Johns Hopkins Medicine’s Susan Peterson, MD, calling for more research funding focused on improving diagnostic accuracy.

Imaging Wire Sponsors

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

Happy to share a first-of-its-kind Imaging Wire Show with Intelerad leaders, Morris Panner and A.J. Watson, performed on-site at RSNA 2022. We discuss Intelerad’s latest initiatives and acquisitions, its expanding cloud focus, and its company strategy heading into 2023, making this a must-watch episode if you’re involved with Intelerad or working on your own enterprise imaging strategy.

Emergency Imaging

Spotlight on ED Diagnostics

A new U.S. federal government study made emergency department diagnostic accuracy a mainstream news story, showing that although ED diagnostic errors are somewhat rare, they occur in high volumes and can carry serious consequences. 

The U.S. Agency for Healthcare Research and Quality and Johns Hopkins University teamed up to analyzed 279 international studies published between 2001 and 2021, finding that:

• Diagnostic errors occur in an estimated 5.7% of ED visits
• Generalized to the U.S., ED diagnostic errors impact 7.4M patients annually
• Those diagnostic errors lead to “preventable harms” in roughly 2.6M patients, and “serious harms” in 371k patients, including 250k deaths
• The top 5 and 15 diseases account for 39% and 68% of “serious misdiagnosis-related harms” 

Although “not all diagnostic errors are preventable,” error rate variations revealed key areas for improvement:  

• Women and people of color were 20% to 30% more likely to be misdiagnosed
• Misdiagnosis is far more common among patients with “atypical” and “subtle” disease presentation
• Hospital and disease-specific error rates varied widely

Imaging played a major role in the study, as most of the top-15 diseases associated with “serious misdiagnosis-related harms” are typically diagnosed with imaging exams (including all of the top-5), and the report mentioned “radiology,” “imaging,” “image,” “scan,” or “ultrasound” a whopping 419 times.

Emergency medicine societies objected to these results, but the consensus among study authors and most observers was that more efforts are needed to understand and address ED diagnostic errors, with a specific focus on the diseases associated with serious misdiagnosis harms.

The Takeaway

Most efforts to improve ED safety over the last 20 years have targeted glaring mistakes (e.g. wrong medications, ED-acquired infections), but this report clearly calls for increased focus on improving EDs’ diagnostic accuracy. 

Those efforts would start at the bedside, but they would definitely involve medical imaging (and potentially error-catching AI tools), especially considering that most of the diseases associated with “serious misdiagnosis-related harms” are diagnosed via imaging.

Solving the AI Adoption Challenge

Imaging AI is evolving fast, but radiology leaders’ expectations for their AI technologies might be evolving even faster. In this Imaging Wire Show with Dr. Charlene Liew of SingHealth and Dr. Nina Kottler of Radiology Partners, we explore radiology leaders’ current and future expectations for AI, and the central role platforms play in their AI roadmaps.

Ready to make MRI more accessible?

See how Siemens Healthineers’ MAGNETOM Free.Max expands MR imaging to more patients, sites, and providers.

The Wire

• PaxeraHealth Series A: PaxeraHealth announced the completion of its Series A funding round (undisclosed value), revealing plans to use the capital to expand its enterprise imaging and AI authoring capabilities and support future commercial growth. It’s rare to see a Series A round from a company this established (founded in 2009, clients in 45 countries). However, this move seems to be inspired by PaxeraHealth’s new “Ark” zero-coding AI authoring platform, which might require outside funding to achieve its full potential.

• Interpretability-Based CBIR: There’s plenty of evidence that content-based image retrieval systems (CBIRs) improve the diagnostic process, but a new Nature study found that CBIRs might be more effective if they retrieved images based on interpretable AI saliency map similarities, rather than overall image similarities. Researchers evaluated four different CBIR approaches to retrieve images with two common CXR abnormalities (pleural effusion & pneumonia), finding that the interpretability/saliency-based approach was most likely to match experienced radiologists.

• MedCognetics FDA: The mammography AI segment welcomed a new player, following the FDA clearance of MedCognetics’ QmTRIAGE solution, which analyzes 2D digital mammograms (not 3D DBT exams) and flags suspicious cases. That’s a typical mammo AI workflow, but MedCognetics suggests that QmTriage is uniquely “unbiased” due to its diverse training data supplied through the company’s alliances with UT Southwestern Medical Center and University of Texas at Dallas. With its first FDA clearance now complete, MedCognetics plans to expand to other areas of cancer detection.

• Standardized Reporting Drives IAM Follow-Ups: Authors of a new JACR study developed a standardized reporting template for incidental adrenal masses (IAMs) that improved IAM reporting, communications, and follow-up evaluations. Among 200 imaging exams with IAMs (54 using the template), the standardized reports led to more PCP follow-ups (53.7% vs. 36.3%), PCP-ordered biochemical tests (35.2% vs. 18.5%), follow-up imaging exams (40.7% vs. 23.3%), and specialist referrals (22.2% vs. 4.8%).

• Ascelia Orviglance’s Post-Meal Enhancement: The latest results from Ascelia Pharma’s Orviglance Phase 3 trial showed that the oral MRI contrast agent produces strong liver imaging enhancement in patients who consumed a light meal, achieving similar enhancement as fasting patients and greater enhancement than patients who consumed a full meal. Ascelia positions its investigative manganese-based MRI contrast as a potential alternative to gadolinium contrast agents to detect liver cancer in patients with reduced kidney function.

• The Cancer Detection Gap: Just 14.1% of U.S. cancer cases were detected through screening in 2017, while most other cases were detected after symptoms emerged or incidentally. A new NORC study found a wide range of screening detection rates for typically-screening cancers (breast = 61%; cervical = 52%; colorectal = 45%; prostate = 77%; lung = 3%), while cancers without recommended screening tests were responsible for 57% of cancer diagnoses and 70% deaths.

• Elucid’s Histological Plaque Evidence: A Elucid-produced study provided solid evidence supporting its histological-based approach to CTA AI atherosclerosis risk assessments. The researchers used the Elucid AI system to analyze 408 CTA vessel cross-sections from 23 patients, demonstrating “excellent agreement” with pathologist-defined histological plaque risk phenotypes (weighted kappa of 0.82), and accurately identifying plaques as unstable, stable, and minimal disease (AUCs: 0.97, 0.95¸ 0.99).

• Predictive Pretreatment 68Ga-PSMA PET: A study in Frontiers in Oncology revealed that radiomics features extracted from pretreatment 68Ga-PSMA PET might help predict the effectiveness of 177Lu-PSMA therapy for patients with metastatic castration-resistant prostate cancer (mCRPC). In the 33-patient study, certain radiomics features were positively correlated (GLCM entropy) and negatively correlated (GLZLM LZLGE) with biochemical response to 177Lu-PSMA therapy, providing strong enough evidence to warrant further research.

• Rezolut’s NY Expansion: Imaging center company Rezolut continued its national expansion, acquiring CNY Diagnostic Imaging Associates (3 Central New York imaging centers, 4 rads). Rezolut has taken advantage of its PE funding, acquiring 11 imaging center companies in the last three years that expanded its presence to roughly 40 locations in six states (NY, NJ, PA, CA, AZ, NM).

• Bedside CXR AI: A new Radiology study detailed an AI model trained using structured bedside CXR reports that improved non-radiologist physicians’ interpretations and rivaled radiologists’ interpretations. A German research team trained, validated, and tested the AI model using 193k CXR reports that featured semi-quantitative disease severity assessments. In testing, the AI model exhibited higher agreements with six expert rads’ majority vote (κ = 0.86) than any of the individual radiologists on their own (κ = 0.81 to ≤0.84), while considerably improving non-radiologists’ preliminary readings (k = 0.87 vs. 0.79 without AI).

• Brainomix to Distribute Pixyl Neuro.MS: Brainomix announced a deal to distribute Pixyl’s Neuro.MS solution in the UK and other key European markets, offering Neuro.MS to its current neuroimaging customers alongside the Brainomix e-Stroke platform. Neuro.MS automatically analyzes brain MRI images, allowing clinicians to identify, quantify, and track abnormalities in under 5 minutes.

Change Healthcare’s Secure Cloud

Did you know one quarter of healthcare organizations have experienced a cyber-attack in the last year? This Change Healthcare animation explains how 3rd-party certified cloud-native enterprise imaging can help secure IT infrastructure that might be exposed with re-platformed imaging systems.

Riverain and the VA’s Precision Alliance

When the VA adopts your technology nationwide, you know you’ve been making an impact. That’s exactly what’s happening with Riverain Technologies’ ClearRead CT, which will be implemented across the VA Lung Precision Oncology Program (22 hub and 87 spoke locations).

The Resource Wire

• Did you know that portable chest X-ray exams are responsible for 69% of patient misidentification errors in radiology? That’s why GE Healthcare’s AMX Navigate portable X-ray system features a handheld barcode reader, allowing technologists to scan a patient’s wristband and automatically match the patient to the worklist.

• Healthcare is rife with data quality issues, creating a range of workflow and financial challenges, and placing increased responsibilities on PACS administrators. See how Enltic’s Curie|ENDEX data governance solution addresses these challenges, while improving radiology teams’ AI adoption readiness.

• See how New Jersey’s Ramapo Radiology Associates overcame their CD burning problems and improved their physician and patient experiences with Novarad CryptoChart. 

• Check out this Imaging Wire Show with United Imaging’s Jeffrey Bundy and Mike Coulter, who detail their unique approach to medical imaging innovations. If you’re trying to figure out a simpler and more scalable way to run your imaging organization, this interview is a great way to start.

• See how Us2.ai cuts echocardiography’s manual work, subjectivity, and turnaround times to automate the fight against heart disease.

• The flow of new AI applications makes it hard for radiology groups to determine which tools would help them and how IT teams can handle increased AI adoption. In this Blackford Analysis white paper, radiology and IT leaders from NYU and Canopy Partners share how a platform approach alongside a curated marketplace can help solve these challenges.

• See how Einstein Healthcare Network reduced its syringe expenses, enhanced its syringe loading, and improved its contrast documentation when it upgraded to Bayer Radiology’s MEDRAD Stellant FLEX CT Injection System.

• Raising awareness about breast cancer is an important mission, but this Intelerad editorial highlights the need to match awareness with action, helped by technology to improve screening workflows.

• Ready to improve your mammography workflows? Arterys is the first and only cloud-native Breast AI provider, and its solution dramatically reduces 3D Mammography reading times, while supporting breast cancer detection, density measurements, and personalized risk assessments.  

• See how Canon AiCE DLR is achieving new levels of cardiovascular imaging in this presentation from Jeannie Yu MD, FACC, FSCCT.

• 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.

• 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.