Figure 1. Example of an Image-Based Deep Learning Model.

Hidden layers signify model architecture that cannot be extracted from the model. Predictive outputs are given per image, and all patient image predictions were pooled for overall patient prediction.