AI Drivers | Purdue’s US + PAT | Developing Market Devices

“Physicians need to actively engage to adapt their practice and to shape the technology.”

Vanderbilt professor, William W. Stead MD, on physicians’ need to adapt to (and adopt) AI as the technology becomes increasingly mainstream.


The Imaging Wire


The Seven Healthcare AI and DL Adoption Drivers
JAMA outlined the seven factors that will drive AI and deep learning adoption in healthcare. They all seem quite logical:

1. Medical imaging is becoming more powerful and more important in healthcare, and AI/DL can interpret images better than humans
2. The accelerating digitization of health-related records and sharing of labeled/specialized data sets
3. Deep learning’s ability to analyze heterogeneous data sets from diverse sources
4. The “enormous capacity” of deep learning for research hypothesis generation
5. Deep learning’s potential to streamline routine clinical workflows and empower patients, making healthcare safer, more humane, and more participatory
6. The rapid-diffusion of open-source and proprietary deep learning programs
7. Today’s basic deep learning technology is sufficient to support each of the above demand drivers (as long as datasets get larger).


Moody’s Troubling Nonprofit Hospital Diagnosis
A report from Moody’s revealed that US nonprofit and public hospitals’ 2017 median expenses increased at a higher rate (+5.7%) than their median revenue (+4.6%) for the second straight year (2016 = expenses +7.1%, revenue +6.1%), putting the sector on a “unsustainable path.” As a result, the share of nonprofit/public hospitals reporting an operating loss in 2017 nearly doubled to 28.4% (vs. 16.5% in 2016), while the number of hospitals reporting lower absolute operating cash flow more than doubled during the last two years to 59% (vs. 24% in 2015). The firm attributed nonprofit hospitals’ struggles to lower reimbursement rates, a shift to outpatient care, growing merger and acquisition activity, and rising ambulatory competition. These are bad signs and Moody’s expects 2018 to be worse. Considering that Morgan Stanley just reported that 20% of ALL hospitals are in poor financial shape, it appears something is very broken in the way hospitals are operating.


Fujifilm and IU Launch AI Partnership
Fujifilm and Indiana University School of Medicine launched a partnership to develop AI-based medical imaging diagnostic support applications. The partnership combines Fujifilm’s image processing and AI technology with Indiana University’s diagnostic and clinical know-how, initially developing applications to measure muscle atrophy (sarcopenia) in body images and detect/ quantify brain lesions in neuroradiology images, while performing research to eventually develop a system to support diagnosis workflow. Fujifilm’s Indiana University alliance is one part of its AI strategy, which also includes in-house development projects and separate partnerships with AI leaders, to support the creation of Fujifilm’s REiLI-branded portfolio of AI technology.


Purdue University Combines Ultrasound and PAT
Purdue University researchers are developing a biomedical imaging system that combines ultrasound and photoacoustic tomography (PAT) technology. Here’s how it works: the PAT system transmits pulsed light that heats body tissue, causing the tissue to expand and generate an acoustic signal that can be detected and visualized by an ultrasound transducer. The system is specifically intended to diagnose cardiovascular disease, cancer, and diabetes, leveraging its ability to capture compositional information to help clinicians understand blood and lipid location.


CMS Takes Price Transparency into its Own Hands
CMS has set out to find a company to build and maintain the department’s new online healthcare price comparison platform as part of its continued effort to give consumers better pricing transparency. Americans are pretty poor at shopping around for healthcare and hospitals aren’t particularly eager to open up about their pricing structure, so although this platform would be a big step in the right direction, CMS has work to do in order to create awareness of the platform and drive adoption once it’s launched.



The Wire

  • Swiss medical imaging startup, Pristem, raised $14.4 million to fund the development of an X-ray system intended for use in low-income regions that generally have limited funding and poor service infrastructure. The Pristem GlobalDiagnostiX U-arm system is on track to launch in two years, combining low costs with an architecture that can withstand tropical climates and electrical supply interruptions, while still offering a user interface and internet connectivity to support teleradiology and remote maintenance.
  • Healthcare deep learning and analytics developer, Vyasa Analytics, joined the NVIDIA Inception program, which helps startups develop AI/DL and data sciences products by providing a range of NVIDIA resources (GPU discounts and early access, AI/DL expertise, AI/DL training, and content marketing). Vyasa’s Cortex deep learning software analyzes different sources of large healthcare datasets (text, images, quantitative data, chemical structures) to find unexpected patterns, relationships, and concepts.
  • A team of researchers found that handheld ultrasound combined with computer-aided diagnosis (CADx) software may allow non-radiologists to triage breast masses, which is particularly valuable in developing countries where access to mammography screening and radiologists are often limited.
  • Not sure if this is surprising to you. A survey from the University of Chicago reveals that 57% of Americans have been surprised by a medical bill, with 35% of respondents reporting being surprised by imaging services bills. However, physician services (53%), lab tests (51%), and other health care facility charges (43%) were the most common sources of surprise medical bills.
  • The University of Maryland’s Institute for Bioscience and Biotechnology Research (IBBR) launched separate collaborations involving its polyphosphazene (PPZ) polymer platform with UPenn and Texas A&M. The Texas A&M collaboration will focus on developing coatings for medical devices (e.g. stents and catheters) and the UPenn partnership will focus on developing medical imaging agent carriers.
  • Alliance HealthCare Services announced its partnership with Northeast Radiology, adding the New York and Connecticut-based practice (5 centers, 9 radiologists) to its Alliance HealthCare Radiology division (46 states, +1,100 hospitals, +100 locations, +600 imaging systems).
  • German researchers developed a compact hyperspectral imaging system able to capture 5D images, measuring multiple wavelengths of light and spatial coordinates as they relate to time, without requiring contact with the object being imaged. The researchers noted that the new system still requires significant research and listed medical imaging as only one potential use, in addition to very non-medical applications like identifying people in airports or measuring fruit ripeness.
  • Research firm Tractica forecasts that the global healthcare AI solution market will grow from $1 billion in 2017 to more than $34 billion by 2025, driven by demand to reduce healthcare costs and achieve care and efficiency improvements.
  • Immuno-oncology imaging company, ImaginAb, and Canada’s Centre for Probe Development and Commercialization (CPDC) launched a partnership, through which CPDC will manufacture and distribute ImaginAb’s clinical-phase PET imaging agent (Zr-89 IAB22M2C) to key global geographies.



The Resource Wire

– Future home of sponsored content.  Maybe yours.



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!