“Artificial Intelligence’s pragmatic laziness is enviable.”
UPenn radiologist and Radiology Twitter star, Suarabh Jha MD, on AI’s tendency to take the easiest path to a finding.
If you haven’t done so already, please take a moment to consider your colleagues who might find this newsletter valuable and let them know about it. Once again, we’re donating $2 to these Australian Bushfires charities for each new subscriber that you refer.
If you’re a subscriber, please consider using your personal referral link to introduce your industry friends to The Imaging Wire (You can find your personal link in your latest newsletter).
If you haven’t received a referral from one of your colleagues yet, just use this link to sign up.
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
- 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.
- 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
- Qure.ai – Making healthcare more accessible by applying deep learning to radiology imaging
The Imaging Wire
Healthcare’s unicorn-status startups (> $1b valuations) may have plenty of funding and buzz, but a recent paper out of Stanford revealed that they are largely ignoring the time-honored step of submitting peer-reviewed papers to validate their breakthrough technologies.
The Data – The 18 current healthcare unicorns published a total of 425 papers (only 8% of which were highly cited), while five of the companies published 0-1 papers (the 29 previous healthcare unicorns had similar publishing records). This may actually be common across all healthcare startups, as the paper suggested that smaller firms publish even fewer peer-reviewed papers.
What’s So Wrong with That? – The Stanford researchers behind this paper were concerned with major healthcare startups ignoring peer reviews, calling the public journals an “indispensable” way to validate new products/technologies and noting that new products that are only supported by internal data may not be “trustworthy.” What else would a team of academics say?
Radiology Twitter Reacts – This may not be an everyday topic for most folks, but Radiology Twitter took an interest in this story. Some called peer-reviews a necessity, others noted how the peer-review process conflicts with the pace of fast-moving startups, and some suggested that this is more of a symbol of scientific journals’ obsolescence rather than a flaw in startup culture.
HALO’s Big Move
Medical imaging center startup HALO Diagnostics (HALO Dx) took its first step towards its goal of “changing the face of radiology from the ground up,” with its acquisition of three Central and Southern California-based imaging center companies (10 total locations). It’s too early to say whether HALO Dx will actually change the face of radiology, but the company is entering the imaging center arena with an intriguing blend of ambition and funding, making it worth a look.
M&A and P – HALO’s big California debut came from simultaneously-announced mergers with Desert Medical Imaging (Indian Wells, Indio, and Palm Springs), Open System Imaging (Chico, El Centro, Palm Desert, Pasadena, and Tustin), and Chico Breast Care Center (Chico), as well as a strategic partnership with major Midwest imaging center group ProScan (26 imaging centers).
HALO’s Big Goals – These could prove to be the first of many acquisitions for the disruption-minded (and PE-backed) company, which unveiled plans to “become one of the top radiogenomics networks in the United States” and significantly expand its number of imaging centers by the end of 2020. HALO Dx is also using its Coachella Valley facility (via the Desert Medical Imaging acquisition) to developing new technologies that it will expand across its imaging center network.
Who is HALO Dx? – Good question. HALO Dx is pretty new (it was incorporated in mid-2019 and just announced “Hello World” to Linkedin last month) and spawned from tech/AI-focused Silicon Valley PE firm, Halo Growth Ventures. HALO Dx is led by a team of radiology outsiders (McKinsey, VMware, and Google, private equity) and most of its radiology leadership appears to have come from these new acquisitions. That said, HALO’s outsider status and disruptive intentions are a big part of what makes this announcement interesting.
- Philips announced enhancements to its year-old Philips Zenition mobile C-arm platform, adding a new Table Side User Interface and the integration of intravascular ultrasound (IVUS) for peripheral vascular procedures. The 12.1-inch Table Side User Interface allows clinicians to operate the C-arm inside the sterile field, while the integration of IVUS “allows physicians to visualize blood vessels from the inside out,” streamlining the OR and improving workflow.
- New research from UPenn reveals that U.S. emergency departments are scanning patients with lower back pain at the “high” rate of one out of every three visits. The researchers reviewed 134,624 ED visits by 18 to 64-year-old patients with low back pain between 2011 and 2016, finding that imaging was used in 44,405 of the visits (33.7% of all visits; 30.9% x-ray, 2.7% CT, 0.8% MRI), although imaging use fell from 34.4% to 31.9% of all visits over the course of the study period.
- Florida health system AdventHealth acquired the assets of Tampa Bay Mobile Mammography, which supports 700 locations around Tampa. The acquisition also included three mobile mammography vehicles equipped with Hologic 3D DBT systems that are collectively capable of scanning 150 people per day.
- The University of Maine received a patent for a computational approach that could help detect early signs of breast cancer by characterizing the “roughness” (a measure of spatial density fluctuations) of each breast tissue region in a mammogram image. UMaine suggests that by measuring each region’s roughness signature, the method could help predict and proactively treat cancer better than current CAD methods.
- A Colorado jury awarded a local woman with a $1.2 million negligence judgment against her previous attorney for mishandling her radiology malpractice lawsuit. The original lawsuit alleged that the woman’s radiologist overlooked her benign brain tumor in an early MRI scan when it could have been removed via a low-impact surgery, leading to three risky brain surgeries. Her attorney then failed to file the malpractice suit before the statute of limitations expired and then repeatedly lied to cover up his mistake, leading to the recent negligence judgment.
- Case Western’s Myeliviz PET agent for the diagnosis of multiple sclerosis gained FDA approval for its first human clinical trials. Myeliviz is the first marker to directly target and bind to myelin (the sheathing surrounding nerves affected by MS), potentially allowing physicians to use PET scans to measure MS’ existence and progression by evaluating myelin damage. Myeliviz PET imaging could also supplement or maybe even replace MRI, which indirectly images myelin, and has traditionally been the standard for MS monitoring.
- A new paper from urgent care services company Experity Health revealed that 92% of U.S. urgent care clinics perform radiology exams and 10.7% of urgent care clinic visits between July 2018 and June 2019 included a radiology exam. X-ray was the most widely used urgent care modality (88% of all scans), followed by EKG (9.2%) and CT (2%). Among other things, Experity Health provides urgent care clinics with teleradiology services, explaining their interest in highlighting this trend.
- New research from MGH found that short-interval breast MRI follow up screenings (e.g. at 6, 12, 24 months) for patients with BI-RADS category 3 lesions help identify cancers while they are still in early stages and help avoid biopsies for benign tumors. The review of 6,672 MRI screening exams (n = 3,214 patients, 202 exams = BI-RADS 3, 13 that developed into cancers) found that by scanning the BI-RADS 3 women in short intervals, 12 of the 13 cancers were identified in stages 0 or 1 and 189 patients with benign tumors avoided biopsy.
- Lithuanian medical imaging solutions company, Oxipit, unveiled its new ChestEye Negative service, which produces preliminary reports for chest X-rays that don’t show abnormalities. The ChestEye Negative service is intended to streamline radiologist workloads and simplify large-scale screening projects (e.g. TB screening, refugee center screening) by allowing radiologists “to focus on X-rays where pathologies are suspected.”
- Philips revealed plans to spin-off or sell its $2.53 billion home appliances unit over the next 12-18 months, increasing its healthcare focus and perhaps giving it more healthcare R&D and M&A funding to work with. This is part of an ongoing trend for Philips, which unloaded its consumer electronics business in 2013 and began the process of selling its lighting business two years later, each time sharpening its focus on healthcare.
- theverge.com joined in on the recent wave of popular media articles focused on healthcare AI’s potential problems, this time digging into how AI could drive cancer overdiagnosis and how many AI studies unnecessarily feed into the “machine versus doctor” narrative. This story comes about a week after wired.com warned against AI exacerbating breast cancer overdiagnosis, and although these two articles don’t exactly include new ideas, they do introduce AI’s challenges to new public audiences.
- A new study from UCSF found that intraluminal thrombus (ILT; a blood clot on the aortic wall) may serve as a marker for future abdominal aortic aneurysm (AAA) growth, suggesting that follow up imaging protocols might need to be changed to include identifying and monitoring AAA patients with ILT (they currently don’t take ILT into account). The retrospective study of 225 men with AAA who underwent contrast–enhanced CT at baseline and CT (n = 207) or black-blood MRI (n = 18) at follow-up, finding that AAAs with ILT started larger (4.1 cm vs. 3.4 cm diameter) and grew faster (2.0 mm/y vs. 1.0 mm/y) than AAAs without ILT.
- Zebra Medical Vision and Clalit Research Institute teamed up to demonstrate how an AI algorithm can identify patients’ risk of osteoporotic fractures using their chest or abdomen CT scans. The new Zebra-trained AI tool calculates fracture risk scores by analyzing three bone imaging biomarkers (along with age and sex from CT metadata) as accurately as the Fracture Risk Assessment Tool (FRAX) without using data from bone mineral density scans (AUCs: 0.709 Zebra vs. 0.691 FRAX), while achieving even greater accuracy when the AI algorithm is combined with the FRAX tool (0.723).
The Resource Wire
– This is sponsored content.
- UPenn radiologist and Radiology Twitter star Suarabh Jha MD’s second installment on thehealthcareblog.com walked through all the discoveries that happened along Qure.ai’s path towards developing their qXR solution for TB screening.
- Nuance recently published a revealing Q&A with Zebra Medical Vision co-founder and Chairman, Eyal Gura, detailing his inspiration to start Zebra-Med, the Zebra-Med “All-In-One” (AI1) bundle’s main benefits (it’s a “one-stop-shop” solution with flat/transparent fee structure), and the benefits of the Nuance AI Marketplace.
- Did you know that 3D images and weight-bearing exams can support faster, more accurate diagnoses? This Carestream blog details the benefits of 3D Extremity CT imaging.
- Sunnybrook Research Institute researchers recently became the first to use focused ultrasound to deliver a molecule to the brain to revive the function of neurons vital to learning and memory in mice with Alzheimer’s 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.
- Yale University research reveals that the average patient drives past six lower-cost 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 what they can afford for their scan, then booking them at a nearby imaging center willing to accept that rate.