Uncategorized

Healthcare Transparency Finalized | A Solid AI Study | Nudged to Screening

“Under the status quo, healthcare prices are about as clear as mud to patients,”

CMMS leader, Seema Verma, introducing a range of ambitious new transparency rules that should make healthcare costs a lot less muddy. For patients anyway.


The Imaging Wire

Healthcare Transparency Finalized

The White House rolled out its finalized healthcare cost transparency rules that will go into effect in January 2021, staying firm on its previous requirements and expanding its focus to insurers. These requirements go beyond what many in the industry expected and could bring notable changes to the way the U.S. healthcare industry works.

Hospital Impact – Hospitals will have to share healthcare cost info, including gross costs, insurer-negotiated rates, and discounts that patients receive for paying for their own care, while forcing all hospitals to publish a list of 300 “shoppable” services.

Radiology Impact – Included among this list of 300 “shoppable” services are 13 schedulable imaging procedures (e.g. mammography, abdominal ultrasound, head CT & MRI, others).

Insurer Impact – A new associated proposal (now open for comment) will require insurers to reveal negotiated rates with hospitals/clinics and create online tools that patients can use to calculate their out-of-pocket rates for all services.

Industry Pushback – The healthcare industry is of course not as excited about lowering healthcare costs as the Trump administration, and more challenges in the press and courts are surely on the way. This is a big enough deal for hospitals and insurers to take the same side, arguing that the ruling will create confusion, reduce competition, hinder innovation, and even increase healthcare costs (that’s a good one).



A Solid AI Study

A team of South Korean researchers found that a commercial deep learning tool outperformed 12 radiologists in detecting malignant pulmonary nodules on chest X-rays, and more importantly improved their performance when they used the solution.

The StudyThe Samsung-funded multicenter study used 800 chest radiographs (600 w/ lung cancer, 200 normal, all marked for regions of interest) to train the model, and then had the 12 radiologists review 2,400 chest X-rays with the assistance of the DCNN software.

The Results – The software improved the radiologists’ average sensitivity (70.3% vs. 65.1%) and decreased false positives per chest radiograph (0.18 vs. 0.20), while resulting in 104 positive changes (from false-negative/positive to true-negative/positive) and 56 negative changes (from true to false).

The Praise – Although AI studies are a daily occurrence at this point, this one was widely praised for the AI tool’s ability to spot inconspicuous cancers and notify radiologists, its solid methodology (removed all benign modules, specific malignant modules, ground truth), and its transparency (they disclosed chance of bias).



Nudged to Screening

A new study from UPenn found that adding an active choice intervention “nudge” to EHR systems that prompts medical assistants to inform patients of their breast and colorectal cancer screening options significantly increased screening orders.

The retrospective study (n= 69,916 patients eligible for breast or colorectal cancer screening) found that the interventions increased breast cancer screening orders by 22 percentage points and increased colorectal cancer screening orders by 13.7 percentage points, although patient screening rates didn’t change.

This UPenn team is becoming an EHR nudge research leader, as it previously used the method to cut unnecessary imaging for palliative radiation therapy and to improve screening decision making.


The Wire

  • Ultromics announced the FDA clearance of its EchoGo Core echocardiography image analysis system, which uses AI to calculate cardiac ultrasound left ventricular ejection fraction, left ventricular volumes, and automated cardiac strain. EchoGo Core is the first AI solution for cardiac strain, which will become reimbursable in January. Next up, Ultromics is planning to launch its new EchoGo Pro solution (predicts cardiac disease from echocardiography) and expand to Europe and Asia.
  • University of Illinois at Urbana-Champaign and Wayne State University researcher landed a four-year NIH RO1 grant to develop an image reconstruction solution that would improve ultrasound tomography’s ability to detect early-stage breast cancer. In addition to ultrasound tomography’s radiation advantages, the team highlighted its sensitivity with certain tissues that will make it easier to detect subtle breast cancers, especially with the help of the image reconstruction solution.
  • An editorial in the Journal of Neurointerventional Surgery suggested that increased compensation may be a very effective way to reduce neurointerventional physician burnout. A survey of neurointerventional physicians (n=293) found that 56% meet the criteria for burnout and revealed that covering multiple hospitals on-call (59% to 67%), compensation (65.2%), and feeling underappreciated were directly correlated with burnout, while physicians who received additional payments for call periods reported far less burnout. Less than 40% of the physicians who received $1k per 24hr call period reported burnouts. This auntminnie.com article provides a lot more details on this study.
  • New research in the International Journal of Cancer provided more evidence that population‐based mammography screening leads to breast cancer incidence and mortality improvements. A review of the German mammography screening program (n=323,719 patients, 2003/04 to 2015/16) found notable 10-year improvements for both the 50‐59yr and 60‐69yr age groups, including declines in stage III and IV BC incidence (50‐59yr: -24.2% and -23.0%; 60‐69yr: -28.3% and -24.2%), BC mortality (50‐59yr: -25.8%; 60‐69yr: -21.2%).
  • Shortly after landing FDA clearance for its CT-guided robotic navigation system, XACT Robotics completed a $36 million funding round that it will use to support its commercial rollout and ongoing R&D. It’s worth noting that the latest round included Chasing Value Asset Management, which previously held positions in an impressive list of recently acquired surgical robotics startups including MAKO Surgical Corp. (acquired by Stryker in 2013), Mazor Robotics (Medtronic in 2018), and Corindus Vascular Robotics (Siemens in 2019).
  • A team of Ontario-based researchers found that adopting patient-centered MRI and CT referral forms improved imaging order completeness, but not appropriateness. The study reviewed 399 requests for MRI/CT lumbar and head scans to set a baseline (51.6% appropriate, 12.0% inappropriate, and 36.3% incomplete) and then reviewed 251 requests from 149 primary care physicians using the new referral forms, finding that incomplete referrals dropped from 36.3% to 13.15%, but didn’t significantly change order appropriateness.
  • MaxQ AI’s ACCIPIO ICH and Stroke Platform is now available on the Arterys Marketplace, adding on to an already impressive distribution network that also includes Nuance, Philips, GE, IBM Watson Health, TeraRecon, and Samsung Neurologica (many added within the last year). The Arterys Marketplace is an internet-based medical imaging AI platform, combining Arterys and partner solutions for the analysis of cardiac MR, lung CT, chest x-ray, and (with the addition of MaxQ AI) non-contrast head CT images.
  • Google’s entering its second consecutive week on the wrong side of the patient privacy spotlight, following reports that the tech giant almost made 100,000 chest X-rays with identifiable info public in 2017 but avoided the breach due to a last-minute warning from the NIH. Google quickly deleted all images and canceled its project with the NIH, but reports reveal that the project lacked legal agreements over patient info and the data wasn’t properly reviewed. Although the key word in this story is “almost,” the narrowly avoided breach is still viewed as a sign of what could go wrong as Google expands deeper into healthcare and the story doesn’t help ease concerns about Google’s ability to keep patient data private.
  • ContextVision introduced its new Rivent ultrasound image enhancement solution, addressing the trend towards higher-resolution ultrasound by optimizing image quality in both near and far field and supporting automatic adaption to line density in real-time. Rivent reportedly supports well-defined borders and noise reduction, supporting smooth tissue imaging while keeping tissue information.
  • New research out of China and published by AJR found that fluorodeoxyglucose PET (FDG PET) combined with high-resolution CT (HRCT) could be a valuable way to predict the histopathologic subtypes and growth patterns of early lung adenocarcinoma, although it doesn’t help differentiate growth patterns. The retrospective analysis of PET/CT data on ground-glass nodules (GGNs) among patients with stage IA lung adenocarcinoma found that FDG uptake was higher in GNNs that also had invasive HRCT signs, while uptake was higher in invasive adenocarcinoma than in adenocarcinoma in situ.
  • Reports from Ireland reveal that patients in the country’s public health system are waiting an average of 126 days for a brain MRI, versus just six days for patients with private coverage, using this difference to highlight the “unfair economic division that is now endemic in Ireland’s health system.”

[ cta]


The Resource Wire

  • Qure.ai is ahead of the AI pack at RSNA 2019 with four research presentations on 1) Evaluating radiology AI models; 2) Comparing AI-based TB screening and bacteriological tests; 3) Segmenting and measuring ventricular and cranial vault volumes; and 4) How clinical context improves AI performance for cranial fracture detection.
  • POCUS Systems is approved as a Veteran Owned Business with the US Government Office of Veterans Business Development, paving the way for partnerships with the federal healthcare delivery systems.
  • By partnering with Medmo, imaging centers can keep their schedules full and their equipment busy. Here’s where to learn more and get started.

You're signed up!

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

-- The Imaging Wire team