IBM Sells Watson Health

IBM is selling most of its Watson Health division to private equity firm Francisco Partners, creating a new standalone healthcare entity and giving both companies (IBM and the former Watson Health) a much-needed fresh start. 

The Details – Francisco Partners will acquire Watson Health’s data and analytics assets (including imaging) in a deal that’s rumored to be worth around $1B and scheduled to close in Q2 2022. IBM is keeping its core Watson AI tech and will continue to support its non-Watson healthcare clients.

Francisco’s Plans – Francisco Partners seems optimistic about its new healthcare company, revealing plans to maintain the current Watson Health leadership team and help the company “realize its full potential.” That’s not always what happens with PE acquisitions, but Francisco Partners has a history of growing healthcare companies (e.g. Availity, Capsule, GoodRx, Landmark Health) and there are a lot of upsides to Watson Health (good products, smart people, strong client list, a bargain M&A multiple, seems ideal for splitting up).

A Necessary Split – Like most Watson Health stories published over the last few years, news coverage of this acquisition overwhelmingly focused on Watson Health’s historical challenges. However, that approach seems lazy (or at least unoriginal) and misses the point that this split should be good news for both parties. IBM now has another $1B that it can use towards its prioritized hybrid cloud and AI platform strategy, and the new Watson Health company can return to growth mode after several years of declining corporate support.

Imaging Impact – IBM and Francisco Partners’ announcements didn’t place much focus on Watson Health’s imaging business, but it seems like the imaging group will also benefit from Francisco Partners’ increased support and by distancing itself from a brand that’s lost its shine. Even losing the core Watson AI tech should be ok, given that the Merge PACS team has increasingly shifted to a partner-focused AI strategy. That said, this acquisition’s true imaging impact will be determined by where the imaging group lands if/when Francisco Partners decides to eventually split up and sell Watson Health’s various units.

The Takeaway – The IBM Watson Health story is a solid reminder that expanding into healthcare is exceptionally hard, and it’s even harder when you wrap exaggerated marketing around early-stage technology and high-multiple acquisitions. Still, there’s plenty of value within the former Watson Health business, which now has an opportunity to show that value.

Duke’s Interpretable AI Milestone

A team of Duke University radiologists and computer engineers unveiled a new mammography AI platform that could be an important step towards developing truly interpretable AI.

Explainable History – Healthcare leaders have been calling for explainable imaging AI for some time, but explainability efforts have been mainly limited to saliency / heat maps that show what part of an image influenced a model’s prediction (not how or why).

Duke’s Interpretable Model – Duke’s new AI platform analyzes mammography exams for potentially cancerous lesions to help physicians determine if a patient should receive a biopsy, while supporting its predictions with image and case-based explanations. 

Training Interpretability – The Duke team trained their AI platform to locate and evaluate lesions following a process that human radiology educators and students would utilize:

  • First, they trained the AI model to detect suspicious lesions and to ignore healthy tissues
  • Then they had radiologists label the edges of the lesions
  • Then they trained the model to compare those lesion edges with lesion edges from an archive of images with confirmed outcomes

Interpretable Predictions – This training process allowed the AI model to identify suspicious lesions, highlight the classification-relevant parts of the image, and explain its findings by referencing confirmed images. 

Interpretable Results – Like many AI models, this early version could not identify cancerous lesions as accurately as human radiologists. However, it matched the performance of existing “black box” AI systems and the team was able to see why their AI model made its mistakes.

The Takeaway

It seems like concerns over AI performance are growing at about the same pace as actual AI adoption, making explainability / interpretability increasingly important. Duke’s interpretable AI platform might be in its early stages, but its use of previous cases to explain findings seems like a promising (and straightforward) way to achieve that goal, while improving diagnosis in the process.

RSNA 2021 Reflections

The first in-person RSNA since COVID is officially a wrap. Hope you had a blast if you made it to Chicago and a productive week if you stayed home. We also hope you enjoy The Imaging Wire’s big takeaways from what might have been both the most special and most subdued RSNA ever.

Crowds & Conversations – We were already expecting 50% lower attendance than RSNA 2019, but the exhibit hall and cab lines looked more like 70% below 2019’s crowds (even less on Sunday & Wednesday). That said, most of the stronger companies had steady booth traffic and nearly every exhibitor emphasized that the attendees who did show up were ready to have high-quality conversations.

Focus on Productivity – Just about every product message at RSNA focused on productivity and efficiency, often with greater emphasis than clinical effectiveness. The modality-based efficiency enhancements seemed to be the most impactful, which is good news for technologist bandwidth and patient throughput, but might be bad news for rad burnout unless informatics/AI efficiency can catch up (it doesn’t seem like that happened this year).

Modality Milestones – The major OEMs did a good job making modalities cool again, debuting milestone innovations across both their MR (low-helium, low-field, reconstruction, coils) and CT (photon-counting, spectral, upgradability) lineups. We also saw the latest scanners take big strides in operator efficiency and patient experience. There weren’t many breakthroughs with X-ray or ultrasound, and most point-of-care ultrasound OEMs stayed home (rads aren’t their market anyway), but attendees seemed okay with that.

AI Showcase – The RSNA AI Showcase had solid traffic and high energy (especially on Mon & Tues), helped by continued AI buzz and the fact that RSNA finally let AI vendors out of the basement. The AI Showcase highlighted many of the trends we’ve been seeing all year, including larger vendors transitioning to AI platform strategies, an increased focus on workflow integration and care coordination, and a greater emphasis on radiologist efficiency. There were also far fewer brand-new AI tools than previous years, as many vendors focused on improving their current products and/or expanding their portfolio via partnerships. 

PACS Cloud Focus – PACS vendors continued to place a major emphasis on their respective cloud advantages, and there was a widespread consensus that cloud is on every imaging IT roadmap. The PACS vendors seemed to talk less about multi-ology enterprise imaging than previous years, and expanding EI beyond radiology/cardiology still seemed pretty futuristic for most players. It was also quite clear that most of the PACS players’ AI marketplaces/platforms haven’t been as prioritized as earlier announcements might have suggested.

Best RSNA Since… 2019 – We’ve heard some folks saying this was the “best RSNA ever” because it was easy to get around and it was great to see everyone, but those seem more like pandemic silver linings than “best ever” qualifications. Still, the imaging industry made the most of RSNA 2021, and everyone seemed truly happy to be together again after two long years of working from home. As long as COVID cooperates, we should be set up for an excellent RSNA 2022.’s Care Coordination Expansion 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’s 3D mobile viewer and clinical communications workflows. It appears that partnered with Avicenna.AI to create these modules, representing a logical way for 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.’s Care Coordination Strategy – 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 could earn “leadership” status across all other imaging-detected emergent conditions.

The Takeaway –’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.

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.

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, & 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, & 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.

Who Owns AI Evaluation and Monitoring?

Imaging AI evaluation and monitoring just became even hotter topics, following a particularly revealing Twitter thread and a pair of interesting new papers.

Rads Don’t Work for AI – A Mayo Clinic Florida neuroradiologist took his case to Twitter after an FDA-approved AI tool missed 6 of 7 hemorrhages in a single shift and he was asked to make extra clicks to help improve the algorithm. No AI tool is perfect, but many folks commenting on this thread didn’t take kindly to the idea of being asked to do pro-bono work to improve an algorithm that they already paid for. 

AI Takes Work – A few radiologists with strong AI backgrounds clarified that this “extra work” is intended to inform the FDA about postmarket performance, while monitoring healthcare tools and providing feedback is indeed physicians’ job. They also argued that radiology practices should ensure that they have the bandwidth to monitor AI before deciding to adopt it.

The ACR DSI Gets It – Understanding that “AI algorithms may not work as expected when used beyond the institutions in which they were trained, and model performance may degrade over time” the ACR Data Science Institute (DSI) released a helpful paper detailing how radiologists can evaluate AI before and during clinical use. In an unplanned nod to the above Twitter thread, the DSA paper also noted that AI evaluation/monitoring is “ultimately up to the end users” although many “practices will not be able to do this on their own.” The good news is the ACR DSI is developing tools to help them.

DLIR Needs Evaluation Too – Because measuring whether DL-reconstructed scans “look good” or allow reduced dosage exams won’t avoid errors (e.g. false tumors or removed tumors), a Washington University in St. Louis-led team is developing a framework for evaluating DLIR tools before they are introduced into clinical practice. The new framework comes from some big-name intuitions (WUSTL, NIH, FDA, Cleveland Clinic, UBC), all of whom also appear to agree that AI evaluation is up to the users.

The Takeaway – At least among AI insiders it’s clear that AI users are responsible for algorithm evaluation and monitoring, even if bandwidth is limited and many evaluation/monitoring tools are still being developed. Meanwhile, many AI users (who are crucial for AI to become mainstream) want their FDA-approved algorithms to perform correctly and aren’t eager to do extra work to help improve them. That’s a pretty solid conflict, but it’s also a silver lining for AI vendors who get good at streamlining evaluations and develop low-labor ways to monitor performance.

Bad AI Goes Viral

A recent mammography AI study review quickly evolved from a “study” to a “story” after a single tweet from Eric Topol (to his 521k followers), calling mammography AI’s accuracy “very disappointing” and prompting a new flow of online conversations about how far imaging AI is from achieving its promise. However, the bigger “story” here might actually be how much AI research needs to evolve.

The Study Review: A team of UK-based researchers reviewed 12 digital mammography screening AI studies (n = 131,822 women). The studies analyzed DM screening AI’s performance when used as a standalone system (5 studies), as a reader aid (3 studies), or for triage (4 studies).

The AI Assessment: The biggest public takeaway was that 34 of the 36 AI systems (94%) evaluated in three of the studies were less accurate than a single radiologist, and all were less accurate than the consensus of two or more radiologists. They also found that AI modestly improved radiologist accuracy when used as a reader aid and eliminated around half of negative screenings when used for triage (but also missed some cancers).

The AI Research Assessment: Each of the reviewed studies were “of poor methodological quality,” all were retrospective, and most studies had high risks of bias and high applicability concerns. Unsurprisingly, these methodology-focused assessments didn’t get much public attention.

The Two Takeaways: The authors correctly concluded that these 12 poor-quality studies found DM screening AI to be inaccurate, and called for better quality research so we can properly judge DM screening AI’s actual accuracy and most effective use cases (and then improve it). However, the takeaway for many folks was that mammography screening AI is worse than radiologists and shouldn’t replace them, which might be true, but isn’t very scientifically helpful.

Unsupervised COVID AI

MGH’s new pix2surv AI system can accurately predict COVID outcomes from chest CTs, and it uses an unsupervised design that appears to solve some major COVID AI training and performance challenges.

Background – COVID AI hasn’t exactly earned the best reputation (short history + high annotation labor > leading to bad data > creating generalization issues), limiting most real world COVID analysis to logistic regression.

Designing pix2surv – pix2surv’s weakly unsupervised design and use of a generative adversarial network avoids these COVID AI pitfalls. It was directly trained with CTs from MGH’s COVID workflow (no labeling, no supervised training) and accurately estimates patient outcomes directly from their chest CTs.

pix2surv Performance – pix2surv accurately predicted the time of each patient’s ICU admission or death and applied the same analysis to stratify patients into high and low-risk groups. More notably, it “significantly outperformed” current laboratory tests and image-based methods with both predictions.

Applications – The MGH researchers believe pix2surv can be expanded to other COVID use cases (e.g. predicting Long COVID), as well as “other diseases” that are commonly diagnosed in medical images and might be hindered by annotation labor.

The Takeaway – pix2surv will require a lot more testing, and its chance of maintaining this type of performance across other sites and diseases might be a longshot (at least right away). However, pix2surv’s streamlined training and initial results are notable, and it would be very significant if a network like this was able to bring pattern-based unsupervised AI into clinical use.

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