Philips’ RSNA MRIs

After two straight solution-centric RSNAs, Philips’ RSNA 2021 booth will be headlined by a pair of new MR systems (plus some MR solutions) and a major focus on operational efficiency.

MR 5300 – The FDA-cleared MR 5300 (1.5T) is Philips’ second helium-free BlueSeal MR scanner, arriving three years after the Ingenia Ambition X, and launching with new AI-powered features intended to automate clinical and operational tasks. Philips also emphasized the image quality advantages of the MR 5300’s 55cm field-of-view and dStream Breeze coils.

MR 7700 – Philips’ forthcoming MR 7700 system (3T) is positioned for both clinical and research use, although its features and messaging largely emphasize its value to researchers. The MR 7700 boasts new multinuclear clinical capabilities, including diffusion imaging and advanced neuroscience sequences, while its XP gradients (65 mT/m) support neuroscience.

MR Workspace – Philips’ new MR Workspace is intended to support technologist productivity, providing a dashboard that automates the planning and execution of many routine scans and supports decision making by suggesting the most suitable Exam Card for each patient. MR Workspace will be included with all new and installed Philips MR scanners. 

Philips SmartSpeed – The SmartSpeed image reconstruction platform is designed to speed up image acquisition and enhance image quality, leveraging Philips’ Compressed SENSE acquisition technique and AI to reconstruct full images from under-sampled data, “while maintaining virtually equivalent image quality.”

The Takeaway – Philips’ MR lineup has been relatively quiet during the last few years, while its main MR competitors made some solid progress (particularly w/ reconstruction, operability, comfort, and low-helium tech). It seems like that won’t be the case at RSNA 2021.

Viz.ai’s Care Coordination Expansion

Viz.ai advanced its care coordination strategy last week, launching new Pulmonary Embolism and Aortic Disease modules, and unveiling its forthcoming Viz ANX cerebral aneurysm module.

PE & Aortic Modules – The new PE and Aortic modules use AI to quickly detect pulmonary embolisms and aortic dissection in CTA scans, and then coordinate care using Viz.ai’s 3D mobile viewer and clinical communications workflows. It appears that Viz.ai partnered with Avicenna.AI to create these modules, representing a logical way for Viz.ai to quickly expand its portfolio.

Viz ANX Module – The forthcoming Viz ANX module will use the 510k-pending Viz ANX algorithm to automatically detect suspected cerebral aneurysms in CTAs, and then leverage the Viz Platform for care coordination.

Viz.ai’s Care Coordination Strategy – Viz.ai called itself “the leader in AI-powered care coordination” a total of six times in these two announcements, and the company has definitely earned this title for stroke detection/coordination. Adding new modules to the Viz Platform is how Viz.ai could earn “leadership” status across all other imaging-detected emergent conditions.

The Takeaway – Viz.ai’s stroke detection/coordination platform has been among the biggest imaging AI success stories, making its efforts to expand to new AI-based detection and care coordination areas notable (and pretty smart). These module launches are also an example of diagnostic AI’s growing role throughout care pathways, showing how AI can add clinical value beyond the reading room.

Siemens’ Healthineers Hardware Evolution

Siemens Healthineers’ Shape 22 pre-RSNA event featured a pair of ambitious hardware announcements that stand to expand what can be done with CT exams and where MRIs can be performed.

NAEOTOM Alpha PCCT – Siemens Healthineers confirmed its pole position in the Photon-Counting CT race, officially launching its NAEOTOM Alpha scanner. Although the NAEOTOM Alpha already received a rare marketing head-start from the FDA, this week’s launch begins its official 2022 rollout, and provides new details about this milestone product:

  • Far higher image quality than CT
  • Provides much more imaging data and new levels of CT-based insights
  • Expands CT to new cardiac, oncology, and pulmonology use cases
  • Allows 50% lower radiation dosage, could shift exams to non-contrast
  • Supports Siemens’ core solutions, including operability and AI-based diagnosis
  • Cleared in US and Europe, 20 systems already installed, 8k patients scanned
  • PCCT expected to become the main CT technology within 10 years
  • Siemens is holding another NAEOTOM Alpha event today (Nov. 18)
  • Siemens might be first, but we’re seeing more PCCT activity from GE Healthcare, and Canon and Philips aren’t far behind

MAGNETOM Free.Star MRI – One year after introducing the MRI-expanding MAGNETOM Free.Max, Siemens continued its MRI accessibility push, revealing the “disruptively simple” MAGNETOM Free.Star. The new Free.Star MRI will inherit much of the MAGNETOM Free.Max’s accessibility-friendly qualities (0.55T, small/light, low helium & installation requirements), and will have the ambitious goal of supporting the half of the world’s population that doesn’t have MRI access. The MAGNETOM Free.Star is still early-stage (it hasn’t begun the FDA process), but it’s massive healthcare ambitions make it worth keeping an eye on.

The Takeaway – The NAEOTOM Alpha is expected to be the start of a major shift towards Photon-Counting CT, while the new MAGNETOM Free.Max and Free.Star could expand where MRIs are used. That makes these extremely significant products.

Right Diagnoses, Wrong Reasons

An AJR study shared new evidence of how X-ray image labels influence deep learning decision making, while revealing one way developers can address this issue.

Confounding History – Although already well known by AI insiders, label and laterality-based AI shortcuts made headlines last year when they were blamed for many COVID algorithms’ poor real-world performance. 

The Study – Using 40k images from Stanford’s MURA dataset, the researchers trained three CNNs to detect abnormalities in upper extremity X-rays. They then tested the models for detection accuracy and used a heatmap tool to identify the parts of the images that the CNNs emphasized. As you might expect, labels played a major role in both accuracy and decision making.

  • The model trained on complete images (bones & labels) achieved an 0.844 AUC, but based 89% of its decisions on the radiographs’ laterality/labels.
  • The model trained without labels or laterality (only bones) detected abnormalities with a higher 0.857 AUC and attributed 91% of its decision to bone features.
  • The model trained with only laterality and labels (no bones) still achieved an 0.638 AUC, showing that AI interprets certain labels as a sign of abnormalities. 

The Takeaway – Labels are just about as common on X-rays as actual anatomy, and it turns out that they could have an even greater influence on AI decision making. Because of that, the authors urged AI developers to address confounding image features during the curation process (potentially by covering labels) and encouraged AI users to screen CNNs for these issues before clinical deployment.

UCSF Automates CAC Scoring

UCSF is now using AI to automatically screen all of its routine non-contrast chest CTs for elevated coronary artery calcium scores (CAC scores), representing a major milestone for an AI use case that was previously limited to academic studies and future business strategies.

UCSF’s Deployment UCSF becomes the first medical center to deploy the end-to-end AI CAC scoring system that it developed with Stanford and Bunkerhill Health earlier this year. The new system automatically identifies elevated CAC scores in non-gated / non-contrast chest CTs, creating an “opportunistic screening pathway” that allows UCSF physicians to identify high-CAC patients and get them into treatment.

Why This is a Big Deal – Over 20m chest CTs are performed in the U.S. annually and each of those scans contains insights into patients’ cardiac health. However, an AI model like this would be required to extract cardiac data from the majority of CT scans (CAC isn’t visible to humans in non-gated CTs) and efficiently interpret them (there’s far too many images). This AI system’s path from academic research to clinical deployment seems like a big deal too.

The Commercial Impact – Most health systems don’t have the AI firepower of Stanford and UCSF, but they certainly produce plenty of chest CTs and should want to identify more high-risk patients while treatable (especially if they’re also risk holders). Meanwhile, there’s growing commercial efforts from companies like Cleerly and Nanox.AI to create opportunistic CAC screening pathways for all these health systems that can’t develop their own CAC AI workflows (or prefer not to).

The MARCA Divide

The American College of Radiology might have a neutral stance on the Medicare Access to Radiology Care Act (MARCA), but a new survey confirmed that most ACR members are far from neutral about non-physicians’ role in radiology. 

MARCA Madness – MARCA would require Medicare to reimburse supervising radiologists for imaging services performed by radiologist assistants, as long as RAs work within physician-led teams. The ACR revealed its neutral position on MARCA in August, enraging some members who are concerned that MARCA will undermine radiologists’ role, and accused the ACR of selling out to PE. 

The Opinion Divide – The ACR survey (n = 4,207, or 16% of members) revealed overwhelming opposition to MARCA, but more balanced views on working with non-physician radiology providers (NPRPs). By NPRPs, they mean radiology assistants, advanced practice registered nurses, and physician assistants.

  • 60% are against MARCA (vs. 19% in favor, 21% neutral)
  • 86% are concerned about NPRP scope creep
  • 55% view NPRPs as a threat to patient care
  • However, just 43% are against using NPRPs in their practice
  • And 62% believe it’s up to practices whether they employ NPRPs

Behind the Divide – A deeper look into the ACR’s (very detailed) survey results revealed that members’ MARCA and NPRP opinions seem largely influenced by their professional situation. 

Career Stage

  • 80% of residents/fellows and 65% of early-career rads view NPRPs as a threat to patient care 
  • 51% of mid-career rads and 41% of late-career rads view NPRPs as a threat to patient care

Practice Type

  • 61% of respondents from academic settings view NPRPs as a threat to patient care
  • 69% of respondents from national and private practices think NPRP use is a practice decision 

Practice Role

  • 61% of non-leaders view NPRPs as a threat to patient care
  • 65% of practices leaders view NPRPs use as a practice decision

NPRP Experience

  • 69% of respondents who do not work with NPRPs view them as a threat to patient care
  • 57% of respondents who work with NPRPs believe they play an important role
  • 84% of respondents who support MARCA currently work with NPRPs

The Takeaway – We now have data confirming what most of you already knew: the majority of radiologists are firmly against MARCA and a small minority support it. However, the data also shows that plenty of radiologists see value in NPRPs, especially if they already work with non-physicians and if their careers are less threatened by them. What’s still unclear is what it will take for the ACR to break its neutrality on MARCA (in either direction).

The PSMA PET/CT + mpMRI PCa Effect

There’s been a growing number of studies comparing 68Ga-PSMA PET/CT and mpMRI’s effectiveness along the prostate cancer pathway, but new research suggests that the modalities might be particularly effective if used together early in the diagnostic process.

The Study – A team of Australian researchers reviewed pre-operative mpMRI and 68Ga-PSMA PET/CT exams from 1,123 men (median PSA = 6), comparing their imaging results and histology findings. Here’s what they discovered:

  • mpMRI identified tumors in 93 men (8%) that were missed by 68Ga-PSMA PET/CT
  • 68Ga-PSMA PET/CT spotted tumors in 117 men (10%) that mpMRI missed
  • The combined modalities identified index tumors with Gleason Scores of ≥3+4 in 92% of men (vs. 80% w/ only mpMRI & 82% w/ only PSMA PET/CT)
  • 68Ga-PSMA PET/CT and mpMRI performed similarly for lesion/tumor detection and localization

Better Assessments and Decisions – In addition to identifying more tumors, mpMRI and 68Ga-PSMA PET/CT’s combined localization accuracy might improve biopsy targeting, leading to more accurate tumor grading and better management decisions.

More Necessary Biopsies – Meanwhile, patients with negative mpMRI and 68Ga-PSMA PET/CT findings would likely have a low risk of clinically significant prostate cancer, making them solid candidates for active PSA monitoring and allowing them to avoid unnecessary biopsies.

The Takeaway – This approach needs a lot more research and PSMA tracers currently only have FDA approval for patients with metastatic prostate cancer or biochemical recurrence, so 68Ga-PSMA PET/CT won’t be combined with mpMRI in early diagnostic exams very soon. That said, this study suggests that we’ll see more future efforts to combine and evaluate mpMRI + 68Ga-PSMA PET/CT as an early diagnostic step.

Chest Pain Imaging Guidance

If it seemed like coronary imaging folks were more excited than usual last week, it’s because the AHA/ACC’s long-awaited chest pain guidelines just set the stage for a lot more imaging.

The Guidelines – The American Heart Association (AHA) and the American College of Cardiology (ACC) released their first clinical guidelines for the assessment and diagnosis of chest pain, outlining a range of new standards, processes, and pathways, while giving coronary imaging a central diagnostic role.

Front-Line Coronary CTA – The new guidelines made coronary CTA a front-line coronary artery disease test, assigning CCTA their highest recommendation level and proposing it for a large group of patients (mid-high risk of CAD, stable chest pain, <65yrs).

FFRct Next in Line – HeartFlow’s FFRct analysis will often serve as the next diagnostic step when CCTA exams reveal obstructive CAD (40-90% stenosis) or are inconclusive, with FFRct results either clarifying diagnosis or supporting treatment decisions. 

Stress Imaging Pathways – The AHA/ACC guidelines also gave stress imaging (e.g. TTE, echo, CMRI, PET, etc.) their highest recommendation level, positioning stress imaging for more serious cases (likely or confirmed obstructive CAD, ≥65yrs) as well as for diagnosing myocardial ischemia and estimating risks of major cardiac events among patients with less severe cases (intermediate risk, no known CAD, acute chest pain).

Takeaway – These new guidelines are a big deal for coronary imaging, given the millions of people who show up at US emergency departments with chest pain each year. It’s also going to require some big changes across EDs, imaging centers, and radiology departments/practices, who will have to retool their imaging protocols/fleets and be able to expertly interpret a wave of coronary imaging exams (and handle a wave of incidentals).

The False Hope of Explainable AI

Many folks view explainability as a crucial next step for AI, but a new Lancet paper from a team of AI heavyweights argues that explainability might do more harm than good in the short-term, and AI stakeholders would be better off increasing their focus on validation.

The Old Theory – For as long as we’ve been covering AI, really smart and well-intentioned people have warned about the “black-box” nature of AI decision making and forecasted that explainable AI will lead to more trust, less bias, and greater adoption.

The New Theory – These black-box concerns and explainable AI forecasts might be logical, but they aren’t currently realistic, especially for patient-level decision support. Here’s why:

  • Explainability methods describe how AI systems work, not how decisions are made
  • AI explanations can be unreliable and/or superficial
  • Most medical AI decisions are too complex to explain in an understandable way
  • Humans over-trust computers, so explanations can hurt their ability to catch AI mistakes
  • AI explainability methods (e.g heat maps) require human interpretation, risking confirmation bias
  • Explainable AI adds more potential error sources (AI tool + AI explanation + human interpretation)
  • Although we still can’t fully explain how acetaminophen works, we don’t question whether it works, because we’ve tested it extensively

The Explainability Alternative – Until suitable explainability methods emerge, the authors call for “rigorous internal and external validation of AI models” to make sure AI tools are consistently making the right recommendations. They also advised clinicians to remain cautious when referencing AI explanations and warned that policymakers should resist making explainability a requirement. 

Explability’s Short-Term Role – Explainability definitely still has a role in AI safety, as it’s “incredibly useful” for model troubleshooting and systems audits, which can improve model performance and identify failure modes or biases.

The Takeaway – It appears we might not be close enough to explainable AI to make it a part of short-term AI strategies, policies, or procedures. That might be hard to accept for the many people who view the need for AI explainability as undebatable, and it makes AI validation and testing more important than ever.

ImageBiopsy Lab & UCB’s AI Alliance

Global pharmaceutical company UCB recently licensed its osteoporosis AI technology to MSK AI startup ImageBiopsy Lab, representing an interesting milestone for several emerging AI business models.

The UCB & ImageBiopsy Lab Alliance – ImageBiopsy Lab will use UCB’s BoneBot AI technology to develop and commercialize a tool that screens CT scans for “silent” spinal fractures to identify patients who should be receiving osteoporosis treatments. The new tool will launch by 2023 as part of ImageBiopsy Lab’s ZOO MSK platform.

UCB’s AI Angle – UCB produces an osteoporosis drug that would be prescribed far more often if detection rates improve (over 2/3 of vertebral fractures are currently undiagnosed). That’s why UCB developed and launched BoneBot AI in 2019 and it’s why the pharma giant is now working with ImageBiopsy Lab to bring it into clinical use.

The PharmaAI Trend – We’re seeing a growing trend of drug and device companies working with AI developers to help increase treatment demand. The list is getting pretty long, including quite a few PharmaAI alliances targeting lung cancer treatment (Aidence & AstraZeneca, Qure.ai & AstraZeneca, Huma & Bayer, Optellum & J&J) and a diverse set of AI alliances with medical device companies (Imbio & Olympus for emphysema, Aidoc & Inari for PE, Viz.ai & Medtronic for stroke).

The Population Health AI Trend – ImageBiopsy Lab’s BoneBot AI licensing is also a sign of AI’s growing momentum in population health, following increased interest from academia and major commercial efforts from Cleerly (cardiac screening) and Zebra Medical Vision (cardiac and osteoporosis screening… so far). This alliance also introduces a new type of population health AI beneficiary (pharma companies), in addition to risk holders and patients.

The Takeaway – ImageBiopsy Lab and UCB’s new alliance didn’t get a lot of media attention last week, but it tells an interesting story about how AI business models are evolving beyond triage, and how those changes are bringing some of healthcare’s biggest names into the imaging AI arena.

Get every issue of The Imaging Wire, delivered right to your inbox.

You're signed up!

It's great to have you as a reader. Check your inbox for a welcome email.

-- The Imaging Wire team