AI researcher Geoffrey Hinton, PhD, in 2016. Hinton this week announced he is leaving Google “so he can freely speak out about the risks of AI.”

Imaging Wire Sponsors

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

We’re pleased to announce the launch of our latest edition of The Imaging Wire Show, “Driving Efficiency in Neuro MRI,” featuring Karen Holzberger of SpinTech MRI and Dr. James Backstrom of Armstrong County Memorial Hospital. Learn how recent advances in technology are helping community hospitals run their MRI service more efficiently – and also see a case review – by visiting our YouTube channel.

Contrast

The Value of CT Contrast

If there was a tool available that would automatically improve your job performance by nearly one-third, you’d want to use it, right? (And no, we’re not talking about ChatGPT.) The tool is contrast media, and a new study in JAMA Surgery found that the use of contrast in emergency CT exams improved the diagnostic accuracy of radiologists by 30%.

Radiology has a love-hate relationship with contrast. Contrast media unquestionably improves medical image quality, but it also adds a layer of complexity to imaging operations. Patients have to be screened and monitored for contrast allergies and reactions, IV lines have to be placed, injectors have to be monitored and maintained. 

Radiology researchers and clinicians have explored a variety of methods for non-contrast imaging, in particular in the emergency setting, where the time spent setting up a contrast exam could delay patient care. 

But while the risks of giving contrast are extensively debated, what are the risks of not using it? Researchers from multiple US institutions sought to answer this question in a study of 201 patients who got CT scans in emergency departments for acute abdominal pain over 3 weeks in 2017. All patients were scanned on Siemens Healthineers’ Somatom Force dual-source dual-energy CT scanner.

The original scans were contrast-enhanced, and a dual-energy technique was used to produce non-contrast images. Both sets of images were read by faculty and resident radiologists. Compared to the gold standard of contrast CT at 100%, findings included:

• Diagnostic accuracy of unenhanced CT was 70%, or 30 percentage points lower than contrast CT
• Faculty radiologists were more accurate than residents for primary diagnoses (82% vs. 76%), but less accurate for actionable secondary diagnoses (87% vs. 90%)
• Faculty made fewer false-negative primary diagnoses than residents (38% vs. 62%)
• False-negative and false-positive results were common (19% and 14%, respectively)

The Takeaway 

Reducing the use of contrast is a worthy goal, but it carries risks of its own, as this study indicates. False-negative interpretations are among the worst kinds of radiology errors with non-contrast exams, and clinicians should weigh the diagnostic penalty of withholding contrast media in the emergency setting, especially given the extremely low rate of contrast reactions in low-risk patients.

Tiny Signs of Hope: Advancing Theranostics

Meet a very special prostate cancer patient, weeks from death, who received 177Lu as a last-ditch effort to save his life. Miraculously, he was a super-responder. Hear from him and the incredible clinical team at BAMF Health who helped restore his life in an episode from a 6-part series on the power of Theranostics from GE HealthCare.

Converting Clinical Documentation into Actionable Data

Is your department struggling with how to convert analog clinical documentation notes into actionable data that can be used in the EHR? Find out how Duke University turned that legacy information into structured data that can be used for registry reporting and analytics in this Intelerad white paper.

The Wire

• AI Guru’s New Warning: The “godfather of AI,” Geoffrey Hinton, PhD, has left Google with a new warning about the dangers of AI’s unrestricted growth. You may remember Hinton’s take on radiology in 2016, when he advised a halt to radiologist training due to AI’s growing expertise. His comments are credited with scaring medical students from radiology for years afterward. If you’re keeping score, radiology has been the strongest medical specialty in the Match since 2021. 

• Automation Bias in AI: Meanwhile, a new study in Radiology details one possible downside of AI – when radiologists rely too much on its analysis. In what European researchers called “automation bias,” an AI system was found to have a significant effect on the BI-RADS category breast radiologists assigned to mammograms – even when the AI was wrong. And inexperienced radiologists – the very ones AI is supposed to help most – were more likely to follow incorrect AI suggestions. Inexperienced radiologists were right 79.7% of the time if AI was right, but only right 19.8% of the time if the AI was wrong compared to moderately experienced (81.3% & 24.8%) and very experienced (82.3% & 45.5%) radiologists.

• Envision Prevails Against UnitedHealthcare: National medical group Envision Healthcare has won a $91.3M arbitration ruling in its dispute with UnitedHealthcare over billing rates for medical procedures. Envision and UnitedHealthcare had an in-network agreement, but Envision charged that UnitedHealthcare unilaterally reduced reimbursement to clinicians in 2017 and 2018 before terminating the contract in 2021. Envision is rumored to be close to filing for Chapter 11 protection from creditors, while UnitedHealthcare has sued Radiology Partners over billing rates.

• ACR Updates Breast Screening Recos: The ACR has updated its breast screening recommendations to consider earlier breast screening for high-risk women, in particular Black and Ashkenazi Jewish women. All women should receive risk assessments by age 25 to determine if screening earlier than age 40 is needed. In addition, MRI surveillance and supplemental breast MRI may be needed for some women. Recent studies have found racial inequities in breast cancer mortality, in particular for Black women. 

• Breast Cancer AI Alliance: Arterys and Therapixel have inked an alliance in which Therapixel’s MammoScreen AI for breast imaging software will be made available on the Arterys platform. MammoScreen analyzes mammograms and provides a score and colored scale of suspected lesions; Arterys will make the application available with its Breast AI platform. Arterys’ Breast AI suite also includes iCAD’s mammo AI portfolio, while Therapixel has been growing its stable of alliances, such as with Qure.ai. 

• EHR Prompts CT Lung Screening: Putting physician prompts about CT lung cancer screening into the EHR boosted screening orders by 20 percentage points in new research in Chest. In a study of over 1k patients, researchers found that the prompts boosted both order rates for lung screening (7.1% vs. 27.3%) and rates for completed screenings (4.4% vs. 17.7%). Better promotion of CT lung screening could boost anemic compliance rates, which remain mired in the single digits.

• FDA Clears Multisource X-Ray: The FDA granted 510(k) clearance to Nanox for the multisource version of its Nanox.ARC digital X-ray system. Nanox.ARC features an array of cold-cathode X-ray tube sources that the company says will produce high-quality 3D tomographic images at a much lower cost than existing radiography systems. Nanox had to jump multiple hurdles at FDA to gain the clearance, and now has the chance to prove whether Nanox.ARC is a revolutionary technology. 

• DL Keeps FFR-CT Data On-Site: A high-speed deep learning algorithm used for on-site processing of FFR-CT data worked well in a new study in American Journal of Roentgenology. Swiss researchers used the XFFR prototype software from GE Research to analyze CCTA exams of 59 patients and calculate FFR, finding that it had an accuracy of 95.9%, sensitivity of 93.5%, and specificity of 97.7%. On-site FFR-CT could be more convenient for clinicians and patients and reduce data security issues. 

• Breast Density Alliance: Breast imaging software developer Volpara is expanding its alliance with advocacy and education group DenseBreast-info.org. Volpara has agreed to donate 5% of the profits from its Volpara Scorecard application to DBI for the next 12 months. Scorecard is an AI-based application that performs automated breast density assessment, and can help breast imaging facilities comply with new FDA rules on breast density reporting. Volpara and DBI are touting the expanded alliance at this week’s SBI 2023. 

• AI’s Better Bedside Manner: Does AI have a better bedside manner than most physicians? That’s the suggestion of a new study in JAMA Internal Medicine in which ChatGPT responded to questions patients might send to their doctors over social media. In 79% of the cases, a panel of healthcare professionals rated the ChatGPT responses as more empathetic and higher-quality than those of real doctors. For follow-up, researchers might survey how many doctors would love to offload patient communications coming in over social media to a chatbot.

• Breast AI Hits Milestone: ScreenPoint Medical’s Transpara breast AI analysis software has now been used on over 4 million mammograms. The milestone case was analyzed in Denmark, where Transpara reviews over 75k mammograms annually at a screening program in the capital region. Radiologists there have used Transpara as an autonomous first reader for 70% of studies. ScreenPoint plans to tout the milestone at this week’s SBI 2023 symposium. 

• DocPanel Adds Cardiac Imaging Expert: Teleradiology practice DocPanel has appointed Scott Flamm, MD, to lead the company’s cardiac imaging department. Flamm comes to DocPanel from the Cleveland Clinic, where he was head of cardiac imaging. He will be available to healthcare facilities through the DocPanel platform for cardiothoracic imaging coverage, protocol assistance, and program development support, such as for cardiac MR and CT. The service will allow providers to add a new cardiac imaging service to their offerings. 

• AI to Democratize Echo: Us2.ai just published what might be the most comprehensive paper we’ve seen on echo AI, detailing the benefits of AI-automated echocardiography, the global need for more scalable and flexible CVD assessments, and how its technology is fit for the future. We don’t usually cover marketing materials in our news roundups, but this paper is worth reviewing if you’re focused on the echo, echo AI, or echo accessibility and efficiency.

Faster PET Scans with SubtlePET

Want to perform PET scans faster, but keep the same image quality? With SubtlePET image enhancement from Subtle Medical, you can conduct PET exams in one-quarter of the original time while preserving image quality. Find out how in this case review.

Annalise.ai’s AI Confidence Bar

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.

The Power of the Platform for Imaging AI

If you’re at SIIM 2023, be sure to attend an InformaticsTECH Talk on Wednesday, June 14 on the Power of the Platform for Imaging AI, by Aaron Sullivan, Head of Business Development for Bayer Digital Solutions.

The Resource Wire

• 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.

• SOIN Soluciones Integrales of Costa Rica turned to Merge enterprise imaging solutions from Merative when it wanted to modernize the imaging environments of 50 hospitals across the country. Download this PDF white paper to find out how they did it.

• How has the implementation of Enlitic’s Curie|ENDEX solution impacted radiology operations at TMC, one of Europe’s leading subspecialty radiology centers? Find out in this on-demand webinar with Ernest Montañà, AI project manager at TMC.

• Visage Imaging’s Visage 7 was named the #1 Universal Viewer in the 2023 Best in KLAS Software and Services Report. Find out what’s behind this award-winning solution and learn more about Visage’s One Viewer philosophy. 

• Check out this Imaging Wire Show featuring Us2.ai’s co-founders – James Hare and Dr. Carolyn Lam – for a great discussion about Us2.ai’s continued clinical and commercial expansion, and their efforts to improve echocardiography accuracy, efficiency, and accessibility.

• Adding cardiac PET to your PET/CT practice may present challenges, but the potential clinical and operational benefits make it worthwhile. Learn how to get started in this free webinar from Siemens Healthineers. 

• 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.

• There’s a lot to learn from St. Michael’s Hospital’s experience implementing the Hyperfine Swoop portable MRI at the point-of-care. See what their hospital leaders had to say about how the ultra-low-field MRI technology impacted their patients and clinicians in this on-demand webinar.

• 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.

• 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.

• Today’s employment environment in radiology has never been more challenging. But there are tools you can use to keep your radiologists happy and engaged, according to this white paper from Medality. 

• If you’re in the business of using or providing AI, there’s a good chance you spend a lot of time managing AI evaluations. But are your evaluations as efficient or effective as they could be? Check out this Imaging Wire Show with Riverain Technologies CEO, Steve Worrell, detailing the best practices for mitigating AI adoption risks, today and into the future.

• Find out what built for the modern world means — and why it matters — in this AuntMinnie profile on United Imaging’s more modern approach to vertical integration, leadership, and culture.

• Creating your AI adoption plan? This Arterys report details what clinical, efficiency, and regulatory factors to look for in radiology AI vendors.