CV19 AI Dos and Don’ts | A Race for Humanity | Cuomo’s Infiltrate

“Not a race to the moon, this is a race for humanity. The rocket ships are GPU supercomputers, and their fuel is scientific knowledge.”

NVIDIA’s Ian Buck, calling the COVID-19 HPC Consortium “the Apollo Program of our time.”


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



CV19 AI Dos and Don’ts

We’re starting to learn to not get too excited each time a COVID-19 AI study suggests that AI will lead to the next CV19 breakthrough… and not to get too disappointed when a study suggests that a CV19 algorithm won’t be as effective as we’d hope. The fact is, we’re still in the learning phase with this disease and both of these kinds of study results represent progress. Here are two of our latest bits of progress:

  • Pneumonia AI Works for CV19 – A team of UCSD researchers found that an algorithm originally intended to diagnose pneumonia in chest X-rays could also be used for early COVID-19 diagnosis and ongoing monitoring. The team analyzed X-rays from five CV19 patients (10 frontal CXRs) using a pneumonia probability map algorithm, finding that the algorithm predicted and consistently localized findings and assigned likelihood with a “surprising degree of generalizability and robustness.”
  • CV19 Predictions, Without CT – A Wuhan, China-based team developed an AI algorithm intended to use a broad combination of clinical data points to predict which CV19 patients would develop severe cases. However, CT images didn’t end up included in the algorithm’s most effective data combination (elevated liver enzymes, body aches, and elevated hemoglobin). The combination of non-imaging data points predicted the development of acute respiratory distress syndrome (ARDS) with 70% to 80% accuracy (n = 53 CV19 patients, 5 who developed ARDS).

The Wire

  • One Month of Revenue: A letter to HHS secretary, Alex Azar, from nearly every major healthcare provider group (including the ACR) asked the U.S. government to provide one month of revenue to each physician, nurse practitioner, and physician assistant to make up for the financial losses and non-reimbursable expenses created by CV19.
  • Breast MRI CAD Features: South Korean researchers found that kinetic features extracted from preoperative breast MRI scans can predict the severity of distant metastasis breast cancers, possibly allowing for more personalized treatment. The researchers evaluated MRIs from 276 women using a commercial CAD system (28 who developed distant metastasis within 79 months), finding that women with higher kinetic heterogeneity were more likely to have distant metastases (0.70 vs 0.43 mean) and also had lower distant metastasis-free survival rates.
  • Cuomo’s Infiltrate: CNN anchor Chris Cuomo didn’t make many fans within the radiology community after he proclaimed in his post-CV19 recovery interview that the second he saw his chest x-ray, “I saw infiltrate on it. And I don’t even know what infiltrate means.” There’s a long list of things that radiology folks didn’t like about this segment, but their main issues were with the fact that CV19 doesn’t need more imaging-based self-diagnosers and “infiltrate” is definitely not the preferred term for what he “saw.”
  • Nanox & Qure.ai: Nanox and Qure.ai announced a partnership that will integrate Qure.ai’s chest X-ray and head CT AI solutions in the Nanox.CLOUD. Nanox’s alliance with Qure.ai comes just a few weeks after Nanox announced a similar deal with CureMetrix, and shows that the well-funded startup is making progress adding partners for its “end-to-end” imaging platform.
  • More UI Mobile CTs: A week after partnering with AMST to construct portable CT scanners intended to support the U.S.’s CV19 fight, United Imaging entered a similar deal with Sea Box to further expand its fleet of portable CT scanners.
  • South Korea’s Secret: An article out of South Korea attributed the country’s low CV19 fatality rate in part to its high number of CT scanners (38.2 CTs per 1m people vs. Spain’ 18.6/1m) and ventilators (19 per 100k people vs. Italy’s 8.3/100k). The story attributed South Korea’s medical device advantage (and its higher personal hygiene standards) to the country’s response to the 2015 MERS outbreak.
  • TTG Acquires NIS: Medical imaging equipment sales/service company TTG Imaging acquired PET & PET/CT company Nuclear Imaging Services (NIS) and its cardiac radiopharmacy, NISotopes. The addition of NIS expands TTG’s reach/capabilities and continues its acquisition pace that included Absolute Imaging Solutions, Northeast Electronics, and Molecular Imaging Specialists during the last year.
  • VUNO’s CV19 AI: South Korean AI developer VUNO adapted its existing lung AI solutions to support COVID-19 diagnosis and monitoring. VUNO’s new CV19 Suite combines its already-available VUNO Med – LungQuant (for CT analysis) and VUNO Med-Chest X-ray, while adding support for remote reading.
  • Grassroots Call for CV19 Protections: A team of physicians launched a change.org petition (~120k signatures so far) supporting their proposed COVID-19 Pandemic Physician Protection Act (CPPPA) to protect healthcare workers. The CPPA includes a range of very reasonable requests including ensuring adequate PPE, malpractice protections, the establishment of a COVID19 Healthcare Workers Fund, tax credits for physicians, expanded telehealth coverage, and provisions to keep physician practices solvent. However, there are also some requests that folks might not view as particularly CV19-related including student loan forgiveness and the abolishment of noncompetes.
  • Blue Eye Gets into CV19 AI: South Carolina-based IT services company, Blue Eye Soft, licensed a deep learning solution developed by a University of Dayton scientist that can reportedly detect COVID-19 in chest X-rays “in seconds” with 98% accuracy. Blue Eye Soft already filed a provisional patent and plans to submit the software for FDA approval “within a matter of days.”
  • How to Take CV19 Patients to Imaging: In general, CV19-positive patients should stay in their hospital rooms. However, if it’s “medically essential” that they are taken out of their rooms for imaging, the CDC’s updated FAQs recommend that: 1. healthcare personnel in the receiving area are notified in advance; 2. Patients should wear a facemask and a clean sheet during transport; 3. Transport personnel should wear all recommended PPE; and 4. After arrival, the receiving personnel and transporter should wash their hands and wear all recommended PPE.

The Resource Wire

  • As the world navigates the unprecedented challenge presented by the COVID-19 pandemic, GE Healthcare supports healthcare providers, partners, communities and patients around the world in addressing it. Find information on GE Healthcare’s COVID-19 Resources here.
  • This Healthcare Administrative Partners blog article details how physicians are now eligible to receive additional funding during the COVID-19 pandemic thanks to the CARES Act.
  • In this quick video, Einstein Healthcare Network’s Peter Wang, MD describes how Nuance’s embedded clinical guidance delivers streamlined radiology workflow and improves report quality.
  • 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. Learn more.
  • Join the Focused Ultrasound Foundation’s webinar today at 2pm Easter, and learn about the latest breakthroughs in focused ultrasound for Alzheimer’s disease, brain tumors, cancer, Parkinson’s disease, pain, psychiatric disorders and more.

Leave a Reply

Basic HTML is allowed. Your email address will not be published.

Subscribe to this comment feed via RSS

Get every issue of The Imaging Wire

Delivered right to your inbox. We will never sell or share your information.