Systematic literature review in cervical cancer staging using neural networks

Abstract

Abstract

Cervical cancer is the number two killer for women and the pap smear method is commonly used in the detection of the disease at an early stage. Cervical cancer is one of the biggest killers for many women globally. Several methods have been used to detect cancer during its early stages so that death figures can be reduced.  The detection of cancer in women has been a challenge and using the Pap smear test because of its slow and accuracy levels.   Scientifical neural networks are strong and reliable tools commonly used for building cancer prediction models for microarray data. Yearly, more than half a million women have been diagnosed with cervical cancer with 300 000 deaths worldwide. The deadly risk of more than a half million women with cervical cancer is the cause of the disease in most cases. Detection of cancer in developing and developed nations must invest a lot of funds in the detection of cancer to reduce mortalities worldwide. Cancer can be cured as long as it has been screened and detected at an early stage. This assists in giving ample time in the treatment of cancer.  The exact location of cancerous cells in thousands of cervical squamous epithelial cells can reduce doctors’ workloads of doctors and improve the accuracy of a cervical cancer diagnosis. Treatment of cancer depends on the disease extent at diagnosis and the availability of resources and might involve chemoradiation or hysterectomy. Conservatively, fertility-preserving surgical procedures are now standard for ladies with low risk at an early stage. Radiotherapy advancements like intensity-modulated radiotherapy have led to less treatment toxicity for ladies with locally advanced diseases. The study aims to review a systematic review of cervical cancer staging using neural networks. The study objectives are to analyse the effects of neural networks in the detection of cervical cancer, evaluation of the limitations in neural networking, assessment of other literature on neural networks and cervical cancer and the implementation of an outstanding method. The manual systematic review search was chosen for this study. The final review literature was 32 after the inclusion and exclusion criteria. The limitations of this study were the use of the manual search technique hence it is recommended that future studies adopted automatic and manual search methods. A systematic review is a very important aspect in the structuring of research studies as well as the gathering of relevant evidence and research findings.

Keywords:   Adenocarcinoma, Cell Carcinomas, Cervical Cancer Squamous, Deep learning, Neural Network, Transfer Learning.