Cloud-Based Hosting Platform for DICOM Medical Images

Confidentiality Notice: This case study is confidential and intended for internal reference or vetted prospect presentations only.

Objective

The goal of this project was to design and deploy a scalable and optimized cloud platform for hosting medical imaging data as a part of a long-term engagement under which our team is developing a med-tech platform capable of interpreting medical scan images. The platform was developed to support the full lifecycle of medical images generated by scanner devices and processed through a core pipeline composed of multiple microservices and AI modules.

Solution Description

The incoming data consisted of high-resolution medical scans, which were processed through an AI-powered pipeline. This pipeline performed various transformations and analyses, ultimately generating DICOM images as the final output.

To meet the performance and reliability needs of clinical use, a robust hosting infrastructure was required. The developed platform enabled secure and scalable hosting of DICOM files, ensuring fast access and integration with healthcare systems.

The hosted images are accessed directly by medical professionals across healthcare institutions, supporting accurate diagnosis and analysis, particularly for detecting abnormalities in different types of tissues.

Key Features

Cloud-native architecture with autoscaling and high availability
Support for DICOM standards
Optimized storage and retrieval workflows for clinical use
Designed for seamless access by healthcare professionals

Frontend Hosting & Deployment

For the frontend component of the application, we used the SLIM (Simple Lightweight Imaging Model) interface.
ref: SLIM on GitHub

SLIM is a lightweight, web-based viewer designed to display DICOM images organized by study, series, and instance. It integrates seamlessly with DICOMweb APIs and is ideal for displaying medical imaging data processed through Google Cloud Healthcare services.

ref: Google Cloud Healthcare API

The SLIM viewer was customized to match the specific structure of DICOM images produced by the PathScience pipeline, enabling efficient access for healthcare professionals.

Architecture

To ensure secure, scalable, and reliable deployment of the PathScience imaging application, the following components and services were used:

• Google Cloud Healthcare API

The Google Healthcare API served as the core infrastructure for storing and managing DICOM images generated by the AI-based image processing pipeline. It provides a fully managed, secure, and HIPAA-compliant environment, with native support for DICOMweb protocols, making it ideal for medical imaging workflows. This allowed the SLIM-based frontend to retrieve and display diagnostic images directly from the cloud, with minimal latency and full compatibility with clinical tools.

• Google Firebase Hosting

The SLIM frontend viewer was hosted on Firebase Hosting, chosen for its fast global CDN support and easy integration. This ensured that healthcare professionals could securely access medical images from anywhere, with high availability and performance.

• GitHub + Google Cloud Build

To streamline development and deployment, the project was stored on GitHub and integrated with Google Cloud Build. Every code change pushed to the repository triggered an automated build and deployment pipeline, ensuring that the latest version of the frontend was always live and reducing manual deployment errors.

• Google Secret Manager

All sensitive configuration data, such as API keys, authentication credentials, and GitHub tokens, was stored securely using Google Secret Manager. This centralized approach to secret management reduced the risk of data exposure and improved compliance with security best practices.

• Google Cloud Run Jobs for Backend Processing

In addition to the frontend deployment, Cloud Run Jobs were used to automate backend data ingestion and processing tasks. A containerized DICOMWeb client was built and deployed using Google Cloud Build and Container Registry, then executed on demand via Cloud Run Jobs.

These jobs were responsible for retrieving medical image studies from the Google Healthcare API and preparing them for use in the frontend viewer. The use of Cloud Run Jobs allowed for scalable, event-driven, and cost-efficient background processing without the need to manage a persistent server environment.

Summary of Google Cloud Services Used

Google Cloud Healthcare API – Secure storage and management of DICOM images
Google Firebase Hosting – Scalable and global hosting for the SLIM frontend viewer
Google Cloud Build – CI/CD automation for deployment from GitHub
Google Secret Manager – Centralized and secure configuration management
Google Cloud Run (Jobs) – Serverless execution of backend DICOM ingestion and processing tasks
Google Container Registry – Hosting and versioning of a containerized DICOMWeb client
Google IAM – Role-based access control for all services and service accounts