Summary of AI/ML Interviews with Industry by the SIIM Machine Learning Industry Liaison Subcommittee

The ML Industry Liaison Subcommittee of the SIIM Machine Learning Committee completed 35 interviews with the intent of collecting insights into the vendor’s perceptions of Artificial Intelligence/Intelligent Automation and SIIM’s role in this technology space.

The team effort allowed the interviews to span across vendor segments of technology companies, imaging/iron vendors, PACS vendors, and industry veterans. The interviews produced qualitative responses that are summarized by the team below.

The survey included companies/people interviewed who were not members of SIIM directly and some were not corporate sponsors as well SIIM members/sponsors.

In the context of the interviews, we did not define the terms of AI, ROI nor standards for the respondents so that they were free to elaborate on the topics. Below are the findings by topic, summarized as themes, with key quotes where the team wanted to call out illustrations

Michael LaChance and Mitchell Goldburgh as co-chairs of the Industry Liaison Subcommittee wanted to personally thank all of the company representatives and our committee members for their participation in this exploration of AI/ML adoption.

Is there an ROI as an industry – in the context of billions of dollars invested in AI software development thus far

  • Without provider reimbursement, direct financial ROI for disease detection algorithms – and particularly so for algorithms targeting a single disease – is generally not perceived to sound business reasoning for vendors or providers. Further, although several algorithms have been granted reimbursement, broad reimbursement is not anticipated.
  • Likewise, direct financial ROI from productivity tools (including triage) is not perceived as a sound business rationale for vendors or providers with the possible exception of large health systems and large reading practices.
  • Providers can realize ROI from the advancements in precision that AI offers in analysis and downstream decisions, therefore, impacting the cost of care pathwaysRespondent Quote: ‘The greatest ROI potential for diagnostic radiology information is outside of the department. It is with AI applied to diagnostic radiology information specifically targeting care pathways, outcomes, population health’
  • Multiple mentions indicated existing reimbursement models assume an acceptable performance and continuous improvement so additional reimbursement for AI won’t be for improvement in triage, detection, diagnostic efficiency.Respondent Quote: ‘It’s about answering two questions: ‘Am I going to die today?’ (better diagnosis and outcomes) and ‘Do you know how I am going to die (in the future)? (better population health/wellness)’
  • Multiple versions of responses indicated that AI can and will deliver clinical and operational benefits. Further, many believe adoption is a must-have to achieve clinical and operational outcomes, but there won’t be direct new revenue and/or increased margin hence direct financial ROI to justify new investment.
  • A couple of technology companies mentioned that they believe that the bulk of AI Vendors are about to hit a cliff of consolidation – somewhere in the 2-4 year horizon

What is the ROI as a company

  • Technology companies or AI software companies confident about achieving productivity savings and efficacy of their AI but many did not have quantifiable cases. Several PACS and AI companies are cooperating around pilots building business cases here.Respondent Quote: ‘Not sure how you produce an ROI on 10M+ in R&D”, perhaps the investment is sunk costs and now the ROI is from this point forward.’
  • PACS companies see AI as a necessity – like seat belts in a car – PACS needs AI and they are funding the use of it within their licensing agreements (not all but mentioned often)
  • The mention of slow rollouts of AI due to the investment/verification as a PACS company – tending to partner with vendors who have more than one algorithm focus.Respondent Quote: ‘Large data companies use AI collaboration to drive providers onto their data platforms (e.g., adopting a data companies cloud platform)’
  • Iron/Modality companies see AI as means to sell more IRON/modalities, not as a feature that has a standalone ROI, but a differentiator making imaging doses less or impacting protocols
  • Existing AI companies have gotten reimbursement for specific areas which has made them viable but many of them have been funded well under the emerging companies.
  • Marketplaces are interesting but not a clear winner as the adoption, evaluation, and ROI of AI is not easily understoodRespondent Quote: ‘Just look at XX company who pivoted to a new platform after a huge investment in AI” – there is no ROI.’

What do you think of Standards in AI

  • Consistently, the industry expressed that as vendors they are ahead of the standards when it comes to integration
  • Technical standards are watered down to the point where the variability makes them difficult to adopt or the use cases that post the inflection where AI can be most usefulRespondent Quote: ‘We know that as a PACS company we have the best platform for integration of AI but the AI companies all have different models so it takes time to bring them to market.’
  • Integration by the PACS vendors of AI vs AI neutral platforms is a battleground that inhibits broad adoption
  • AI standards are weak in how AI results are integrated into reporting and viewing
  • Qualified, curated data sets are key for training, evaluation, and regulatory approval
  • IHE is only on some roadmaps for product managers – a must-have but not a priority today

Respondent Quote: ‘It’s about integrating into the workflow’

Written by

Mitchell M. Goldburgh

Publish date

Sep 13, 2021


  • Artificial Intelligence

Media Type

  • News & Announcements

Audience Type

  • Clinician
  • Developer
  • Imaging IT
  • Researcher/Scientist


PIIRT Chapter 10: Ancillary Services

May 7, 2024

Dive into the fascinating world of imaging informatics with SIIMcast, the official podcast of the Society for Imaging Informatics in…


Between Two Fellows with Dr. Elizabeth Krupinski and Dr. Tessa Cook

Apr 16, 2024

Welcome to another special episode of the SIIMcast, recorded live at the SIIM Annual Meeting in Austin, Texas. Join us for a…


Effective Strategies for Post-Production Monitoring of AI Algorithms in Radiology

Apr 1, 2024

The integration of artificial intelligence (AI) algorithms into radiology practices is rapidly increasing, with many promising applications on the horizon.…