Artificial intelligence and data sharing have been two of the biggest topics this year – and for good reason, too. These two tools have a significant impact on the world of healthcare, such as improving patient care, easing provider burnout, and much more! But these topics are so vast that the only way to properly discuss their ins and outs is to do so category by category. Today, we are going to focus our attention on their impact on radiology and imaging technology. We reached out to our brilliant Healthcare IT Today Community with a variety of questions on this subject. The following is what they had to share.
Improving the Storage, Retrieval, and Sharing of Medical Imaging Data to Enhance Provider Collaboration
Josh Russell, Chief Medical Officer at UCP Merchant Medicine
Radiology solutions that are powered by AI have the ability to greatly enhance provider collaboration. The use of AI produces higher-quality images, which reduces the need for room to store multiple (low-quality) scans. AI has the ability to prioritize cases it identifies as urgent by tagging them for expedited review. Providers also benefit from the automated – and standardized – reports that AI can generate. More streamlined reports free up time radiologists typically spend on interpretation.
Overall, AI automates many of the manual elements associated with storing, retrieving, and sharing medical imaging data, which increases speed, improves workflow, and ultimately allows clinicians to work at the top of their licenses.
Optical Character Recognition functionality that uses AI to screen for abnormalities in radiographic images (ie, X-rays) is becoming increasingly powerful. The value of this tool for urgent care is that it can, with high accuracy, screen for abnormalities and, with confidence, say when there are none. Urgent care is typically staffed with APCs with limited clinical experience and discomfort with X-ray interpretation is nearly universal. Such a feature would allow for more rapid disposition of patients, faster door-to-door times, and better patient experience compared to the status quo of waiting for interpretation from a human radiologist.
As the vast majority of X-rays obtained in urgent care are normal, most patients could be discharged within minutes of X-ray acquisition. For patients with likely abnormalities, which is a small minority, these individuals would be triaged to a human radiologist who is less burdened by urgent cases because the AI has siphoned off the normal ones that can be reviewed by a human in a less pressing fashion.
Sandra Johnson, Senior Vice President, Client Services at CliniComp
Radiology is at the center of the healthcare data ecosystem, yet too often imaging systems operate in silos. By unifying imaging, clinical, and operational data within a single, integrated EHR platform, providers gain instant access to the insights they need without disruption. A fully integrated solution suite, including PACS with native AI, enables seamless data sharing, accelerates image interpretation, optimizes workflows, and delivers proactive decision support to enhance collaboration and improve patient outcomes.
Morris Panner, President at Intelerad
Healthcare’s digital transformation has made enterprise imaging fundamental to care coordination. Picture archiving and communication systems (PACS), when paired with a vendor-neutral archive (VNA) and image exchange, offer far more than image storage. Together, they function as an integrated platform orchestrating workflows across sites and systems, moving images securely and quickly wherever and whenever they’re needed. The solutions maintain seamless data and reporting channels between systems and care teams to enhance interoperability and information access. When collaboration is effortless, care is administered faster, exam duplication is reduced, and patient outcomes improve. Clinicians can also work and coordinate confidently knowing PACS and VNA protect patient data with built-in access controls and consent management, ensuring privacy while delivering the right image to the right provider at the right time.
AI is a powerful tool that can automatically triage urgent cases, flag potential abnormalities, and even generate preliminary reports. The technology won’t replace radiologists, but it is easing their cognitive load, speeding turnaround times, and giving clinicians more time to focus on complex diagnostic analyses and consultations. Beyond workflow efficiency, AI has already begun to improve patient outcomes and holds enormous promise to set new care standards.
Thanks to its usefulness across a wide range of important tasks, from detecting cancer to classifying brain tumors, AI is quickly becoming an indispensable tool for both diagnosis and prognosis. Still, as with any new technology, success depends on balance. AI is helping radiologists automate procedures, but it’s an assistive tool that requires human supervision, not an autonomous practitioner. Radiologists provide the clinical context, judgment, and accountability that AI fundamentally lacks.
Beyond AI, radiologists are embracing more efficient systems and protocols to streamline the entire imaging process. Cloud-based solutions, in particular, are transforming accessibility and interoperability by centralizing radiologists’ tools and data, which enables them to work effectively and collaboratively from virtually anywhere. Cloud adoption in enterprise imaging also allows information to be shared instantly and confidently with external providers, partners, and patients, creating a truly connected care experience.
Cloud PACS further enhances operations and supports teleradiology by ensuring fast, secure availability of images and reports from any location. It also eliminates the need for expensive, on-site servers, which reduces IT complexity and cost. Most importantly, cloud infrastructure’s computing power is essential for integrating the AI algorithms that empower radiologists to improve diagnostic accuracy, efficiency, and overall clinical performance.
Dr. Scott Schell, Chief Medical Officer at Cognizant
Modern radiology solutions improve collaboration by adopting DICOMweb standards such as QIDO-RS (query), WADO-RS (retrieve), and STOW-RS (store), which use standard HTTPS protocols for secure, real-time image exchange. These enable zero-footprint viewing and teleradiology without custom connectors. When paired with IHE imaging-exchange profiles (XDS-I.b and MHD-I), providers can discover and retrieve studies across organizations in alignment with TEFCA’s push for interoperable data exchange.
Luke Barré, MD, MPH, FACP, Contractor Medical Director at Noridian Healthcare Solutions
In recent years, the rapid shift toward remote radiology work and decentralized imaging interpretation has transformed radiology operations. It is more flexible and efficient, but it has also unintentionally weakened the spontaneous, real-time dialogue that used to happen in the reading room. What once began with, “Here’s why I’m worried about this patient,” now arrives as a requisition stripped down to checkboxes. When an EHR order replaces a narrative clinical question with a template, the radiologist never receives the same richness and nuance of clinical context that a discussion provides: diagnostic uncertainty, suspected differentials, and the clinical “why” behind the study.
Next-generation imaging platforms should deliberately rebuild that context into the workflow. Within PACS and the EHR, lightweight, integrated chat and structured-narrative fields can capture the ordering clinician’s reasoning, including what they’re ruling out, what’s changed since the last scan, or which labs are concerning. The goal is to restore the highly valuable context signal that guides interpretation. Equally important, storage and retrieval must evolve beyond “archive the pixels.”
Imaging systems should index and surface contextual metadata alongside the study itself—recent notes, key lab values, relevant priors—so the radiologist can launch a read with the patient story at hand, not buried across systems. When images, priors, and clinical context are retrieved together, we close the collaboration gap created by decentralization and return to a more informed, timely, and clinically aligned read, digitally recreating the conversations that improve care.
AI’s Role in Radiology Tasks and Gaps that Need to be Filled
Luke Barré, MD, MPH, FACP, Contractor Medical Director at Noridian Healthcare Solutions
AI already helps with triage, anomaly detection, and quantification of findings, but its highest value is sharpening—not replacing—radiologist judgment. To realize that value, tools must be co-designed with practicing radiologists and specialty societies, validated on representative data sets, and address the needs of real-world workflows. Gaps remain in algorithm transparency, dataset diversity, and rigorous post-market monitoring to ensure accuracy and effectiveness across broad, real-world patient populations.
Too many models learn from narrow cohorts, yielding biased or brittle performance. Priorities for development and application of AI in radiology should include continuous real-world validation, clinician feedback loops, and published performance metrics (e.g., sensitivity/specificity vs. radiologist benchmarks), aligned with the FDA’s AI/ML-Based SaMD Action Plan and ACR Data Science Institute guidance, and informed by recent JAMA accuracy studies.
Josh Russell, Chief Medical Officer at UCP Merchant Medicine
Artificial Intelligence plays a significant, valuable, and growing role in a variety of tasks across the entire radiology process. In particular, AI-assisted X-ray workflow tools provide real-time guidance, ensuring diagnostic-quality images on the first attempt. This enables higher-quality images and reduces the need for retakes, which limits unnecessary radiation from repeat imaging. AI verifies appropriate/diagnostic quality images have been acquired prior to transmission.
AI can use Optical Character Recognition (OCR) for a preliminary interpretation as Normal vs. Abnormal. For the many normal preliminary reads, these can be triaged to be completed less urgently, decreasing the demands for human radiologists tasked with STAT reads to focus on the images that are more likely to show pathology.
AI augments human radiologists in screening for incidental findings needing follow-up (eg, lung nodules). AI doesn’t fatigue like people do, meaning the systematic screening of image functions, which are better done by AI than humans, can be delegated. This can increase radiologists’ satisfaction and confidence and improve patient safety.
Dr. Scott Schell, Chief Medical Officer at Cognizant
AI is now mainstream in radiology for triage, detection, quantification, and report support. The next frontier is not better algorithms, but reliable workflow integration, safety, and monitoring. The largest remaining gaps include:
- Evidence of prospective clinical utility at scale
- Governance and continuous performance monitoring for bias and drift
- Seamless integration into PACS/RIS worklists and structured reporting (DICOM SR)
Together, AI’s future will lie in becoming invisible, embedded in the workflow, and measured by throughput, accuracy, and trust.
Blake Richards, COO at Elucid
AI in radiology should serve as a force multiplier: enhancing, not replacing, the expertise of radiologists. As Dr. Curtis Langlotz famously said, “AI won’t replace radiologists, but radiologists who use AI will replace those who don’t.”
Today, AI is helping automate routine tasks, accelerate image analysis through pre-reads and annotations, and generate quantitative insights, some beyond human perception. These tools can reduce time burdens and improve consistency, enabling radiologists to focus on higher-value clinical interpretation and patient care. However, not every AI solution fits every practice. Organizations must honestly assess their clinical and operational needs to identify tools that truly enhance patient care, boost efficiency, and improve outcomes—rather than adopting technology for its own sake.
Deepak Prakash, Co-Founder and CTO at Sonio
AI in radiology is moving beyond simple pattern recognition. It now supports clinicians by automating repetitive tasks, flagging anomalies, and ensuring guideline adherence. In prenatal imaging, for instance, AI can suggest required views, highlight missed steps, and reduce variability across practitioners. This doesn’t replace the clinician’s expertise—it augments it. However, gaps remain: AI adoption is still fragmented, data quality and bias continue to be challenges, and integration into clinical workflows is often overlooked. What’s missing is not more algorithms, but thoughtful design that fits AI into the day-to-day practice of radiologists and sonographers without adding friction.
Finding and Vetting Radiology AI Solutions
Dr. Scott Schell, Chief Medical Officer at Cognizant
Organizations should begin with a formal governance and intake process:
- Define the clinical problem
- Verify FDA clearance and training dataset provenance
- Assess DICOMweb/PACS compatibility and workflow fit
- Establish post-deployment monitoring metrics
Health systems can identify candidates through neutral sources such as the ACR Data Science Institute’s AI Central or via cloud AI marketplaces. While imaging vendors may bundle algorithms, the provider organization owns validation, fit, and safety monitoring.
Luke Barré, MD, MPH, FACP, Contractor Medical Director at Noridian Healthcare Solutions
Health systems should carefully consider the most appropriate method for validating and selecting AI tools. Imaging vendor add-ons and third-party solutions each offer distinct benefits and disadvantages, but ultimately, any decision should depend on how cleanly it plugs into current PACS/RIS/EHR systems. To get it right, health systems should start with a multidisciplinary panel that includes radiologists, referring clinicians, IT, compliance, and patient representatives to define use cases and guardrails. A structured vetting framework is important, and it should ensure consideration of:
- Evidence of clinical validity and peer-reviewed outcomes
- Regulatory status (e.g., FDA clearance as SaMD)
- Integration and cybersecurity compatibility
- Transparency in algorithm updates and performance monitoring
Vet AI tools with a structured rubric that evaluates each of the above criteria objectively. Favor tools that strengthen clinical communication and decision-making at the point of care, and be wary of efficiency claims, as that alone shouldn’t drive procurement.
Blake Richards, COO at Elucid
Healthcare organizations should evaluate radiology AI with the same rigor applied to any clinical technology—starting with a clearly defined clinical need, demanding transparent and robust validation, and ensuring seamless fit into the clinical workflow. Many AI tools address real gaps in workflow, decision-making, or care delivery, but others exist mainly for the “wow” factor. Focus on solutions that solve meaningful problems.
When vetting technology, prioritize the quality of the ground truth used for training, the rigor and representativeness of validation studies, and the interpretability of the model: whether clinicians can understand and visualize how it works. Even the best AI won’t succeed if it doesn’t integrate easily into daily practice or make economic sense. Adoption depends on both clinical value and practical feasibility.
Deepak Prakash, Co-Founder and CTO at Sonio
Healthcare organizations should approach AI adoption strategically. First, define the clinical and operational problems you want to solve. Then evaluate solutions based on three pillars: clinical validation, regulatory compliance, and integration feasibility. Some AI comes embedded within imaging vendor platforms, but there is also a growing ecosystem of specialized AI companies that bring unique expertise. In many cases, hospitals will need to combine both approaches—leveraging AI from their vendor while sourcing best-of-breed solutions that are vendor-neutral. What matters most is interoperability and ensuring the AI solution can fit into your existing IT and clinical environment without creating silos or additional complexity.
Solutions Being Leveraged to Improve the Efficiency and Effectiveness of Radiologists
Josh Russell, Chief Medical Officer at UCP Merchant Medicine
As the population ages and patients increasingly seek care at sites outside of hospitals – including urgent care facilities – the need for radiology technicians grows. There is a glaring gap between patients who seek – and need – radiology services and the ability to meet the escalating demand. Artificial Intelligence is one of the top solutions to fill that gap. Among its many benefits, it has the potential to reduce the number of practitioners leaving the field due to burnout.
Of the many tools that improve radiologists’ efficiency and effectiveness, AI-assisted interpretation is particularly notable. The technology provides immediate feedback on abnormal findings and enables rapid point-of-care decision making. This is invaluable for busy clinicians and those with limited independent clinical experience. When specialist radiologists aren’t immediately accessible, AI-assisted interpretation multiplies diagnostic capacity. Radiologists also appreciate that using AI lessens time spent charting and alleviates overall administrative burden, allowing for more time for their number one priority: patient care.
Dr. Scott Schell, Chief Medical Officer at Cognizant
Efficiency gains are coming from:
- Enterprise imaging architectures based on DICOMweb and IHE profiles, which remove barriers between facilities and systems
- AI-assisted triage and decision support, which prioritize urgent studies and accelerate report generation
- Operational analytics that optimize turnaround time, case routing, and quality metrics
The unifying principle: standards-based infrastructure paired with monitored, integrated AI that enhances both accuracy and velocity.
Luke Barré, MD, MPH, FACP, Contractor Medical Director at Noridian Healthcare Solutions
Radiologist efficiency is improving through AI-assisted triage, speech-to-text reporting, and automated structured reporting templates that reduce repetitive clerical tasks and give time back to radiologists so they can focus on interpretation. Natural language processing (NLP) and generative models can draft preliminary impressions from dictated findings and surface relevant priors, but their outputs must stay editable and ultimately be signed by the radiologist to preserve accountability and quality of care.
When it comes to AI image-analysis, early experience with first-generation tools is mixed, reinforcing the need for rigorous validation and active clinician oversight in solution development. AI should serve as a second reader or workflow accelerator, not as an unsupervised interpreter, and patients deserve transparency about its use in their care.
Deepak Prakash, Co-Founder and CTO at Sonio
Several strategies are converging. Workflow automation—such as auto-population of structured reports, intelligent hanging protocols, and automated quality checks—reduces time spent on administrative tasks. AI tools help triage urgent cases, suggest relevant prior studies, or improve the completeness of imaging protocols. Cloud platforms further reduce inefficiencies by enabling distributed reading and tele-collaboration. Importantly, the next wave of solutions isn’t just about efficiency; it’s about sustainable practice. Reducing burnout, minimizing repetitive tasks, and allowing radiologists and sonographers to focus on patient care is just as critical as improving throughput.
Such great insights and responses to think about here! Thank you to everyone who took the time out of their day to submit a quote to us! And thank you to all of you for taking the time out of your day to read this article! We could not do this without all of your support.
What are your thoughts on radiology and imaging technology? Let us know over on social media, we’d love to hear from all of you!
No comments:
Post a Comment