Uncategorized

Virtual Inaction | Broken Hearts Explained | IBM Lessons

“A monumental task, but we were not afraid.”

Vysioneer CEO Jen-Tang Lu on the company’s development of a tumor-focused auto-segmentation tool, rather than focusing on normal organ contouring like everyone else.


Imaging Wire Sponsors

  • Arterys – Reinventing imaging so you can practice better and faster.
  • Bayer Radiology – Providing a portfolio of radiology products, solutions, and services that enable radiologists to get the clear answers they need.
  • Canon Medical Systems – Delivering innovative imaging solutions and services through industry leading partnerships to improve the quality of life for all people.
  • GE Healthcare – Enabling clinicians to make faster, more informed decisions through intelligent devices, data analytics, applications and services.
  • Healthcare Administrative Partners – Empowering radiology groups through expert revenue cycle management, clinical analytics, practice support, and specialized coding.
  • Hitachi Healthcare Americas – Delivering best in class medical imaging technologies and value-based reporting.
  • Novarad – Transformational imaging technologies that empower hospitals and clinicians to deliver clinical, operational and fiscal excellence.
  • Nuance – AI and cloud-powered technology solutions to help radiologists stay focused, move quickly, and work smarter.
  • Riverain Technologies – Offering artificial intelligence tools dedicated to the early, efficient detection of lung disease.
  • Siemens Healthineers – Shaping the digital transformation of imaging to improve patient care.
  • United Imaging – Our mission, Equal Healthcare for All, pushes us beyond conventional boundaries to help clinicians expand modern, digital, intelligent care to more people within their communities.
  • Zebra Medical Vision – Transforming patient care with the power of AI.



The Imaging Wire


Virtual Inaction

When the COVID-19 pandemic made many healthcare visits virtual, Cedars-Sinai cardiologists proved to be significantly less likely to prescribe medication or order diagnostic tests (including imaging) if patients didn’t come into the office. That might make sense if the virtual patients had less severe issues, but these patients were also more likely to have cardiovascular comorbidities.

  • The Study – Cedars-Sinai researchers reviewed their ambulatory cardiology visits before COVID (April-Dec 2019; 87k in-person) and during COVID (April-2020, 74k in-person, 4.7k video, 10.3k telephone).
  • Virtual Inaction – A much lower percentage of COVID-era virtual visits resulted in medication being prescribed (phone 23.5%, video 33.6%, in-person 55.7% of visits) or tests being ordered. Although the test order variations were most notable for electrocardiograms (phone 1%, video 1.7%, in-person 27.8%), there were also significant drop-offs in imaging tests like echocardiograms (phone 0.76%, video 1.1%, in-person 1.74% of visits) and coronary CTAs (phone 0.39%, video 0.49%, in-person 1.51% of visits).
  • Virtual Patients – The study also found that virtual patients were more likely to be minorities, use private insurance, and have cardiovascular comorbidities.
  • The Takeaway – At first glance, this story is only marginally related to medical imaging, but considering the massive growth in telehealth taking place, this study is worth nearly all specialties’ attention.


The Wire

  • Vysioneer’s Tumor Contouring FDA: Vysioneer announced the FDA clearance of its VBrain brain tumor auto-contouring solution for radiation therapy, making it the first FDA approved tumor-focused segmentation tool (all others focus on normal organ contouring). VBrain uses AI to auto-contour tumors in CT/MR images (metastasis, meningioma, acoustic neuroma) and to detect any additional lesions, improving radiotherapy accuracy and shortening treatment times. Vysioneer might not be a household name in AI, but folks seem impressed with VBrain, and Vysioneer is already planning to expand its unique tumor-focused contouring approach to the rest of the body.
  • Explaining a Broken Heart: MGH Researchers analyzed brain imaging from 104 patients, including 41 who subsequently developed “broken heart syndrome” (aka Takotsubo syndrome or TTS), finding that people with higher stress-related brain activity could be at greater risk of developing “broken heart syndrome” in the future. The people who showed higher activity in the brain’s stress-associated amygdala area were more likely to develop “broken heart syndrome” from relatively common stressors such as “a routine colonoscopy or a bone fracture,” not just because of “dreadfully disturbing” personal events.
  • Learning from IBM: A new MedCity News editorial used the “entirely predictable” demise of IBM Watson Health’s marketing-first AI strategy to outline how to implement clinical AI the right way. Here’s some of its key suggestions: 1) Expect clinician resistance and try to reduce it; 2) Involve clinicians from the start to build trust; 3) Create internal clinician champions; 4) Make sure data is representative of their patients; 5) Make sure AI complements their existing workflows, rather than adding to them; 6) Actually deliver the promised results and make sure clinicians see that it worked.
  • POCUS Prime: GE Healthcare and a trio of ultrasound training/certification groups just launched a new all-in-one package intended to streamline POCUS adoption. The POCUS Prime package combines GE’s new Vscan Air handheld POCUS system, education courses from Gulfcoast Ultrasound Institute or 123sonography, and certifications from the POCUS Certification Academy.
  • MRI ML for MS Type: A UCL-led team used a machine learning technique to categorize patients’ multiple sclerosis (MS) subtype using brain MRI scans, suggesting that an approach like this could be used to predict individual patients’ MS progression and treatment response. The researchers trained the unsupervised ML model using 6,322 MS patients’ MRIs, and then defined MS subtypes using MRIs from 3,068 other patients, finding that the model was able to identify three distinct MS subtypes using multidimensional MRI data (groups: cortex-led, normal-appearing white matter-led, and lesion-led).
  • HeartFlow’s Medicare Expansion: HeartFlow expanded its Medicare coverage to 38 U.S. states after three more Medicare Administrative Contractors (MACs) decided to cover its HeartFlow FFRct Analysis solution (analyzes coronary CTAs for blood flow / blockage). HeartFlow already had local coverage through two other MACs, while the two remaining MACs cover HeartFlow FFRct on a case-by-case basis. HeartFlow is on a public health roll, as it also received an adoption mandate from the UK NHS a few months ago.
  • Overestimating AI Accuracy: A npj Digital Medicine systematic review (n = 503 studies) found that imaging DL algorithms generally achieve high diagnostic accuracy in studies. However, these high scores might not translate to real world accuracy given that many similar studies in the review used very different approaches and most featured “poor design, conduct, and reporting.” The authors proposed developing AI-specific EQUATOR guidelines to address these issues and to avoid overestimated diagnostic accuracy in AI research. This study might not surprise some of you, as it closely follows other critical study reviews in recent weeks.
  • Breast MRI Preference: Many women are open to breast MRI-based screening, despite the associated risks and costs. That’s from a new study in Academic Radiology (n = 1,011 asymptomatic women) that found just 34.7% of the women were satisfied with mammography screening, while 54.7% were willing to pay at least $250-500 for MRI screening. Women with dense breasts or previous CEM/MRI exams were most open to screening with breast MRI.
  • Varian’s Google AI Alliance: Varian will use Google Cloud’s Neural Architecture Search (NAS) technology to build an AI-based automatic organ segmentation engine trained with Varian’s own image data. Varian will integrate these auto-segmentation tools into its treatment planning software, helping its cancer center clients streamline the labor-intensive radiation therapy targeting process.
  • Quantitative Ultrasound Advantages: A UPenn-led study revealed that ultrasound-based quantitative image analysis could support COVID‐19 diagnosis and make the modality less operator-dependent. Using ultrasound images from 20 patients (10 w/ COVID), the researchers compared computer‐based pleural line (p‐line) ultrasound features with traditional lung texture (TLT) features, finding that p‐line thickening and irregularity allowed them to identify the COVID‐19 patients with perfect sensitivity and specificity (vs. 90% & 70%).
  • A Call for Medical Image Analysts: An Academic Radiology editorial called for the creation of a new medical image analyst role responsible for radiology teams’ imaging data, informatics, and AI workflows, while allowing radiologists and radiographers to focus on image interpretation and acquisition.

The Resource Wire

– This is sponsored content.

  • United Imaging’s approach to brain imaging in molecular imaging puts the patient first, with a focus on reducing scan times and correcting for patient motion to avoid repeat studies, while supporting interpretation with high resolution images and quantitative values.
  • 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 here.
  • Learn how Yale New Haven Health improved its radiology efficiency, communications, and turnaround times when it adopted Nuance’s PowerScribe Workflow Orchestration and PowerConnect Communicator solutions.
  • This AI economics overview from Healthcare Administrative Partners details the various AI ROI scenarios and ways that AI can contribute to radiology practices until reimbursements become more of a reality.
  • This NIH study details how Imaging Biometrics MR (available via Arterys) can support brain tumor treatment management decisions.
  • This Bayer Radiology case study details how Einstein Healthcare Network reduced its syringe costs, enhanced its syringe loading, and improved its contrast documentation when it upgraded to the MEDRAD Stellant FLEX CT Injection System.
  • Zebra-Med’s AI1 Imaging Analytics Engine can be fully embedded in PACS/RIS systems, making clinical insights natively accessible in the PACS viewer to support interpretations or seamlessly in the worklist to support prioritization.
  • This Riverain Technologies case study details how Einstein Medical Center adopted ClearRead CT enterprise-wide (all 13 CT scanners) and how the solution allowed Einstein radiologists to identify small nodules faster and more reliably.

You're signed up!

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

-- The Imaging Wire team