Case Study: Johns Hopkins University BIOS accelerates brain scan insights and improves accuracy with Google Cloud Platform

Johns Hopkins University BIOS Division Advancing intracerebral hemorrhage treatments through AI

Johns Hopkins University BIOS Division needed a faster, more accurate way to analyze brain injury CT scans and improve decision-making for stroke and hemorrhage patients. Using Google Cloud Platform, along with partner support from Quantiphi, the division moved its clinical research and medical image analysis to the cloud and applied machine learning to large-scale imaging data.

Google Cloud Platform enabled BIOS to use services like Compute Engine, Cloud Dataflow, Healthcare API, AI Platform, and Google Kubernetes Engine to automate and accelerate its imaging workflows. The results were dramatic: scan review time dropped from five hours to 30 seconds, analysis of about 500 patients fell from 2,500 hours to 90 minutes, accuracy reached a dice coefficient of 0.93, and research and infrastructure costs were reduced, with some estimates indicating costs could be cut by 50%.

Johns Hopkins University BIOS

Daniel F. Hanley

Director