fbpx

Digital Twins | Roaring 30s | Generalizable Evidence

“We’re inching toward that watershed moment.”

Stanford’s Fei-Fei Li on imaging AI’s slow progress towards the moment when it improves diagnosis AND radiologists’ jobs.


Imaging Wire Sponsors

Arterys | Bayer Radiology | Canon Medical Systems | GE Healthcare
Healthcare Administrative Partners | Hitachi Healthcare Americas
Novarad | Nuance | Riverain Technologies | Siemens Healthineers
United Imaging | Zebra Medical Vision



The Imaging Wire


Q Bio’s Big Debut

Q Bio made quite a splash this week when the Silicon Valley imaging startup unveiled its “digital twin” platform, its whole-body scanner, and another round of funding. Here are some details:

  • The Digital Twin – The new Q Bio Gemini platform calculates patients’ anatomy/biochemistry data and displays their “physiological state in the form of a digital twin,” potentially allowing clinicians to proactively monitor and treat patients.
  • Q Bio Mark I – The starting point for these digital twins is the company’s Q Bio Mark I whole-body scanner, which captures patients’ health and personal risk data in just 15 minutes, and is reportedly far less cost prohibitive than previous digital twin scanners.
  • Q Bio’s $80m – After raising $58m between 2016 and 2020 (most of it in 2020), Q Bio completed another funding round that brought its total above $80m. Q Bio’s funding sources lend it quite a bit of health and tech credibility, as this latest round came from Kaiser Foundation Hospitals and its previous Series B round was led by Andreessen Horowitz.
  • The Takeaway – Proactive imaging is still pretty niche and the use of digital twins still feels very futuristic. Q Bio seems to be making progress towards changing that.

Bayer Cuts Contrast in East Texas

See how East Texas Medical Center reduced its abdominal CT contrast volumes by 30% after adopting Bayer Radiology’s contrast dose management software.
– This is sponsored content.



AI’s Roaring 30s

AI heavyweights, Fei-Fei Li and Andrew Ng, just broke the news that they don’t expect any major healthcare AI breakthroughs in the next few years but they still see it transforming healthcare in the 2030s. Here’s why:

  • First Slow, Then Fast – Technology “progress tends to happen slowly and then very quickly,” and the rate of healthcare AI progress is still slow.
  • Prove it First – Healthcare AI still has to prove that it can make a fundamental impact on patients’ care and healthcare workers’ work, which is why they’re having more AI students shadow clinicians.
  • Generalization Barrier – Another clear step towards “providing it” is performing consistently across different hospital systems, which is why they advocate a shift from model-centric AI to data-centric AI.
  • Slow is OK – Even though they see healthcare AI at an earlier stage than some might prefer, they do see AI “blossoming” over the next decade before it has a chance to truly “transform healthcare.”

Canon’s Spectral Heartbeat CT

Canon Medical’s Aquilion ONE / PRISM Edition CT system now supports one-beat cardiovascular imaging, allowing healthcare providers to acquire whole-heart spectral images in one heartbeat (0.275 seconds).
– This is sponsored content.



The Wire

  • Avicenna.ai’s Generalizable Evidence: A new UCI-led study found that Avicenna.ai’s stroke detection tools can effectively help radiologists detect ICH and LVO strokes across “a wide variety of hospital systems” and multiple scanner brands. In this retrospective/multicenter study, the researchers used the Avicenna.ai tools to analyze 814 non-contrast CTs for ICH (w/ 95.6% accuracy, 91.4% sensitivity, 97.5% specificity) and 378 CTAs for LVO (w/ 98.1% accuracy, 98.1% sensitivity, 98.2% specificity).
  • POCUS for Dyspnea: The American College of Physicians just recommended the use of point-of-care ultrasound for patients with acute dyspnea (shortness of breath) in both emergency and in-patient settings. The recommendation calls for POCUS use, in addition to the standard diagnostic pathway (patient history, physical exam, blood labs, chest or cardiac imaging, and ECG), when there is diagnostic uncertainty.
  • The Pandemic’s Screening Backlog: The COVID-19 pandemic left the United States with a massive 9.4m-exam cancer screening deficit, requiring new public health efforts to catch up. That’s from a new JAMA Oncology study that measured pandemic-related screening declines and rebounds, estimating current deficits of 3.9m breast cancer screenings, 3.8m colorectal cancer screenings, and 1.6m prostate cancer screenings.
  • X-Ray ARDS AI: University of Michigan researchers developed a DL algorithm that detects Acute Respiratory Distress Syndrome (ARDS) in chest X-rays “with expert-level accuracy.” Faced with a limited number of labeled ARDS scans, the team first used 450k CXRs to develop a CNN to identify common radiologic findings (but not ARDS) and then trained it to detect ARDS using 8k CXRs that were annotated for ARDS (aka transfer learning). Using an internal test set of 1,560 CXRs from 455 patients with acute hypoxaemic respiratory failure, the CNN detected ARDS with a 0.92 AUROC (vs. >6 physicians’ 0.93 AUROC).
  • Vantage Galan 3T’s AiCE Expansion: Canon Medical Systems announced that its Vantage Galan 3T MRI can now use Canon’s AiCE deep learning image reconstruction solution for body applications (previously only brain and knee). The expanded coverage adds a range of body exams “from prostate to shoulders,” including all joints, cardiac, pelvis, abdomen, and spine applications, allowing the Vantage Galan 3T to use AiCE reconstruction with 96% of exams.
  • Butterfly & Sientra: Butterfly Network announced a partnership with breast implant company, Sientra, with the goal of making the scanner part of plastic surgeons’ implant monitoring workflows. The partnership will make the Butterfly IQ+ available through Sientra, while also integrating the handheld ultrasound into Sientra’s education platform and peer-to-peer network. Butterfly has actively worked to expand the Butterfly IQ+’s use cases, but this might be the first specialist-targeted channel partnership we’ve seen from them.
  • Linking TBI and AD: USC researchers found that patients with mild traumatic brain injuries (mTBI) and patients with Alzheimer’s Disease have similar brain MRI patterns, suggesting that post-TBI cognitive impairment could predict Alzheimer’s risk. Using MRI scans and a machine learning algorithm (n = 33 w/ TBI, 66 w/ AD, 81 healthy), the researchers found “substantial similarities” in the two patient groups’ white matter and gray matter.
  • Appendicitis’ Independent Outcomes: A new JACR study found that suspected acute appendicitis outcomes are not affected by the skill/experience of the individual sonographer that performs an appendix ultrasound or the radiologist that reads it. That’s from a new Cincinnati Children’s Hospital-led study (n = 9,283 exams, 31 sonographers, 31 radiologists) that showed plenty of variability among the individual sonographers and radiologists, but found that other non-clinician factors actually predicted hospital admission and appendectomies (e.g. temperature, white blood cell count, male gender, tenderness, hospital type, acute appendicitis in ultrasound scan).
  • PIXYL’s €2.2m: French neuroimaging AI startup, PIXYL, completed a €2.2m funding round that included contributions from 32 radiologists. PIXYL will use the funding to support the European launch of its Pixyl.Neuro product, its FDA regulatory efforts, and ongoing product development.
  • MRS Breakthrough: University of Illinois researchers developed an open-access tool that could make it easier to evaluate infants’ brain health by streamlining brain metabolite measurements. The team used proton magnetic resonance spectroscopy scans from 140 infants to define normal/abnormal metabolite levels and key metabolite relationships, allowing clinicians to quickly assess metabolite concentrations.

How Hemet Global Medical Center Ditched the Disk

Learn how Hemet Global Medical Center improved efficiencies, safety, and security when it moved to the Nuance PowerShare Network and #ditchedthedisk.
– This is sponsored content.



The Resource Wire

– This is sponsored content.

  • Whether you’re exploring new research questions or identifying the best clinical approach, a 106 cm axial view could be a game changer for you and your clinical team. See how with Siemens Healthineers’ Biograph Vision Quadra PET/CT.
  • Easy access to patient records, reduced inefficiencies and costs, improved collaboration and compliance, and enhanced security. These are just a few of the benefits of Novarad’s enterprise imaging solution detailed right here.
  • This Hitachi Healthcare blog outlines the criteria providers should consider for their image and reporting platforms, and how the Hitachi VidiStar platform’s features, service, and vendor collaboration meet providers’ needs.
  • United Imaging took another step in its growth strategy, announcing its first U.S. channel partnerships with Radon Medical Imaging (mid-Atlantic), Imaging Solutions (mid & northern U.S.), and Medimax (Puerto Rico).
  • A new study in European Radiology highlighted Riverain Technologies’ ClearRead Xray – Detect as one of just two imaging AI products to achieve the FDA’s most stringent premarket approval level. See how they measured up against the other 99 AI tools here.
  • Check out this Imaging Wire Q&A, where Arterys CEO John Axerio-Cilies, PhD discusses medical imaging’s AI and cloud evolution and how Arterys works with its Center of Excellence partners to make AI real.