Another AI Milestone | Great Reporting | Defining POCUS


“The current CPT framework forces the square POCUS peg exclusively through the round hole of ‘diagnostic imaging’. . .”

A new JACR paper suggesting that POCUS should be defined as a patient evaluation and management process (not diagnostic).


Nominations are now open for the 2020 Imaging Wire Awards, so if you or a colleague made outstanding contributions to radiology practice this year, please take three minutes to nominate them (or you).


Imaging Wire Sponsors

  • Bayer Radiology – Providing a portfolio of radiology products, solutions, and services that enable radiologists to get the clear answers they need.
  • GE Healthcare – Enabling clinicians to make faster, more informed decisions through intelligent devices, data analytics, applications and services.
  • Healthcare Administrative Partners – Empowering radiology groups through expert revenue cycle management, clinical analytics, practice support, and specialized coding.
  • Hitachi Healthcare Americas – Delivering best in class medical imaging technologies and value-based reporting.
  • Nuance – AI and cloud-powered technology solutions to help radiologists stay focused, move quickly, and work smarter.
  • Riverain Technologies – Offering artificial intelligence tools dedicated to the early, efficient detection of lung disease.
  • Siemens Healthineers – Shaping the digital transformation of imaging to improve patient care.

The Imaging Wire



Another Medicare AI Milestone

RapidAI’s ‘Rapid LVO’ stroke platform just became the second AI software granted a Medicare New Technology Add-on Payment (NTAP), representing a major milestone for the company and for the business of imaging AI.

  • RapidAI’s NTAP – With the NTAP approval, providers who use Rapid LVO to triage suspected stroke patients (that meet certain criteria) can receive Medicare payments for each use. RapidAI didn’t reveal details about its reimbursement structure, but Viz.ai’s NTAP went up to $1,040 per use.
  • This is a Big Deal for AI, Right? – When Viz.ai announced its NTAP early last month, the radiology internet exploded with excitement, calling it a game-changer for AI adoption and for the AI business model. When RapidAI announced its NTAP last week, it seemed like the folks in the radiology media and on radiology social media barely noticed.
  • Right. – Two AI products getting an NTAP feels like a much bigger deal than one. RapidAI’s NTAP shows that Viz.ai’s add-on payment was not an anomaly, while suggesting that once CMS sees clinical value in a given use case (e.g. faster stroke diagnosis & treatment) multiple AI tools that support this use case (e.g. LVO triage & notification) could be in line for an NTAP.
  • It’s A Big Deal for RapidAI, Too – RapidAI now gets to add CMS reimbursements to its Rapid LVO value proposition (in addition better care and increasing thrombectomy volumes) and it would have been in a tough spot if Viz.ai’s ContaCT was the only LVO tool that CMS would reimburse.


Creating a Great Radiology Report

A new RSNA paper (and helpful video) from a global team detailed “how to create a great radiology report,” which might seem like an elementary topic, but it’s getting a lot of praise. Here are the big takeaways:

  • Clarity is King – There are plenty of training, workflow, and psychological reasons that some radiology reports are vague or complicated, but ‘great’ radiology reports are “readily understood and provide relevant, clear, and actionable information.” This is especially important given radiology reports’ diverse audience (referring physicians, other specialists, other rads, patients) and the high-stakes decisions they have to make.
  • Factual Findings – Radiology report findings sections should include “short, informative, and factual observations,” while avoiding interpretations (save that for impressions section), terms of perception (e.g. “is visualized”), and unnecessary words. They also encouraged radiologists to give “incidentals a proper burial,” so if an incidental isn’t serious, say it isn’t serious.
  • Impressions – The impression section should start with the diagnosis and use language that is “understandable, memorable, and actionable,” while avoiding restating evidence (leave that in findings), vague or hedging language (doesn’t protect rad or help w/ patient care), technical language (doesn’t help non-rads or impress them), or clinically insignificant observations (they could be mistaken as significant). They also encouraged radiologists to consider creating a separate recommendations section, following a checklist format.



Defining POCUS

A team of radiologists proposed “paradigm”-level changes to how point-of-care ultrasound (POCUS) is defined and reimbursed, segmenting ultrasound into four categories and positioning POCUS as separate from diagnostic ultrasound.

  • Four US Segments – The JACR paper suggested that POCUS exams are performed for “patient evaluation and management” and should be defined and reimbursed accordingly. They also suggested that diagnostic ultrasound exams should be defined by their level of comprehensiveness (Non-Comprehensive DXUS, Partial-Comprehensive DXUS, and Complete-Comprehensive DXUS), allowing for more precise coding.
  • POCUS Payments – If this proposal became a reality it could bring a major change to POCUS reimbursements, allowing multiple clinicians to perform POCUS evaluations on a given patient (and get reimbursed for it), since most payors only cover one US scan per patient.

The Wire

  • PET Positron Timer: A Japanese research team designed a ‘positron lifetime timer’ that allows PET systems to detect oxygen concentration in patient tissue, potentially helping physicians identify tumors with faster cell growth (shorter positron lifetime = more oxygen). The researchers were confident in the new timer’s real world potential, revealing that the project was “quick” and they expect the PET timer’s future adoption to be both fast (within a decade) and economical.
  • The State of AI: AI players, Nathan Benaich and Ian Hogarth, just published the most comprehensive report on artificial intelligence that we’ve ever seen, addressing everything from AI research, talent, politics, and future predictions. They cover “our” type of AI on slides 33-36, 60, and 88-92, although we recommend reviewing the whole report if you’re in the AI space.
  • Nanox’s Moment of Truth: Nanox and its critics have another reason to look forward to RSNA 2020, when the controversial would-be imaging disruptor is scheduled to demonstrate its Nanox.ARC System for the first time. The demonstration will include “a range of 2D and 3D medical imaging procedures,” followed by a Q&A and a series of clinical presentations. Although virtual trade show demonstrations aren’t usually that newsworthy, Nanox’s critics often point out that few people have ever seen the Nanox.ARC System or the images it creates.
  • DECT’s ED Advantage: Dual-energy CT improves emergency radiologists’ diagnostic confidence, reducing recommended follow-up study volumes and associated costs. That’s from a new AJR study that reviewed 3,159 ED reports (887 with DECT), finding that DECT altered management in 298 cases (9.4%) and increased diagnostic confidence in 455 cases (14.4%). DECT also avoided up to 191 MRI exams, 28 CTs, and 25 ultrasound exams compared to conventional CT, eliminating $53k – $61.5k in healthcare costs.
  • SOPHiA GENETICS’ $110M: The well-funded startup, SOPHiA GENETICS, completed a $110m Series F round (increasing its total to $250m) that it will use to expand its presence in the U.S. and Asia. SOPHiA GENETICS combines imaging/radiomics and genetic sequencing with machine learning tools to identify how diseases spread through the body.
  • MIT’s Image + Text PE AI: The MIT CSAIL team developed an AI system that can measure pulmonary edema severity using chest X-rays and radiology report language (0-3 severity scale), potentially improving physicians’ ability to predict heart failure. The researchers used 377,110 CXRs from 227,835 radiology reports to create a training dataset with a 247,425 image-text pairs, then developed a model able to identify “level 3” pulmonary edemas with 90% accuracy (level 1 and 2 = 82% & 81%). MIT CSAIL plans to integrate the PE AI system into Beth Israel Deaconess’ ED workflow next fall.
  • Breast MRI Up, But Some Underscreened: A new study in the Cancer journal revealed a significant increase in breast MRI screening, but finding that younger women with BRCA mutations are still underscreened. The study (n = 37,447 MRIs, 25,617 women) found that breast MRI screening rates increased from 2.9 to 12.1 exams per 10k women between 2006 and 2016, and MRI screening rates for women with a BRCA mutation saw much greater increases (210.8 per 10k to 1562 per 10k). However, screening rates were far lower for younger 20‐29yr women with BRCA mutations (1198.4per 10k) compared to 30‐39yr (1519.1), 40‐49yr (1567.2), and 50-64yr women with BRCA mutations (1669.6 per 10k).
  • Liver CEUS Update: The major ultrasound societies updated their guidelines for liver contrast-enhanced ultrasound (CEUS) for the first time since 2012. The new guidelines include 38 recommendations with added components focused on differentiated contrast agents, intracavitary CEUS, and CE intraoperative ultrasound.
  • 986K Missed Mammograms: UK breast cancer group, Breast Cancer Now, estimates that 986k women across the UK missed their mammogram screenings during the COVID lockdown, suggesting that roughly 8,600 women who missed these scans have undetected breast cancer. Although the NHS restarted screenings at a reduced capacity, the organization urged the NHS to “press play” on their plan to get through its backlog, despite an ongoing undersupply of radiologists. The article didn’t mention AI as a potential solution, but probably could have.
  • Structured Follow-Ups: A German radiology team developed a structured reporting software program that allowed radiologists to produce “profoundly structured” cancer treatment follow-up imaging reports, finding that the software was both effective and appreciated. The software produced 7,471 cancer treatment radiology reports within a year (processing 95% of all oncology reports within “a few months”), while 89% of the hospital’s oncologists rated the reports positively (vs. 67% for conventional reports).
  • MRI Guided Biopsy Evidence: A new Emory study detailed the effectiveness of in-bore MRI-guided prostate biopsies after positive transrectal US-guided biopsy (TRUS), revealing that the procedure increases disease upgrade rates by 40% and improves risk assessments. The study evaluated 40 men who were found to have prostate cancer via a TRUS–guided biopsy and then underwent an in-bore MRI-guided biopsy. The MRI-guided biopsies had a 65% cancer detection rate (so 35% false negatives), spotting 14 previously unidentified cancerous lesions (57% clinically significant), while upgrading 18 lesions and downgrading one.
  • Private AI: A new JACR paper from Radiology Partners’ Nina Kottler, MD shared some lessons from the mega practices’ experience adopting imaging AI to help the many private practices who haven’t yet scaled up their own AI operations. Here are some of her main points: 1) AI isn’t plug-and-play, and implementation will take work and collaboration (IT, data, radiologists, vendor); 2) AI that makes radiologists more efficient will get the most buy-in; 3) Appointing an AI leader at each practice will help drive acceptance; 4) AI can improve performance and even provide unexpected benefits (e.g. prioritization app spotting missed findings).

The Resource Wire

– This is sponsored content.

  • Vessel suppression from Riverain Technologies’ ClearRead CT software was found to significantly improve nodule detection, interreader agreement, and reading time with oncologic chest CT scans.
  • This GE Healthcare Insight post details how improving the radiology workflow with AI are improving radiology workflows across a range of modalities and clinical roles.
  • Siemens Healthineers’ c.cam cardiac camera allows you to perform high-quality scans in a small space without costly renovation fees. Discover how quickly you could be up and running with a budget-friendly cardiac camera.
  • Radiology will see a significant cut in Medicare reimbursement in 2021 if the MPFS Proposed Rule is applied without a change to the budget neutrality requirement in the law. Get all the details in this Healthcare Administrative Partners blog post.
  • This Bayer Radiology case study details how its Certegra P3T Software automates contrast enhanced abdominal CT injection protocols based on patient characteristics and contrast concentration.
  • This Nuance white paper shares insights from radiologists across nine U.S. healthcare systems on how AI is supporting their evolving roles and changing needs.

Leave a Reply

Basic HTML is allowed. Your email address will not be published.

Subscribe to this comment feed via RSS

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

Join thousands of imaging professionals.