Leveraging Machine Learning for Detection and Deep Learning for Prediction

Introduction and Hypothesis

Cancer is a serious public health issue worldwide and the second leading cause of death in the United States. According to the International Agency for Research on Cancer (IARC), about 18.1 million new cases and 9.6 million deaths caused by cancer were reported in 2018. [1] Breast cancer accounts for the largest share of cancer types world-wide and is on the rise.

False Negative Results

In cancer screening, a negative result means no abnormality is present. False-negative results occur when mammograms appear normal even though breast cancer is present. Overall, screening mammograms miss about 20% of breast cancers that are present at the time of screening. [2] While a false positive result may lead to undue stress and worry, the end result is no cancer. False negatives are far more alarming, as the result in this case is a woman who believes she is cancer-free when she is not.

Hypothesis

Great strides have been made in both Machine and Deep Learning in various medical fields regarding the prediction and detection of certain diseases. One of these areas showing promising results is breast cancer; both prediction based on observable measurements and detection regarding whether a tumor is benign or malignant. In this paper I hope to show two examples of models I trained that offer impressive results in this field.