Premium Reports
Contact KLAS
 Download Report Brief  Download Full Report    Zoom in charts

Preferences

   Bookmark

Related Series

Artificial Intelligence in Imaging 2018
|
2018

Related Segments

Related Articles

 End chart zoom
Artificial Intelligence in Imaging 2020 Artificial Intelligence in Imaging 2020
* A page refresh may be necessary to see the updated image

Artificial Intelligence in Imaging 2020
Shifting Mindshare

author - Monique Rasband
Author
Monique Rasband
 
March 5, 2020 | Read Time: 7  minutes

Current Time Inside Cache Tag Helper: 11/29/2022 4:55:46 AM and Model.reportId = 1531

Many provider organizations are exploring the deployment and adoption of artificial intelligence (AI) in imaging, and market competition has begun to increase among vendors who offer this technology. Based on feedback from 81 organizations with advanced imaging strategies, this report explores what healthcare organizations have planned for AI in imaging and their perceptions of which vendors are best positioned to help (responses could apply to modalities or software). It also identifies best practices related to AI deployment and adoption as reported by some of the healthcare industry’s most successful AI users. In the near future, KLAS plans to publish in-depth case studies from customers of some of the most frequently mentioned vendors to explore what is currently being achieved with AI in imaging.

HtmlReportContent Current Time Inside Cache Tag Helper: 11/29/2022 4:55:46 AM and Model.reportId= 1531 and Model.HtmlReportContent_LastWriteTimeUtcInTicks=637798632195926660

Defining AI in Healthcare

In collecting this research, KLAS used the following broad definition of AI: algorithms and software that approximate human cognition in the analysis of complex data. However, since this research was collected, KLAS has solidified a more specific definition for AI in healthcare (see the graphic to the right). Moving forward, KLAS intends to validate AI in imaging using this latter definition, with consideration for the nuances specific to imaging.

KLAS Definition of Healthcare AI Software

Software that provides machine learning (ML) or natural language processing (NLP) capabilities for healthcare-related clinical, operational, or financial areas.

ML tools (for structured data) study and learn computer systems’ algorithms and statistical models to effectively perform tasks without requiring explicit instructions, relying instead on patterns and inference to determine results.

NLP (for unstructured data) enables software solutions to understand, process, and analyze natural language (whether speech or text).

KLAS does not use the term AI to mean technology with all capabilities of human intelligence.

† This definition comes from KLAS’ 2019 Healthcare AI report.

State of the Market: How Have Provider Organizations’ Plans Changed since 2018?

organizing status with ai in imaging

AI in Imaging Has Accelerated

In 2018, many interviewed organizations felt more time and research were needed to identify use cases and prove an ROI for AI in imaging. Now, nearly half of interviewed organizations are live with or, more commonly, piloting an imaging AI program (30% are piloting; 16% are live). Those in pilots report use cases related to nonspecific algorithms that enhance workflows and help physicians diagnose conditions. Those who are live report multiple use cases, including natural language processing of radiology/clinical notes and post-processing of images to make it easier for radiologists to diagnose.

when do organizations making plans expect to go live

Deployment Is Closer Than Ever

Organizations are closer than ever to deploying AI in imaging. In 2018, over one-third of interviewed organizations with AI plans reported being at least three years out from deployment. That percentage has been cut in half, with no organizations reporting in 2019 that they are more than five years away. Those with the furthest timelines report hesitation caused by an influx of AI technology that lacks a proven use case. Even with this uncertainty, many organizations mention a willingness to partner with vendors and pilot new technology.

depth of anticipated adoption for ai in imaging

Depth of Adoption Expected to Increase

Organizations today expect deeper AI adoption (e.g., more use cases and more imaging departments) than did organizations interviewed in 2018. One contributing factor is less expected pushback from physicians. A radiology VP explained, “A year or two ago, . . . my radiologists and cardiologists thought that AI signified doomsday. Today, those same physicians are asking when they can get their hands on AI. They see the value that AI can bring to them and their patients.”

IBM Watson Health Leads in Mentions; Considerations of Other Vendors Increase

what vendors are best positioned to deliver ai in imagingIn 2018, IBM Watson Health was mentioned three times more than any other vendor when report participants were asked which vendor is best positioned to deliver AI in imaging. While IBM Watson Health still receives the most mentions, their total mindshare has decreased by almost half due to increased interest in a variety of other vendors. Some respondents mention IBM Watson Health’s head start in AI in general, and indeed the vendor’s general AI offerings have a larger customer base than their imaging AI offerings. Referring specifically to these imaging offerings, respondents (who may or may not be actual customers) report tempered confidence, noting that despite the vendor’s large pool of data and resources, development has been slow.

Aidoc and Nuance have seen the biggest increases in mindshare, though the vendors differ widely. With multiple solutions that already have FDA clearance, Aidoc is described as being ahead of the game and as having a fast-moving, targeted approach. Nuance customers have not yet seen concrete AI deliverables, but respondents feel the vendor’s existing speech recognition customer base may give Nuance an advantage in imaging AI.

Several respondents also feel that provider organizations and niche/start-up vendors will be well positioned to facilitate AI success in imaging. Provider organizations—specifically large organizations and academic health centers—are seen as having the experience to know how AI can benefit patients and the data resources needed to make it successful. Though not specified by name, niche/start-up vendors were noted to be flooding the AI market and are seen as innovative but potentially lacking long-term viability.

Zebra Medical Vision Inspires Confidence While Gaining Visibility; GE Healthcare and Nuance Also Grow in Mindshare

mindshare vs. provider confidence in deliveryAfter IBM Watson Health, the next most frequently mentioned vendors were Zebra Medical Vision, GE Healthcare, and Nuance. Zebra Medical Vision is described by respondents as having a strong understanding of imaging and as being neutral, nimble, and willing to partner with other vendors. Many also mention that several Zebra solutions already have FDA approval. GE Healthcare and Nuance are seen as well positioned to deliver AI in imaging based on their existing installs of other technology. Respondents highlight GE Healthcare’s access to capital and data and mention the vendor’s AI efforts related to imaging modalities (e.g., flagging potentially critical images); a couple report concerns about GE (one based on previous experiences and one related to cost). Respondents who mentioned Nuance are optimistic about the early information and promises Nuance has shared. 


iSchemaView was mentioned by just three respondents but receives the highest average confidence rating. All three respondents report excitement about the vendor’s RAPID technology and its ability to help providers deliver faster, more precise care to stroke patients.

Other Mentioned Vendors

The following vendors were mentioned by one or two respondents each. They may offer AI technology only through partnerships with other vendors or they may not currently have an AI offering.

One Mention

  • Change Healthcare
  • Ferrum Health
  • iCAD
  • Intelerad
  • LUMEDX
  • Riverain Technologies
  • Vital
  • Viz.ai (through Medtronic)
  • ZoftSolutions

Two Mentions

  • Invivo, a Philips Company
  • Epic
  • MModal
  • TeraRecon EnvoyAI
  • Visage Imaging

Provider Perspectives on Top-of-Mind Vendors

Includes only vendors with three or more mentions; vendors ordered alphabetically

Given the uncharted territory of the AI market, provider organizations place great weight on the references of their colleagues when making technology purchase decisions. Below are early insights from respondents regarding the top-mentioned vendors for imaging AI. These insights represent general market perceptions of the vendors, not feedback from actual clients.

provider perspectives on top of mind vendors

Note: A historical look at several vendors’ general ability to deliver new technology can be found in KLAS’ 2018 Imaging AI report.
‡ Respondents were asked to name the vendors best positioned to deliver AI in imaging and to rate their confidence in the vendors’ ability to deliver.

How Healthcare Organizations and Technology Vendors Can Drive Success with AI

As reported by some of the healthcare industry’s most successful AI users

KLAS’ other two recent reports on healthcare AI (a 2018 report on AI in imaging and a 2019 report on AI in healthcare in general) offer several best practice suggestions healthcare organizations and vendors should consider for the deployment and adoption of AI.

Best Practices for Healthcare Organizations

Embed AI into the workflow

  • Observe clinician workflows and find the right way and place to share AI insights. AI should work with clinicians and not be extra hoops to jump through.

Bring together experts on AI

  • AI should be an interdisciplinary collaboration to promote successful rollout across all stakeholders.

Take ownership of change management

  • Take a social engineering approach to engaging staff in changes and report progress and success to staff to encourage adoption.

Look for vendors with the following attributes

  • Crystal clear expectations
  • Proactive strategic relationships
  • A central focus on training
  • Strong data governance

Best Practices for Vendors

Deliver comprehensive services for AI platform

  • Customers benefit when vendors have healthcare experts on staff in roles such as client success managers, dedicated data scientists, and field engineers.

Provide strong training for customers

  • Whatever form training takes (e.g., at-the-elbow, train-the-trainer), customers are most satisfied when organization analysts are given the data-science training needed to build, deploy, and maintain their own models.

Be a humble, active partner

  • Due to the immaturity of healthcare AI, customers benefit when vendors are transparent about their progress, treat customers as development partners, and encourage constructive feedback.

† Additional information on partnership ratings for some vendors can be found in KLAS' 2018 imaging AI report.

author - Elizabeth Pew
Writer
Elizabeth Pew
author - Madison Moniz
Designer
Madison Moniz
 Download Report Brief  Download Full Report

This material is copyrighted. Any organization gaining unauthorized access to this report will be liable to compensate KLAS for the full retail price. Please see the KLAS DATA USE POLICY for information regarding use of this report. © 2022 KLAS Research, LLC. All Rights Reserved. NOTE: Performance scores may change significantly when including newly interviewed provider organizations, especially when added to a smaller sample size like in emerging markets with a small number of live clients. The findings presented are not meant to be conclusive data for an entire client base.