“While I had imagined what the images would look like for years, nothing prepared me for the incredible detail we could see on that first scan.”
UC Davis Department of Biomedical Engineering professor, Simon Cherry, on the first images produced by UC Davis’ new EXPLORER total-body PET/CT system.
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
- OpenMarkets – A marketplace for healthcare equipment, used by hundreds of hospitals and suppliers to buy and sell imaging equipment in the most efficient way possible.
- 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
Ezra’s Prostate MRI Plan
AI startup, Ezra, raised $4 million to support the launch of a new service that intends to replace painful prostate cancer biopsies with MRI scans. Although funding announcements for medical imaging AI startups is a weekly occurrence these days, Ezra’s business is unique, as it adopts some of the biggest trends seen in consumer industries (a subscription-based model, direct-to-consumer channel mechanics), and provides an imaging-based alternative for an important, but uncomfortable, screening process. Here’s how it works: At-risk men pay a $999 annual subscription fee that gets them a radiologist-reviewed MRI scan each year, plus access to medical staff and educational guides, and ongoing support if the MRI reveals signs of cancer. Ezra is doing this with a 50% gross margin using actual human radiologists, but will rely on its FDA-pending AI software (90% accurate) to automate the prostate cancer diagnosis process in the future, which will play a key role in making this $999 annual rate possible and allowing the company to scale (including scaling to other cancer types).
Full Body Image
After 13 years of planning and development, the first human images from UC Davis’ EXPLORER total-body PET/CT system debuted at RSNA this week, and early reports reveal an “astonishing” level of detail. These first images mark a key milestone for the EXPLORER system, which will be followed by its installation at UC Davis by mid-2019 and its eventual commercial launch by development partner United Imaging Healthcare of Shanghai. EXPLORER reportedly “captures radiation far more efficiently than other scanners,” allowing it to produce an image in as little as one second (40x faster than current PET scans) with a far lower dose (40x lower radiation). The ability to view what is happening in all the organs and tissues of the body simultaneously is pretty amazing, and could support a range of new diagnostic and research applications, making the EXPLORER an interesting system to watch.
Washington U’s Alzheimer’s Predictor
Researchers at Washington University developed a way to predict if patients will develop Alzheimer’s disease with MRI scans using diffusion tensor imaging (DTI). The researchers used DTI to access data on 61 patients’ white matter integrity, specifically studying fractional anisotropy (FA), which measures how well water molecules travel along the brain’s white matter tracts. As you may expect, the team found “quantifiable differences” in patients who later developed Alzheimer’s Disease (lower FA). It’s worth noting that this is the second MRI-based Alzheimer’s prediction breakthrough to come out of Washington University Mallinckrodt Institute of Radiology in as many months.
Google Absorbs DeepMind Health
Google announced plans to absorb Alphabet’s DeepMind Health business, as Google reorganizes and ramps-up its overall healthcare efforts. The star of the integration is DeepMind’s Stream mobile app, an AI-based clinical assistant that DeepMind developed for the UK NHS, and is about to be scaled to a much wider audience. The integration got the press’ controversy-watchers excited, as it gives Google access to a new level of personal information from largely unknowing patients. However, the integration is more notable for its role in Google’s growing focus on healthcare, coming as part of new Google Health CEO David Feinberg’s decision to centralize all of Google’s healthcare efforts — apparently including Alphabet subsidiaries like DeepMind. The integration also represents a more-forceful move from Google to control and then commercialize the output of its DeepMind subsidiary (at least the healthcare side of it), as a key part of its 2014 acquisition was an agreement that DeepMind would remain autonomous.
Enough Images
Research from Stanford University found that convolutional neural networks (CNNs) can accurately classify chest X-rays with a relatively modest number of labeled images (95% accuracy with 20k images), but a significant increase in labeled images only yielded “marginal” improvements (96% with 200k images). Before we get carried away and try to go too far below 20,000, the researchers’ CNN test with just 2,000 images resulted in a much lower 84% accuracy rate. The researchers suggested that this is good news for healthcare providers who would benefit from trained CNNs but don’t have access to hundreds of thousands of images for training, citing previous research that found CNNs plateau after approximately 60,000 images.
The Wire
- Clarius Mobile Health announced the launch of its Clarius AI “collective intelligence ultrasound platform,” intended to support the development of AI-based solutions for use with its Clarius app and Clarius Cloud. Noting the challenges with developing and using AI for ultrasound, Clarius AI focuses making it easier to acquire, manage, and label large amounts of ultrasound data that is then fed into AI frameworks such as TensorFlow or PyTorch for training and eventually used to develop Clarius-based solutions.
- Researchers from Detroit’s Henry Ford Hospital may have discovered a way to predict type 2 diabetes based on the brightness of ultrasound images of the deltoid muscle, which they believe shines brighter due to low glycogen levels caused by diabetic patients’ insulin resistance. In the 186-patient study, researchers were able to use shoulder ultrasound scans to identify 77.4% of 137 diabetic patients as either “suspected diabetic” or “diabetic,” while 100% of the 13 prediabetic patients were identified as either “suspected diabetic” or “definite diabetic.”
- Life Image and Dicom Systems announced a new interoperability partnership to improve access to clinical content within and across healthcare providers/systems. Linked together through their Google Cloud partnerships, Dicom Systems’ Enterprise Imaging Unifier Platform (used to ensure data compatibility in information systems) will be offered as part of Life Image’s Interoperability Suite (allows providers to exchange images, reports, data).
- Mayo Clinic researchers found that children may be better off not using lead aprons during pediatric chest CT scans, as the downside of using lead aprons (artifacts, infection, discomfort, repeat scans) may outweigh the minimal radiation dose reduction (0.2% – 0.7%) that the aprons provide.
- Philips announced the launch of IntelliSpace Discovery 3.0, the latest version of its radiology research platform intended to support the development and deployment of AI assets. The latest update adds research applications and tools to aggregate, normalize, and anonymize data, which can be visualized and annotated to train and validate deep learning algorithms. These algorithms can then be deployed as plug-in apps into research workflows to analyze new datasets and help facilitate clinical research.
- Radiology AI-focused and radiologist-led investment fund, Bold Brain Ventures, just launched with the goal of connecting AI startups with venture funding and radiologist expertise. The fund is managed by a pair of radiologists, includes solid list of radiology and AI advisors, and is currently seeking fund investors with a specific focus on (you guessed it) radiologists.
- Samsung NeuroLogica and MaxQ AI announced an agreement to offer MaxQ’s Accipio Ix intracranial hemorrhage (ICH) detection software integrated with NeuroLogica’s medical imaging platforms in the US and Europe. The software will specifically integrate with Samsung NeuroLogica’s OmniTom mobile 16-slice CT scanner as well as mobile stroke unit ambulances equipped with the company’s CereTom CT scanners. MaxQ AI previously revealed pending agreements with Samsung NeuroLogica, GE, and IBM Watson, so although not a surprise, this announcement at least gives more details on the partnership.
- Novarad launched its new NovaMG-Pro mammography software, intended for multi-modality radiologists that read both mammography and non-mammography studies, allowing them to use the same workstation and PACS interface for all breast-related studies (mammography, MRI and ultrasound).
- Konica Minolta announced the Japan launch of its Kinosis motion X-ray analysis workstation, used to view and analyze “X-ray in Motion” videos created with the vendor’s Dynamic Digital Radiography (DDR) technology. The Kinosis workstation’s image processing technology allows clinicians to view X-ray animations from rapidly-captured X-ray images, quantifies “movement” using graphs to support diagnosis (e.g. tracking diaphragm movement in COPD diagnosis), and captures signal value changes in lung tissue for screening before a CT or MRI scan (e.g. alveoli and pulmonary blood vessels).
- Intelerad Medical Systems expanded its partnership with national medical group and relatively major radiology services provider, Envision Physician Services (10m interpretations annually, 500 hospitals). The expanded partnership appears to make Intelerad’s enterprise imaging solutions available across Envision’s US hospital sites.
The Resource Wire
– This is sponsored content.
- A new study in JACR revealed that the rise of high-deductible health plans has led to greater patient concerns over imaging costs than ever before, while patient price shopping often leads to “confusion, misinformation, and opaqueness.” These are the exact patients who can be helped by the Medmo platform, which connects high-deductible patients with radiology centers, ensuring the lowest costs for patients and a profitable revenue stream for imaging centers.
- OpenMarkets recently welcomed several new ultrasound refurbishing companies, including Trisonics and Redstone Healthcare, giving radiology directors who work with OpenMarkets the flexibility to order new ultrasounds from brands like Philips and Mindray or evaluate lower-cost refurbished options using the same platform. To get connected to OpenMarkets’ ultrasound suppliers, Imaging Wire readers just need to describe their current and future needs here.
- Carestream highlighted the ongoing adoption of its OnSight 3D Extremity Systems, which provide orthopaedic specialists with high-resolution 3D images for more accurate diagnostic and treatment decisions, providing testimonials from a number of OnSight 3D customers.
- The Focused Ultrasound Foundation announced that Taiwan-based focused ultrasound manufacturer, NaviFUS, began its first clinical trial for glioblastoma, already treating its first three patients by using focused ultrasound to open the blood-brain barrier and treat the deadly cancer.
- 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 systems.
