AWS HealthImaging

At the absolute core of the AWS medical imaging strategy is AWS HealthImaging (AHI). AHI is a purpose-built, HIPAA-eligible cloud service specifically engineered to ingest, store, analyze, and share medical imaging data at a petabyte scale.

By treating medical images as a cloud-native construct rather than a flat file on a hard drive, AHI fundamentally alters the traditional DICOM storage methodology through a proprietary data structure and highly advanced compression algorithms.

View full size

Source: AWS HealthImaging developer guide overviewLast verified: 2026-03-11

This official overview is most useful as a service-boundary map. The custom tables and diagrams that follow are intentionally more detailed: they unpack how the vendor-level picture translates into image-set organization, metadata extraction, progressive frame retrieval, and operational controls.

AHI Architecture Components

A well-architected AHI deployment consists of multiple AWS services working together to provide secure, scalable medical imaging storage and retrieval.

ImageSet Architecture Flowchart

The following flowchart illustrates how AWS HealthImaging decouples metadata from pixel data, enabling independent access patterns:

Traditional on-premises PACS store DICOM files exactly as received from the edge modality—as massive, contiguous binary blobs containing both pixels and metadata (DICOM Part 10 files). AHI dismantles this rigid paradigm.

During ingestion, AHI decodes the traditional DICOM P10 file and mathematically decouples the heavy binary pixel data from the structured textual metadata. The service automatically analyzes and normalizes the metadata at the Patient, Study, and Series levels.

ImageSet Hierarchy

AHI organizes imaging data into a hierarchical structure that mirrors the DICOM information model while enabling efficient cloud-native access patterns.

A well-architected cloud pattern utilizes an Amazon API Gateway to expose a secure endpoint, directing uploaded DICOM files into a designated Amazon S3 staging bucket.

The arrival of a new object triggers an event notification via Amazon SQS. This queue event invokes an AWS Lambda function or initiates an AWS Step Functions state machine, performing rigorous virus scanning and structural DICOM validation.

During ingestion, AHI converts legacy pixel data into High-Throughput JPEG 2000 (HTJ2K), a state-of-the-art, mathematically lossless compression codec. This reduces overall enterprise storage costs by up to 40% while maintaining absolute diagnostic fidelity.

HTJ2K Progressive Loading Flow

HTJ2K encoding enables a revolutionary progressive rendering workflow that fundamentally transforms the radiologist experience over wide-area networks.

HTJ2K Progressive Loading Sequence

The following sequence diagram illustrates the TLM request and tile streaming workflow that enables sub-second initial image display:

Radiologists demand instantaneous, zero-latency interactions when panning, zooming, and manipulating images. Traditional DIMSE protocols (like C-MOVE) suffer from severe latency degradation over WAN.

AHI resolves this by offering native DICOMweb APIs: WADO-RS for image retrieval and QIDO-RS for metadata searching. By utilizing HTJ2K encoding, AHI supports progressive resolution data access.

WADO-RS and QIDO-RS Sequence

The typical interaction sequence leverages QIDO-RS for instant metadata queries and WADO-RS for retrieving pixel data over standard HTTP:

Cloud architecture offers inherent cost optimization through intelligent storage tiering. In traditional on-premises environments, migrating older studies from high-speed SAN to cheaper archive tapes is manual and labor-intensive.

AHI automates clinical lifecycle management natively. Freshly imported, active ImageSets automatically reside in a high-performance Frequent Access Tier. After 30 consecutive days with no access requests, AHI automatically transitions ImageSets to a highly cost-effective Archive Instant Access Tier.

Crucially, this transition yields significant financial savings but does not incur any retrieval latency penalty; when an older prior study is requested, it is delivered with the same sub-second speed as an active study.

Intelligent Tiering Timeline

AHI Intelligent Tiering automates storage cost optimization without any operational overhead or retrieval latency penalties.

AWS HealthImaging provides a comprehensive set of APIs for secure, authenticated access to medical imaging data. Understanding the authentication flow and API endpoints is critical for integrating AHI with clinical viewers like OHIF.

Access Token Workflow

The access token workflow leverages Amazon Cognito as an OpenID Connect (OIDC) identity provider, integrating with existing hospital Active Directory via SAML federation.

DICOMweb API Endpoints

AHI implements standard DICOMweb RESTful APIs, enabling interoperability with any DICOMweb-compliant viewer including OHIF, Weasis, and commercial PACS workstations.

The StartDICOMImport API orchestrates the conversion of traditional DICOM P10 files into optimized ImageSets. Proper S3 bucket configuration and IAM role permissions are critical for successful ingestion.

StartDICOMImport Workflow

S3 Bucket Requirements

AWS HealthImaging imposes specific requirements on S3 buckets used for DICOM imports to ensure security, compliance, and operational reliability.

The following table summarizes the key differences between traditional on-premises monolithic storage and AWS HealthImaging cloud-native architecture.

Australian healthcare organizations face unique challenges: extreme geographic dispersion, stringent data sovereignty laws, and a critical radiologist workforce maldistribution concentrated in metropolitan areas.

IRAP PROTECTED Certification

The Information Security Registered Assessors Program (IRAP) is the Australian Cyber Security Centre (ACSC) framework for evaluating cloud security against the Information Security Manual (ISM). AWS HealthImaging, along with foundational services (S3, Lambda, EC2, KMS), has been independently assessed and certified at the IRAP PROTECTED classification level.

Network Infrastructure for Teleradiology

Effective teleradiology requires both high upload throughput (for regional clinics to push DICOM studies) and ultra-low latency (for metropolitan radiologists to interact with cloud viewers). The recent deployment of Low Earth Orbit (LEO) satellites has transformed regional Australian connectivity.

Disaster Recovery Strategies

Medical imaging is a tier-one mission-critical system. AWS enables Australian healthcare organizations to implement robust cross-region disaster recovery without data leaving Australian borders.

AWS Elastic Disaster Recovery (DRS) provides a bridging solution for organizations with legacy on-premises PACS, continuously replicating block-level data to a staging area in the recovery region for rapid failover.

For further reading on AWS HealthImaging and medical imaging viewers: