Health systems beware: Digital imaging data is the sleeping giant

Digital health imaging has become the most urgent and underestimated issue in health IT because it is growing at a pace that outstrips current storage, interoperability and cost‑management strategies, said Prashant Akolia, senior director, head of enterprise architecture at Memorial Sloan Kettering Cancer Center in New York City.

"As imaging devices become more advanced – higher‑resolution CT, MRI, PET, digital pathology and multimodal imaging – the volume of data multiplies exponentially," he explained. "This creates a silent but massive pressure on storage infrastructure, data transport, and the ability to use imaging data effectively in both clinical and research settings."

Why imaging is the big challenge

Akolia outlined five reasons digital imaging is health IT's sleeping giant:

• Data volume is exploding as modalities shift from megabytes to gigabytes per study, and digital pathology introduces terabyte‑scale datasets.
• Storage costs rise continuously, especially for organizations that must retain images for long periods due to clinical, legal or research requirements.
• Transporting imaging data between clinical systems, research environments and cloud platforms strains networks and slows workflows.
• AI models require high‑quality imaging data, which further increases the need for scalable, accessible and standardized storage.
• Vendor‑specific formats and PACS silos make it difficult to share or repurpose imaging data across departments or institutions.

"Imaging is foundational to diagnosis and treatment, yet the infrastructure supporting it is often decades behind the clinical demand," he noted.

A common scenario involves managing the flow of imaging data between clinical PACS systems and research environments. For example, when high‑resolution MRI or CT datasets are exported for research, the file sizes can overwhelm existing storage tiers or slow down network transfer, Akolia explained. Teams must manually compress, de-identify or segment data to make it usable, which adds operational burden and delays, he said.

"Another frequent challenge is balancing clinical storage requirements with research needs," he added. "Clinical systems prioritize fast retrieval and reliability, while research teams need flexible access to large datasets for training AI models. Coordinating these needs often requires dedicated pipelines, additional storage tiers and careful governance to avoid duplication and unnecessary cost."

What health IT leaders should do now

So, what should health system leaders be doing now? For one thing, they should build a long-term enterprise imaging strategy, Akolia advised.

"Executives must recognize imaging as a strategic asset and develop a comprehensive plan that accounts for multi-petabyte growth," he said. "This strategy should encompass robust lifecycle management, intelligent tiered-storage solutions and future-proof cloud-ready architectures.

"Also, leaders should be investing in modern, scalable storage," he continued. "Prioritize investments in hybrid cloud systems, object storage and intelligent archiving technologies. These approaches can significantly reduce costs while maintaining high accessibility. Leaders should evaluate storage systems not solely on raw capacity, but on scalability, retrieval speed and total cost of ownership."

Akolia further advised that leaders should:

• Standardize imaging data for interoperability. Adopting industry standards such as DICOMweb, FHIR Imaging Study and Vendor Neutral Archives is crucial.
• Strengthen governance for clinical and research imaging. Establish clear and robust policies for data retention, de-identification, access control and data movement.
• Prepare imaging infrastructure for AI. Given AI models' dependence on large, high-quality imaging datasets, leaders must ensure their imaging strategy actively supports essential AI enablement tools, he concluded.