Seminars

Value Added CT with Opportunistic Screening: Automated biomarkers, Applications, and Future Directions

John Garrett, PhD, DABR
Associate Professor (CHS) of Radiology, Medical Physics, and Biostatistics and Medical Informatics; Director of Imaging Informatics
University of Wisconsin–Madison

Computed Tomography (CT) is a cornerstone of diagnostic imaging, providing critical anatomical and pathological insights. Beyond its diagnostic value, CT also offers untapped potential for opportunistic screening, leveraging existing scans to extract quantitative biomarkers that inform patient health. This presentation explores the concept of value added CT, highlighting how automated analysis tools can unlock this latent value.

We will review state-of-the-art automated biomarkers derived from CT, including body composition analysis, vascular calcifications, and bone density, discussing their implications for screening and risk stratification in chronic diseases like cardiovascular disease, osteoporosis, and cancers. We will explore the impact of CT scan protocol variability and will also discuss the concept of biological aging and how imaging markers might help shed some light on that. Additional case studies will illustrate real-world applications, focusing on technical methods, clinical integration, and the benefits of opportunistic analysis for population health.

Finally, we will discuss future directions, including emerging AI-driven techniques, extension to other modalities, and integration with electronic health records. By repurposing routine CT imaging data, we can enhance the value of radiological studies, driving innovation in preventive care and multidisciplinary collaboration.

Beyond constraint satisfaction: how genomics can usher us into true radiation plan optimization

Jacob Scott, MD, DPhil
Professor of Medicine and Physics, Staff Physician-Scientist and Radiation Oncologist
The Cleveland Clinic and Case Western Reserve University

While we say we are making optimal plans for patients in radiation oncology, what we are doing is actually choosing from a number of plans which satisfy constraints. In the absence of continuous functions of outcome, this is the best we can do. However, with the advent of functions relating physics radiation dose to tumor genomics, these continuous functions do exist, opening the door for true multicriteria optimization. In this lecture I will go through the derivation of these continuous functions from tumor genomics using canonical equations of radiation biology, and review the data supporting their mapping to outcome.

Can Near Infrared Nerve-Specific Imaging Improve Surgical Outcomes?

Summer Gibbs, PhD
Douglas Strain Endowed Professor of Biomedical Engineering
Oregon Health & Sciences University

Fluorescence guided surgery (FGS) is a nascent field, however with ~15,000 clinical FGS system distributed worldwide, its potential to specifically highlight tissues to be resected (e.g., cancer) and avoided (e.g., nerves) has been recognized and there are >125 ongoing clinical trials with novel contrast agents to leverage this clinical imaging technology. Iatrogenic nerve damage is arguably one of the most feared surgical complications as nerve injury is often permanent leaving patients with pain, loss of function and disability. Near infrared (NIR) nerve-specific contrast agent(s) that are spectrally matched to the existing clinical FGS infrastructure have a direct path to clinical translation with broad surgical applicability. However, development of NIR nerve-specific probes has been a substantial challenge as these probes must be small enough to cross the tight blood nerve barrier, but have a sufficient degree of conjugation (i.e., double bounds, which by definition increase the molecular weight) to reach NIR wavelengths. Through a directed fluorophore medicinal chemistry approach, we have designed and developed first-in-kind, small molecule NIR nerve-specific fluorophores. Our team is currently working towards clinical translation of our novel probes as we explore the utility of nerve imaging for a variety of surgical indications including prostatectomy, neurosurgery, endocrine surgeries, head and neck surgeries and orthopedic indications.

Expanding Horizons in Diagnostic Ultrasound: New Modalities and Their Clinical Impact

Ivan Rosado Mendez, PhD
Assistant Professor, Departments of Medical Physics and Radiology
University of Wisconsin–Madison

Diagnostic ultrasound is at a pivotal juncture. The increasing computational power of ultrasound scanners has enabled the development of novel imaging modalities, revealing new sources of contrast that enhance clinicians’ understanding of disease onset, progression, and response to therapy. Concurrently, the advent of handheld and wearable transducers is democratizing access to diagnostic tools, thereby promoting more equitable access to medical imaging. These advancements present new challenges for medical physicists.

This talk will provide an overview of the biological and clinical motivations behind the development of new ultrasound imaging techniques that uncover novel sources of contrast in diagnostic ultrasound. Emphasis will be placed on their physical foundations and the technical aspects of their implementation. Specific applications from the Quantitative Ultrasound Lab will be presented. The talk will conclude with a critical discussion on the challenges these new techniques face and on potential research avenues to address these challenges and facilitate clinical adoption.

Modern Applications of Quantitative Ultrasound for Biomedical Research

Jonathan Mamou, PhD
Professor of Electrical Engineering in Radiology
Department of Radiology, Weill Cornell Medicine, New York, NY

Quantitative ultrasound (QUS) is an active research field focused on obtaining quantitative tissue properties (i.e., system- and user-independent) from ultrasound data. Conventional ultrasound imaging is commonly used to visualize soft tissue morphology. During scanning, a gray-scale B-mode image is displayed on screen from which a trained clinician can evaluate tissue states. However, B-mode ultrasound image formation discards valuable information in the raw backscattered echo signal that encodes information about tissue microstructure. Therefore, microstructural changes in soft tissues that accompany disease processes, but do not directly affect tissue morphology, may not be visible in B-mode images. QUS methods use the raw ultrasound data to reconstruct parametric maps that are representative of tissue microstructure. In this talk, I will review conventional ultrasound imaging and QUS methods based on analyzing the backscatter coefficient and envelope statistics. I will present recent in vivo results from lymph nodes (cancer), thyroids (cancer), and eyes (myopia). I will also briefly present results to improve and use quantitative acoustic microscopy systems used to image thin sections of tissues at ultra-high ultrasound frequencies (i.e., 250-1000 MHz).

Optimizing Ultrasound with AI: Physically-Inspired, Semi-Supervised, and Self-Supervised Learning for Efficient Aberration Correction and Beamforming

Hassan Rivaz, PhD
Professor & Concordia University Research Chair, Electrical and Computer Engineering

Ultrasound is a safe, fast, cost-effective, and portable imaging modality. With the introduction of pocket-sized, battery-powered ultrasound devices, it has the potential to become the modern-day stethoscope, providing clinicians with an accessible tool for critical patient insights. However, interpreting ultrasound images, which are often noisy, requires the expertise of trained clinicians. This presentation explores how advances in AI can simplify ultrasound interpretation by addressing two key areas: enhancing ultrasound image quality and extracting objective, quantitative tissue properties, such as tissue elasticity.

A significant challenge in training AI models for ultrasound analysis is the absence of ground truth in real ultrasound data. While simulated ultrasound images can be used, they often oversimplify the complexity of real-world conditions, leading to the domain shift problem, where model performance degrades on real test data. Additionally, even if ground truth could be estimated in real ultrasound data, variations in imaging settings, such as center frequency, further complicate model generalization. Our datasets frequently contain thousands of real ultrasound images, but without corresponding ground truth labels. In this talk, I will introduce methodologies based on self-supervision and unsupervised learning to unlock the value of these large, unlabeled datasets and address the challenges of domain shift and data variability in ultrasound imaging.

Radiopharmaceutical Therapy – How to Treat What You Cannot See

Roger W. Howell, PhD
Chief, Division of Radiation Research
Rutgers University

Radiopharmaceutical cocktails have been developed over the years to treat cancer. Cocktails of agents are attractive because one radiopharmaceutical is unlikely to have a curative therapeutic effect due to nonuniform uptake by the targeted cells. Therefore, multiple radiopharmaceuticals targeting different receptors on a cell and different degrees of diffusion into micrometastases is warranted. However, past implementations of radiopharmaceutical cocktails in vivo have not met with convincing results due to the absence of quantitative optimization strategies. We have developed artificial intelligence (AI) tools, within our recently released MIRDcell V4 application, that can optimize cocktails of radiopharmaceuticals. The AI tool determines the molar activities for each radiopharmaceutical in the cocktail that minimize the total disintegrations of the therapeutic radionuclide(s) that are required to achieve a specified tumor cell kill. Tools are provided for populations of cells that represent circulating or disseminated tumor cells, and for micrometastases. The capabilities of these tools will be demonstrated using both model and experimental data to show that this approach could be used to analyze samples of cells from cultures, animals, or patients to formulate the best combination of radiopharmaceuticals for maximum therapeutic effect with the least total number of disintegrations.

The Role of Imaging in CNS Drug Discovery and Development

Sjoerd J. Finnema, PhD
Research Fellow
AbbVie

The discovery and development of central nervous system (CNS) drugs demand substantial resources, extended timelines, and significant costs, all entailing high risks. Positron emission tomography (PET) and magnetic resonance (MR) imaging have become indispensable tools in CNS drug development, facilitating decision-making in early-phase studies by assessing tissue exposure, target engagement, and pharmacological activity. As CNS drug development shifts towards disease modification, neuroimaging has emerged as a critical translational tool, optimizing the stratification of clinical trial participants and enhancing decision-making throughout the entire drug development process. In this seminar, we will explore various examples of how PET imaging plays a pivotal role in the CNS drug-development process.

Reducing Uncertainty, Increasing Confidence: Ionizing Radiation Metrology at NRC

Bryan Muir, PhD
Research Officer (NRC), Adjunct Research Professor (Carleton University)
National Research Council Canada

The National Research Council is Canada’s largest research and technology organization, founded in 1916 to carry out applied research to address industrial and societal needs. A specific mandate of the NRC is to fulfil the role of Canada’s National Metrology Institute (NMI) and this is realized through NRC’s Metrology Research Centre. NRC Metrology is responsible for the realization of the international system of units (SI) and disseminates measurement standards through calibration and technical services to a wide range of end users.

The Ionizing Radiation Standards group within the Metrology RC is very active in addressing measurement issues in radiation protection, radiation therapy and industrial applications of ionizing radiation. Through i) the development of primary measurement standards for dosimetry, radioactivity and neutron fields, ii) the dissemination of these standards through calibrations, and iii) participation in research collaborations and best-practice guidance, it supports users across all applications of ionizing radiation to ensure that point-of-use measurements are accurate, traceable and equivalent.

This presentation will explore the history and mandate of the NRC, and describe the worldwide system of metrology that provides the basis for all measurements. I will describe activities in ionizing radiation metrology in Canada with a focus on research projects in linac dosimetry.

Listening to light: Photoacoustic imaging and its applications

Manojit Pramanik, PhD
Northrop Grumman Associate Professor
Iowa State University

Photoacoustic imaging is an emerging novel imaging technique combining both optical and ultrasound imaging. A short-pulsed laser illuminates the tissue to generate sound waves called the photoacoustic wave. These sound waves are used to get high contrast, high resolution deep tissue imaging of various parameters. Photoacoustic imaging, like ultrasound imaging, is a multi-scale multi-depth imaging modality.

In this presentation, the basics of photoacoustic imaging and several imaging systems will be covered. For example, photoacoustic computed tomography (PACT), photoacoustic microscopy (PAM). Few clinical/pre-clinical applications of photoacoustic imaging will be discussed, such as, non-invasive sentinel lymph node detection for breast cancer staging, intracranial hypotension detection in small animal, photoacoustic aided ultrasound-guided needle tracking etc. Recent development of deep learning applications in photoacoustic imaging and contrast agents will also be discussed.

Optical Microscopy: Faster and Gentler Imaging

Abhishek Kumar, PhD
Assistant Professor
Cell and Regenerative Biology and Center for Quantitative Cell Imaging
UW-Madison

Optical imaging is widely used in biomedical sciences. Despite the wide range of optical imaging techniques available, there is no one universal modality suitable for all length scales and types of biological samples. Also, development of new optical imaging systems that can bridge existing gaps between cellular and organismal imaging is vital. Confocal and light sheet microscopes are two popular optical imaging modalities that are routinely used for imaging live and fixed samples at multiple spatial and temporal scales.

For the first portion of this talk, I will discuss our group’s efforts to advance these technologies for imaging large or rapidly growing samples at high spatial and temporal resolutions. Specifically, I will present our implementation of a confocal microscope for improving imaging speed by a few orders of magnitude for large samples. I will then show the use of this imaging technique, in collaboration with your colleagues, to answer important biological questions in systems ranging from living zebrafish embryos to whole cleared mammalian organs.

In the second half of my talk, I will discuss the computational tools we are developing and using to compliment imaging modalities. I will conclude by sharing our lab’s vision for the future and plan to integrate nanophotonics to augment imaging modalities.

Combining Technology and Use Design for High Value MRI

Jim Pipe, PhD
Professor of Radiology & Medical Physics
UW-Madison

Clinical MRI scanners were not designed de novo, but have their origin
in the NMR scanners used in laboratory science starting in the mid 20th century. The complexity of the equipment and the richness of the information that can be obtained makes MRI an incredibly useful diagnostic tool. However, some of this legacy includes unnecessary complications that pervade clinical MRI, which in turn can produce errors, image variability,
inefficiency, and poor access.

The Magnetic Resonance Technology and Use Design (MRTUD, pronounced “MR2D”) group at UW Madison produces next-generation MR technology with a strong emphasis on the interaction of the user. As we improve technical quality and efficiency, we embrace simplicity, operational clarity, conceptual models, and improved user experience. Examples of the technology we are designing, along with examples of the way in which strong design principles are applied, will be discussed. Our framework of “Value” will also be discussed, which is used as a guide for improving the impact our work on each individual patient as well as the global healthcare system in general.

Imaging synaptic vesicle protein 2A in the living brain

Jason Cai, PhD
Associate Professor of Radiology & Biomedical Imaging
Yale School of Medicine

Extensive preclinical and postmortem data have shown that the brain region specific synapse loss is a robust biomarker of a variety of neurodegenerative diseases. Traditional techniques for quantifying synapses exclude their uses in the clinic, limiting their translation and direct clinical impact. PET imaging of synaptic vesicle protein 2A (SV2A) for the first time allowed us to quantify synapses in living systems noninvasively, opening up the possibility of using the spatial-temporal synapse density information in the early diagnosis of neurodegenerative diseases and the evaluation of treatment effects of novel therapeutics in the clinic.

In this presentation, we will talk about the development of SV2A PET ligands and their preclinical, translational, and clinical applications in neurodegenerative disorders.

Ultrasound and acoustic cavitation for therapy

Aarushi Bhargava, PhD
Assistant Professor, Department of Biomedical Engineering
University of Wisconsin-Madison

Although ultrasound is conventionally recognized as a diagnostic modality, it is emerging as a therapeutic strategy for various medical conditions. Ultrasound can non-invasively and wirelessly deliver thermal or mechanical energy via waves and induce cavitation to manipulate tissues and injected synthetic materials inside the body. This manipulation can be used for drug delivery and cell stimulation at low ultrasound intensities. At higher intensities, it can ablate tissues. In this talk, I will briefly discuss my research on using different acoustic cavitation regimes of microbubbles at varying intensities of focused ultrasound to treat retracted blood clots and alter dense fibrous environments at microscale levels. We will also be looking at the ability of ultrasound to stimulate synthetic microsystems for performing spatiotemporally controlled drug delivery and high-frequency cell-specific neurostimulation.

The Infinite Game of Research

Marty Pagel, PhD
Professor of Medical Physics & Radiology
University of Wisconsin-Madison

This presentation will discuss a variety of molecular imaging methods and research applications. We have developed MR Fingerprinting, PET/MRI, Photoacoustic Imaging (PAI) to quantitatively measure extracellular pH, oxygenation, and vascular transport rates in solid tumors. More recently, we have developed optical agents that detect extracellular proteases for eventual use during image guided surgery. We apply these molecular imaging methods to preclinical tumor models, and we have also applied our imaging methods to evaluate wound healing models.

Furthermore, this presentation will discuss the Infinite Game of research. Although biomedical research has often become a Finite Game in the USA, this type of game is unsustainable. Instead, a return to an Infinite Game of research has long-term benefits for our society, which can be accomplished by focusing on the fundamentals of the Infinite Game.

Milestones in CT: Past, Present, and Future

Cynthia H. McCollough, PhD
Brooks-Hollern Professor of Medical Physics and Biomedical Engineering
Mayo Clinic, Rochester, MN

Since the first patient CT scan in 1971, CT technology has continued to improve, both in evolutionary and revolutionary ways. This presentation will describe the technical and clinical innovations that have occurred, including the introduction and clinical adoption of spiral, multi-slice, cardiac, dual-source, dual-energy, and photon-counting-detector CT. Reconstruction of images from the projection data has also undergone major changes, moving from filtered back-projection to iterative reconstruction and deep learning approaches. The consequences of these advances in technology include the ability to diagnose and monitor an ever-expanding list of clinical indications. Several of these will be highlighted in this tour of CT imaging.

Educational outreach is essential for increasing the visibility of medical physics and enticing the next generation of brilliant and motivated professionals. To encourage engagement with outreach efforts, the Medical Physics Graduate Student Outreach committee has worked with the Department of Medical Physics to offer scholarships for the creation of activities demonstrating interesting concepts in medical physics to non-expert audiences. After a short presentation detailing the current state of outreach in the Medical Physics program, each of the scholarship awardees will present their demonstrations. This seminar will be a fun and interactive opportunity to consider how we communicate our field’s value to the public.