CADC is a neural network trained on more than a thousand chest CT scans. Point it at a nodule and it estimates how suspicious that nodule looks, and shows you where it looked.
Learn how lung cancer develops, see how the AI was built and tested, or try it on a CT scan.
An interactive 3D model shows the lungs, airways and how a tumour grows and spreads from stage I to IV.
Upload a chest CT, click on a nodule, and get a malignancy estimate from five models with a heatmap of their attention.
ROC curves, learning curves and an honest look at accuracy, overfitting and limitations.
A chest CT is a stack of hundreds of X-ray slices that together form a 3D image of the lungs.
A 48 mm cube around the nodule you select is cut out and resampled to 1 mm voxels.
Five 3D convolutional neural networks each study the nodule's shape, edges and density.
Their votes are averaged, and a heatmap shows which regions drove the decision.
This tool can't answer that, but a doctor can. See which symptoms mean you should get checked.