Radiology sits at the center of modern medicine and often serves as the bridge between clinical symptoms and treatment decisions. Globally, an estimated 3.6 billion diagnostic imaging exams are performed yearly. In the U.S., annual volume exceeds 250 million procedures, with spending topping $100 billion, according to the National Institute for Health Care Management Foundation.
Physician burnout is real. In April 2026, the American Medical Association reported a 45% burnout rate in radiology, the fifth-worst among specialties, behind emergency medicine, urological surgery, hematology/oncology and OB/GYN.
The radiology landscape is shifting
All this is happening amidst a rapidly changing industry landscape. In Top 2026 Radiology Trends, The Imaging Wire’s Brian Casey shares perspectives from thought who suggest these trends will shape the industry in the coming year: a.) AI-based workflow becoming more widespread and harder to differentiate, b.) technology and services increasingly bundled into care pathway solutions, and c.) intensifying competition, with partnering and M&A accelerating as healthcare spending tightens.
Healthcare leaders have long viewed radiology performance through the familiar lens of volume growth as well as turnaround time and staffing levels. While each of these challenges is real, treating them as separate issues risks missing the bigger point: radiology interpretation itself is approaching a structural breaking point.
As digital radiology has evolved, the technology environments supporting radiologists have become fragmented collections of systems assembled over decades. To speed radiology interpretation, many organizations opt to incrementally layer on new tools including fast-changing AI offerings. Instead of approaching the challenge with a bolt-on mindset, a better question to ask is whether the underlying architecture still fits the realities of modern healthcare.
The growing gap between imaging demand and capacity
Radiology has become a victim of its own success. As imaging becomes increasingly central to diagnosis, treatment planning and population health, utilization is expanding dramatically in part due to chronic illness burdens and the greying of America.
Workforce growth hasn’t kept pace. Research from the Harvey L. Neiman Health Policy Institute forecasts continued pressure through at least 2055, with imaging demand expected to rise 17 to 25% depending on modality, while workforce growth remains constrained by residency capacity, attrition and retirement. Radiologist attrition more than doubled between 2014 and 2022, from 1.1% to 2.5% annually.
Organizations have responded with familiar strategies: improving recruitment and retention, outsourcing reads, expanding teleradiology, upgrading speech recognition and deploying targeted AI. These can help but mostly treat symptoms rather than root causes.
The hidden cost of workflow fragmentation
Healthcare organizations frequently talk about system interoperability, but less often about the operational costs associated with the sheer number of systems clinicians must navigate. Radiology may be an extreme example: a radiologist interpreting a complex study moves between PACS, RIS, EHRs, prior reports, lab systems, AI applications, communication platforms, reporting software, and quality tools. Information exists, but it’s scattered.
The challenge isn’t simply finding data. It’s assembling and presenting the right context to the radiologist at the very moment interpretation begins. Every time a radiologist is required to leave the primary reading environment to search for history, verify findings or complete downstream tasks, cognitive load increases. Small inefficiencies accumulate thousands of times per day across an enterprise, driving higher integration costs, heavier support and training burdens, greater security and governance challenges and reduced technology ROI.
Too often, organizations optimize for individual application updates or additions while leaving the broader workflow untouched. In this “advanced environment,” radiologists must still manually coordinate their own work.
Why AI alone won’t solve the problem
AI is generating excitement in radiology. In research published in the November 2025 International Journal of Computer Assisted Radiology and Surgery paper, Michael Friebe evaluated four workflows and concluded that AI has produced measurable gains in accuracy, efficiency and standardization, but key limitations remain. These include algorithm generalizability, clinical interpretability, organizational readiness and regulatory uncertainty. “AI will augment rather than replace human expertise, with collaborative human-AI workflows being essential. Future integration efforts must address interoperability, workforce adaptation and ethical considerations to ensure safe, equitable, and clinically impactful deployment,” he stated.
Radiology’s dominant use of AI has centered on image analysis, detection algorithms, and report generation. These create real value, but AI applied to a fragmented workflow delivers limited gains. A more accurate algorithm doesn’t surface clinical context within the reading environment, a faster dictation engine doesn’t reduce the number of systems a radiologist must navigate, and a smarter report doesn’t eliminate fragmentation.
This is not an argument against AI. In fact, AI will likely be one of the most important enablers of next-generation radiology operations. Its greatest impact will not come from improving individual tasks, but instead from orchestrating the entire interpretation workflow: from assembling clinical context to prioritizing work intelligently to routing studies dynamically to initiating downstream actions to supporting quality processes to reducing the manual decisions required throughout the interpretation journey.
From reporting systems to interpretation systems
The technology infrastructure of most radiology organizations evolved as a collection of specialized applications: PACS managed images, RIS managed operations, reporting systems generated reports, quality systems handled peer review, communication platforms managed critical results. Each were developed independently. This model made sense when volumes were lower and workflows simpler.
Today, interpretation is a continuous process spanning multiple systems, data sources and decisions. The radiologist’s final report is the output of a much larger clinical reasoning workflow.
The industry’s spotlight on reporting has intensified in recent months since Microsoft’s early 2026 announcement that it plans to sunset PowerScribe 360, a solution that’s held more than 70% market share across radiology practices for the past decade. Many organizations are now deciding whether to stay with Microsoft or move on. Instead of asking “What reporting tool should we implement?” the wiser question is: “What interpretation environment should we build?” This distinction matters because one optimizes a step and the other optimizes the entire process.
If radiology is approaching a structural breaking point, the path forward requires much more than whack-a-mole point solution replacement. It requires a more critical look at redesigning the entire workflow architecture with a goal of holistically advancing operational, clinical and strategic improvements.
Photo: athima tongloom, Getty Images
Mike Moore is co-founder and COO of NewVue, a health IT company building cloud-native software that unifies how radiologists work. He was previously co-founder of PeerVue, which created the radiology workflow orchestration category before being acquired by McKesson. A commercial architect, market builder and strategic adapter, he’s held leadership roles in organizations known for pioneering the first B2B internet services (BBN), expanding hosted services via global joint ventures (British Telecom) and creating the broadband roaming capability (iPass.)
This post appears through the MedCity Influencers program. Anyone can publish their perspective on business and innovation in healthcare on MedCity News through MedCity Influencers. Click here to find out how.
