What Happens in Brazil | Google’s AI | Agfa’s In Imaging

“The most pressing problem with the US health care system is not a lack of data or analytics but changing the behavior of millions of patients and clinicians.”

A JAMA Viewpoint from U Penn’s Ezekiel J. Emanuel, MD, PhD and UCSF’s Robert M. Wachter, MD, suggesting that AI may promise diagnostic breakthroughs, but the healthcare system would benefit the most if physicians and patients change their behavior.


Imaging Wire Sponsors

  • Carestream – Focused on delivering innovation that is life changing – for patients, customers, employees, communities and other stakeholders
  • Focused Ultrasound Foundation – Accelerating the development and adoption of focused ultrasound
  • Medmo – Helping underinsured Americans save on medical scans by connecting them to imaging providers with unfilled schedule time
  • Nuance – AI and cloud-powered technology solutions to help radiologists stay focused, move quickly, and work smarter
  • Pocus Systems – A new Point of Care Ultrasound startup, combining a team of POCUS veterans with next-generation genuine AI technology to disrupt the industry
  • Qure.ai – Making healthcare more accessible by applying deep learning to radiology imaging


The Imaging Wire


What Happens in Brazil, Stays in . . .
The Brazilian medical device scandal that brought down GE Healthcare’s Latin America CEO, along with Rio’s ex-Governor and Health Secretary, might become a lot more serious.

  • Here’s Why it’s Serious – The FBI is now reportedly investigating some of the over 20 medical device companies involved in the Brazilian kickback and price-fixing scheme (including GE, Siemens, Philips, and Johnson & Johnson) for potentially violating the U.S. Foreign Corrupt Practices Act (prohibits U.S. citizens and companies or foreign companies listed in the U.S. from bribing foreign officials). And it’s not just the FBI, as Brazilian federal prosecutors are also collaborating with the U.S. Department of Justice and SEC to support their investigation.
  • Here’s What’s AllegedBrazilian prosecutors originally alleged that the “cartel” of companies bribed public officials to win medical equipment contracts and then charged the government inflated prices (up to 8x higher) to cover the cost of the bribes. However, Brazilian prosecutors (and potentially FBI) now have evidence from court testimonies that additional schemes existed, “including alleged bribes paid by multinationals to obtain public contracts in Brazil.”

Some might say that this is how business is done in Latin America, and there’s some truth to that, but keep in mind that the U.S. and Brazilian joint prosecution of construction conglomerate Odebrecht in 2016 resulted in a $3.5 billion fine (world’s largest-ever compliance penalty) and years of bad press.


Google’s AI
The “AI beats radiologists” storyline got some serious attention this week, as the results of Google’s lung cancer CT screening algorithm quickly spread across mainstream media (Fox, NTY, WebMD, LinkedIn News, The Imaging Wire – all the big ones). The study, which Google first discussed at its I/O conference earlier this month, found that Google’s algorithm can use patients’ prior and current CT studies to assess cancer risks as well as or even better than radiologists. Here’s how the study worked out:

  • Google used 42,290 Low-dose CT scans from 14,851 patients (578 w/ confirmed cancer within 1yr) to create cancer detection and risk prediction models
  • 70% of the patients were used for training, 15% for tuning, 15% for testing versus six radiologists
  • The Google model achieved a 0.944 AUC on 6,716 NLST cases and performed “similarly” with a 1,139-case independent clinical validation set
  • The Google model performed as well as the six radiologists when patients’ prior and current CT images were both available (Google model = 0.96 AUC)
  • The Google model’s advantage was particularly strong when patients’ prior CTs were unavailable, beating all six of the radiologists (Google model = 0.93 AUC), with 11% fewer false positives and 5% fewer false negatives

Google is much more likely to serve as an AI technology provider than an end-to-end medical AI company and its AI team made it clear that this solution is not ready for clinical use. However, news like this can certainly get the attention of the healthcare industry (and the general public), especially given the growing belief that Silicon Valley is about to invade medicine, and considering that these results are really impressive.


Agfa’s In Imaging
Agfa-Gevaert confirmed that it plans to sell its Healthcare IT group’s Hospital IT and Integrated Care (HCIS) businesses, in a deal that will also include Afga’s Imaging IT operations in regions where the solutions are too integrated to separate (mainly Germany, Austria, Switzerland, France, and Brazil). After the acquisition, the Agfa Healthcare IT division will consist of its Imaging IT businesses in all remaining regions, which contribute about 83% of the company’s current Imaging IT revenue.

This deal will make Agfa-Gevaert once again a completely imaging-focused company. This may feel good from a nostalgia perspective, but it’s debatable whether this is the final step in the restructuring of Agfa’s healthcare business or if more healthcare-related deals are coming. Considering that the two medical imaging units are now separated from each other, from the HCIS unit, and from Agfa’s print divisions, additional deals are now a lot more feasible.


The Wire

  • Philips continued its image-guided portfolio expansion this week with the launch of its new IntraSight interventional applications platform, used to integrate intravascular imaging (iFR, FFR, IVUS), physiology, and co-registration tools. IntraSight integrates with Philips’ Azurion image-guided therapy platform, adding tableside touchscreen control, systems integration, data management, and remote service diagnostics.
  • A Thomas Jefferson University Hospital research team used data from 1,000 Medicare beneficiaries and found that elective myocardial perfusion imaging (MPI) volume fell by 36% between 2006 and 2016 (to 48 studies per 1k beneficiaries). The researchers suggested that these declines may have been due to “inappropriate or unnecessary self-referral” early in the study period (when reimbursements were higher), as private cardiologist offices were the source of the majority of the MPI utilization declines since 2006 and volumes from cardiologists in hospital outpatient settings increased.
  • Research from IMV Medical Information Division and published on AuntMinnie.com found that PET procedure volume in the US hit an all-time high last year (up 7% YoY to ~2,086,000 clinical PET scans), continuing PET’s 6% annual volume growth rate since 2013. However, comparisons versus 2013 don’t tell the entire story, as volumes were particularly low that year (believed to be due to reimbursement and program changes), with PET volume up a more modest 12.5% since 2011. With that, perhaps the more notable take-away from this research is the rapid adoption of PET/CT scanners, which grew from a 2% share of the fixed U.S. PET system installed base in 2001 to a 96% share in 2018 (the other 4% are PET-only and PET/MR scanners).
  • Life Image and Bialogics Analytics announced a partnership that will make Bialogics’ medical imaging business intelligence analytics platform available across Life Image’s global network of hospital and health systems. The partnership will give Life Image’s customers the ability to analyze their imaging operations and manage their operational efficiency.
  • The U.S. Senate just welcomed a companion bill to the U.S. House’s Medicare Access Radiology Care Act (MARCA) that seeks to allow radiologists to submit Medicare claims for imaging and non-imaging services performed by radiologist assistants in hospital and office settings. Given that CMS now allows RAs to perform a wider range of imaging and non-imaging services under radiologist supervision (as long as rads are in the same facility and available for assistance), billing for these services makes sense and it appears that congress agrees.
  • Philips announced the FDA clearance of its MRCAT MR-only radiation therapy planning solution for pelvic cancers (it’s already approved in Europe and Canada). The MRCAT (Magnetic Resonance for Calculating ATtenuation) solution provides both soft tissue and density maps using a single MR exam, making the time/cost-intensive use of both MRI and CT for radiotherapy planning unnecessary. Philips plans to expand its MR-only simulation portfolio to more applications in the future.
  • An Australian research team developed a device that splits light beams into modes (vs. pixels), allowing for more data to travel through optical fiber and potentially leading to faster internet and more detailed medical and space imaging.
  • ControlRad’s Trace dose-reduction fluoroscopy tech gained FDA clearance, making it the “only radiation reduction technology that integrates into mobile C-arms” available in the US. Trace can be retrofitted on existing mobile C-arms and combines a semi-transparent filter, tablet, and image processing technology to reduce unnecessary radiation by up to 89% without compromising image quality.
  • A new study from the American Cancer Society found that the first six states that expanded Medicaid experienced an 8.8% increase in colorectal cancer screening rates among low-income adults between 2012 and 2016 (42.3% to 51.1%), while the states that didn’t expand Medicaid saw a 3.8% increase in CRC screening (44.2% to 48.0%). However, the study found only a modest difference in breast cancer screening between the states that expanded Medicaid early and didn’t expand at all. These results suggest that screening rates do not increase quickly once Medicaid expansion is adopted, but greater improvements may occur later.

The Resource Wire

  • Did you know that imaging patients are most likely to no-show for their procedures on Mondays and Saturdays? By partnering with Medmo, imaging centers can keep their schedules full, despite the inevitable Monday no-shows.
  • This Qure.ai blog details the challenges it overcame with its qER Head CT algorithm, including developing the algorithm to support CT’s 3D images and high resolution as well as the validation hurdles it faced due to the low prevalence of abnormalities, the need to create a dataset enriched with positives, and steps required to support radiologist reading.
  • Focused Ultrasound Foundation founder and Chairman, Neal F. Kassell, MD, took to the foundation’s blog to emphasize the critical role of industry collaboration in its success and to encourage ongoing collaboration.
  • Nuance’s updated PowerScribe One radiology reporting platform now provides access to AI-powered diagnostic and decision-support tools within radiologists’ familiar workflow, giving them tools to improve efficiency and throughput, increase accuracy and specificity, and ensure evidence-based follow-up.

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.