Art and the Radical Act of Diagnostic Excellence

This session makes links between core principles of diagnosis and expe- riential learning in art. Diagnosis is the core art of medicine, an intentional practice, a team sport, and particularly complex to teach and learn. It draws deeply from both scientific and humanistic knowledge, and requires skillful interactions to maximize the opportunity for findings to emerge. This session highlights specific features and challenges of diagnosis in radiology to frame an experience with art and to underscore key aspects of careful, disciplined visual inspection and communication. All participants will have the opportu- nity to experience the Visual Thinking Strategies – an approach increasingly applied and researched in the health professions – and to deconstruct that experience. The session will conclude with an approach to looking based on Miller’s research of adaptive approaches to uncertainty in clinical masters and in artists. Join us to highlight the role of observation in patient care excel- lence – no previous experience with art required.

Looking Back, Looking Ahead: Trends in Radiology

Among various disciplines of medicine, Radiology stands out as a hub for innovation in academic health systems that constantly drives changes in practice. The resulting changes have over many years of evolution spurred tremendous growth and shaped the work, workload, workflow, and workforce in Radiology. The work has become more complex, the workload has escalated to unprecedented levels, workflows require much more interaction with much larger volumes of data, and the workforce is facing cultural transformation while the workplace shifts from centralized and in person to distributed and remote. In this talk, we will examine several key trends in our clinical and research systems, discuss the emerging challenges that these trends present to academic Radiology, and outline strategies to thoughtfully and intentionally steer future changes in practice.

High Frequency Optical Coherence Tomography in Neurointerventional Radiology/Surgery

Diagnosis of cerebrovascular diseases is currently limited to CT angiography, MR angiography and Digital subtraction Angiography. Vessel wall characterization is possible with MR but is limited to bigger lesions and not possible beyond the vessel wall. Intravascular ultrasound and currently available OCT techniques and technology is limited to peripheral and coronary circulation. We will present the development, pre-clinical results of high frequency OCT specifically for the Neurovascular anatomy.

Rethinking Failure: Core Lessons for Diversity, Equity, Initiatives in Radiology

In recent years, much attention has rightly been given to the importance of diversity, equity, and inclusion (DEI) efforts in radiology, as a means to further organizational excellence, improve quality, and help address issues of social justice. But what happens when such efforts fail? How should an academic radiology department respond to these moments? How do we create and rebuild viable paths forward? In this session, we will revisit failure as a valuable resource to be exploited, through the lens of common DEI initiatives used in academic medicine departments in general and radiology departments in particular. By applying core lessons from organizational psychology, we will elucidate a set of key strategies and frameworks for recovery when we encounter the inevitable moments of failure.

Point-of-care Ultrasound and MR-based Diagnostics for Fatty Liver Disease

Fatty liver disease has become the most common cause of chronic liver disease in children and adults worldwide, affecting over one billion human beings. Fatty liver disease is associated with obesity, metabolic syndrome, alcohol consumption, and other factors, and it can progress to cirrhosis, liver cancer, and end-stage liver failure. It is also associated with and may contribute to the development of diabetes, cardiovascular disease, and cancer originating outside the liver. Early diagnosis and intervention can reverse the disease and prevent its downstream complications. The gold standard method for diagnosis of fatty liver disease is percutaneous biopsy with histology analysis. To avoid the risks and other shortcomings of invasive biopsy, noninvasive imaging-based diagnostics for fatty liver disease have been developed, chief among them ultrasound- and MR-based methods. These methods have improved considerably over recent years and are now available on clinical ultrasound scanners and MR systems. Despite these technical advances, however, clinical ultrasound scanners and MR systems do not have enough capacity to screen, diagnose, and monitor a disease that afflicts about one quarter of the earth’s population. Emerging point-of-care ultrasound and MR-based diagnostics may be part of the solution.

Quantitative Viscosity of Abscess Fluid

Image-guided percutaneous drainage of intra-abdominal abscesses is critical to modern medicine, reducing mortality, morbidity, length of hospital stay, and hospital costs. At the time of abscess drain placement, however, it is difficult to confidently predict whether a particular abscess will resolve soon after drain placement or will be challenging to drain and require multiple drain exchanges. Although the presumed or qualitative viscosity of abscess fluid regularly guides drain management, such as the choice of drain size, there are minimal published quantitative measurements of abscess fluid viscosity. Due to the challenges of quantitatively measuring abscess fluid viscosity with commercially-available methods, our research group has explored novel measurement methods with the potential to smoothly integrate into the clinical workflow. We will present our preliminary findings examining the quantitative physical properties of clinical abscess fluid samples and the possible implications of these results for both the initial selection and management of abscess drains.

Synthetic Cancer Reporters

[18F]Fluorodeoxyglucose positron emission tomography ([18F]FDG-PET) has high sensitivity for lung nodule detection but low specificity in distinguishing malignant from benign nodules, and cross-sectional imaging for routine lung cancer screening is expensive. An ideal cancer screening test would be sensitive and specific as well as inexpensive, such that only those individuals who test positive would then undergo expensive cross-sectional imaging for accurate tumor localization. Breath analysis holds enormous promise for inexpensive, rapid, and noninvasive early lung cancer detection and surveillance; however, clinical implementation has been limited by low signal from cancer cells and high background noise from nonmalignant tissue. To address this issue, we are developing a novel breath-based “synthetic” reporter system in which cancer cells are induced to express D-limonene, a volatile organic compound (VOC) from citrus fruit that is not produced in humans and thus has low background signal. In vivo gene delivery and tumor-specific expression of limonene could enable sensitive, specific, and inexpensive cancer screening via exhaled breath, but tumor localization would still be needed in the event of a positive test. This could be achieved via a dual-reporter system encoding a volatile reporter along with a PET imaging reporter, both driven by a tumor- activatable promoter and delivered in vivo within DNA minicircles (MC). PET imaging could also be performed using an emerging radiotracer, (4S)-4-(3-[18F]fluoropropyl)-L- glutamate [18F]FSPG, a glutamate analog which has shown greater sensitivity than [18F] FDG in detecting lung cancer based on preliminary results of our phase II clinical trial in patients with solitary pulmonary nodules. These strategies could permit a two-step screening process in which individuals who test positive for cancer via an inexpensive limonene breath test would then be triaged to PET imaging for tumor localization.

Real-world Utilization, Practice Patterns and Efficacy of Off-label 90Y Radioembolization in Liver Metastases

Liver-directed embolization therapy is increasingly utilized for liver tumors in patients ineligible for surgical resection or transplantation. Early and prominent use of 90Y radioembolization therapy for liver tumors include inoperable HCC and mCRC, which resulted in FDA approvals. As the liver is one of the most common sites of metastatic disease, the concept of using the tumor vasculature for anatomically targeted 90Y internal radiation holds promise for treating a broader range of tumors. Here we present the utilization and efficacy data of 90Y radioembolization from 1600 patients in a multicenter, observational, registry-based study. The results of our study demonstrate that 90Y radioembolization is frequently utilized as a palliative treatment of liver metastases of non- colorectal cancer origin. We will discuss potential candidate tumor types that have promising treatment responses to 90Y radioembolization, which may incentivize future randomized trials to expand the current indications of this treatment.

Exploring the Value of Radiology and Radiologists

Medical imaging has transformed health care, but defining and measuring value in radiology has posed a challenge. In recent years, the peer-reviewed literature has explored a range of approaches to understanding how radiologists add value to patient care and to health systems. Such work indicates that value in radiology has varying meanings from the perspectives of patients, referrers, and payors. Radiology departments and practices can pursue various strategies to measure and improve value. In addition, initiatives from Medicare and other federal agencies are seeking to measure and influence the value delivered by the specialty. Radiology practices are also facing growing demands for transparency in components of value including quality and cost. Awareness and reflection on these considerations will help radiology practices thrive in an increasingly complex healthcare framework.

Photon Counting Detector CT: What We’ve Been Waiting 50 Years For

In 2014, the first high-flux whole-body photon-counting detector (PCD) CT sys- tem capable of human imaging was installed as an investigational device at the Mayo Clinic in Rochester, MN, USA. In 2020, a single-source PCD-CT proto- type with a newer detector design was installed, and in 2021, an investigational dual-source PCD-CT was installed – a system that has now received CE and FDA clearance for clinical use. To date, our research program in PCD-CT has examined approximately 1200 patients. The clinical system, the NAETOM Al- pha from Siemens Healthcare, is inherently spectral in nature, allowing tem- porally and spatially synchronous acquisition of multi-energy data. The lack of optical reflectors between detector pixels coupled with an active-focus x-ray source yields 125-micron limiting spatial resolution without a dose penalty. Also, the dual-source geometry allows spectral cardiac imaging with 66 ms tempo- ral resolution. To date, studies covering the range of clinical applications have demonstrated measurable benefits relative to scintillating, energy-integrating detectors. In addition to the greatly improved spatial resolution and simultane- ous spectral information, benefits include the ability to reject electronic noise and the ability to enhance signal from low energy photons, which strongly in- creases the iodine signal-to-noise ratio. The ability to set energy thresholds adjacent to k-absorption edges also enables the possibility to perform k-edge spectral imaging of high atomic number materials, which may result in the use of novel CT contrast agents. Based on the accumulated evidence of clinical benefit, all major CT manufacturers have invested in research and develop- ment programs in photon counting CT detectors. This presentation will describe the state of the art in PCD-CT scanners and present numerous examples of the clinical benefit derived from this latest innovation in CT technology.

The Economics of Breast Cancer Screening

The Economics of Breast Cancer Screening Chair, Department of Radiology Associate Medical Director for the Medical Specialties Ochsner Medical Center - New Orleans Friday, March 10, 2023 12:00pm - 1:00pm LKSC 101/102 and via Zoom Healthcare payment policy in the United States is complicated, and nowhere is this more apparent than for screening services. When Congress created Medicare under Title XVIII of the 1965 Social Security Act, screening services were not contemplated, and only diagnostic services and treatments for illnesses were included. Coverage differs among payers, and the process to obtain reimbursement for new services and procedures is also variable. This lecture will review the current state of breast imaging reimbursement, including breast cancer screening. We will review the impact of government payers on breast imaging reimbursement and opportunities for advocacy. Because payment policy varies from payer to payer and can change over time, physicians and patients are strongly encouraged to develop a strong understanding of coverage by different Global Learning Objectives insurers and out of pocket responsibility.

Artificial Intelligence Methods Bridge the Gap between Radiology and Pathology To Improve the Radiology Image Interpretation

Artificial Intelligence methods, specifically Machine Learning, have been extensively used in medical imaging for a variety of tasks, e.g., segmenting organs, detecting and localizing diseases, and many more. There are many challenges associated with training such methods, associated with (1) availability of medical data, (2) availability of accurate labels, and (3) generalization to other cohorts beyond the training data. In this lecture, I will present my team’s efforts to address these challenges by bridging the gap between radiology and pathology images. We develop artificial intelligence methods focused on bringing information from pathology into radiology images, either as labels (through multi-modal registration) or radiology imaging signatures (through multi-modal correlation learning). Along with the data, the labels and imaging signatures are used to train models to detect cancer and distinguish aggressive from indolent cancers in different types of radiology images. Most of the talk will be focused on our studies in prostate cancer using MRI or b-mode ultrasound images, with some examples on renal and breast cancers.