#441 – The Wire

  • Ultra-Portable MRI Validation: A team of Spanish scientists developed and tested an ultra-portable and low-cost extremity MRI (72mT, 70cm wide, 250kg, <€50k), finding that it could produce diagnostic-quality exams in “previously unrealistic” settings. The researchers used the MRI to scan a volunteer’s knee in five different environmental and power source settings (the lab, a home, an office, outdoors with electricity, open air with a gas generator), producing diagnostic images in clinically viable scan times.
  • Automated Cancer Coordination: A PLOS Digital Health study of 127 liver cancer patients at a VA hospital found that an automated system for reviewing radiology reports improved the timeliness of care. The EHR-linked system generated a queue of abnormal cases for review while allowing care coordinators to create lists of next steps / reminders. Over the course of the 5-year study, the system reduced mean time from diagnosis to treatment by 36 days, time from imaging to diagnosis by 51 days, and time from imaging to treatment by 87 days.
  • ‘Staggering’ CVD Rise: A study published in JACC projects steep increases in cardiovascular disease in the coming decades, with major patient care (and likely imaging) implications. By the year 2060, the number of people with ischemic heart disease is forecast to increase by 31% (22M to 29M), heart failure by 33% (10M to 13M), myocardial infarction by 30% (12M to 16M), and stroke by 34% (11M to 15M). 
  • Using Thyroid AI with Breast Lesions: Just as AI is starting to work with pathologies it’s trained to detect, a team of South Korean researchers explored using a thyroid ultrasound AI tool to assess breast cancer lesions. The thyroid ultrasound AI tool originally performed poorly with 1,042 breast lesions (0.678 AUC), but improved after fine-tuning training with 1,084 separate breast lesions (0.841 AUC). Despite these “not bad” results after fine-tuning, the authors concluded that an organ-specific approach to AI training “guarantees better diagnostic performance.”
  • Mammography’s Healthcare Cost Barriers: A new JACR study highlighted a number of patient-reported healthcare cost barriers that appear to reduce mammography screening utilization. Survey data from 7.5k American women (69% screened within prev. 2yrs) revealed that women are less likely to attend mammography screening if they worry about the cost of medical bills, find affording dental care or eyeglasses “challenging,” or skip/reduce/delay their medication refills (odds ratios: 0.86, 0.65, 0.67, 0.69/0.63/0.71).
  • Mayo Clinic + Mercy: Mayo Clinic and Mercy Health entered a 10-year partnership to share deidentified patient data from over 500M visits to develop novel AI algorithms focused on “personalized, predictive, and proactive medicine.” Most larger health systems do similar work with their own data, but this collaboration is designed to take advantage of Mayo Clinic and Mercy’s different populations and geographies to help reduce model bias while improving the accuracy of treatment recommendations.
  • MGB’s CT Contrast Intervention: An AJR study detailed how Mass General Brigham’s EHR-based interventions helped it reduce contrast-enhanced CT volumes during the recent shortage. The first EHR intervention on May 10th informed referrers of the shortage and recommended alternative modalities, while the second intervention on May 16th required referrers to enter detailed clinical information. As a result, MGB’s contrast-enhanced body CT volumes fell from 561 patients per weekday, to 531 after the first intervention, and 491 after the second, even though its overall CT volumes remained steady (1,350, 1,323, 1,314).
  • Siemens Supports AvoMD: AvoMD named Siemens Healthineers as a supporter of its app that will make the CAD-RADS 2.0 guidelines available to referring clinicians at the point of care. AvoMD’s app combines the Society of Cardiovascular Computed Tomography’s clinical expertise with AvoMD’s decision support technology in an effort to streamline CV care. Siemens is providing financial backing to offset costs associated with app development. 
  • Aidence Efficiency & Consistency: An EJR study highlighted Aidence’s Veye Lung Nodules solution’s ability to improve radiologists’ lung nodule reporting efficiency and consistency. The researchers had two radiologists assess 50 chest CT cases, finding that Veye Lung Nodules reduced the radiologists’ reading times by 33.4 % and 42.6 % (226min to 151min, 321min to 184min), while improving their agreement rate (kappa: 0.61 to 0.84).
  • Four Strikes and You’re Out: A 2,000 consumer survey from ModMed found that 73% of patients keep a “mental scorecard” of their doctor visits, and allow an average of four strikes on poor experiences before picking a new provider. Timeliness and modern technology use were the top factors contributing to a good experience, although 67% said friendliness of staff influences their provider choice and one-third said they’ve switched providers due to an unpleasant encounter with office staff. This study might not apply to most radiologists, but it seems very applicable to imaging centers.
  • Inappropriate ENT Imaging: A new Insights into Imaging study revealed particularly high inappropriate otolaryngology imaging volumes at two Spanish hospitals. A review of 538 otolaryngology imaging orders found that just 42% were appropriate, while all other exams were inappropriate, not adequately justified, or not included in the guidelines (34%, 12%, 12%). Primary care physicians were most likely to order inappropriate exams (78%), while sinus X-ray orders were most likely to be inappropriate (60.7%).

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!