“This is not peanuts.”
Philips CEO, Frans van Houten, on the $69 million profit hit the company is expecting in 2019 as a result of the US-China trade war and the “hundreds of millions” of euros worth of production the company is shifting to-and-from the US and China to minimize the impact of the tariffs.
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.
The Imaging Wire
AI Investments Still Surging
A new report from Signify Research confirms that the growing buzz around medical imaging AI has coincided with significantly increased investments. Signify Research reported that investments in AI-enabled medical image analysis companies between 2014 and 2018 surpassed $1.2 billion, including $580 million in 2018 alone. HeartFlow continues to dominate fundraising, generating $476.6 million over the four years and $240 million in 2018, followed by VoxelCloud (~$80m 4yrs, $50m 2018), InferVision (~$73m, $46m), Deepwise (~$51m, $22.6), and Zebra ($50m, $30m). As expected from 2018’s hefty investment numbers, Signify found that company launches and associated early-stage investments have declined since a 2015-2017 peak (29 early-stage deals in 2017 vs. 15 in 2018), while later-stage investments doubled in 2018 and are expected to continue to lead 2019’s funding activity. The most investor dollars are flowing to US and Asia-based companies and to companies focused on general imaging and cardiovascular imaging applications (with the US and CV numbers helped by HeartFlow’s massive investments), while investments in EMEA AI companies appear to be lagging (especially outside of Israel). The Signify report provides some solid details and helpful visuals and is worth a look if you’re interested in AI investments.
The Toronto Star’s Opinion section became home to a debate on whether a move to protect the income potential of Ontario’s radiologists was justified. The op-ed faceoff started with a piece by former Ontario Deputy Health Minister (and former surgeon), Dr. Bob Bell, criticizing local radiologists for trying to leave the Ontario Medical Association to protect their own financial interests even though they’ve already enjoyed extraordinary income growth (+ 163% avg.) and labor declines (- 12% days) over the last 20+ years, owing their reduced workload to taxpayer-funded technologies. Within two weeks Canadian radiologist, Dr. Andrew Chung, clapped back against what he views as the media’s frequent and unnecessary targeting of radiologists, suggesting that an overall misunderstanding of the specialty (beyond its high-income potential) has made radiologists “convenient bogeymen against whom it is easy to sway the public and even our fellow physicians against.” Dr. Chung pointed to radiologists’ growing responsibilities for diagnostic interpretations and image-guided procedures (both volume and complexity) as evidence of radiologists’ importance, which can be underestimated because of rads’ behind-the-scenes status and misunderstandings regarding how/whether new technologies are actually simplifying radiology. This type of specialist rivalry may be common in cafeterias and physician message boards, but taking it to the op-ed page is unique and its probably a path worth avoiding.
Imaging Order Errors
Imaging order errors landed in the middle of a debate over self-reporting, EHR functionality, and clinician accountability, after emergency department doctors at Rhode Island Hospital were served with subpoenas for medical misconduct and the hospital received a consent decree requiring it to report any incidents or near-misses. The Rhode Island Department of Health’s efforts to discipline the provider and the ED doctors for four imaging errors (wrong-side mammogram, wrong-patient angiogram and brain angiography, wrong-vertebra vertebroplasty… all in under a month) gained national attention. The medical community was quick to point out that the RI DOH’s penalties may dissuade clinicians from self-reporting in the future and don’t address the underlying issues that led to these errors (EHR complexity, high ED volume), while others highlighted the need to reduce the high number of EHR-related imaging order errors.
Russian researchers developed a new AI-based CT scan lung cancer diagnostic system that can mark images for pathology and distinguish between malignant and benign tumors in 20 seconds. The appropriately named Doctor AIzimov solution (get it, “AI”zimov) uses randomly selected segments of CT images connected by chords, then analyzes nodules “in the form of compact and simple histograms” (vs. entire complex CT images), to reduce the diagnostic heavy lifting (CTs can be 1GB). Doctor AIzimov will be tested at the St. Petersburg Clinical Research Center this year and expand to other medical institutions going forward, with plans to eventually support other organs and additional modalities (specifically ultrasound and X-ray).
Two Sides on AI’s Med Student Impact
The debate over how the threat of AI is influencing the career decisions of would-be radiologists heated up over the last few weeks. It started when Ross University medical student, Kush Purohit, published a rebuttal to a previous study from a group of University of British Columbia radiologists that revealed a high level of concern about the future of radiology among Canada’s med students. Purohit assured that although AI may cause anxiety, interest in radiology among medical students is still growing due to greater awareness, the appeal of IR, and radiology’s lifestyle benefits. Purohit backed-up this claim with the fact that 6.9% of all National Residency Match Program applicants applied to diagnostic radiology programs in 2017/2018, marking the specialty’s highest percentage since 2010 and representing a 38.8% increase since 2014/2015. The Canadian group was quick to issue a response to Purohit, noting that the 2014/2015 comparison period followed five years of application declines, and metrics such as radiology job competitiveness (71% in 2018 vs. 89% in 2009) and the radio of students placing radiology as their top specialty (1.4 in 2018 vs. 0.9 in 2001 – lower is more competitive) are far below their historical peaks.
- RadNet expanded its California presence with its acquisition of Bakersfield-based radiology practice Kern Radiology (4 offices, 19 radiologists), expanding its already-massive presence (341 locations, 500 radiologists, 6 states). Kern will apparently merge with RadNet’s existing Bakersfield area practices (specifically Truxton Radiology), leading to the layoffs of 173 Kern employees, although some may be rehired. The acquisition comes about nine months after RadNet made a similar move to expand in inland California, when it acquired five imaging centers in Fresno and merged them with its two existing Fresno centers.
- Philips Healthcare is partnering with South Korean imaging AI firms, Lunit and Vuno, with the goal of developing AI-based solutions that they will eventually commercialize globally, using the Philips IntelliSpace Discovery research platform to develop and deploy the algorithms. In addition to serving as an example of how IntelliSpace Discovery expands Philips’ AI presence, it is also notable that Lunit has been effective in partnering with some of the biggest names in imaging, as the company announced partnership with Fujifilm in Mexico just a few months ago.
- Lithuanian medical imaging solutions company, Oxipit, announced the CE clearance and pilot availability of its AI-based ChestEye chest x-ray computer aided diagnosis platform. ChestEye produces preliminary radiology reports for 75 common radiological findings (at 93% AUC), localizing the findings on a radiograph as a heatmap and producing a standardized text report that they claim cuts time per patient by 30% and reduces errors by up to 50%.
- University of Michigan researchers quantified the benefits of well-built price transparency resources, finding that a New Hampshire-run healthcare cost site helped reduce medical imaging costs by 5% for patients (~ $7.9m) and 4% for insurers (~ $36m) over five years, with high-deductible patients achieving “almost double the savings.”
- Canon Medical Systems launched its new Global Illumination medical imaging rendering technology at Arab Health last week, intended to help clinicians visualize and manipulate photorealistic anatomical images in real-time for improved anatomical understanding (by both clinicians and patients). The new technology is part of Canon’s Vitrea enterprise imaging platform produced by Canon subsidiary, Vital Images (acquired by Canon in 2016).
- Researchers at Texas Children’s Hospital were able to classify acute and non-acute pediatric elbow x-rays using a convolutional neural network approach by viewing an entire series of patient images together, similar to how a human radiologist reviews multiple images of each patient. The study (n= 20,350 training reports, 1,106 validation reports) achieved an AUC of 0.95 with 88% accuracy, 91% sensitivity, and 84% specificity, with the greatest challenges being with elbow effusions (62/77) and “other” abnormalities (147/184).
- Just over two months after first suggesting that GE may look to sell as much as 49.9% of GE Healthcare stock at its IPO, GE CEO Larry Culp added more weight to this suggestion, revealing that the company now expects “to monetize just up to under 50 percent of our healthcare business,” raising about $50 billion. This far exceeds the 20% stake GE originally planned to make available, but still keeps the IPO’s tax-free status, and is an understandable move given GE’s liquidity issues and immense shareholder pressure.
- Singaporean researchers unveiled an object detection CNN that can detect and localize fractures on wrist x-rays, representing an evolution from traditional CNNs that are only able to detect fractures (not localize them). The study used 7,356 wrist radiographs (90% for training, 10% validation) and was able to correctly detect and localize 91% of all radius fractures (310/340, frontal view) and 96% of all ulna fractures (236/245, lateral view).
- iSchemaView announced upgrades to its RAPID stroke imaging software, including an update to the RAPID APP (allows viewing of raw Non-Contrast CT images and CTA source data), new custom notification features for any RAPID module (text, email, app notifications), the added ability to customize the delivery of critical Personal Health Information (limits content based on regulations, policies, or preferences), and notification support for RAPID’s latest indication for thrombectomy.
- Philips revealed that the trade war between the US and China has forced it to shift “hundreds of millions” of euros worth of production from China to the US… and from the United States to China… to avoid tariffs during the first half of 2019. Even after these “serious changes to [its] supply chains,” Philips still expects to lose $69 million from its 2019 profits as a result of the tariffs.
The Resource Wire
–This is sponsored content.
- A new Gallup poll revealed that 13.7% of American adults are uninsured, representing the highest rate since 2013 when it was 18% (before most of ACA went into effect). These are the exact patients who can be helped by the Medmo platform, which connects high-deductible or uninsured patients with radiology centers, ensuring the best value for patients and a profitable revenue stream for imaging centers.
- Carestream cleaned up in KLAS’ 2019 awards, earning Best in KLAS VNA/Image Archive, Best in KLAS PACS Middle East/Africa, and Category Leader in Universal Viewer for the second consecutive year.
- POCUS Systems is approved as a Veteran Owned Business with the US Government Office of Veterans Business Development, paving the way for partnerships with the federal healthcare delivery system.
- Focused Ultrasound Foundation-funded researchers at the Ohio State University College of Medicine have begun a clinical trial investigating the use of focused ultrasound to control medication-refractory lobe focal onset epilepsy in adult patients.