UVA Neurorad Jason Druzgal, MD, PhD on how physicians decide to work for insurance companies.

Imaging Wire Sponsors

annalise.ai  •  Arterys  •  Bayer  •  Blackford  •  Canon Medical  •  CARPL.ai  •  Change Healthcare  •  Enlitic  •  GE HealthCare  •  Hyperfine  •  Intelerad  •  Merge by Merative  •  Novarad  •  Nuance  •  Riverain Technologies  •  Siemens Healthineers  •  Subtle Medical  •  United Imaging  •  Us2.ai

Artificial Intelligence

Radiology NLP’s Efficiency and Accuracy Potential

The last week brought two high profile studies underscoring radiology NLP’s potential to improve efficiency and accuracy, showing how the language-based technology can give radiologists a reporting head-start and allow them to enjoy the benefits of AI detection without the disruptions.

AI + NLP for Nodule QA – A new JACR study detailed how Yale New Haven Hospital combined AI and NLP to catch and report more incidental lung nodules in emergency CT scans, without impacting in-shift radiologists. The quality assurance program used a CT AI algorithm to detect suspicious nodules and an NLP tool to analyze radiology reports, flagging only the cases that AI marked as suspicious but the NLP tool marked as negative.

• The AI/NLP program processed 19.2k CT exams over an 8-month period, flagging just 50 cases (0.26%) for a second review.
• Those flagged cases led to 34 reporting changes and 20 patients receiving follow-up imaging recommendations. 
• Just as notably, this semi-autonomous process helped rads avoid “thousands of unnecessary notifications” for non-emergent nodules.

NLP Auto-Captions – JAMA highlighted an NLP model that automatically generates free-text captions describing CXR images, streamlining the radiology report writing process. A Shanghai-based team trained the model using 74k unstructured CXR reports labeled for 23 different abnormalities, and tested with 5,091 external CXRs alongside two other caption-generating models.

• The NLP captions reduced radiology residents’ reporting times compared to when they used a normal captioning template or a rule-based captioning model (283 vs. 347 & 296 seconds), especially with abnormal exams (456 vs. 631 & 531 seconds). 
• The NLP-generated captions also proved to be most similar to radiologists’ final reports (mean BLEU scores: 0.69 vs. 0.37 & 0.57; on 0-1 scale).

The Takeaway

These are far from the first radiology NLP studies, but the fact that these implementations improved efficiency (without sacrificing accuracy) or improved accuracy (without sacrificing efficiency) deserves extra attention at a time when trade-offs are often expected. Also, considering that everyone just spent the last month marveling at what ChatGPT can do, it might be a safe bet that even more impressive language and text-based radiology solutions are on the way.

Building a Mobile Lung Cancer CT Screening Program

The number of patients eligible for low-dose CT lung cancer screening has expanded, and so has the need to reach at-risk patients closer to where they live. That’s why Siemens Healthineers’ Mobile Lung Screening Solution combines the quality, ease of use, and flexibility needed to create a program that meets the real-life needs of your community.

COPC Ditches the Disk with Novarad

See how Novarad’s CryptoChart solution allowed Central Ohio Primary Care (COPC, 70 practices, 400 physicians) to make the transition to digital imaging sharing in this Healthcare IT News case study.

The Wire

• Lantheus Acquires Cerveau: Lantheus Medical Imaging expanded its radiopharmaceutical pipeline into Alzheimer’s disease with its acquisition of Cerveau Technologies. Cerveau’s flagship MK-6240 F18-labeled PET agent is designed to target Tau protein tangles in patients with known or suspected Alzheimer’s Disease, and has the potential to be used for Alzheimer’s diagnosis, staging, treatment selection, and pharmaceutical trials.

• MCB Radiology Selects HAP: McClow, Clark, and Berk, PA Radiology Services (MCB Radiology) selected Healthcare Administrative Partners to be its full-service radiology revenue cycle management provider, including billing, coding, carrier credentialing, business intelligence, and MIPS Measure Assurance Services. The well-established Jacksonville, Florida area practice (23 radiologists, serves four hospitals and 10 imaging centers, in business over 70-years) highlighted HAP’s broad capabilities, radiology knowledge, and level of service in the announcement.

• CXR AI’s Real-World Impact: Radiology Journal just published what might be the first prospective randomized controlled trial for a CAD AI solution, finding that Lunit INSIGHT CXR improves radiologists’ lung nodule detection in real-world and workflow-integrated screening settings. Radiologists with and without AI analyzed 10.5k patients’ screening CXRs, finding that AI assistance helped rads detect more actionable nodules (0.59% vs. 0.25%; odds ratio: 2.4) and malignant nodules (0.15% vs. 0.0%), while producing similar false-referral (45.9% vs. 56.0%) and positive-report rates (2.3% vs. 1.9%).

• January Jobs Report: The healthcare job market still appears to be firing on all cylinders, with the latest US employment data showing that the industry added over 58k jobs in January. The gains were led by ambulatory healthcare services at 30k jobs, followed by nursing / residential care facilities (17k each), and hospitals (11k). The job growth represented an acceleration from last year’s average of 47k jobs per month during the pandemic recovery.

• Avicenna.AI Adds $7.5M: Avicenna.AI closed a $7.5M Series A round (total funding now nearly $10M) that it will use to scale its global presence and diversify its product portfolio. Avicenna.ai has big plans for 2023, forecasting deployments at 30 new sites every month of the year, and adding to its growing list of four FDA clearances and six CE marks. 

• DDR Rotator Cuff Exams: Emory University researchers highlighted dynamic digital radiography’s (DDR) potential to help diagnose rotator cuff injuries and monitor patients’ treatment response. Among 121 shoulders (w/ 40 controls), DDR showed that scapulohumeral rhythm was significantly lower during shoulder abduction with patients who experienced massive rotator cuff tears (44%), adhesive capsulitis (-54%), and glenohumeral osteoarthritis (32%). 

• DiA’s LVivo IQS Clearance: DiA Imaging Analysis announced the FDA clearance of its LVivo IQS (Image Quality Score) AI-based ultrasound image quality guidance solution, expanding the echo AI company’s portfolio beyond analysis and reporting. The vendor-neutral software uses a color and numerical scoring system to provide clinicians with real-time feedback during left ventricle exams, with the goal of improving both image and interpretation quality.

• HHS Health Coverage Boost: HHS data showed that the percentage of Americans without health coverage decreased to 10.5% in 2021 (down from 11.1% in 2019), due in part to changes in federal and state coverage policies for Medicaid and the Marketplace. The largest improvements were seen among young adults (percentage without coverage dropped to 15% from 16%), Latinos (down to 19.1% from 20.2%), American Indian/Alaska Natives (down to 21.5% from 22.4%), and non-English speakers (down to 27.3% from 28.8%).

• Breast Screening Education Impact: A study led by DenseBreast-info found that web-based education can reduce gaps in providers’ knowledge about breast cancer risk models. Researchers tested 177 women’s health clinicians before and after receiving online breast cancer risk education, finding that while some knowledge gaps persisted, online education resolved other gaps. For example, the ratio of clinicians who knew whether the Gail model should be used to identify women who meet MRI screening guidelines improved from 48.6% to 66.1%.

• Imaging’s Transparency Challenge: Transparency mandates might be among the biggest recent developments in U.S. healthcare, but complex billing practices for non-emergency outpatient imaging might keep radiology price transparency initiatives from achieving their goals. Emory University analysis of 5.2M outpatient imaging encounters revealed that contrast agents were usually billed separately from imaging procedures at hospitals, offices, and imaging centers (55.9%, 71.5%, 55.3%), while patients who had hospital-based outpatient exams often received bills from multiples entities (70.9%, 4.5%, 7.6%).

• Cerebriu’s Series A: Danish MRI AI startup Cerebriu launched a €10M Series A round (€5.5M already raised) to fund its Apollo brain MRI triage solution’s upcoming clinical trials, commercialization, and FDA clearance efforts. The CE-marked solution performs real time brain MRI triage and automated protocol optimization, ensuring proper image acquisition and improving patient prioritization.

Deidentifying and Anonymizing Healthcare Data

Proper patient data anonymization and deidentification is a must, but it can be challenging to do while still retaining clinical relevance. See why Enlitic proposes an AI-based approach to deidentify and anonymize healthcare data (both pixel data and metadata), and how it would be valuable to your organization.

Unlocking the Value of AI

Despite significant interest, there’s still confusion about the value of imaging AI. This Blackford Analysis white paper explores the key cost considerations and ROI factors that radiology groups can use to figure out how to make AI valuable for them.

The Resource Wire

• An NIH paper detailing how portable MRI can aid in stroke assessments and democratize imaging access sparked this spirited conversation with Hyperfine’s leadership, providing two insiders’ perspectives on how portable MRI can achieve these improvements – and more.

• When one of Precision Imaging Centers’ 3T MRIs wasn’t meeting their requirements, they implemented Subtle Medical’s SubtleMR solution, rather than purchasing a new scanner or an expensive upgrade. See how SubtleMR enhanced Precision’s patient throughput and comfort, without compromising image quality in this case study.

• See how Thomas Jefferson University relied on CARPL.ai to accelerate its AI validation and clinical adoption in this presentation by informatics and AI leader, Dr. Paras Lakhani.

• See Dr. Brian Goldner of UC Davis Sacramento detail his experience with Canon’s Ultra High Resolution CT and how it can be applied to cardiothoracic interpretations.

• HealthPartners understands “the domino effect of imaging,” so they worked with Merative to move to enterprise imaging that’s specifically built for high availability. See how the major Minnesota-based health care system decided to make this change, the steps they took to make sure they implemented it the right way, and the impact this transition had on their imaging operations.

• Curious how certain your AI is about its own finding? annalise.ai’s confidence bar displays the likelihood of each finding and the AI model’s level of certainty, helping clinicians perform their interpretations with greater confidence.

• See how Valley Radiology’s decision to make Intelerad IntelePACS its single reading environment helped the independent practice gain control of its growing volumes and rising case complexity, improve its efficiency and radiologist experience, and deliver better patient care.

• This Riverain Technologies case study details how Einstein Medical Center adopted ClearRead CT enterprise-wide (all 13 CT scanners) and how the solution allowed Einstein radiologists to identify small nodules faster and more reliably.

• We may be on the verge of the next revolution in how cancer is detected, diagnosed, and treated and this editorial by GE HealthCare Imaging leader, Jan Makela, details how GE and its partners are using AI and imaging data to drive this revolution.

• Is your organization ready to move enterprise imaging to the cloud? Check out this Change Healthcare and ACHE webinar detailing cloud-native imaging’s benefits, best practices, and ROI.

• Imaging AI’s clinical and productivity benefits are becoming increasingly clear, but selecting and implementing the right solution can be difficult. This Arterys paper details how an AI platform strategy allows providers to efficiently and accurately evaluate AI applications, so they can start realizing their targeted AI benefits.

• What if AI could produce echo measurements that are comparable to expert physicians, but with less variability? That’s exactly what this Nature study revealed about Us2.ai’s solution, finding that its measurements had fewer and smaller differences compared to three human experts than when the experts were compared with each other.

• AI automates what radiologists can’t stand, surfaces what radiologists can’t see, and identifies what radiologists can’t miss. But only if it’s implemented in the way radiologists work. See how Nuance helps radiologists achieve these results through a single, streamlined, end-to-end AI experience.

• See how United Imaging’s new uCT ATLAS combines advancements in image quality, patient comfort, and operator efficiency to bring you one step closer to your masterpiece.

• Radiology is leading healthcare’s AI revolution, and yet many people in radiology are just starting to build their understanding of AI. That’s why Bayer published its truly Complete Guide to Artificial Intelligence in Radiology, detailing how AI can address radiology’s challenges, AI’s core use cases, and AI’s path towards adoption.