- MRI Scan-a-Van: Researchers from Brown University and Hyperfine developed a proof-of-concept mobile MRI by installing a Hyperfine Swoop low-field MRI in a modified cargo van. After finding that their “Scan-a-Van” might be able to perform quick and accurate point-of-patient brain MRI exams (e.g. at patient homes or assisted living facilities), the authors suggested that this approach might improve imaging access for patients unable to travel to imaging centers or hospitals (e.g. mobility challenges, rural locations). The study also comes just a few months after MUSC Health similarly debuted the first stroke ambulance mobile MRI.
- FDA LVO AI Reminder: The US FDA published a “reminder” that AI-based LVO (CADt) detection tools are intended to triage and prioritize suspected large vessel occlusion cases, not provide diagnostic information or eliminate cases from radiologists’ worklists. That might seem like an unnecessary reminder, but this is actually a strong message from the FDA (who rarely publishes these reminders) that clinicians should not take clinical action based on CADt output alone. Considering the growing role of CADt care coordination tools that deliver their output via smartphones to non-rad clinicians, there are plenty of opportunities for clinicians to misunderstand CADt’s intended use case and act without confirming radiologists’ interpretations.
- Developing a PET Supplement: A Memorial Sloan Kettering-led team developed a Cerenkov light imaging (CLI) technique that could evolve into a PET supplement for patients in low-resource settings. The researchers performed CLI on 96 patients with existing or suspected tumors, finding that the technique localized tumors with ‘acceptable’ or higher agreement with PET exams among 90% of the patients. The researchers aren’t looking to replace PET, but CLI’s speed and cost benefits could make it a viable PET supplement for smaller/rural providers without an on-site PET system or in countries without nearly enough PET scanners.
- Arterys’ Neuro AI Expansion: Arterys continued its Neuro AI platform expansion, officially launching its Combinostics-based cNeuro cMRI solution, which automates white matter lesion quantification to support multiple sclerosis and dementia assessments. cNeuro joins a growing list of solutions on the Neuro AI platform, which also includes LVO/ ICH stroke detection, 4D Flow, and tumor diagnostics tools.
- De-innovation: Becker’s published an interesting article exploring Cleveland Clinic and other organizations’ “de-innovation” processes for discontinuing practices that no longer benefit the health systems. The piece included recommendations such as ensuring that “data from the legacy applications is archived and the technology is sunset” and that “hospitals should apply the same enthusiasm they have when launching new projects to asking the hard questions about which operations might need to be eliminated.” The article didn’t mention radiology, but we’re sure some readers can think of imaging processes that belong on their de-innovation list.
- Emory Adds Fujifilm Cardiology PACS: Fujifilm Healthcare installed its 7x Synapse Cardiology PACS across Emory Healthcare, including the Georgia system’s 11 hospitals and 23 heart and vascular clinics. The installation comes just a year after Sectra announced an enterprise imaging deal supporting Emory’s radiology and orthopedics departments.
- RP Takes the Stand: The federal criminal conspiracy case against DaVita and its CEO named Radiology Partners among the three companies that allegedly had anti-poaching agreements with the dialysis giant. RP seems to have played a smaller role, which is so far limited to an informal “ground rules” email after RP already hired several DaVita execs. That said, this is a high-profile antitrust case with major healthcare labor implications, and RP’s hiring practices might be subject to more public scrutiny as the trial continues.
- All in on Admin AI: Clinical AI might get all the headlines, but a recent MIT Sloan Management Review editorial stated a compelling case for why “administrative AI may be a better bet.” The authors detailed clinical AI’s long list of adoption and ROI barriers (regulatory approval, physician acceptance, workflow integration, reimbursements, inefficiency), arguing that admin AI solutions avoid many of these barriers and target problems that are far more solvable and more likely to deliver economic ROI. After all, administrative tasks are responsible for 34% of US healthcare costs and probably an even higher share of healthcare waste.
- Surveillance Mammography Declines: A new JNCCN study revealed significant declines in surveillance mammography adherence since 2009. Analysis of 142k commercially insured breast cancer survivors (aged 40-64yrs) showed that annual surveillance mammography use declined by 1.5% annually from 2009 to 2016 (73.7% to 67.1%) after remaining stable from 2004 to 2009 (~74%). These declines occurred across healthcare and demographic segments, although they were greatest among younger women (40-49yrs fell by 2.8%/yr after 2009) and women with lower deductibles (-2,1%/yr after 2010).
- Encord’s Auto-Annotation Launch: Computer vision startup Encord announced the launch of its CordVision DICOM image annotation solution, which uses deep learning to automate manual labeling processes. AI development teams would label a small/representative set of images that are used to train CordVision to automatically label a much larger database – and that data would be used to train AI models. Equipped with $17M in 2021 VC funding, Encord joins a growing group of startups targeting the “picks and shovels” side of AI, alongside labeling competitors like Centaur Labs and Gradient Health, as well as AI validation companies like Gesund and CARPL.ai.
- Hospital M&A Slowdown: Only 12 US hospital acquisitions were announced in Q1 2022, the lowest total since Kaufman Hall began tracking the metric in 2016. On top of the low total, there were no mega transactions where the acquisition target had revenues over $1B, driving down the average revenue of the acquired hospital to $246M (vs. $619M in 2021). Kaufman Hall referred to the Q1 results as “an interesting anomaly,” and expects activity to pick back up given the clear benefits of health system scale.
