•  
  •  
 

Abstract

Ovarian cancer is considered the fifth most common cancer type among females in the United States. Furthermore, ovarian cancer accounts for 25% of all gynecologic cancers, and usually, this cancer is diagnosed at a late stage. A patient can live at least five years longer if ovarian cancer is diagnosed early. Therefore, the early diagnosis of ovarian cancer is essential. This study aims to classify ovarian cancer using biomarkers such as ovarian cancer antigen (CA125), tumor necrosis factor alpha (TNF-α), interleukin 6 (IL-6), human epididymis protein 4 (HE4), and anti-TP53 antibodies. Rising or persistent CA125 blood levels provide a highly specific biomarker for epithelial ovarian cancer, but not an optimally sensitive biomarker. Addition of HE4, CA 72.4, anti-TP53 autoantibodies and other biomarkers can increase sensitivity for detecting early stage or recurrent disease. It also uses three data-classifying models called Decision Tree (DT), kth Nearest Neighbor (kNN), and Logistic Regression (LR) to compare their performances. We computed various model performances, such as accuracies, precision, and recall values. Based on the findings, the LR model shows the highest performance compared to the other two models. Furthermore, it records 87% accuracy and 99% recall in classifying ovarian cancer.

Download

Share

COinS