Self-Referrals Act | AI-Aided Breast US | CDS’ Modest Savings

“Let’s cut & paste each other & say nothing useful.”

Radiologist, Saurabh Jha, on the wave of national “AI perspective” working papers that seem to be getting published on a weekly basis.



We’re honored to announce that Qure.ai is now an Imaging Wire Sponsor. The Qure.ai team and the deep learning tools they’ve developed are really impressive and we’re excited to share their message and achievements with you. The Imaging Wire is made possible by Qure.ai and the rest of our sponsors, so make sure to take a moment to check them out.


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.
  • Pocus Systems – A new Point of Care Ultrasound startup, combining a team of POCUS veterans with next-generation technology to disrupt the industry.
  • Qure.ai – Making healthcare more affordable and accessible by applying deep learning to radiology imaging.

The Imaging Wire

Curbing Self-Referrals
New federal legislation introduced to congress last week looks to address a Medicare self-referral loophole responsible for over $3 billion in wasteful spending each year, potentially impacting how advanced medical imaging services are ordered (along with anatomic pathology, radiation therapy, and physical therapy).

The Promoting Integrity in Medicare Act (PIMA) would prevent physicians from referring patients to receive these services if they have “ownership interest” and provide the services in their offices. This would adjust a current law that makes an exception for in-office services, but was primarily intended to allow self-referrals for less-complex services like X-ray and blood tests. PIMA argues that these four complex services (again: advanced imaging, pathology, RT, PT) are not typically performed at the time of the patient’s initial office visit and suggests that it will help ensure that decisions are made in patients’ best interests.

This may be a law that radiologist groups should also consider supporting. We’ve seen previous research that the highest-earning cardiologists attribute over 18% of their billings to imaging (and high-earning vascular surgeons, OBGYNs, and orthopedic surgeons attribute 9.6% to 11.6% to imaging). It’s quite possible that the threat of Medicare fraud could shift some previously self-referred orders back to radiologists.

AI-Aided Breast Ultrasound
Swiss researchers developed a deep convolutional neural network (dCNN) algorithm that can detect and classify lesions on breast ultrasound images as effectively as radiologists, noting that dCNNs like this may help reduce variability and false positives, while potentially replacing CAD systems. The researchers trained the dCNN with 1,019 breast ultrasound images from 582 patients and tested it against a pair of internal (101 images) and external (43 images) datasets that were also evaluated by two radiologists. The algorithm outperformed the radiologists when:

  • Differentiating BI-RADS 2 from BI-RADS 3–5 lesions – dCNN: 87.1% internal set and 93% external set; Rads: 79.2% internal and 95.3% external
  • Differentiating BI-RADS 2–3 from BI-RADS 4–5 lesions – dCNN: 93.1% internal set and 95.3% external set; Rads: 91.6% internal and 94.1% external
  • AUC – dCNN: 83.8 internal set and 96.7 external set; Rads: 84.6 internal and 90.9 external

This new dCNN joins a growing list of algorithms intended to support breast imaging, but it’s positioning as a breast ultrasound tool is notable, especially following new research suggesting that supplemental breast ultrasound is in need of significant accuracy improvements in order for its benefits to outweigh its risks.

CDS’ Modest Savings
Research from a team of MIT Sloan and Aurora Health Care economists found that the adoption of clinical decision support (CDS) systems may reduce certain advanced medical imaging scans by about 6%, but may only “modestly increase the appropriateness of high-cost imaging.”

The yearlong CDS trial at Aurora Health Care involved 3,511 providers (physicians, DOs, NPs, PAs), with half randomly assigned to use the CDS tool and the other half operating without any decision support guidance. As noted above, the CDS group reduced their “targeted” (CT and MRI) order volume about 6%, with CT representing nearly 80% of all eliminated scans. However, the test group’s high-cost and low-cost imaging order volumes each only fell by 2% (statistically insignificant), due in part to a CDS software function that suggests alternative imaging procedures in the event of an inappropriate order.

There’s no shortage of imaging overuse and CDS effectiveness studies (like this, this, and this), but the new MIT study represents the first large-scale randomized research on this topic. The study is also pretty timely, given that CMS will require CDS use for imaging orders to qualify for Medicare reimbursement starting next year.

Four Steps to Price Transparency
In light of the ongoing push towards healthcare price transparency and the challenges associated with clearly communicating medical imaging costs, radiology consultant Dr. Nicolas Argy took to Radiologybusiness.com to share four ways that radiology departments can improve price transparency. Argy suggested that radiologists should:

  • Focus on making sure quality outweighs costs.
  • Establish clear direct-to-consumer pricing for common imaging procedures so patients know what to expect . . . or create sliding scale pricing based on patients’ need and budget.
  • “Invest in dedicated personnel and resources to educate patients” and encourage shared decision making.
  • Negotiate pricing contracts with large employers and insurance brokers.

There’s a lot that goes into imaging costs, significant price variability between providers, and plenty of room for patients to improve their role as healthcare consumers, making efforts like this worth considering. It’s better than relying on online price lists for transparency, anyway.


The Wire

  • Fujifilm Canada announced the creation of its new Fujifilm Canada Medical Systems division, revealing plans to “greatly expand the commercialization” of the company’s imaging systems in Canada and add to its Canadian product portfolio with new DR and women’s health solutions. Fujifilm Canada Medical Systems will continue to work with the company’s longtime Canadian distributor, Christie Innomed, although it’s quite possible that this Northern expansion would also include a greater direct sales presence.

  • Parascript rolled-out a new payment model for its AccuDetect CAD system that it states cuts mammography CAD costs by “literally 90%,” while also improving performance. That’s a pretty compelling value proposition. Parascript is rightfully trying to make sure its new cost advantage is clearly communicated, combining this announcement with new “Superior CAD at 10% of the Cost” banner ads on AuntMinnie.com and perhaps other sites. However, at least for now, it’s not clear how this new payment model achieves these cost savings.

  • Butterfly Network’s Butterfly iQ portable ultrasound system gained CE Mark clearance, allowing the much-hyped system to expand to Europe (plus Australia and New Zealand), where its now available for pre-order. The sub-$2,000 Butterfly iQ has generated impressive buzz inside and outside of healthcare since launching in the U.S. last year and the system’s availability in the 30 CE Mark countries represents another major milestone.

  • Konica Minolta announced the Japan launch of its new SONOVISTA GX30 pelvic ultrasound, replacing the SONOVISTA FX premium edition after five years on the market with a new self-developed design (believed to be previously Siemens-developed). The new SONOVISTA GX30 launches with a range of improvements including a smaller size, a new rotating control panel (allowing wide-angle viewing), improved image quality with a new 220 degree viewing angle, and an enhanced user interface. Konica Minolta’s SONOVISTA ultrasound systems have traditionally only sold in Japan, and this may remain true for the SONOVISTA GX30, however the adoption of a new self-developed design could create opportunities to expand the system to other markets.


The Resource Wire

This is sponsored content.

  • Yale University research reveals that the average patient drives past SIX lower-priced providers on the way to an imaging procedure, due in large part to patients’ and physicians’ limited cost consciousness. Medmo helps address this issue by letting patients enter the price they can afford for their scan, then books them at a nearby imaging center willing to accept that price.

  • POCUS Systems has developed a low-cost, hi-resolution, and AI-enabled POCUS unit, designed to diagnose specific MSK abnormalities and eliminate cost, size, and training barriers.

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.