#444 – The Wire

  • Radiologist Monitoring: The New York Timesrecent article on the growing use of employee productivity monitoring systems (including for radiologists) caught the attention of quite a few rads who clearly oppose this trend. The article inspired several Twitter and Aunt Minnie forum “conversations” that criticized monitoring’s negative impact on the profession (role commoditization, declining job satisfaction/retention, burnout), while blaming corporate leadership for prioritizing efficiency over efficacy.
  • CXR AI Demographic Predictions: Almost exactly a year after we learned that imaging AI can predict patient race, the authors of a new JACR study developed a series of chest X-ray AI models that accurately predicted patient demographics. After training four AI models on more than 55k CXRs, tests against an external dataset found that the models accurately predicted patients’ gender, age, and ethnicity (AUCs: 0.997, 0.892, 0.832), although they only achieved a “moderate” 0.675 AUC for predicting the patients’ insurance type.
  • Transfer Discrepancies: A new study in JACR highlighted the importance of reviewing emergency transfer patients’ prior radiology reports. Researchers analyzed data from 5,834 patients who were transferred to a level 1 trauma center after receiving either a CT or MRI at the first ED they visited, finding that trauma center radiologists flagged 12% (669) of the outside radiology reports for discrepancies, with 92% (613) of those discrepancies affecting decision making.
  • Nuance & Covera’s Quality Collaboration: Nuance and Covera Health announced the launch of their Quality Care Collaborative (QCC) program, combining Covera’s clinical intelligence platform and Nuance’s Precision Imaging Network. The alliance will use Covera’s quality assessment capabilities and Nuance’s analytics and infrastructure to give payors, providers, and self-insured employers a foundation for their radiology quality improvement initiatives
  • Multimedia Reporting Gets Technical: The HIMSS and SIIM team that first introduced interactive multimedia reporting (IMR) took another step towards their goal, releasing their new IMR technical white paper. The extremely detailed paper is worth a deep dive if you’re in this space, describing the standard-based infrastructure and communications that would be required in order to adopt IMR.
  • Probo Acquires Mi Healthcare: Probo Medical continued its M&A spree, using its new private equity funding to acquire UK-based medical imaging equipment and service company, Mi Healthcare. This is Probo’s fourth European acquisition since first expanding to Europe in 2020, representing nearly half of its global acquisitions since 2018.
  • What Rads Want from Mammo AI Tools: A JACR-published survey of 66 radiologists suggests that 46% to 60% of those radiologists intend to adopt mammography AI tools, but 26-33% would be deterred if AI features/functionality don’t align with their preferences. Highlighted among the rads’ mamo AI preferences were balanced sensitivity and specificity (94% sensitivity with <25% of examinations marked), tools that show AI findings after radiologists performed their review, and AI models that incorporate both mammography images and clinical data.
  • Dyad’s AI Echo Clearance: Cardiac image analysis company Dyad Medical has secured FDA clearance for its echo AI application Echo:Prio. The software, which is part of its complete cardiac platform called Libby (FDA-cleared), provides operators and physicians with a computer-assisted decision-support tool for index quantification of cardiac function. With the new clearance, Dyad will soon be able to offer physicians an immediate AI-powered second opinion when evaluating echos.
  • Faster CTs & Fewer Sedated Kids: Emory researchers revealed that when their two pediatric EDs upgraded to faster dual-source dual-energy CTs (1-3 vs. 12 seconds/scan), fewer children required sedation. Analysis of 15k patients who underwent head CTs either before or after installation of the faster CT scanners showed that the upgrade reduced their sedation rate from 8% to 7%. That one percentage point improvement is notable given these kids’ average age (7yrs) and the types of sedatives that they were able to avoid.
  • bpMRI AI for PCa Surveillance: A team of Dutch researchers demonstrated that using AI-assisted biparametric MRI (bpMRI) surveillance might improve detection of clinically significant prostate cancer. In the 73-man study, the AI model performed better when analyzing the patients’ prior and current bpMRIs compared to only analyzing their current exams (AUCs: 0.81 vs. 0.73), and the model was even more accurate when its analysis also incorporated clinical data (AUC: 0.86). Meanwhile, all AI calculations outperformed the study’s radiologists who weren’t using AI (AUC: 0.69).
  • Same-Day Feedback: A BMC-published survey of 185 women who attended a mammography screening and received their results during the same appointment found that this practice improved 48% of the women’s breast cancer screening experiences. Interestingly, 47% reported no significant difference, while 5% reported that the process made their experience worse (potentially due to long wait times for results). Based on these findings, the authors called for new research into patient experiences with same-appointment reporting, focusing on workflows that have shorter reporting wait times.

You might also like

You might also like..

Select All

You're signed up!

It's great to have you as a reader. Check your inbox for a welcome email.

-- The Imaging Wire team

You're all set!