#388 – The Wire

  • Stockholm3 Effectiveness: A new study out of Sweden added evidence that the novel Stockholm3 prostate cancer risk model (uses PSA, plasma biomarkers, and other clinical variables) reduces unnecessary exams while maintaining detection levels. The study found that using Stockholm3 with PSA ≥2 ng/ml as an MRI threshold reduced lifetime prostate MRI exams (-60%) and unnecessary biopsies (-9%) versus PSA-based screening, with comparable quality life year gains.
  • ThedaCare’s Great Resignation: A legal battle involving a team of interventional radiology techs went viral last week. Wisconsin hospital ThedaCare sued to stop seven of its stroke team members (out of 11) from leaving to work for Ascension, but a judge allowed the team to leave, noting that Ascension didn’t recruit them and ThedaCare declined to match Ascension’s offer. While national news coverage focused on healthcare labor dynamics, radiology social platforms focused on ThedaCare’s CEO’s previous attempt to force radiology practice members to become hospital employees.
  • CTC DL Evidence: A European Radiology study detailed a DL model that effectively differentiated premalignant and benign colorectal polyps in CT colonography scans, while suggesting that CTC models might not require expert polyp segmentation. The researchers used CTCs from 63 patients (w/ 107 polyps) to train DL models with and without polyp segmentation. Against an external test set (59 patients, 77 polyps), the models detected the same number of premalignant nodules, but the DL model trained with segmentations had fewer false positives (sensitivity 80% vs. 80%, specificity 69% vs. 44%). Heatmaps revealed that the models also used the same image regions for decision making.
  • VUZE’s FDA: VUZE Medical announced the FDA 510(k) clearance of its VUZE System, which would allow minimally invasive spinal surgery teams to reference CT-based views in the OR. The VUZE System overlays a graphical representation of spinal surgery tools captured with intraoperative 2D X-ray systems onto axial and sagittal views generated from the patient’s preoperative 3D CT scans. 
  • INSIGHT CXR for Pnx: A new Radiology Journal study showed that Lunit’s INSIGHT CXR solution improved pneumothorax detection in CXRs after patients receive percutaneous transthoracic needle biopsies. A Seoul National University Hospital team analyzed their pneumothorax detection results before and after adopting INSIGHT CXR (676 CXRs w/ both groups), finding that the AI tool achieved higher sensitivity (85.4% vs. 67.1%), negative predictive value (96.8% vs. 91.3%), and accuracy (96.8% vs. 92.3%), while reducing the proportion of patients with pneumothorax who required a drainage catheter (2.4% vs. 5%).
  • KPMG Healthcare Survey: KPMG’s 2022 Healthcare and Life Sciences Investment Outlook Survey of 300 industry executives suggests we might be in for another record-setting year of M&A and investments. Over 70% of respondents expect to increase their M&A activity in 2022, while 30% plan on increasing investment activity by 10% or more. The most attractive investment areas for the next 12 to 24 months were telehealth, EHRs, and clinical workflow solutions – all of which can be tied to addressing staffing issues and workflow inefficiencies.
  • Imaging’s OOP Impact: A new JACR study highlighted the significant impact that out-of-pocket imaging costs have on some patients. A survey of 671 patients undergoing CT or MRI exams at five different Midwest outpatient imaging centers revealed that 22% experienced financial anxiety due to OOP imaging costs, 35% resorted to financial coping strategies (e.g. reducing other spending or using credit), and 15% didn’t follow their care plan because of imaging costs (e.g. taking less medication or skipping imaging exams). 
  • Mammography AI Concerns: A growing field of studies suggest that mammography AI can improve cancer detection, but design flaws undermine these results. That’s from a JACR review of 13 external validation studies (77% using commercial AI) that found AI improved accuracy when used as a standalone system (AUC improvement: 0.02-0.13) or when rads have AI support (AUC improvement: 0.028-0.115). However, all of the studies were retrospective, simulated, or both (n = 6, 5, 2), most studies had bias risk or applicability concerns (9 & 9), and only two studies used ground-truth cancer outcomes.
  • Medical Device Vulnerabilities: According to a report from security firm Cynerio, 53% of internet-connected hospital devices are vulnerable to cyberattacks. Ultrasound systems, IV pumps, and cardiac monitoring systems were among the most vulnerable devices, as well as any equipment running Linux-based software (unfortunately that’s most medical equipment) since traditional Windows-based cybersecurity is incompatible. Cynerio notes that most vulnerabilities were due to simple problems such as weak default passwords, which are easily fixable if acted upon.
  • Fixing England’s Sonographer Shortage: The NHS awarded over £1M to the Universities of Cumbria and Salford to help address North West England’s sonographer shortage. The funding will be used to build upon the universities’ new virtual ultrasound academy, including improving their training facilities, adding new ultrasound training rooms across the region’s hospitals, and creating new educator and clinical coordinator roles.
  • Intracranial Aneurysm AI: An EJR study detailed a CTA AI tool that improved less experienced radiologists’ ability to detect intracranial aneurysms, while boosting all radiologists’ efficiency. Using CTAs from 212 patients (179 w/ aneurysms) the AI system had mixed performance as a standalone system (sensitivity 84.9%, specificity 18.2%, accuracy 74.5%), but it allowed junior physicians to rival experienced rads without AI (sensitivity 95% vs. 95%, specificity 48.5% vs. 57.6%, accuracy 87.7% vs. 89.2%). More notably, the tool significantly improved the junior and senior rads’ reporting efficiency (-20.7% to 112 seconds & -18.8% to 92 seconds).

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!