal carcinoma in-situ (DCIS), known as Tissue Associated Collagen Signature Level 3 (TACS3), has been shown to have a direct relationship with the migration of tumor cells and the low survival rates of patients. Our instrumentation integrates CAMM, a custom-designed second harmonic generation imaging system, with software such as ct-FIRE and CurveAlign, used to compute features that describe collagen interactions with epithelial cells. This CAMM system will be able to detect collagen fiber characteristics and alignment with respect to the tumor cell boundary. As we work further to improve it, CAMM should become a valuable tool for both breast cancer research and diagnostics.
Better Radionuclide Therapy for Difficult Tumors
In targeted radionuclide therapy, toxic doses of radioactivity are beamed into a patient to treat cancerous tumor cells. The primary treatment goal is the conformal and isolated delivery of the radiation, to maximize the destruction of the tumor cells while sparing healthy tissue. However, this goal is difficult to achieve in the radiation therapy of small tumors and of metastatic disease where cancer cells have spread. Our radionuclide therapy project addresses the challenge of treating difficult tumors. Its aim is to identify and develop an agent for dual diagnosis and radiotherapy that would target multiple tumors and destroy them. We are investigating various radioisotopes, molecular agents, and fluorescent agents for this purpose. In our preliminary research efforts, we are conducting simulations using Monte Carlo N-Particle tracking to evaluate the dose delivery of several promising radioactive candidates.
Improving Diagnostics for Ovarian Cancer
Ovarian cancer is a major health threat for women, causing more deaths than all other gynecologic cancers combined. More effective approaches for diagnostics and treatment are urgently needed. Recent discoveries have shown that the remodeling of the extracellular matrix (ECM) can indicate the progression of ovarian cancer. We are developing a new technique to acquire meaningful 3D extracellular matrix images to better monitor this cancer. We start with a reconstructed 3D image of a tissue structure, generated by specially focused laser beams. We then quantify the tissue’s ECM remodeling, using 3D texture analysis to delineate the collagen fiber morphology observed in normal and high grade malignant ovarian tissues. We next create a library of texture features and calculate a representative model for each tissue type. This procedure enables us to classify images of normal and high grade malignant ovarian tissues. Our algorithm is a more general method to probe rapidly changing fiber morphologies than existing global analyses, and is more versatile than other texture approaches as it can be tailored to different disease states.
Custom Prosthetics through 3D Printing
Currently, upper limb prosthetics often present more hindrances than benefits to people who need them. Given the bulky construction and slow actuation of available products, many amputees go without one altogether. Eric Ronning, a UW–Madison undergraduate engineering student affiliated with Morgridge, is designing a next-generation prosthetic hand using the capabilities of our Fab Lab. This technology implements both 3D printing and scanning technologies to replicate an amputee’s lost hand for a custom fit. The prosthetic, which now incorporates a number of inventions and design iterations, provides superior functionality, and is aesthetically innovative and cost-effective to manufacture.
Vertical CT: More Versatile Imaging for Veterinary and Human Medicine
The Equine Standing Computed Tomography (CT) project is a partnership with the UW–Madison School of Veterinary Medicine. Every year hundreds of horses are euthanized, often without any prior symptoms, when they break a leg and fall. Often a rider is also injured. The medically ideal solution for scanning horse legs is to acquire the images under the horse’s own standing load, thereby enhancing the ability to detect sub-clinical leg fractures. The Equine Standing CT can also be repositioned to scan a horse’s head and neck. The system is based on a three-axis robotic CT support system and utilizes diagnostic quality multi-row fan beam CT detectors. The system also has potential use in human treatments. For example, proton radiotherapy patients can sometimes be best treated in the seated position but this requires a CT capable of scanning in the treatment position.
New Technology to See Around Corners
What if you could visualize the speed of light to see around corners? We have built a camera system fast enough to see light move through a scene. Captured information about the light’s time of flight can reveal images hidden to other imaging devices. Our ultra-fast transient imaging camera, along with new algorithms, can process and understand its images. The combination of ultra-fast hardware and advanced signal processing enables use of scattered light to reconstruct images of an object that we cannot see directly—to effectively “see around the corner.” This is a platform technology with many potential applications. We are exploring the potential of this method for biomedical imaging, for example in scattering tissues, and for endoscopy. We are also developing a new transient imaging system to image the inside of a room from outside it, and are designing a much larger system for the satellite exploration of caves on the moon.
SETI: Imaging Deeper into Tissue Samples
Our Sequential Erosion Tissue Imaging (SETI) project addresses a key limitation of optical imaging: the depth to which it is possible to image, which is constrained by how far light can penetrate into tissue and other samples. This limit is typically on the order of a few hundred microns, and means that only samples within this depth range can be effectively imaged. SETI is an innovative technique that can bypass this limitation. During the SETI process, we embed biological samples in hard plastic. The surfaces of these plastic-embedded samples are then imaged. After imaging, a layer of the sample is machined off using a mill, revealing a fresh surface. This new surface is then imaged. By repeating these alternating imaging and machining steps, we can image samples of much larger size. While this process is destructive, it allows for the creation of 3D datasets of samples that are much larger than what optical imaging typically allows.