Enterprise Imaging Summit 2024
Advancing into a New Age of Imaging

KLAS defines enterprise imaging as the ability to store and/or view images across the enterprise in one place from more than one service line and/or from multiple PACS or long-term storage solutions, enabling improved patient care through integrated image access throughout the organization.

Expanding Enterprise Imaging Strategies

In KLAS’ Enterprise Imaging 2024 report, 66% of the 131 interviewed organizations report having plans to expand their EI strategies into additional service lines over the next two years. Currently, radiology and cardiology are the core service lines most commonly incorporated into EI strategies; looking forward, many organizations are working to incorporate point-of-care ultrasound (POCUS), with high interest in additional areas, including scopes, pathology, and ophthalmology.

At the summit, speakers focused on how to expand strategies into these additional service lines. The key insights that follow are ideas compiled from (1) small-group discussions among all summit participants; (2) a keynote address delivered by Dr. Cheryl Petersilge of UPMC and Vidagos; (3) and success stories presented by Jeff Agricola and Dr. Johann Hertel of UNC Health, Neil Singh of UCSF Health, and Karissa Rothkopf of Froedtert Health.

EI Expansion Requires Strong Foundational Governance

Strong governance is crucial to the successful expansion of EI strategies, especially when moving into service lines such as POCUS, where there is high variation in how images are acquired and stored. Governing bodies should help to drive change management and guide decision-making to ensure that solutions are scalable, solve pertinent problems, align with organizational strategies, and meet clinical and technical standards.

Expanding to Point-of-Care Ultrasound Will Be Challenging but Is Necessary for EI

Dr. Cheryl Petersilge’s keynote address emphasized the need for POCUS in an EI strategy. Because clinical decisions are made based on POCUS studies, the information needs to be accessible, ideally through the EHR. POCUS consolidation can also result in benefits such as vendor consolidation, workflow efficiencies, stronger competitive advantage, and better revenue recognition. However, challenges may emerge, as POCUS devices are often used across multiple specialties with varying workflows and little governance.

Other Key Thoughts Regarding Expanding EI Strategies

• Unified worklists are key to driving workflow efficiencies for EI: AI-driven study distribution with smart routing, prioritization, screening, and load balancing is highly needed for efficient EI workflows. Opinions are split between whether such study distribution should be ultimately driven by EHR or imaging vendors.
  
  
• DICOM routing is a needed component of the EI ecosystem: It can have significant implications on workflow, especially as AI becomes more widespread. Organizations can decrease time to diagnosis and improve patient outcomes by using a DICOM router to ingest images from modalities and then simultaneously send the images to their PACS and AI solution.
  
  
• Compression and cropping are needed for encounter-based scopes and surgical videos:  Encounter-based scopes and surgical videos capture more than is clinically relevant. Organizations should outline workflows to compress videos/images and crop to the clinically relevant pieces to avoid excessive storage.
  
  
• Digital pathology requires more than a scanner and an image management system (IMS):  There is no one correct way to approach digital pathology, and going digital brings many benefits. Organizations should involve IT early while elements beyond the adoption of a scanner and an IMS are being considered, including lab workflows, monitoring, scalability, integration, and a cost-efficient archive for very large pathology images. (Read more in KLAS’ US Digital Pathology 2023 report.)
  
  
• Don’t forget the patient: EI data needs to support the trend of consumerism and patient empowerment. Image exchange and patient image access (often through patient portals) should be key areas of focus as organizations expand their strategies.

Approaching AI in Imaging

After years of being a buzzword in the imaging market, AI is now gaining real traction. Though AI technology is still in its early stages, a number of organizations have already adopted individual algorithms for specific use cases, and provider organizations and vendors express high optimism for the technology, specifically its ability to drive efficiency and offset radiologist shortages.

AI was a key topic throughout the EI summit. Insights in this section come from the following: (1) tabletop discussions among all attendees; (2) a keynote address by Dr. Nina Kottler of Radiology Partners; and (3) a panel discussion with Dr. Douglas Gentile of the University of Vermont Health Network, Dr. Jason Wiesner of Sutter Health, and Dr. Nina Kottler.

What Is the Value Proposition of AI for Imaging?

AI has significant potential to improve efficiency: Most attendees see improved efficiency as the top value driver for AI in imaging. AI has significant potential to improve workflows across an exam’s life cycle, such as image acquisition, QA, study distribution, and hanging protocols. It is anticipated that generative AI’s ability to standardize unstructured data will reduce the administrative burden of reviewing/reporting. This will be key for handling the growing imaging volume amid radiologist shortages.

AI can drive better patient outcomes: Computer vision AI can enhance radiologists’ detection and catch incidental findings within studies. While this type of AI may not have the same financial ROI implications as workflow-focused solutions, it can have a profound impact on patient outcomes, especially for use cases such as breast cancer detection. Further AI development for use cases such as care coordination, patient follow-up, and precision medicine has strong promise to improve patient outcomes.

ROI from AI differs between hospitals and radiology groups: For hospitals, ROI from AI is driven by increased income (e.g., increased procedure rates, higher DRGs) and decreased costs (e.g., decreased length of stay, faster ED turnaround). Radiology groups, on the other hand, realize the most ROI from AI through efficiency gains for radiologists. Across hospitals and radiology groups, malpractice avoidance is also a value driver from AI.

How Can Organizations Set Themselves Up for Success When Adopting AI?

Perform validations up front: Before adopting AI, organizations need to validate the algorithms to ensure they deliver the desired outcomes. Algorithm outputs should be compared to radiologist results for detection AI, and if detection levels are not enhanced, do not spend the money to adopt it.

Invest in a team that is led by clinicians: AI should ultimately be seen as a clinical tool, and leading clinicians should work closely with data scientists to drive AI. Invest in dedicated, multifunctional individuals who can communicate with IT, operations, and clinical stakeholders.

How Can Organizations Manage Bias When Using AI?

Test for bias before deployment: Although bias testing is done by the FDA, things may differ in a given organization’s clinical environment; this indicates a need to perform necessary testing before adopting the AI.

Train radiologists to avoid automation bias and complacency: Radiologists need to know where and when AI might make a mistake and what they should look for. Develop a robust feedback loop that allows radiologists to report when AI output seems wrong.

Perform ongoing monitoring of algorithms: Continually validate algorithms by running random samples of studies through AI, comparing them to radiology reports (potentially using a large language model), and keeping track of how frequently the AI report differs from the radiologist report to catch AI drift.

Consistently provide AI education for all team members: Educate team members in accordance with their involvement with AI. Clinical leaders should receive the most education so they can act as points for their team. Engage end users and create opportunities for peer learning about AI.

What Is Needed for AI to Advance?

Shared industry best practices: By more broadly sharing best practices (especially for managing AI bias for end users and performing continuous monitoring), organizations and vendors can help the industry move forward more quickly and securely with AI.

A growth mind-set: Organizations looking into AI should focus on uses cases that can deliver outcomes, even if those outcomes start small. These types of use cases often relate to efficiency and can be used to learn, educate, and grow into deeper use.

DICOM standardization: To make it easier to integrate AI into clinicians’ workflows, provider organizations and vendors need to move away from proprietary formatting and push toward DICOM and standard APIs.

Platform approaches: Provider organizations cannot manage hundreds of algorithms and vendor relationships, so platforms/AI hubs are critical for decreasing the management burden and enabling higher adoption.

Moving to the Cloud for Imaging

Summit participants agreed that the move to the cloud for imaging is inevitable—a sentiment that was also noted in KLAS’ Imaging in the Cloud 2024 report. Still, vendor development of cloud-native technologies and provider adoption is early. While vendor attendees broadly expressed excitement over moving customers to the cloud, provider attendees showed more caution due to questions about cloud performance, cost, and security.

The following insights on how to best move to the cloud come from (1) tabletop discussions and (2) a panel discussion with Amy Radonich of UC San Diego Medical Center, Audrius Polikaitis, PhD, of the University of Illinois Hospitals & Health Sciences System, and Mark Logan of Radiology Partners.

What Is the Value Proposition of the Cloud for Imaging?

The cloud has often been presented as a cost saver, but the reality is that the cloud can be cost neutral at best or even slightly more expensive than an on-premises solution. Some speculate about long-term cost-savings from the cloud, but actual savings are yet to be determined. Organizations can drive down costs through significant cloud management, but when considering the overall value proposition of the cloud, they should analyze other factors, including the following:

• The cloud provides future-proofing benefits: Moving to cloud-native solutions prepares organizations to more rapidly take advantage of technological advancements, especially AI. Likewise, organizations will be able to more easily scale with increasing volumes and computing needs, allowing them to remain competitive and efficient. This benefit will be more fully realized in the future as more truly cloud-native systems are developed and adopted.
  
  
• The cloud delivers consistency: Being in the cloud allows organizations to move away from managing multiple system versions and brings consistency to end users. Additionally, the cloud allows for the management of a single point of failure.
  
  
• The cloud can eliminate the need for data center management: With the cloud, organizations can stop paying for data center management (resources, power, etc.). Still, organizations should invest in cloud experts on their staff.
  
  

What Does a Healthcare Organization Need to Do to Successfully Move to the Cloud?

Get clinician buy-in: Moving to the cloud should not be an IT-only decision. Ensure clinicians understand the road map and reasons for moving to the cloud (especially in terms of efficiency and future-proofing) so they can support and promote the move.

Test and invest in network infrastructure: Organizations should work closely with their network team when making cloud connections, ensure there are redundant pathways for when systems go down, and practice failovers to ensure network failure won’t result in system failure.

Invest in cloud experts: Even with SaaS models, someone on staff needs to understand the cloud environment to be able to effectively communicate with the organization and vendor if it does not perform as expected.

Rigorously evaluate cloud security: The cloud is not more or less secure than on-premises data centers, so the cloud requires the same security precautions as an on-premises solution. It is critical to have a CISO that understands the cloud.

What Will Enable a Strong Provider-Vendor Partnership during a Move to the Cloud?

Provider-vendor trust and flexibility is crucial: The cloud requires a long-term partnership built on trust. Vendors and provider organizations need to work together to become aligned on long-term visions where much of the value of the cloud will be delivered. Flexibility is required in the short term—vendors need to understand that not all provider organizations are ready to fully migrate to the cloud, and providers need to understand that there may be bumps as vendors support cloud and on-premises customers.

Vendors need to guide customers: Vendors need to share best practices and ensure customers are set up for success. Best practices will be needed around infrastructure requirements, resource requirements, and budgeting/planning, especially when moving from a capital expenditures (CapEx) to an operating expenses (OpEx) model.

Improving Provider-Vendor Relationships to Enable Progress

Shrinking budgets across the healthcare industry make it difficult for healthcare organizations to invest in new EI technologies, hire needed staff, and advance their EI strategies. As a result, where organizations used to hire consulting companies, they now rely on EI vendors to provide strategic guidance and develop technology quickly, and this has raised their expectations for their vendor partners. Simultaneously, vendors feel their customers are more reluctant to engage and adopt the technology they develop due to budgetary challenges.

Attendees at the summit collaborated in small tabletop discussions to identify what vendors and provider organizations need to do and understand to make progress amid challenges; below are key insights from those discussions.

Vendors Need to Help Provider Organizations Justify Investments in New Technology

Vendors need to enable their provider customers to procure funds for new investments by clearly communicating ROI/efficiency gains from new technology, presenting a clear long-term road map for how the technology will be developed, and working with providers to forecast costs related to upgrades, optimizations, and new features to allow for better budgeting and planning. It will become even more critical for vendors to own the financial vision for their solutions as they help customers transition to a cloud environment.

Vendors Need to Engage as Strategic Partners, Not Just Sellers 

With change happening quickly, provider organizations need their vendors to more continually engage with them as a partner rather than episodically engage with them as a seller. Both parties need to focus on setting realistic expectations, co-developing strategies, understanding technology ecosystems, and engendering mutual understanding around decisions. Provider organizations would like vendors to visit more on-site and provide dedicated support personnel who deeply understand the organization.

Provider Organizations Need Help Deploying and Adopting Technology 

Staff shortages and a lack of expertise in provider organizations have required vendors to take a more active role in bringing their expertise to the table and ensuring successful deployments and integrations. Vendors need to guide the change management process and provide training required for true adoption. Provider organizations can enable this by selecting physician champions who can drive and promote decisions to adopt new technology.

2024 Summit Attendees

Healthcare Organizations

Vendors/Consultants