Mayo Clinic Florida neuroradiologist Erik Middlebrooks, MD, pushing back against the extra work radiologists are asked to do when AI-based findings are wrong.

Imaging Wire Sponsors

Arterys  •  Bayer Radiology  •  Blackford Analysis  •  Canon Medical Systems  •  Fujifilm Healthcare  •  GE Healthcare  •  Novarad  •  Nuance  •  Riverain Technologies  •  Siemens Healthineers  •  United Imaging  •  Zebra Medical Vision

Artificial Intelligence

Who Owns AI Evaluation and Monitoring?

Imaging AI evaluation and monitoring just became even hotter topics, following a particularly revealing Twitter thread and a pair of interesting new papers.

Rads Don’t Work for AI – A Mayo Clinic Florida neuroradiologist took his case to Twitter after an FDA-approved AI tool missed 6 of 7 hemorrhages in a single shift and he was asked to make extra clicks to help improve the algorithm. No AI tool is perfect, but many folks commenting on this thread didn’t take kindly to the idea of being asked to do pro-bono work to improve an algorithm that they already paid for. 

AI Takes Work – A few radiologists with strong AI backgrounds clarified that this “extra work” is intended to inform the FDA about postmarket performance, while monitoring healthcare tools and providing feedback is indeed physicians’ job. They also argued that radiology practices should ensure that they have the bandwidth to monitor AI before deciding to adopt it.

The ACR DSI Gets It – Understanding that “AI algorithms may not work as expected when used beyond the institutions in which they were trained, and model performance may degrade over time” the ACR Data Science Institute (DSI) released a helpful paper detailing how radiologists can evaluate AI before and during clinical use. In an unplanned nod to the above Twitter thread, the DSA paper also noted that AI evaluation/monitoring is “ultimately up to the end users” although many “practices will not be able to do this on their own.” The good news is the ACR DSI is developing tools to help them.

DLIR Needs Evaluation Too – Because measuring whether DL-reconstructed scans “look good” or allow reduced dosage exams won’t avoid errors (e.g. false tumors or removed tumors), a Washington University in St. Louis-led team is developing a framework for evaluating DLIR tools before they are introduced into clinical practice. The new framework comes from some big-name intuitions (WUSTL, NIH, FDA, Cleveland Clinic, UBC), all of whom also appear to agree that AI evaluation is up to the users.

The Takeaway – At least among AI insiders it’s clear that AI users are responsible for algorithm evaluation and monitoring, even if bandwidth is limited and many evaluation/monitoring tools are still being developed. Meanwhile, many AI users (who are crucial for AI to become mainstream) want their FDA-approved algorithms to perform correctly and aren’t eager to do extra work to help improve them. That’s a pretty solid conflict, but it’s also a silver lining for AI vendors who get good at streamlining evaluations and develop low-labor ways to monitor performance.

Validating ClearRead CT

It says a lot when a solution works so well for a radiology department that they decide to perform a study to quantify its benefits. In this Imaging Wire Q&A, University Hospital of Zurich’s Thomas Frauenfelder discusses his experience and study on Riverain Technologies ClearRead CT.

Ramapo Radiology’s Case for Novarad CryptoChart

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

The Wire

• Adaptix’s Multi-Modal Funding: Portable imaging startup Adaptix completed a £12.9m funding round that it will likely use to develop and commercialize its digital tomosynthesis imaging technology. Interestingly, the funding round included a £2.5m investment from Avingtrans, cementing Adaptix’s alliance with point-of-care MRI company Magnetica (an Avingtrans subsidiary). The companies will develop a multi-modal imaging system that combines Magnetica’s Cryogen-free MRI and Adaptix’s digital tomosynthesis technology (initially for orthopedics and urgent care).

• Image Search Works: A new Radiology Journal study detailed how a content-based image retrieval system (CBIR) can significantly improve interstitial lung disease (ILD) diagnosis by searching an image database for similar chest CTs and providing them to radiologists. The retrospective study had eight radiologists interpret scans from 80 patients with different types of ILDs (with and without the CBIR, two weeks apart), finding that the CBIR significantly improved diagnostic accuracy (60.9% vs. 46.1%) and interreader agreement (Fleiss κ 0.47 vs. 0.32).

• vRad Ups the Ante: vRAD announced that it will introduce “the largest pay hike” in the company’s history, increasing radiologist compensation by up to 25% in January 2022 (it’s fourth increase since early 2019). We don’t typically cover pay increase announcements, but that’s because we’ve literally never seen an announcement like this, illustrating the hiring/retention pressures vRAD and other rad/telerad practices are facing.

• Radiation Unawareness: A new JAMA survey (n = 2,866 patients in Italy, 2019-2020) revealed a long list of patient misunderstandings about medical radiation (e.g. 55% didn’t know chest CT exams have more radiation than CXR), which makes sense considering that 44% admitted to having an inadequate knowledge of radiation risks. The patients largely preferred to be informed about radiation risks by medical staff (80% agreed), and a corresponding editorial agreed with this approach, calling for a “systemic and seismic shift” in how physicians educate patients about radiation.

• MSU’s Contrast Process: A team of Michigan State University scientists are using a $2.2m NIH grant to create new MRI contrast agents that use biocompatible peptides and proteins to target specific body parts. That’s pretty cool on its own, but MSU placed the greatest emphasis on the fact that this project is run by a multidisciplinary team. The project’s computer scientists built/operate a machine learning system to identify promising molecules, and the team’s biomedical engineers run lab experiments on the ML-vetted molecules, potentially making their contrast discovery process far more efficient. 

• 2022 ICD-10 Code Impact: With 2022’s ICD-10 codes now public, Healthcare Administrative Partners detailed how the 159 new codes (and many cancelled codes) will impact radiology. The good news is relatively few are important to radiology (most are just more specific versions of previous codes), although practices still need to carefully review the codes and adjust their templates and EHR systems, paying particular attention to deleted codes (can’t bill for those…). The 2022 ICD-10 also brings new codes for post-COVID patients and a greater emphasis on complete and consistent documentation.

• Mount Sinai AI: The Icahn School of Medicine at Mount Sinai announced the opening of its new Department of Artificial Intelligence and Human Health, the first department of its kind within a US medical school. The new Mount Sinai AI department aims to drive healthcare transformation through AI research, leveraging data from the health system’s eight hospitals to build AI-enabled healthcare tools.

• RAYUS Acquires Again: RAYUS Radiology (formerly CDI) continued its nationwide expansion, acquiring South Florida’s Diagnostic Centers of America (DCA, 8 imaging centers). The acquisition comes with DCA’s existing partnership with Boca Radiology Group (BRG, a Mednax/RP practice), and BRG will continue to read studies performed at DCA’s imaging centers. RAYUS/CDI has been an active acquirer since its own PE acquisition in 2019, and has become even more aggressive since June with acquisitions in Washington, Orlando, Pittsburgh, and now across South Florida.

• COVID’s Abnormality Plateau: A new Radiology Journal study found that 25% of once-hospitalized COVID patients still have abnormal chest CTs a year after recovery. The study analyzed a series of chest CTs from 209 patients, finding that lung abnormalities resolved at a slower rate after three and seven months (at discharge = 0%, 3mo = 61%, 7mo = 73%, 12mo = 75%). Patients with persistent lung abnormalities were more likely to be older or have acute respiratory distress syndrome. 

• New Interoperability Standards: The ONC is partnering with CMS and the CDC on updated interoperability standards that build upon the United States Core Data for Interoperability (USCDI) initiative. The new program, aptly titled USCDI+, will work towards interoperability on datasets, certification criteria, and implementation specifications for hospital IT teams and federal partners.

• PSMA PET/MRI Superiority: A new Weill Cornell study found that Ga-68 PSMA-11 PET/MRI is far superior to multiparametric MRI for monitoring men with biochemical recurrent prostate cancer (n = 108), achieving far higher sensitivity (95.5% vs. 63.6%), PPV (87.5% vs. 84.9%), and detection rates (82.4% vs. 54.9%). Although PSMA PET/MRI was more effective than mpMRI across all patients, it was particularly effective among men with PSA values below 0.2 ng/mL (40% vs. 9% detected).

MRI AI Opportunities

How can artificial intelligence (AI) help improve clinical and operational performance in MRI? Listen to the latest Siemens Healthineers podcast to learn more.

The Resource Wire

• When Birmingham Radiological Group-GV adopted Nuance PowerScribe One, the practice eliminated 60-75 minutes in daily reporting time and reduced calls to the radiology reading room by 80% by getting its reports to clinicians faster. See how in this Nuance Case Study.

• Canon’s Cartesion Prime is the only air-cooled digital PET/CT system, which means big reductions in service, chiller, and infrastructure costs. Check out Canon’s case for the air-cooled Cartesion Prime.
• Learn how leveraging the right cardiology image and reporting platform drives performance and outcomes in this Fujifilm Healthcare white paper.

• Thinking about AI ROI? Check out this AIMed conversation featuring Blackford CEO, Ben Panter and Lahey Hospital & Medical Center’s radiology Chairman, Dr. Christoph Wald discussing how to demonstrate the value of healthcare AI.

• See how and why Zebra Medical Vision sees a much bigger future for public health AI than many of us imagine in this Imaging Wire Q&A with company CEO, Zohar Elhanani.

• Check out this 4D flow user infographic detailing how Arterys Cardio AI has constantly added meaningful clinical diagnostic value.

• Being so highly vertically integrated is unusual in medical imaging, but gives United Imaging the flexibility to control quality standards and create savings that turn into customer benefits. Learn more in this video.

• Check out this Imaging Wire Show featuring GE Healthcare’s US & Canada MRI leader, Brian Murphy, discussing MRI’s evolution and how AIR Recon DL is eliminating MRI’s signal, speed, and resolution compromises.

• With radiation dose management now largely considered best practice, this Bayer white paper details the top five benefits of adopting contrast dose management.