|
Non-Doctors Read More Images | FDA Eyes Gen AI
November 21, 2024
|
|
|
|
|
Together with
|
|
|
|
“As I talk with clinicians across the country, I’m hearing increasing concern that the criteria for adopting an AI are almost purely financial. I hope this committee will help us figure out how to balance patient outcome benefit with financial return.”
|
|
FDA Commissioner Robert Califf, MD, speaking at this week’s meeting of the FDA Digital Health Advisory Committee.
|
|
|
AI for radiology is evolving, from point sources that analyze medical images for a single pathology to solutions that are able to detect a wide variety of health conditions. We talked about this evolution with Dimitry Tran of Annalise.ai in this edition of The Imaging Wire Show.
|
|
|
Image courtesy of Harvey L. Neiman Health Policy Institute. |
|
|
Non-physician practitioners are reading more medical images in U.S. medical offices. That’s according to a new study in JACR by researchers who found that the share of images interpreted by NPPs has doubled in the last 10 years.
In the U.S., radiologists consider themselves to be the primary interpreters of medical images, but inroads have been made not only by other physicians but also by non-physician practitioners like nurses and physician assistants.
- NPPs are supposed to receive specialized training in image interpretation, but radiologists question whether such training is adequate, especially compared to the years of training that radiologists receive.
Previous research has documented the rise in NPP image interpretation, but the new study takes a longer view, examining the period 2013-2022.
- It also specifically focuses on the medical office setting, where it’s believed NPP interpretation is growing faster than in hospitals, where radiologists still dominate interpretations.
In their analysis of Medicare claims, researchers from the ACR’s Harvey L. Neiman Health Policy Institute found …
- NPPs’ share of office-based image interpretations grew 9% annually (2.5% to 5.5%).
- Growth rates varied by modality, with MRI growing at 9.9% annually, followed by CT (9.4%), ultrasound (9.4%), radiography (8.9%), and nuclear medicine (7.2%).
- Despite the growth, just 5.6% of NPPs were interpreting images.
- By specialty, the share of NPP interpretation was most common with primary care (40%) and orthopedic offices (34%).
The researchers also tracked variability in NPP interpretation rates by state, finding the highest rates ( ~13%) in Western states with large rural areas like Montana, Alaska, and Idaho, where presumably there are fewer radiologists available to read images.
The Takeaway
The findings provide a good news/bad news look at non-physician image interpretation. The good news for radiologists is that NPP interpretation is still pretty rare; the bad news is that rates are growing quickly. And given the ongoing radiologist shortage, there is sure to be continuing pressure to allow allied health staff to read images on their own.
|
|
|
Book an RSNA Meeting with CARPL
Explore radiology AI like never before by booking a meeting with CARPL at RSNA 2024, or drop by booth #5733 to tune in to any of seven research presentations across the show.
|
|
Breakthroughs in Healthcare. For Everyone.
Join Siemens Healthineers at RSNA 2024 booth #2529 to experience the future of healthcare. Explore cutting-edge innovations, connect with leading experts, and collaborate with industry professionals in live demonstrations.
|
|
Fully Automated AI Echo vs. 3D and Human Readers
While 3D echo is becoming more accurate, 2D still dominates clinical care. A new study evaluates agreement in measures of LV volume and function between human readers, echo AI from Us2.ai, and the 3D Heart Model.
|
|
- FDA Eyes Generative AI Regulation: The FDA on Wednesday opened the first meeting of its new Digital Health Advisory Committee, formed to guide the agency’s regulation of digital technologies in medicine. This inaugural meeting is focusing on generative AI – the FDA has not yet granted marketing authorization for a generative AI product, and the outcome of the discussion could guide its future decisions. The two-day meeting continues at 9 am ET on Thursday and interested parties can follow along on a livestream on YouTube.
- Intelerad Enhances AI Integration: Intelerad is making it easier for its customers to integrate AI algorithms into their software with the debut of InteleGence. The solution is an integration platform with an open API that lets users integrate their AI algorithms into the company’s InteleOrchestrator and InteleWorkflow applications, and users can integrate with the AI marketplaces of their choice. Intelerad is also launching AI Lab for customers to share their experiences, and a curated AI partner ecosystem. Intelerad will maintain its relationship with AI platform company Blackford.
- What If AI Advice Is Wrong? What happens if AI gives radiologists bad advice? That question is explored in a new paper in Radiology in which 220 radiologists and other physicians read chest X-rays alongside AI advice. Physicians often aligned their diagnostic decisions with AI guidance, and when AI was correct they had high accuracy for both local and global AI explanations (93% and 85%). But accuracy plummeted if AI was incorrect (24% and 26%). The findings align with recent studies on AI’s impact on physician behavior.
- AI Detects Missed Breast Cancers: ScreenPoint Medical’s Transpara AI algorithm was able to detect breast cancers that radiologists missed on screening mammograms in a study in Journal of Breast Imaging. In 30k mammograms, AI detected 54% of 13 false-negative cancers on 2D mammograms and 40% of 10 missed cancers on DBT exams. Transpara detected false-negative cancers much sooner than radiologists for both 2D and DBT exams (272 and 248 days sooner). All the cancers detected by Transpara were invasive and 82% were luminal A, the most common subtype.
- AI Supports Mammography Screening: Another new study on AI for mammography – this one in Radiology – also supports deep learning for breast screening. Researchers used three commercially available AI algorithms (they didn’t say which ones) for breast screening in the U.K. and compared them to first-reader radiologist ratings for sensitivity (63%) and specificity (97%). They found that two algorithms were not inferior to radiologists in sensitivity (59% to 65%) while specificity was higher for one (98%) and close for another (97%). The findings echo other research on mammography AI.
- CAC Scores Predict Mortality: AHA 2024 wrapped up this week, and a study presented at the meeting confirms using coronary artery calcium scores for predicting mortality. Intermountain Health researchers derived CAC scores from PET/CT scans of 40k people, finding after two years of follow-up women with CAC scores of 0 had lower rates of all-cause death or non-fatal myocardial infarction compared to those with CAC>0 (3.3% vs. 9.5%) while rates for men were also higher for those with elevated CAC (3.3% vs. 10.2%).
- CCTA Diabetes Screening Falls Short: In another study from AHA 2024, researchers found that using coronary CT angiography to screen diabetes patients for heart problems may not be that effective. Researchers used CCTA to screen 452 patients from the FACTOR-4 study of people with diabetes, finding no statistically significant difference in all-cause mortality or non-fatal heart incidents compared to those who didn’t get screened after up to 12 years of follow-up. The findings show the importance of managing diabetes medically rather than relying on CT screening exams.
- AI-ECHO Shows Promise: Notching a big win at AHA 2024 for automated echocardiography, the AI-ECHO RCT trial found that AI helps improve echo exam efficiency, quality, and operator fatigue. Using Us2.ai’s software and traditional methods on alternating days, the AI led to improvements in exam times (14.3 vs. 13.0 minutes) and exams per day (14.1 vs. 16.7), while increasing the share of “excellent” quality exams (31% to 41%) and decreasing sonographer mental fatigue (from 4.7 to 4.1 points out of 5).
- AI Analyzes Multiple Echo Views: A new echocardiography AI algorithm developed at Yale University that analyzes multiple echo views turned in good performance in another talk at AHA 2024. Yale researchers reported results using the PanEcho algorithm independently for detecting all major findings from echo studies. The algorithm had AUC scores above 0.91 for multiple classification tasks, ranging from assessing ventricle function to identifying mitral regurgitation. PanEcho was developed by Yale’s Cardiovascular Data Science (CarDS) Lab.
- FDA Clears Echo AI for Amyloidosis: The FDA cleared Ultromics’ EchoGo Amyloidosis software for detecting cardiac amyloidosis from echo AI scans. The application is part of the company’s EchoGo platform and complements its EchoGo Heart Failure solution. Cardiac amyloidosis is a type of infiltrative cardiomyopathy that can cause heart failure and is caused by the buildup of abnormal proteins. Ultromics notes the application is the first of 15 breakthrough cardiac devices in the FDA’s Total Product Lifecycle Advisory Program to get marketing authorization.
- TeraRecon Gets Cardiac MRI Clearance: TeraRecon received FDA 510(k) clearance for auto-contouring on cardiac MRI exams. The company believes the clearance will help improve workflows in cardiac MRI, which historically has been more challenging to perform than other cardiac imaging modalities. The solution helps users achieve more efficient and precise results, and it’s another sign of TeraRecon’s new push into cardiac imaging, such as with the introduction of new cardiac MRI modules in its Intuition 4.10 release.
- 4DMedical Signs UC San Diego: 4DMedical signed a contract with UC San Diego that will see UCSD Health adopt the company’s XV technology for respiratory imaging. In the initial phase, UCSD Health will deploy 4DMedical’s CT-based imaging solutions, and in later phases could install the company’s XV Scanner, which uses fluoroscopy-based imaging to provide 4D color-coded images of lung function using the company’s Lung Ventilation Analysis Software (XV LVAS). The agreement continues 4DMedical’s recent string of success with academic institutions, such as Vanderbilt University.
- Paragon Gets a Dell Deal: Paragon Health IT landed a big fish for its Strings radiology analytics software: computer giant Dell Technologies, which will offer the solution as part of its Medical Imaging as a Service offering. Dell will integrate Strings into its own software to give users insights into performance metrics like applications, network performance, and infrastructure components to identify bottlenecks and performance issues before they become critical. Paragon launched Strings in 2023 as part of the company’s pivot from consulting to healthcare IT.
- Akumin Allies with United: Imaging center operator Akumin signed a partnership with United Imaging to outfit what the company is calling “its most advanced facility to date,” in Plantation, Florida. United Imaging will supply its uMR Omega 3T MRI scanner and uMI Panvivo PET/CT system to the site, which will serve as a clinical center of excellence for Akumin. The facility will also be a showcase of Akumin’s Radiology Command Center for hospitals and healthcare providers.
- Flexible X-Ray Firm Funded: A U.K. startup company that’s developing flexible X-ray detectors landed £4M ($5.1M) in seed funding to further advance its technology. Silveray was spun out of the University of Surrey to commercialize methods for creating a sensitive, flexible material that converts X-rays into digital images at the point of use. Silveray is targeting a 2025 product launch for industrial imaging applications, with clinical products such as a flexible mammography detector to follow.
- EchoLight Signs with Siemens: Italian bone densitometry developer EchoLight signed an agreement to supply its technology to Siemens Healthineers. Siemens will sell EchoLight bone densitometers, which calculate density measurements based on ultrasound scans of the spine and femur. The company believes its radiation-free REMS technology has advantages over traditional DEXA for monitoring bone health over time, and Siemens will offer the technology as part of its women’s health business.
- ‘AI Radiology Resident’ Firm Raises $19M: It’s long been the dream of radiologists and AI developers alike – an AI algorithm smart enough to assist radiologists like radiology residents do. A California company called New Lantern emerged from stealth to realize that dream, raising $19M to commercialize a platform that automates 25% of radiology functions so far, such as key imaging and reporting workflows. The company plans to invest the funds in continued R&D and grow its engineering and sales teams.
|
|
Leading Radiology Forward at RSNA 2024
DeepHealth is leading radiology forward through AI-powered informatics and smart technology. Book a meeting today at RSNA 2024 booth #1340 to learn how they can help you achieve increased staff productivity and a better experience for patients and staff.
|
|
Your Trusted Enterprise Imaging Cloud Partner
Visit Optum at RSNA 2024 booth #6750 to learn how their cloud-based enterprise imaging solutions can enhance the potential of your imaging data for efficient, cost-effective care.
|
|
Imaging Workflows that Actually Work
Not a fan of medical image exchange on discs? Then check out Clearpath and find out how it’s removing obstacles to better radiology workflow. Request a demo today.
|
|
- Meet Merge at RSNA 2024: Merge is looking forward to sharing a wide range of news, innovations, and updates at RSNA 2024. Join them at booth #6100 to explore how their imaging solutions put your workspaces, clinicians, and transformation into focus.
- Connect with Intelerad at RSNA 2024: Connect with Intelerad at RSNA 2024 to see how they can empower your radiology team to deliver faster, more accurate diagnostics. Visit them at booth #6113 to see how they’re solving everyday radiology workflow challenges with scalable, cloud-based solutions.
- Discover the Power of an AI Ecosystem: Get the insights you need to fuel your AI strategy in the Bayer Visionary Theater at RSNA 2024. Preview the theater sessions today and plan to attend at booth #2514 – and don’t forget to experience the Tunnel of Innovation!
- Meet with Medality at RSNA 2024: Book a visit with Medality at RSNA 2024 or swing by their booth #3140 to see how they are at the forefront of practice growth and training in radiology.
- Learn about Gleamer at RSNA 2024: Visit Gleamer at booth #5144 at RSNA 2024 to learn about their novel AI solutions for musculoskeletal imaging, including measuring Cobb angles on X-ray and the beneficial impact of BoneView for fracture detection. Schedule a meeting today.
- 2 Questions about AI for Radiology Leaders: Are today’s radiology AI solutions solving the right problems? And are there other solutions available for AI of brain MRI? Read this article from SpinTech MRI to learn how their STAGE solution can optimize MRI utilization.
- Deliver Care without Gaps – Join PocketHealth at RSNA: Discover how PocketHealth closes critical gaps in patient care to modernize the healthcare experience. Stop by for a demo at booth #2601, South Hall.
- Book an RSNA Meeting with Riverain: Riverain’s ClearRead technology with Clear Visual Intelligence can improve your ability to detect lung cancer with either CT or X-ray. Book at meeting with them at RSNA 2024 or drop by for a visit at booth #4523.
- Connecting Patients to Exceptional Care: Mach7 Technologies’ enterprise imaging solutions connect patients and clinicians across the globe to enable exceptional care. As modern healthcare evolves, so should your strategies. Schedule a demo at RSNA 2024 or stop by booth #3953 to learn more.
- Come See Enlitic at RSNA 2024: Visit Enlitic at RSNA 2024 at booth #4365 to learn how the company is revolutionizing radiology with its data standardization solutions, including data migration technology through its acquisition of Laitek. Book a demo today.
- The Road to Cloud-Based PACS: Radiology facilities are turning to cloud-based PACS like Visage’s Visage 7 to solve their medical image management needs. Learn about their experiences in this Imaging Wire Show with Amy Thompson of Signify Research and radiologist Marc Kohli, MD.
- Connect with United Imaging at RSNA 2024: United Imaging will be celebrating the theme of Building Intelligent Connections at RSNA 2024. Come visit the company at booth #1929 to learn about their imaging solutions and how they connect to United’s mission of Equal Healthcare for All.
- AI-Driven Enterprising Imaging: Learn how to unlock the full potential of AI-driven enterprise imaging in this Lunch & Learn session with AGFA HealthCare at 12 pm on December 4 at RSNA 2024. Save your spot and RSVP today!
- Image Sharing to Improve Patient Outcomes: In this case study, learn how AdventHealth improved treatment planning and delivery because of Microsoft’s PowerShare’s immediate connection to other sites in the network for medical image sharing.
- TeraRecon Doc Talks at RSNA 2024: Visit the TeraRecon booth #8134 at RSNA 2024 to hear captivating presentations from radiology thought leaders every day of RSNA week. Schedule a meeting today.
|
|
|
|
|
