54 — CHIPS – A Service for Collecting, Organizing, Processing, and Sharing Medical Image Data in the Cloud

CHIPS (Cloud Healthcare Image Processing Service) is a cloud-based service that authenticates and authorizes browser-based interaction with medical imagery. It enables 3D visualization and realtime collaboration, which is becoming critical in a clinical environment.

CHIPS both manages data collection from typical EHR-linked sources, and supports data-processing through a library of containerized compute resources. The frontend uses XTK to visualize medical imagery.

Based on user preference, CHIPS dispatches compute jobs to a family of compute resources, and enables the tracking of data and manipuations via the web visualization platform.

This work is interesting because this sort of containerized, distributed cloud-compute is difficult to pull off — more-so in a clinical environment. Running these sorts of analysis pipelines will become a fundamental part of clinical work in the future (I suspect), so it’s important to lay that groundwork now.

I’m interested in how these containers’ interfaces will generalize: How will all clinical-resource developers keep a consistent IO model, so that any framework can run any container in any pipeline or workflow. If that interface cannot be standardized, I think we’ll have some challenges adopting containers in the clinic.