In order for machine learning algorithms to work robustly and produce accurate results, large amounts of diverse data are necessary. The amount of data available for radiological applications at any single center is insufficient and thus pooling of data across centers is critical for robust AI. Traditional approaches for sharing imaging data involves each hospital transferring its curated and harmonized data sets to a central system who then releases the trained model back to the participating hospitals. However this approach is burdened by regulatory and privacy concerns, even with...

We evaluated deep learning algorithms’ segmentation of acute ischemic lesions on heterogeneous multi-center mutli-vendor clinical diffusion-weighted magnetic resonance imaging (MRI) data sets and explored the potential role of this tool for phenotyping acute ischemic stroke. The ensemble consisting of a mixture of data from 12 international stroke genetics research center  and single-center convolutional neural networks performed best, demonstrating the need for diverse training data sets for creating robust machine learning algorithms. 

You can read our...

In the MRWITNESS study, led by investigators from the CCNI laboratory, we showed that patients with unwitnessed acute ischemic strokes may be safely treated with a "clot busting" drug, alteplase. Under current guidelines, treatment with intraveous alteplase is approved up to 4.5 hours from when the patient was last known to be well. Approximately 25% of ischemic stroke patients have stroke onsets that are unwitnessed. Our trial used an MR witness to identify patients who might be able to be treated safely with alteplase...