fbpx

Aduhelm Approved | AI Education | Undisclosed Interests


“This might be the worst approval decision that the F.D.A. has made that I can remember,”

Harvard Medical School professor and recently-resigned FDA advisory panelist, Dr. Aaron Kesselheim, describing the FDA’s approval process for Biogen’s Aduhelm Alzheimer drug.


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


Aduhelm Approved

Biogen’s Aduhelm Alzheimer’s drug gained FDA approval last week, creating optimism among Alzheimer advocates and imaging providers, while also prompting extremely strong criticism.

About Aduhelm – Aduhelm (aka Aducanumab) is the first new FDA-approved Alzheimer’s treatment in 18 years and the first ever to attack the disease process (vs. just addressing symptoms). The monthly intravenous infusions would be provided to patients with early-stage Alzheimer’s and above-normal amyloid levels, reducing these patients’ brain amyloid levels with the goal of slowing cognitive decline. Biogen believes 1.5m people in the U.S. qualify for the drug, and that number would likely increase as our population ages.

Imaging’s Role – Patients will have to undergo a PET scan or lumbar puncture in order to qualify for Aduhelm treatment, while qualified patients would receive ongoing brain MRIs throughout their treatment (baseline, before 7th & 12th treatment, potentially more).

Aduhelm’s Rough Start – A new FDA-approved Alzheimer’s treatment would typically be cause for celebration, but instead Aduhelm’s approval has been widely criticized.

  • Medical leaders questioned Aduhelm’s effectiveness (no actual clinical evidence, debatable trial results) and suggested that the FDA was far too quick to approve the drug.
  • Health policy folks questioned how the $56k/yr treatment would be paid for (not to mention diagnostics / imaging costs) and which patients would be approved to receive the drug.
  • Health equity watchers warned that underdiagnosed minorities would miss out on treatment.
  • Imaging lobbyists reminded CMS that Medicare still doesn’t cover the PET scans required to qualify for Aduhelm.

Imaging Upside – Assuming Aduhelm proves to be highly effective, appropriately covered, and widely adopted, this approval would lead to a lot more brain imaging, likely requiring more scanners and more trained radiologists. There are other new Alzheimer’s drugs expected in the coming years too.



Doubling Follow-Up at Saint Joseph Mercy

The Saint Joseph Mercy Health System doubled its follow-up recommendation identification/tracking when it adopted Nuance PowerScribe Follow-up Manager, achieving ROI within the first year. Find out how in this Nuance case study.

– Sponsored.



Zebra-Med Thinks Big

See how and why Zebra Medical Vision sees a much bigger future for public health AI than many of us imagine in this Imaging Wire Q&A with company CEO, Zohar Elhanani.

– Sponsored.


The Wire

  • AI Education Reform: A new JACR opinion paper warned that AI will redefine physicians’ skill set requirements, calling for medical educators to adapt their curriculums accordingly. Envisioning a future with far more AI tools and patient data sources, the paper suggests that medical schools and academic centers should train future physicians to: 1) Understand and be able to evaluate algorithms; 2) Integrate and analyze multiple data sources (both imaging and non-imaging); 3) Effectively communicate about data; and 4) Guide how AI is developed and integrated.
  • Qure.ai POqUS: Qure.ai expanded into the ultrasound AI segment with its new POqUS AI tool for handheld ultrasound-based carotid artery stenosis detection. Although not yet commercially available, POqUS will support image acquisition (guides navigation & auto-capture), interpretation (automates IMT measurements), and risk detection (estimates stroke and cardiovascular risk).
  • X-Ray Dumping Violation: A Pennsylvania court ruled that Wayne Memorial Hospital violated a federal law designed to prevent “dumping” when it didn’t X-ray a patient with an open fracture. The prison inmate had his finger broken when he was bitten by a fellow prisoner, but Wayne Memorial only gave him a Tetanus shot and antibiotics. The plaintiff’s finger was X-rayed over a month later, revealing that he required surgery.
  • MDCT For ICI-Colitis: Multi-detector CT (MDCT) can help radiologists evaluate patients with immune checkpoint inhibitor-related colitis (ICI-colitis). The European Radiology study found that MDCT isn’t effective for diagnosing ICI-colitis (n = 118, 52% specificity), but found that MDCT severity scores are valuable for predicting ICI-colitis outcomes (need for intravenous steroids, prolonged inpatient stays, and endoscopic mucosal ulcerations).
  • ML for RT Treatment Planning: University of Toronto researchers developed a ML algorithm that generated radiation therapy treatment plans better than human physicians. The researchers had physicians and an AI model generate 100 treatment plans (50 retrospective, 50 prospective), finding that more AI generated plans were clinically acceptable (89% vs. 72%) and revealing that ML-based RT planning reduced the RT planning process by 60% (118 to 47 hours).
  • Spaceflight Bone Loss Ultrasound: Chinese researchers found that quantitative ultrasonic backscatter could be used to identify bone loss after long-term bedrest or microgravity space travel. The researchers examined 36 participants after 90 days of head-down-tilt bedrest (mimicking low-gravity) and 180-days of recovery, finding that ultrasonic backscatter measurements effectively identified decreases in body mass index and compartmental bone densities.
  • CXR Lung Cancer AI: A European Radiology study detailed a deep learning model that could help radiologists and residents detect lung cancer on chest X-rays. In a reader study (n = 173 CXRs w/ cancer, 346 healthy CXRs), the AI model helped residents detect lung cancer with similar sensitivity as radiologists who weren’t using AI (72% vs. 75%) and improved the residents’ accuracy when recommending follow-up CTs (54.7% without AI vs. 70.2% with AI), while reducing experienced radiologists’ false positive rates (24% vs. 17%).
  • Radiology’s Payment Problem: A new Radiology Journal paper found that 47% of physicians who presented at the 2018 RSNA Annual Meeting did not disclose at least one >$200 industry payment from the previous year (n = 310 total physicians, 198 received payments, 93 didn’t disclose at least 1 payment). These undisclosed payments carried a sizable $6,516 median value (range: $219 – $376,634), were most likely to be related to research (64% of undisclosed payments), and were most likely among abdominal radiologists (only a 12.9% disclosure rate).
  • Weakly Supervised, But Accurate AI: New research in Academic Radiology detailed a “weakly supervised” algorithm (trained with labels, not pixels) that was able to accurately assess breast MR images. The algorithm (trained/validated w/ 80k images, tested w/ 13k) distinguished benign and malignant tumors with an AUC of 0.92, 74.4% sensitivity, and 95.3% specificity.
  • POCUS Confidence: University of New Mexico researchers found that providing family medicine residents (FMRs) with POCUS training “significantly” increases their confidence performing and interpreting POCUS exams. The yearlong resident training program included a 3hr hands-on workshop, weekly 30-minute exam-based POCUS sessions, and access to online POCUS resources.
  • Lumbar MRI Scoring: Noting that RVUs don’t accurately account for the wide variations in lumbar spine MRI study complexity, UCSF researchers developed an NLP-based lumbar MRI composite severity score (CSS) that accurately predicted radiologist read times, and could potentially help with workflow planning and performance evaluations. To create these scores, researchers used an NLP tool to extract and categorize radiologists’ severity language and reporting timestamps from 13,388 lumbar spine MRI exams.
  • Surgalign & Inteneural Networks Partner: Surgalign Holdings launched a partnership with brain AI company Inteneural Networks, revealing plans to evaluate how Inteneural’s neural structure segmentation / detection tools could integrate into Surgalign’s digital surgery portfolio.
  • Preoperative fMRI Benefits: A Johns Hopkins study review (68 studies, 3,280 patients) found that when preoperative fMRI is used for brain tumor resection planning, patients experience significantly fewer negative outcomes (adverse events: 11% vs. 21%). Negative outcomes were especially reduced when fMRI was paired with other imaging techniques such as diffusion-tensor imaging, intraoperative MRI, or cortical stimulation.

AI Workflow Integration Success

Discover how to make your next AI workflow integration a success in this upcoming webinar featuring AI leaders from DASA, Signify Research, and GE Healthcare.

– Sponsored.


The Resource Wire

  • New ergonomics and enhanced auto-positioning are just a few of the ways that Canon Medical System’s new OMNERA 500A DR system improves technologist workflow and patient care. Check out the rest here.
  • United Imaging’s CT agreement with Vizient is a solid example of United Imaging’s “Equal Healthcare for All” mission, providing Vizient members with quality and true business value that can help them expand the care they provide.
  • With radiation dose management now largely considered best practice, this Bayer white paper details the top five benefits of adopting contrast dose management.
  • Independent and staying that way? Healthcare Administrative Partners just released a helpful set of guidelines that radiology practices can follow to stay private despite ongoing consolidation pressures.
  • Tune in to Riverain Technologies’ on-demand webinar demonstrating how its AI solutions integrate into LucidHealth’s radiology workflow and sharing best practices on how to combine AI with radiologist expertise.
  • See how Arterys LungAI matched (and actually exceeded) radiologists’ accuracy measuring lung nodule volumes in CT scans.