“Why not? Why aren’t doctors standing up for themselves and their patients?”
Eric Topol questioning why doctors aren’t organizing and taking a stand against the problems with the medical system.
Imaging Wire Sponsors
- Carestream – Focused on delivering innovation that is life changing – for patients, customers, employees, communities and other stakeholders
- Focused Ultrasound Foundation – Accelerating the development and adoption of focused ultrasound
- Medmo – Helping underinsured Americans save on medical scans by connecting them to imaging providers with unfilled schedule time
- Nuance – AI and cloud-powered technology solutions to help radiologists stay focused, move quickly, and work smarter
- Pocus Systems – A new Point of Care Ultrasound startup, combining a team of POCUS veterans with next-generation genuine AI technology to disrupt the industry
- Qure.ai – Making healthcare more accessible by applying deep learning to radiology imaging
The Imaging Wire
Why Doctors Should Organize
Physician and AI leader, Eric Topol, took to The New Yorker this week to voice his concerns over physicians’ “appalling working conditions and the deteriorating doctor-patient relationship,” and state his case “why doctors should organize.” Here’s some highlights:
- There’s No Action – There’s been plenty of talk of physician burnout, but Topol argued that the million-plus physicians in the U.S. aren’t taking advantage of the influence they have.
- Why Not – Despite serious concerns, Topol points out that “there have been no marches on Washington, no picket lines, no social-media campaigns,” and asks: “Why not? Why aren’t doctors standing up for themselves and their patients?”
- One Reason is – Outside of the AMA (with “only” about 250k members), most other physician organizations are relatively small and specialty-focused (and largely focused on protecting doctors’ earnings), leaving doctors without a central or care-centric voice.
- Physicians Need More – Topol argued that doctors would be a lot better-off today if physicians and industry groups took a stand against EHRs and RVUs in previous decades, while pointing out that most medical associations are more interested in generating revenue than fighting for change.
- Imagine – Topol asked readers to imagine a “doctors’ organization devoted to patients” and “restoring the human factor” of medicine, rather than “the business of medicine.” This organization would theoretically bring important changes to healthcare that benefit both clinicians and patients.
A lot of changes will have to happen in order for an organization like this to become a reality, but it’s easy to see that Topol struck a nerve with this piece. The combination of the always-popular anti-EHR/burnout storyline with the premise of physician empowerment (along with Topol’s and The New Yorker’s huge platforms) is giving this story some serious online momentum.
Exo’s Big Goals
Exo Imaging announced the launch of a $35 million Series B round (adding to $50m from previous rounds) that it will use to fund the development, FDA clearance, and commercialization of its Exo Ultrasound Platform. However, Exo’s goals go far beyond bringing just another handheld ultrasound to market:
- Big Goals, Big Words – Exo aims to “up-end the economics and accessibility of medical imaging” and develop imaging and therapeutic applications that scale “from low-cost handheld imaging to the highest end equipment.”
- Beyond Ultrasound – Exo’s “highest end” and “therapeutic” goals may have some ultrasound industry insiders wondering how the company will do this with a handheld system. That’s the twist… Although Exo Imaging will initially lead with its handheld ultrasound, the company plans to bring its technology to other applications (e.g. catheters, wearable devices, CT, and focused ultrasound) through technology partnerships.
- Until Then – Exo’s forthcoming ultrasound is still pretty intriguing, as it can reportedly image the entire body (CV, MSK, abdomen) in 3D with a single handheld probe.
Mednax Steps Back
- The Step Back – Mednax disclosed that it took “a step back on radiology acquisitions” during Q2 due to soaring practice valuations, driven by demand private equity-backed firms. Mednax is now facing practice valuations with “higher than double-digit multiples” that it views as high enough to avoid radiology M&A “for the foreseeable future.”
- Reverse Course – Given the role of acquisitions in Mednax’s history and ongoing strategy, this is a pretty big shift. Especially considering that the company significantly increased its radiology practice M&A in 2017 and 2018 (5 radiology practice acquisitions). Mednax will continue to make some Maternal Fetal and Pediatric practice acquisitions, but this shift reduced its 2019 acquisition budget from $100m to $40m-$50m.
- Abstinence as a Strategy – The “buy low, sell high” crowd will applaud Mednax for abstaining from M&A given current multiples. Meanwhile, some in the PE-backed crowd might argue that radiology remains extremely fragmented (this is actually true) and the firms that take advantage of this consolidation will come out on top in the long run.
More Mammography DL Evidence
New research from an MIT-led team that’s already produced some compelling breast imaging AI work (see: this and this) found that mammography deep learning can be used to improve sensitivity (already supported by research) without limiting radiologists’ specificity or efficiency (this is a bigger deal).
- The Study – The simulation study (n= 223,109 mammograms, 66,661 women, 7,176 women test set) used a DL-triage workflow to identify negative mammograms that radiologists could skip and identify suspicious mammograms that rads should read.
- The Results – The DL-simulated workflow reduced radiologist reading volume by almost 20% (21,420 vs. 26,540), while attaining a sensitivity of 90.1% (vs. 90.6%) and specificity of 94.2% (93.5%). The model also performed well across a variety of populations.
Although only based on a simulated workflow, this is a first step towards using deep learning to triage mammograms, potentially overcoming the sensitivity challenges of CAD and the inefficiency of double reading. It also comes closer to what many radiologists want, which is for AI to save them time and help them focus on the most important studies.
- RadNet took its biggest step yet into the AI arena, fully acquiring imaging AI firm Nulogix (it already had a 25% stake in Nulogix) and creating a new division dedicated to “developing, acquiring and investing in technologies that focus on image interpretation and radiology business processes.” This is an efficiency play for RadNet, which noted that it spends almost 20% of its net revenue on radiologist interpretations, and suggested that solutions that make radiologists more productive/accurate will have a major impact on its financial performance.
- A team of Belgian scientists developed a bimodal agent that could allow clinicians to simultaneously use MRI and photoacoustic imaging (PAI), improving the sensitivity of both modalities with a single agent. The team linked a gadolinium-based agent (Gd-PCTA) with an organic PAI dye (ZW800-1) using lysin to create the initial two-in-one agent, but noted that it might be possible to create a trimodal agent in the future by adding peptides that recognize specific biological disorders.
- British radiology groups are putting new UK Prime Minister Boris Johnson to work right away, publishing a public letter urging Johnson to address concerns over how Brexit may impact the country’s radioisotope supplies and costs. The letter asked Johnson to guarantee that radioisotopes are not delayed by UK customs, that there’s sufficient domestic logistics support, and that tariffs are reimbursed for future tests and treatments.
- Research from a UCLA-led team found that PSMA PET-CT (the European standard) is superior to 18F fluciclovine PET/CT (the U.S. standard) for localizing prostate cancer recurrence after radical prostatectomy, calling it “the PET tracer of choice.” The study looked at 50 adults who underwent radical prostatectomy, with PSMA detecting recurrence in 56% of scans and 18F fluciclovine detecting recurrence in only 26% of scans.
- Handheld ultrasound startup Dolphin Medical Imaging and mixed reality glasses company ThirdEye Gen announced a partnership to combine their products, initially to make it easier for clinicians to find blood vessels for catheter placement and eventually expanding to other applications (pre-hospital, inpatient, emergency, and critical care). ThirdEye’s glasses would give Dolphin Medical ultrasound users an alternative to mobile devices for image viewing (the current standard) while supporting POCUS’ low-cost value proposition (ThirdEye’s glasses are $1,950 vs. Microsoft HoloLens’ $3,500).
- A new JACR survey (n=367) found that larger radiology practices are particularly prone to burnout, with 71% of practice leaders citing burnout as a “very significant problem” (vs. 50% overall and 37% of practices with ≤5 rads). The silver lining is that larger practices were better equipped to deal with burnout, with 34% having mechanisms in place to deal with burnout (vs. 19% overall and 3% of practices with ≤5 rads).
- LG’s IT services subsidiary, LG CNS, is preparing to launch an AI-based X-ray analysis service across 255 South Korean public health centers, following a September-December trial at Seoul’s Eunpyeong Public Health Centre. The LG CNS solution is developed through its partnership with Lunit and will at least initially focus on chest X-ray analysis for lung disease (one of Lunit’s specialties) delivered via the cloud in a software-as-a-service package. There are some uniquely-Korean components to this story (LG, Lunit, Korea’s public health centers) that may be hard to replicate in other top-11 GDP countries, but it still serves as a useful example of how a solution from an AI startup could expand across an entire country.
- Research from an Ottawa-based team found that focused cardiac ultrasonography (FoCUS) used at the patient’s bedside improves diagnosis. The team analyzed nine previous studies finding that using FoCUS for bedside cardiac exams had greater sensitivity but similar specificity compared to clinical assessment for identifying left ventricular dysfunction (sensitivity: 84% vs. 43%; specificity: 89% vs. 81%) and aortic or mitral valve disease (sensitivity: 71% vs. 46%; specificity: 94% vs. 94%).
- ScImage announced the launch of PICOM ModalityGuard, a new device that isolates unsecure legacy DICOM and non-DICOM devices in a private network, giving DICOM and HL7 traffic a similar level of protection as other informatics systems.
- A pair of Georgetown University radiologists developed a machine learning algorithm that classified radiographs according to laterality (left or right) with an AUC of 0.999, potentially helping to prevent wrong-side medical mistakes and improving efficiency. The team used 15,405 X-rays with laterality-specific info to develop machine learning models that could identify left and right lead markers as well as patients’ left and right sides. This seems like a straightforward algorithm that even healthcare’s AI pessimists could get behind.
- Cleveland Clinic researchers found that a novel 29 MHz high-resolution micro-ultrasound system outperformed conventional ultrasound systems for both systematic sampling and real-time targeting of suspicious regions during prostate biopsy. The study looked at 67 subjects who underwent prostate biopsy using the micro-ultrasound system (19 using prostate MRI imaging, 38 subjects with prostate cancer), finding that micro-ultrasound increased detection rate on prostate biopsy from 44.8% to 56.7%, giving it advantages over both MRI fusion and systematic biopsy.
The Resource Wire
– This is sponsored content.
- This Nuance blog details how VIDA’s AI-driven LungPrint Discovery tool, integrated with Nuance’s PowerScribe One reporting platform, can cut the time it takes to interpret a study with a greater understanding of the underlying patient condition.
- Qure.ai’s Rohit Ghosh takes the Tedx stage to discuss using artificial intelligence to tackle India’s TB problem.
- Carestream’s first OnSight 3D customer, Resurgens Orthopaedics (24 locations, 104 physicians), shared some of the benefits they’ve experienced from the cone beam CT system in this video, including the importance of weight bearing in surgery decisions, and the system’s image quality, ease of use, and fast study time.
- The Focused Ultrasound Foundation and MITA shared key highlights from the first year of their alliance and outlined their plans for their second year working together.
- How much does a CT scan cost? According to Medmo, that depends. Scans made with the exact same device on the exact same body part could cost $225 at one facility and $2,500 at another. Medmo also provides some advice to make sure patients don’t overpay for their scans, including using the Medmo Marketplace, where the average CT costs between $225 and $700.
- POCUS Systems’ forthcoming ultrasounds will combine ease of use, durability, and reliability, allowing clinicians to focus on their patients.