Uncategorized

Imaging Wire #200

“Thanks to the helpers. Let’s take care of ourselves and each other.”

Tom Hanks from quarantine this weekend.



Happy 200th Imaging Wire issue, everyone. It doesn’t feel like a great day to celebrate, but it’s a good day for gratitude and I’m super grateful for all of the readers, sponsors, friends, and family who’ve made Insight Links & The Imaging Wire possible. Massive thanks to all the medical professionals doing great work on the frontlines right now, too. Looking forward to properly celebrating issue number 300 with all of you next year.



Imaging Wire Sponsors

  • Focused Ultrasound Foundation – Accelerating the development and adoption of focused ultrasound.
  • GE Healthcare – Providing point of care ultrasound systems, from pocket-sized to portable consoles, designed to support your clinical needs and grow along with your practice.
  • Healthcare Administrative Partners – Empowering radiology groups through expert revenue cycle management, clinical analytics, practice support, and specialized coding.
  • Nuance – AI and cloud-powered technology solutions to help radiologists stay focused, move quickly, and work smarter.
  • Qure.ai – Making healthcare more accessible by applying deep learning to radiology imaging .
  • Riverain Technologies – Offering artificial intelligence tools dedicated to the early, efficient detection of lung disease.

The Imaging Wire



COVID-19

There are a thousand COVID-19 stories to tell as its gravity has quickly become a reality in Western countries and across the globe. Here’s the latest imaging-related COVID-19 news:

  • CT Not Recommended The ACR came out with a recommendation that CT shouldn’t be used for COVID-19 screening or as a first-line test, and should only be used for hospitalized patients with specific clinical needs (same goes for CXR if CT isn’t available). There’s been a lot of focus on imaging’s role in COVID-19 diagnosis, but the ACR decided that chest CT’s low specificity (COVID-19 looks like the flu, SARS, MERS, etc.) and the risk of transmission from the scan procedure to other patients or staff makes it a secondary option after viral testing.
  • COVID-19 TeleradUSARAD announced a COVID-19 CT screening program in partnership with blockchain radiologist marketplace company, Medical Diagnostic Web. The COVID-19 screening program will use a network of chest CT-trained radiologists and other experts (pulmonologists and infectious disease experts), along with AI support, for CT screening diagnostics and consultations.
  • CT StratificationA new study in the European Journal of Radiology analyzed CT scans from 73 patients with COVID-19, identifying the imaging characteristics of the disease’s various levels of severity. Here’s the breakdown: 1. “Mild” patients (n = 6, no abnormal CT findings); 2. Patients with “common” severity (n = 43, unique or multiple ground-glass opacities in the periphery of the lungs); 3. “Severe” patients (n = 21, 16 w/ extensive GGO, 5 w/ pulmonary consolidation); 4. “Critical” patients (n = 3, extensive “white lung”, with atelectasis and pleural effusion).
  • Pediatric Covid-19A new NEJM study detailed the characteristics of pediatric COVID-19 cases (n = 366 hospitalized children near Wuhan, 6 with SARS-CoV-2). Among the six children, the study revealed the following characteristic: high fevers (6/6), coughs (6), vomiting (4), CT images showing typical viral pneumonia patterns (3 w/ patchy shadows in both lungs, 1 w/ patchy ground glass opacities in both lungs), and below-normal levels of lymphocytes (6), white cells (4), and neutrophils (3). The good news is, all of the children recovered after hospitalization (7.5 day median).
  • ACR’s COVID-19 AI Use Case – The ACR Data Science Institute (DSI) published an AI use case for COVID-19 diagnosis, which suggests that AI could help identify early onset of COVID-19 on chest CT scans. The use case workflow involves: 1. Sending the chest CT image to the PACS and the AI engine; 2. The AI engine returns a COVID-19 likelihood score, plus predictions and other clinical info; 3. The AI engine sends an alert to the PACS that highlights the features that contributed to the score/prediction.
  • A Call for Home Teleradiology – Noting the importance of working from home as we try to flatten the coronavirus curve (and keep radiologists healthy/working), radsresident.com issued a call for universal home radiology. The editorial noted how COVID-19 changed home teleradiology from a luxury to a necessity, while suggesting that it’s now hospitals and imaging centers’ responsibility to support this shift.

The Wire

  • MEG Enhancement: Researchers at Sandia National Laboratories landed a $6m NIH grant to develop a brain imaging device that would make magnetoencephalography (MEG) imaging “more comfortable, more accessible and potentially more accurate.” Made possible by new sensors that don’t require cooling, the adjustable helmet-based system could allow for improved brain imaging (relaxed patients, more positions, tighter fit) and expand applications to patients who struggle staying still (e.g. Parkinson’s disease).
  • NY PACS Hack: New York-based radiology practice, Northeast Radiology, revealed that its PACS system was breached by “unauthorized individuals” who accessed 29 patients’ records (but potentially more) including their health and personal data. The Alliance HealthCare-connected practice informed all patients whose info could have been accessed, is working with a forensic security firm to investigate the breach, and is addressing its security system to ensure there are no future breaches.
  • RefleXion X1: RefleXion Medical announced the FDA approval of its X1 machine, combining fan-beam CT imaging with a linear accelerator to deliver radiation to a tumor while avoiding nearby normal structures. Although this FDA approval allows the RefleXion X1 to treat a single tumor, the company plans to eventually use the new system to treat any stage of cancer.
  • Telemammography: A new University of Cincinnati study added to what will likely prove to be a watershed month for telemedicine, revealing that screening mammography patients may prefer a video message from their radiologist rather than traditional notification methods (mailed letters, phone calls, getting results from PCPs). The prospective study provided 94 women with their screening results either in video format or by traditional methods. Of the 80 women who responded, 73% preferred a video message and it was even higher among 40-60yr-old women.
  • Philips’ Fiber Cath Lab: Philips is developing new light-based fiberoptics technology that could someday replace or supplement X-ray-based angiography in the cath lab. Philips’ Fiber Optic RealShape (FORS) uses fiberoptic light traveling through custom catheter and/or guidewires. These fiberoptic images combine with preoperative imaging scans to produce 3D images used to guide cath lab procedures.
  • VIDA Funded: Lung imaging AI firm, VIDA Diagnostics, announced an $11m Series C round that it will use to fund the commercialization of its LungPrint chest CT solution suite and expand LungPrint’s clinical portfolio. VIDA has now raised $20.4m over six rounds.
  • Image-Guided IVs: Rutgers engineers developed a near-infrared/ultrasound-based image-guided tabletop device that could be used to guide vein and artery access for drawing blood and/or inserting IVs. The new device would introduce a technology-based approach to a historically manual/intuitive nursing practice, using AI to identify blood vessels in the images and then instructing the robot how to perform the insertion/blood drawing procedure (location, depth, etc.).
  • behold.ai Cleared: The U.S. FDA approved behold.ai’s red dot ‘instant triage’ algorithm, which identifies collapsed lung (pneumothorax) in chest X-rays as soon as the image is captured and then alerts radiologists. The red dot algorithm will launch stateside later this year (it’s already available in EU), using a per exam billing model.
  • Seismic Imaging: Imperial College London and UCL researchers developed a new sound wave-based imaging technique that could produce “fast, finely detailed brain imaging.” Based on technology used in seismic imaging, the new sound wave-based technique can be used for continuous monitoring of severe patients, delivered via a relatively small device, and used in small practices (all unlike MRI, CT, PET), while its helmet form factor allows imaging to penetrate bone (unlike ultrasound).
  • Transformational Reality: A new paper from UPenn detailed how augmented and mixed reality (AR & MR) are emerging as potentially transformative parts of image-guided interventions, despite IR’s relatively slow AR/MR adoption so far. The team’s bullish view is due to AR/MR’s ability to: 1. Make medical imaging more accessible, 2. Display 3D medical images to enhance guidance during procedures; 3. Overlay holograms onto objects within the surgical suite and interact with them without looking away; 4. Support real-time training or remote assistance; and 5. Help guide C-arm placement for technologists.
  • Video Pretraining: A Stanford team found that pretraining imaging AI models with videos can improve algorithm accuracy. The study pretrained an AI algorithm for identifying appendicitis using nearly 500k video clips and just 438 CT scans annotated for appendicitis. Pretraining the 3D model on natural videos improved the model’s performance from a 0.724 AUC to an 0.810 AUC.
  • Siemens Teamplay Platform: Siemens Healthineers announced the launch of its Teamplay Digital Health Platform, intended to support healthcare organizations’ operational, clinical (diagnostic & therapeutic), and shared decision making, as well as their overall digital transformation. The platform is vendor, system, and device-neutral, allowing cross-departmental and cross-institutional interoperability (it’s connected to over 5k institutions) and enabling broader data aggregation and analytics.
  • AI Stroke Predictor: A Stanford-led team developed a deep learning model that successfully predicted infarct lesions from pre-stroke imaging as effectively as current methods. The researchers trained the DL model on acute and follow up MR images from 182 patients with acute ischemic stroke to predict the size and location of infarct lesions at 3 to 7 days. In patients with minimal and major reperfusion, the model showed comparable performance to current clinical methods (0.92 AUC).

The Resource Wire

– This is sponsored content.

  • ClearRead Xray from Riverain Technologies includes the first FDA-cleared software solution to transform a chest x-ray into a soft-tissue image, providing unprecedented clarity for efficient, accurate, early detection of lung disease.
  • The GE Healthcare Venue Go features a uniquely adaptable design, a simple interface, and streamlined probe layout so you can go through your day quickly, efficiently, confidently.
  • Qure.ai’s qXR tool was included in a study that found that deep learning algorithms can help identify TB-associated abnormalities in chest radiographs and are recommended for TB programs with limited resources.
  • Learn how and why Seattle Children’s Hospital, Duke University Health System, and HCA Healthcare chose to ditch the disk by adopting Nuance’s PowerShare Network.

You're signed up!

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

-- The Imaging Wire team