Echo AI Coronary Artery Calcium Scoring

A Cedars-Sinai-led team developed an echocardiography AI model that was able to accurately assess coronary artery calcium buildup, potentially revealing a safer, more economical, and more accessible approach to CAC scoring.

The researchers used 1,635 Cedars-Sinai patients’ transthoracic echocardiogram (TTE) videos paired with their CT-based Agatston CAC scores to train an AI model to predict patients’ CAC scores based on their PLAX view TTE videos. 

When tested against Cedars-Sinai TTEs that weren’t used for AI training, the TTE CAC AI model detected…

  • Zero CAC patients with “high discriminatory abilities” (AUC: 0.81)
  • Intermediate patients “modestly well” (≥200 scores; AUC: 0.75)
  • High CAC patients “modestly well” (≥400 scores; AUC: 0.74)

When validated against 92 TTEs from an external Stanford dataset, the AI model similarly predicted which patients had zero and high CAC scores (AUCs: 0.75 & 0.85).

More importantly, the TTE AI CAC scores accurately predicted the patients’ future risks. TTE CAC scores predicted one-year mortality similarly to CT CAC scores, and they even improved overall prediction of low-risk patients by downgrading patients who had high CT CAC scores and zero TTE CAC scores.

The Takeaway

CT-based CAC scoring is widely accepted, but it isn’t accessible to many patients, and concerns about its safety and value (cost, radiation, incidentals) have kept the USPSTF from formally recommending it for coronary artery disease surveillance. We’d need a lot more research and AI development efforts, but if TTE CAC AI solutions like this prove to be reliable, it could make CAC scoring far more accessible and potentially even more accepted.

Incidental Findings and Low-Value Care

A whopping 15% to 30% of diagnostic imaging exams reveal at least one incidental finding. Each of those findings might seem like blessings to radiology outsiders, but a popular new AJR editorial argues that imaging incidentals are far more likely to drive low-value care.

Michigan Medicine’s Matthew Davenport, MD led-off his editorial by suggesting that early cancer detection “is not always an ideal outcome,” because most of those cancers won’t affect patient health, while incidental follow-ups always require resources and often negatively impact patients (physically, financially, and emotionally).

He identified numerous reasons for radiology’s incidental overdiagnosis challenges…

  • Screening low-risk patients inherently uncovers low-risk incidentals
  • There’s a lack of understanding of incidental risks (clinically and downstream)
  • Many early cancers don’t or shouldn’t require treatment
  • Radiologists face significant pressure to recommend follow-ups

Although many incidental findings significantly improve patient outcomes, and those positive examples have established incidentals as a “secondary benefit of imaging,” the editorial suggests that incidentals will have a negative overall impact on radiology’s value until current practices change. 

So, what should we do? Dr. Davenport encourages radiologists to…

  • Become more aware of the harms of incidental management
  • Advocate for guidelines that emphasize high-value care
  • Support research on incidental management practices
  • “Avoid being alarmist” about incidentals in radiology reporting
  • Adopt solutions to help rads assess incidental patients’ risk factors
  • Balance diagnostic sensitivity with minimizing follow-up risks 

The Takeaway

If you scroll through the Imaging Wire archives, you’ll find plenty of stories that depict incidentals as a net positive for patient care. In fact, most suggest that radiology’s research and business leaders are actively trying to find ways to detect more incidentals. However, efforts to better understand or to reduce incidentals’ negative impacts are far less common. 

That divide is pretty notable given how many radiologists agree with Dr. Davenport, and it suggests that the barriers to solving incidental findings’ value problems are quite high.

Acute Chest Pain CXR AI

Patients who arrive at the ED with acute chest pain (ACP) syndrome end up receiving a series of often-negative tests, but a new MGB-led study suggests that CXR AI might make ACP triage more accurate and efficient.

The researchers trained three ACP triage models using data from 23k MGH patients to predict acute coronary syndrome, pulmonary embolism, aortic dissection, and all-cause mortality within 30 days. 

  • Model 1: Patient age and sex
  • Model 2: Patient age, sex, and troponin or D-dimer positivity
  • Model 3: CXR AI predictions plus Model 2

In internal testing with 5.7k MGH patients, Model 3 predicted which patients would experience any of the ACP outcomes far more accurately than Models 2 and 1 (AUCs: 0.85 vs. 0.76 vs. 0.62), while maintaining performance across patient demographic groups.

  • At a 99% sensitivity threshold, Model 3 would have allowed 14% of the patients to skip additional cardiovascular or pulmonary testing (vs. Model 2’s 2%).

In external validation with 22.8k Brigham and Women’s patients, poor AI generalizability caused Model 3’s performance to drop dramatically, while Models 2 and 1 maintained their performance (AUCs: 0.77 vs. 0.76 vs. 0.64). However, fine-tuning with BWH’s own images significantly improved the performance of the CXR AI model (from 0.67 to 0.74 AUCs) and Model 3 (from 0.77 to 0.81 AUCs).

  • At a 99% sensitivity threshold, the fine-tuned Model 3 would have allowed 8% of BWH patients to skip additional cardiovascular or pulmonary testing (vs. Model 2’s 2%).

The Takeaway

Acute chest pain is among the most common reasons for ED visits, but it’s also a major driver of wasted ED time and resources. Considering that most ACP patients undergo CXR exams early in the triage process, this proof-of-concept study suggests that adding CXR AI could improve ACP diagnosis and significantly reduce downstream testing.

Bayer Establishes AI Platform Leadership with Blackford Acquisition

Six months after becoming radiology’s newest AI platform vendor, Bayer accelerated its path towards AI leadership with its acquisition of Blackford Analysis.

The acquisition might prove to be among the most significant in imaging AI’s short history, combining Blackford’s many AI advantages (tech, expertise, relationships) with Bayer’s massive radiology presence and AI ambitions. 

After closing later this year, Blackford will operate independently through Bayer’s well-established “arm’s length” model, allowing Blackford to preserve its entrepreneurial culture, while leveraging Bayer’s “experience, infrastructure and reach” to drive further expansion.

Bayer’s Calantic platform and team will operate separately from Blackford, providing Bayer customers with two distinct AI platforms to choose from, while giving Bayer two ways to drive its AI business forward. 

Although few would have predicted this acquisition, it makes sense given Bayer and Blackford’s relatively long history together and their complementary situations. 

  • Blackford was part of Bayer’s 2019 G4A digital health accelerator class
  • The companies have been working together to develop Calantic since 2020
  • Bayer has big AI goals, but its AI customer base and reputation were unestablished
  • Blackford’s AI customer base and reputation are solid, but it needed a new way to scale and a positive exit for its shareholders

Even fewer would have predicted that imaging contrast vendors would be the driving force behind AI’s next consolidation wave, noting that Guerbet invested in Intrasense just last week. However, imaging contrast and imaging AI could serve increasingly interrelated (or alternative) roles in the diagnostic process, and there’s surely advantages to being a leader in both areas for Bayer and Guerbet.

Speaking of AI consolidation, it appears that all those 2023 AI consolidation forecasts are proving to be correct, while bringing some of radiology’s largest companies into an AI segment that’s historically been dominated by startups. It wouldn’t be surprising if that trend continued.

The Takeaway

Bayer and Blackford have been working on their AI strategies for years, and this acquisition appears to give both companies a much better chance of achieving long-term AI leadership. Considering that AI is still in its infancy and could eventually play a dominant role in radiology (and across healthcare), AI leadership might be a far more significant market position in the future than many can imagine today.

CXR AI’s Screening Generalizability Gap

A new European Radiology study detailed a commercial CXR AI tool’s challenges when used for screening patients with low disease prevalence, bringing more attention to the mismatch between how some AI tools are trained and how they’re applied in the real world.

The researchers used an unnamed commercial AI tool to detect abnormalities in 3k screening CXRs sourced from two healthcare centers (2.2% w/ clinically significant lesions), and had four radiology residents read the same CXRs with and without AI assistance, finding that the AI:

  • Produced a far lower AUROC than in its other studies (0.648 vs. 0.77–0.99)
  • Achieved 94.2% specificity, but just 35.3% sensitivity
  • Detected 12 of 41 pneumonia, 3 of 5 tuberculosis, and 9 of 22 tumors 
  • Only “modestly” improved the residents’ AUROCs (0.571–0.688 vs. 0.534–0.676)
  • Added 2.96 to 10.27 seconds to the residents’ average CXR reading times

The researchers attributed the AI tool’s “poorer than expected” performance to differences between the data used in its initial training and validation (high disease prevalence) and the study’s clinical setting (high-volume, low-prevalence, screening).

  • More notably, the authors pointed to these results as evidence that many commercial AI products “may not directly translate to real-world practice,” urging providers facing this kind of training mismatch to retrain their AI or change their thresholds, and calling for more rigorous AI testing and trials.

These results also inspired lively online discussions. Some commenters cited the study as proof of the problems caused by training AI with augmented datasets, while others contended that the AI tool’s AUROC still rivaled the residents and its “decent” specificity is promising for screening use.

The Takeaway

We cover plenty of studies about AI generalizability, but most have explored bias due to patient geography and demographics, rather than disease prevalence mismatches. Even if AI vendors and researchers are already aware of this issue, AI users and study authors might not be, placing more emphasis on how vendors position their AI products for different use cases (or how they train it).

Guerbet’s Big AI Investment

Guerbet took a big step towards advancing its AI strategy, acquiring a 39% stake in French imaging software company Intrasense, and revealing ambitious future plans for their combined technologies.

Through Intrasense, Guerbet gains access to a visualization and AI platform and a team of AI integration experts to help bring its algorithms into clinical use. The tie-up could also create future platform and algorithm development opportunities, and the expansion of their technologies across Guerbet’s global installed base.

The €8.8M investment (€0.44/share, a 34% premium) could turn into a €22.5M acquisition, as Guerbet plans to file a voluntary tender offer for all remaining shares.

Even though Guerbet is a €700M company and Intrasense is relatively small (~€3.8M 2022 revenue, 67 employees on LinkedIn), this seems like a significant move given and Guerbet’s increasing emphasis on AI:

What Guerbet was lacking before now (especially since ending its Merative/IBM alliance) was a future AI platform – and Intrasense should help fill that void. 

If Guerbet acquires Intrasense it would continue the recent AI consolidation wave, while adding contrast manufacturers to the growing list of previously-unexpected AI startup acquirers (joining imaging center networks, precision medicine analytics companies, and EHR analytics firms). 

However, contrast manufacturers could play a much larger role in imaging AI going forward, considering the high priority that Bayer is placing on its Calantic AI platform.

The Takeaway

Guerbet has been promoting its AI ambitions for several years, and this week’s Intrasense investment suggests that the French contrast giant is ready to transition from developing algorithms to broadly deploying them. That would take a lot more work, but Guerbet’s scale and imaging expertise makes it worth keeping an eye on if you’re in the AI space.

Radiology in 2040

A new Radiology Journal editorial shared a radical vision for how the specialty will operate in 2040, warning that “seismic” changes will require radiologists to overhaul their roles in order to thrive, or even stay relevant.

Here’s what the authors expect:

Super Reporting – Radiology reporting will become far more automatic and dynamic, as reports embrace multimedia formats, become far more accessible and patient-friendly, and integrate into automatic follow-up systems.

Disease Focus – The growth of at-home care and the emergence of mobile and self-examination imaging technologies will force radiology workflows to become organized by diseases, rather than by patients’ “location” (ED, ICU, etc.).

Inevitable AI – “AI will not replace radiology,” but it will “profoundly affect [radiologists’] relevance and workflow” as algorithms become more comprehensive, autonomous, and accurate.

The AI Threat – AI will eliminate many current radiologist tasks, but its greatest threat to radiology would come from referring physicians using imaging AI independently. 

Multi-Diagnostics – The rise of non–imaging precision diagnostics (ie, “liquid biopsies”) and multimodal/multiomic diagnostics will reduce imaging’s role in disease detection, and lead to a more-integrated diagnostic and treatment planning process.

Future Therapy – Major advances in precision imaging, image-guided technology, and theranostics would allow radiology to increase its clinical value by owning image-related procedures.

Those are some major changes, and would require radiologists to take similarly major actions in order to thrive in 2040 and beyond:

  • Understand that image interpretation will become a commodity, and maybe “obsolete”
  • Maintain a “laser-sharp” focus on adding value across the healthcare continuum 
  • Actively embrace radiologists’ role as AI’s primary users, owners, and managers
  • “Extensively cultivate” radiology’s interventional and theranostics capabilities

The Takeaway
It’s impossible to accurately predict how medicine will evolve over the next two decades, and there’s surely plenty of readers who are growing tired of obsolescence warnings.

That said, the authors are very well-respected and each of their forecasts can be directly linked to today’s emerging trends, suggesting that radiologists who follow their advice might be more likely to “thrive” in 2040 regardless of how the future unfolds.

Medical Imaging in 2023

Happy New Year, and welcome to the first Imaging Wire of 2023. For those of you getting started on your annual gameplans, here are some potential imaging trends that you might want to consider.

Provider Strain – Many providers limped into 2023 with shaky finances, workforce shortages, and burnout problems, and now they have to navigate an economic downturn. That means they might only be open to “must have” initiatives/technologies that address that list of challenges.

Startup Strain – A market full of strained healthcare providers is generally bad news for medtech startups, especially considering that many of those startups are still in search of their next funding round (from a smaller / more selective group of VCs) or are trying to make their previous rounds last longer than they initially planned.

Paying Down Imaging IT Debt – Southwest’s holiday shutdown prompted many radiology leaders to reevaluate which of their long-delayed tech updates should be now viewed as “must haves” in 2023. These leaders will have a lot to choose from, given how many imaging IT infrastructures are patched together and how many tech initiatives were delayed by COVID.

The Year of AI (again) – There’s a lot of activity in imaging AI and perhaps even more attention, qualifying the AI segment for a wide range of 2023 predictions:

  • AI tools will continue to become more comprehensive
  • Pharma companies will play a growing role in AI funding and strategies
  • AI research will shift towards evaluating commercial tools
  • New chatGPT-inspired reporting/communications use cases will emerge
  • The AI consolidation wave will peak
  • AI adoption will expand among mid-sized hospitals and practices
  • Administrative/operational AI solutions will have another big year
  • The list of reimbursable AI solutions will continue to expand

Diversified Diagnostics – 2022 brought more imaging informatics players into digital pathology, welcomed ambitious new theranostics efforts, and saw a surge of intriguing multi-omics/olgy studies. Those trends should intensify in 2023, as traditionally separate diagnostic areas slowly converge with imaging technologies and teams.

Non-Imaging Biomarkers – Speaking of which, 2023 will bring more progress towards the development and adoption of imaging-related (or imaging-alternative) biomarker tests, including three brand new biomarker techniques that we covered below in today’s newsletter.

More Home Imaging – Medical imaging will continue its shift beyond hospital walls, as home and outpatient care boom, and mobile and DIY imaging technologies evolve.

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

You might also like..

Select All

You're signed up!

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

-- The Imaging Wire team

You're all set!