Heavily weighting malignancy

It is hypothesized a neural network that is custom crafted and fine-tuned to weight malignant false negatives heavily negative will improve breast cancer diagnostic results. Furthermore, it is hypothesized that by increasing the number of training samples, the success rate will increase. With modern technology, networks should be able to handle data outliers. Neural networks learn by example, so with more samples to learn from, the neural network should become more accurate.

Ultimately, the neural network implementation needs to be efficient enough to be delivered in a cloud service to enable global collaboration.  This service will facilitate the future acquisition of additional data samples, however, the impact of training sample sizes can be proven without receiving more data.