Ethical Issues for Radiology and Radiologists

While it may seem that as radiologists, we perform and interpret the imaging studies that are ordered, in fact, we are making ethical decisions every day. This becomes even more notable in our research, our teaching, and in how we understand the challenges that we are presented with. Ethical decisions are often based in our moral codes, which may not be universal. This talk will provide historical context, current issues, and cases for discussion. 

The Real Intelligence in Radiology - Understanding & Protecting Our Greatest Resource - Radiologists

Radiology is changing in many ways due to such factors as deep learning, artificial intelligence, reimbursement, policy and the recent COVID pandemic. These changes necessitate consideration of how radiologists and radiology trainees can adapt and perhaps redefine their role in patient care as well as a consideration of the impact of these factors on training methods, productivity, the interpretation process, and the environment in which radiology is practiced. Understanding the perceptual and cognitive mechanisms underlying image interpretation and skill acquisition are fundamental to creating work environments that facilitate and enhance patient care without leading to fatigue and burnout. This talk will summarize some of the research approaches develop to accomplish this.

Promote Your Academic Career Using Social Media

Scientific knowledge doubles every 90 days. Physicians must quickly learn about recent discoveries to remain current in their chosen specialties. How does tech help doctors stay up-to-date? TikTok, Twitter, Snapchat, Instagram, and Facebook are all used to teach physicians and their patients. Alt Metrics now measure the impact of individual scientists instead of the journals in which they publish. Social media outlets give physicians a platform to advocate for public health, social justice, and equity in our medical schools and society at large. This session will discuss the use of social media to advance the academic careers  and impact of physicians and scientists.

Lessons Learned from 20 years of Venous Stenting

Over the past 25 years, venous stent placement has become a mainstay in treating symptomatic iliofemoral venous thrombosis with a favorable safety and efficacy profile compared to surgical interventions. In the past decade, venous stent placement has also gained traction in the treatment of venous stenosis (most commonly secondary to May Thurner syndrome) in the absence of thrombosis. Despite its popularity, very little research has been done to identify best practices in venous stenting including pre-procedural workup, procedural techniques, and post-procedural follow-up. This talk will provide an overview of venous stenting history and also discuss research performed at Stanford to debunk commonly held myths related to venous stent placement.

PSMA PET in Prostate Cancer Biochemical Recurrence

¹⁸F-DCFPyL is a urea-based small molecule inhibitor of prostate specific membrane antigen (PSMA) that was recently approved by the FDA for evaluation of prostate cancer prior to definitive therapy and for biochemical recurrence. I will discuss our experience in the prospective evaluation of ¹⁸F-DCFPyL in a single center study for detecting lesions in prostate cancer patients with biochemical recurrence (BCR). We prospectively enrolled 247 men with biochemical recurrence (after primary definitive treatment with prostatectomy or radiotherapy. The ¹⁸F-DCFPyL positive lesions compatible with prostate cancer were identified. We showed that ¹⁸F-DCFPyL PET/CT had an overall positivity rate of 81%, which increased with higher prostate specific antigen (PSA) levels, and high Gleason score and shorter PSA doubling time in prostatectomy patients. We also showed that ¹⁸F-DCFPyL PET/CT impacted management of up to 61% patients.

Innovative Use of Deep Learning in Radiology

 In the recent years deep learning has become a very powerful tool for physician researchers in a large variety of tasks such as classification and segmentation, enabling higher productivity with large data sets. Fundamental technical progress is mainly driven by computer science, often within large private companies. Physician researchers often apply proven technology and evaluate feasibility, usefulness, or performance for various clinical questions. But in addition to this, the physician perspective also provides opportunities to identify interesting technical applications at the intersection of the deep internals of technology and the direct application of known algorithms. Examples like the use of CycleGANs for data augmentation or variational autoencoders for automatic segmentation quality evaluation will be presented.

— Juergen K. Willmann Lectureship —

Better with Bubbles: New Roles for Ultrasound in Imaging and Therapy

Juergen Willmann devoted much of his professional energy, both in research and clinical practice, to the combination of ultrasound and bubbles. HIs fascination with a microbubble oscillating in a sound field has been shared by a long list of distinguished scientists, starting more than a century ago with Lord Rayleigh. That a resonating bubble is able to concentrate sound energy and re-radiate an amplified echo back to the transducer, like a ringing bell, is an appealing concept which has given rise to a new generation of blood pool contrast agents for radiology with an exceptional profile for sensitivity, safety and tolerance. More extraordinary, however, is that a bubble can do the converse: transform a burst of sound into an ultra-brief event that releases energy in such a short time at such a small point in space that instantaneous temperatures exceeding that of the sun’s surface are easy to attain. Such events can blow up whole ships, provide shrimp with means to communicate with peers, porate a cell to deliver a drug through its membrane, and open the blood brain barrier focally and reversibly under MR guidance. For imaging, new strategies that produce several thousand images per second can be used to track individual bubbles through vessels too small to resolve anatomically. By imaging the path of these tracks, rather than the vessels themselves, the resolution of the image, usually limited by the wavelength of sound to a fraction of a millimetre, can be cheated, showing organ vasculature and flow velocity at the level of tens of microns. For therapy, bubbles can both porate cell membranes and open junctions between endothelial cells, allowing delivery of drugs, genes and nanoparticles into otherwise inaccessible tissue such as the brain. New methods for the creation of bubbles within the interstitium may have particular relevance to breaking down the extracellular matrix in solid tumors, potentiating cancer therapies. The ubiquity of sound and bubbles in the natural environment has meant that many of these techniques have been pre-empted by members of the animal kingdom: examples will be shown.

Molecular Imaging of Brain Tumor Metabolism with Positron Emission Tomography

Despite an extensive history of research, the complex connection between metabolism and cell proliferation remains an exciting area of investigation. Understanding and visualizing this critical aspect of tumor biology is likely to have a major impact on our understanding of cell proliferation and cancer. Pyruvate kinase M2 (PKM2) catalyzes the final step in glycolysis, a key process of cancer metabolism. PKM2 is preferentially expressed by glioblastoma (GBM) cells with minimal expression in healthy brain, making it an important biomarker of cancer glycolytic re-programming. The talk will describe the bench-to-bedside development, validation, and translation of a novel positron emission tomography (PET) tracer to study PKM2 in GBM.

MRI of the Brain Tumor Microenvironment in the Current Clinical Setting

The brain tumor microenvironment, especially in the post-treatment setting, is complex and is the result of interactions between tumor cells, immune cells such as macrophages, microglia, and T cells, components of the extracellular matrix, and endothelial cells and blood vessels.  Given these complex interactions, it is challenging to determine effective tumor response to therapy.  The reference standard for assessing post-treatment tumor response remains gadolinium-enhanced MRI, which in itself is nonspecific and has limitations.  We have developed and implemented a pipeline that leverages quantitative perfusion imaging to improve our accuracy in assessing brain tumor response to therapy.  This talk will highlight the pipeline from inception to implementation and its ultimate integration into the clinical setting for informing real-time treatment-related decision making.  We will also briefly highlight the use of ferumoxytol-enhanced MRI as a potential alternative technique for assessing brain tumor immune response.

Brain Molecular Imaging of Dementia

There have been exciting developments in brain molecular imaging in the field of neurodegenerative cognitive disorders. In addition to widely available FDG PET in the clinic, amyloid PET and anti-amyloid treatment have been approved by the US FDA. Non-amyloid PET imaging, such as tau PET, and clinical trials of non-amyloid treatments are moving forward. New neurodegenerative cognitive disorders as well as frequent occurrence of dementia with co-pathologies have been recognized recently, and brain molecular imaging plays a major role in the characterization of such conditions. We will review and discuss the current status and future directions of brain molecular imaging in dementia.

Data Science and Machine Intelligence in Medical Imaging of Breast Cancer and COVID-19

Artificial Intelligence in medical imaging involves research in task-based discovery, predictive modeling, and robust clinical translation.  Quantitative radiomic analyses, an extension of computer-aided detection (CADe) and computer-aided diagnosis (CADx) methods, are yielding novel image-based tumor characteristics, i.e., signatures that may ultimately contribute to the design of patient-specific cancer diagnostics and treatments. Beyond human-engineered features, deep convolutional neural networks (CNN) are being investigated in the diagnosis of disease on radiography, ultrasound, and MRI.  The method of extracting characteristic radiomic features of a lesion and/or background can be referred to as “virtual biopsies”.  Various AI methods are evolving as aids to radiologists as a second reader or a concurrent reader, or as a primary autonomous reader.  This presentation will discuss the development, validation, database needs, and ultimate future implementation of AI in the clinical radiology workflow including examples from breast cancer and COVID-19.  In addition, aspects of MIDRC (midrc.org) will be discussed.

The Role of Ultrasound in Evaluation of Musculoskeletal Pathology

There are a number of common applications for the use of ultrasound in the musculoskeletal system, the most common being evaluation of the rotator cuff. Tendinosis will appear as hypoechoic enlargement of a tendon and tendon tear will appear as a well-defined hypoechoic or anechoic defect. Tendon retraction is an indicator of full-thickness tear. Ultrasound can also effectively evaluate inflammatory arthritis, but identifying joint effusion, synovitis, and erosions. The presence of flow on color Doppler imaging indicates the activity of synovitis. Another application is evaluation for ligament tear. Hypoechoic or anechoic disruption of normal ligament fibers indicates tear, and evaluation while stressing the ligament further indicates severity of ligament disruption. Ultrasound can also evaluate peripheral nerves, showing hypoechoic enlargement with entrapment neuropathies, nerve trauma, and nerve sheath tumors. Soft tissue foreign bodies are well seen with ultrasound given high resolution. One last application is evaluation for soft tissue masses, most commonly confirming the presence of a presumed benign mass, such as Baker cyst, ganglion cyst, or lipoma.

Microfluidic Label Free Sorting Technologies for Applications in Medicine: From Isolation of Extracellular Vesicles in Biological Fluids to Microswimmers

Micro- and nano-scale technologies can have a significant impact on medicine and biology in  the  areas of biofabrication, cell manipulation, diagnostics and monitoring. At the convergence of these new technologies and biology, we research for enabling solutions to real-world problems at the clinic. Emerging nano-scale and microfluidic technologies integrated with biology offer innovative possibilities for creating intelligent, mobile, medical tools and devices that could transform diagnostics and monitoring, microrobotics, tissue engineering and regenerative medicine. We will present interesting applications of microfluidic technologies in IVF, cancer detection via the isolation of extracellular vesicles and microswimmers. Some of these innovative microfluidic devices have been translated into FDA approved and CE marked products, where they have been widely used by clinics around the world serving patients.

— Gary M. Glazer Lectureship —

Towards A Deeper And More Human-centric Medicine

The use of deep neural networks to transform healthcare and medicine is in the early stages, but substantial progress has been forged for enhancing image interpretation. Much more validation is needed to provide compelling evidence for changing medical practice and giving patients more autonomy. The major challenges that lie ahead are to achieve liberation of keyboards, use of multimodal AI to set up the potential for remote monitoring (reducing need for hospitals) and the virtual health assistant, attempts to eradicate bias and reduce inequities, and use of the powerful tools to restore the patient-clinician relationship.