Minnies finalists, page 3
Image of the Year
The first finalist for Best Radiology Image is from research suggesting that gallium-68 (Ga-68) prostate-specific membrane antigen (PSMA)-11 PET imaging can provide clues on how men with prostate cancer may respond to Pluvicto treatment.
In a study published August 20 in Radiology, lead authors Phillip Kuo, MD, PhD, of the City of Hope in Duarte, CA, and Michael Morris, MD, of Memorial Sloan-Kettering Cancer Center in New York City, analyzed baseline Ga-68 PSMA-11 PET scans of 826 participants from the VISION trial, a phase-III trial conducted between June 2018 and October 2019 to determine Pluvicto’s benefit.
“The left MIP image shows the segmentation in red of all the prostate cancer throughout the body on the pre-therapy PSMA-PET scan," Kuo told AuntMinnie.com. "The right MIP image shows the segmentation of the prostate cancer metastases categorized and color-coded by organ system. The segmentation was performed preliminarily through a combination of artificial intelligence and semi-automated manual segmentation, and then a radiologist checked and adjusted as needed all segmentation for final approval."
Whole-body anterior coronal PSMA-PET maximum intensity projection images in a 63-year-old white male participant with tumors in the liver, bone, and lymph node (LN) who had an initial prostate-specific antigen level of 181.9 ng/mL and Eastern Cooperative Oncology Group performance score of 0/1. The images show all PSMA-positive (PSMA+) disease as a single whole-body volume in red (left) and segmented according to anatomic region (right; bone lesions in blue, liver lesions in green, lymph node lesions in red). Image courtesy of RSNA.
The second finalist for Best Radiology Image is from a study by a group at Columbia University that used an MRI technique called diffusion tensor imaging (DTI) to link soccer heading to a decline in the microstructure and function of the brain over a two-year period.
“There is enormous worldwide concern for brain injury in general and in the potential for soccer heading to cause long-term adverse brain effects in particular,” said senior author Michael Lipton, MD, PhD, last November at the 2023 RSNA meeting, where the research was presented.
The study included 148 young adult amateur soccer players (mean age 27, 26% women). The players were assessed for verbal learning and memory and underwent diffusion tensor imaging (DTI) at the time of enrollment and two years later.
Compared with the baseline test results, the high-heading group (over 1,500 headers in two years) demonstrated an increase of diffusivity in frontal white-matter regions and a decrease in orientation dispersion index (a measure of brain organization) in certain brain regions after two years of heading exposure. The analysis adjusted for variables including age, sex, education, and concussion history.
Researchers used diffusion tensor imaging, an MRI technique, to study the impact of soccer heading on the brain. This method tracks the microscopic movement of water molecules through brain tissue to characterize the brain's microstructure. Image and caption courtesy of the RSNA.
Scientific Paper of the Year
Generative Artificial Intelligence for Chest Radiograph Interpretation in the Emergency Department. Huang, et al, JAMA Network Open, October 5, 2023. To learn more about this paper, click here.
First up in the finalists that the Minnies expert panel picked for Scientific Paper of the Year is a study by a team at Northwestern University in Chicago that developed an AI model for interpreting chest x-rays.
The model is a transformer-based encoder-decoder that takes chest x-ray images as input and generates radiology report text as output. The model was trained on 900,000 chest x-ray reports that included radiologists’ text on findings and then tested on 500 patient images acquired in an emergency department.
“The emergency department, which often relies on teleradiology in addition to in-house radiology services to make timely care decisions, provided a natural setting to evaluate this model," first author Jonathan Huang, a fifth-year MD/PhD program student, told AuntMinnie.com. "We were able to show, for the first time, that the model performed on par with human radiologists, even outperforming teleradiologists."
Implementing the model in the clinical workflow could enable timely alerts to life-threatening pathology while aiding imaging interpretation and documentation, Huang added.
Radiologist Workforce Attrition from 2019 to 2024: A National Medicare Analysis. Rosenkrantz, et al, Radiology, July 23, 2024. To learn more about this paper, click here.
The second finalist for Scientific Paper of the Year presented findings suggesting that attrition isn’t to blame for the shortage of radiologists in the U.S.
Andrew Rosenkrantz, MD, and Ryan Cummings, MD, of NYU Langone Health in New York City, delved into Medicare billing data from 2015 to 2024 by 23,496 radiologists. They found an attrition rate among the group of 13.2%, the second lowest among the eight specialties included in the analysis.
“The workforce shortage has been of the recent top issues in radiology, with implications for burnout and care quality, but its drivers remain debated," Cummings told AuntMinnie.com. "By better understanding what is leading to the shortage, we can better develop solutions and targeted interventions for potential relief.”
Ultimately, the findings suggest that attrition – defined as switching to a different career or retiring early – is not a primary driver of the current national radiologist workforce shortage and that other potential drivers should be explored, the pair added.
Best New Radiology Device
NeuroLF dedicated brain PET imaging system, Positrigo
NeuroLF is an ultracompact brain PET system cleared for use in the U.S. in July and in Europe in early October.
An early prototype began showing promising results back in 2021, in a study of 10 patients at the University Hospital Zurich. The device is designed to help diagnose and monitor brain disorders, such as Alzheimer's disease, tumors, epilepsy, and Parkinson's disease. Significantly, the scanner allows patients to undergo a brain scan in a seated position, with participants in early studies indicating they preferred that position to lying down, which is required in a traditional PET scan, Positrigo says.
Positrigo's NeuroLF brain PET scanner. Image courtesy of Positrigo.
A clinical trial of NeuroLF is underway, which is comparing the quality of NeuroLF images to those from a full-body PET/CT scanner. First images from the trial were presented in late March at a neuroimaging conference in Germany by principal investigator Henryk Barthel, MD, a nuclear medicine physician at the University of Leipzig.
The images show noninferiority compared with conventional PET/CT devices, according to the vendor.
Aquilion One/Insight Edition CT scanner, Canon Medical Systems
The Aquilion One/Insight Edition is Canon’s new flagship CT scanner, first unveiled at RSNA 2023.
Canon's Aquilion One/Insight Edition CT scanner. Image courtesy of Canon.
The Aquilion One/Insight Edition is a premium system that pairs the company's Precise IQ Engine (PIQE) deep-learning reconstruction technology with Instinx AI-enabled workflow software – translating into more efficient exams, according to Canon. It produces super-resolution 1024 matrix images for cardiac and body exams and features a 0.24-second rotation speed, which allows for one-beat scanning in patients with high heart rates.
At ECR in 2024, Canon highlighted hardware upgrades to the scanner, including CoolNovus x-ray tube for heat dissipation and Canon's PureInsight detector for reduced noise.
Best New Radiology Software
MIM Symphony HDR Prostate software, MIM Software
Launched in July by GE HealthCare subsidiary MIM Software, Symphony HDR Prostate is designed to support the use of MRI for planning high-dose-rate (HDR) brachytherapy procedures, according to the vendor.
The software aligns contours on preoperative MRI with live ultrasound, enabling the prostate, lesions, and critical structures to be visualized during HDR brachytherapy, the company said. Clinicians can use the software to define the lesion and track changes to guide needle placement, correct differences in lithotomy orientations between MRI and ultrasound, and automatically digitize needles on CT or ultrasound planning images.
PowerScribe Smart Impression AI dictation tool, Nuance Communications
First unveiled at the 2023 RSNA meeting, PowerScribe Smart Impression utilizes generative AI technology to automatically create draft impressions and recommendations for radiology reports.
The software also offers reminders of important information to include and adapts to radiologists’ reporting style and preferences, according to the Microsoft company. Smart Impression was trained on millions of radiology reports, Nuance said.
Best New Radiology Vendor
Radpair
Radpair is a radiologist-led startup that has used natural language processing and generative AI technologies to develop a software platform aimed at streamlining the process of creating radiology reports.
The zero-footprint reporting platform automatically generates complete reports, including findings and impressions, according to the firm. It works out of the box without any need for training, but it can also be fine-tuned to suit individual users, Radpair said.
Radpair’s automatic impressions tool is designed to deliver concise summaries of key findings, according to the vendor. Image courtesy of Radpair.
Over the last year, the vendor launched version 2.0 of the Radpair platform, which incorporated new AI inference capabilities from AI technology provider Groq. Radpair 2.0 also included Wingman, a new feature that enables real-time editing and interactive support that adapts to each radiologist’s unique style, according to the firm.
In addition, Radpair formed a partnership with Radiopaedia.org to launch Pair Insights, which integrates Radiopaedia’s radiology reference library into the Radpair platform. The company said it has also integrated its platform with commercial PACS software and workflow orchestration systems.
Springbok Analytics
Our second finalist for Best New Radiology Vendor offers AI-powered image analysis of MRI exams to quantify muscle properties and provide 3D visualization of musculoskeletal data.
Springbok's interactive viewer can show, muscle symmetry, for example from analysis of MRI exams. Image courtesy of Springbok Analytics.
Its AI software produces a digital twin of a person’s musculature to enable the creation of personalized and progressive care plans. A color-coded 3D model of the individual aids in assessment of strengths and weaknesses, according to Springbok.
Recently, the company garnered U.S. Food and Drug Administration (FDA) clearance for MuscleView, its AI algorithm for analyzing muscle health from MRI scans. In 2024, the company has been involved in several clinical trials related to muscle health, ranging from sarcopenia to muscle-wasting disease facioscapulohumeral muscular dystrophy.
The vendor said it has also released new patellar and Achilles tendon analysis capability as an add-on to its muscle analysis technology. Furthermore, it has received a grant from the U.S. National Institutes of Health to support the development of its rotator cuff analysis tool for improving surgical outcomes.
Best Educational Mobile App
MRI Made Easy... well almost, BestApps (iOS)
A first-time finalist for Best Educational Mobile App, MRI Made Easy aims to serve as a “classic introduction to MR physics, reimagined for iOS.”
Although it can be read cover to cover, MRI Made Easy also features a searchable, “dynamic” index designed to serve as a reference for quickly looking up MR phenomena and underlying physics principles. What’s more, the animated app can quiz users on their physics knowledge; all questions are answered directly and with links to the full-text background information.
MRI Made Easy... well almost is an animated app designed to provide an introduction to MR physics. Image courtesy of BestApps.
MRI Made Easy and the other finalist, Radiology Assistant 2.0, are part of the doRadiology family of educational mobile apps.
Radiology Assistant 2.0, BestApps (iOS)
Radiology Assistant 2.0 features articles written by expert radiologists on a variety of categories. Image courtesy of BestApps.
Radiology Assistant 2.0 is a finalist for Best Educational Mobile App for the second year in a row and also won this award in 2018 and 2021.
Radiology Assistant 2.0 features peer-reviewed articles contributed by expert radiologists and is the result of a collaboration between Wouter Veldhuis, MD, PhD, of the University Medical Center Utrecht and Robin Smithuis, MD, of Rijnland Hospital Leiderdorp.
Multiple new articles have been added in 2023 and 2024.
