Ovarian cancer identification using machine learning techniques
Principal Investigator
Name
Julius Ting
Degrees
BSc (Hons) in Software Engineering
Institution
Asia Pacific University
Position Title
Student
Email
About this CDAS Project
Study
PLCO
(Learn more about this study)
Project ID
PLCO-452
Initial CDAS Request Approval
Feb 11, 2019
Title
Ovarian cancer identification using machine learning techniques
Summary
Project Summary
Ovaries are reproductive organs in females that produces ovums, and reproductive hormones. Ovarian cancer occurs when abnormal cells in the ovary rapidly multiplies, forming a tumour. Statistics has shown that early detection allows the patient more choices of treatment and a higher chance of survival, up to 93. However, only 20% of patients are diagnosed in the early stages.
Early detection of ovarian cancer is insufficient and requires more exploration into new methods of diagnosis. In 2011, The National Cancer Registry of Malaysia found out of 714 cases of ovarian cancer, 56% were diagnosed with late stage ovarian cancer. A similar standpoint was reported by the American Cancer Society in 2018, they found that 80% of ovarian cancer patients were diagnosed with late stage cancer.
With the application of machine learning and its ability to identify patterns, early stage ovarian cancer can be detected with a better accuracy than doctors. Besides that, early diagnosis would decrease the mortality caused by ovarian cancer.
Tangible Benefits
Reduce cost of diagnosis for women with healthy ovaries.
Reduce time for detection as the analytical model can provide fast screening of ovaries.
Reduce workload of pathologist as the machine learning model acts as a pre-screening tool.
Intangible Benefits
Screening for ovarian cancer will lessen the number of ovarian cancer biopsies cases, reducing the workload for pathologists.
Provide more accessible ovarian screening for women.
Decrease the chances of missed diagnosis.
Raise awareness for ovarian cancer by providing a cheap alternative.
The findings of the research is expected to help software engineers create an enhanced machine learning model.
This research is expected to show evidence of the capabilities of machine learning in the field of healthcare and convince hospitals to work with the latest technology as it is more accurate, faster diagnosis, and cheaper.
The researcher will produce a functioning machine learning model capable of distinguishing diseased ovaries and healthy ovaries using various machine learning techniques such as convolutional neural network.
Ovaries are reproductive organs in females that produces ovums, and reproductive hormones. Ovarian cancer occurs when abnormal cells in the ovary rapidly multiplies, forming a tumour. Statistics has shown that early detection allows the patient more choices of treatment and a higher chance of survival, up to 93. However, only 20% of patients are diagnosed in the early stages.
Early detection of ovarian cancer is insufficient and requires more exploration into new methods of diagnosis. In 2011, The National Cancer Registry of Malaysia found out of 714 cases of ovarian cancer, 56% were diagnosed with late stage ovarian cancer. A similar standpoint was reported by the American Cancer Society in 2018, they found that 80% of ovarian cancer patients were diagnosed with late stage cancer.
With the application of machine learning and its ability to identify patterns, early stage ovarian cancer can be detected with a better accuracy than doctors. Besides that, early diagnosis would decrease the mortality caused by ovarian cancer.
Tangible Benefits
Reduce cost of diagnosis for women with healthy ovaries.
Reduce time for detection as the analytical model can provide fast screening of ovaries.
Reduce workload of pathologist as the machine learning model acts as a pre-screening tool.
Intangible Benefits
Screening for ovarian cancer will lessen the number of ovarian cancer biopsies cases, reducing the workload for pathologists.
Provide more accessible ovarian screening for women.
Decrease the chances of missed diagnosis.
Raise awareness for ovarian cancer by providing a cheap alternative.
The findings of the research is expected to help software engineers create an enhanced machine learning model.
This research is expected to show evidence of the capabilities of machine learning in the field of healthcare and convince hospitals to work with the latest technology as it is more accurate, faster diagnosis, and cheaper.
The researcher will produce a functioning machine learning model capable of distinguishing diseased ovaries and healthy ovaries using various machine learning techniques such as convolutional neural network.
Aims
To develop a machine learning generated analytical model to identify ovarian cancer with high accuracy by classifying pathology images to decrease the chances of missed diagnosis which often lead to difficulties in effective treatment in the future.
To reduce the cost of diagnosis
To fill the gap of insufficient pathologist or specialised doctors
To provide more accessible ovarian screening for women
To tune the machine learning algorithm to get the optimum prediction
To train the model to pick up patterns of all stages of ovarian cancer
To produce a research paper on machine learning and ovarian cancer
Collaborators
Julius Ting