Developing a medical scoreboard for prediction of breast cancer recurrence / Nurul Husna Jamian

Jamian, Nurul Husna (2013) Developing a medical scoreboard for prediction of breast cancer recurrence / Nurul Husna Jamian. Masters thesis, Universiti Teknologi MARA.

[img] Text
TM_NURUL HUSNA JAMIAN CS 13_5 1.pdf

Download (1MB)

Abstract

Breast cancer is the commonest diagnosed cancer with substantial high proportion of female globally. Incidence of breast cancer in Asia is lower than in the west but the incidence is increasing rapidly. In Malaysia, the most common cancer occurred among women is breast cancer. Breast cancer recurrence can be defined as the return of breast cancer after primary treatment and can recur within the first three to five years. The Preliminary Report by National Cancer Patient Registry (NCPR) in 2008 reported that 108 out of 154 patients with available follow-up information (1st June - 31st Dec 2008), 94.4 percent patients are disease free while 5.6 percent patients are recurrent cases. There is still lack of studies on the risk factors of breast cancer recurrence in Malaysia. Most studies focus on risk factors and survival rate of breast cancer patients. Breast cancer recurrence studies have been mostly conducted in developing countries such United Stated, Japan and Canada. Thus, the primary aim of this study is to identify the risk factors of breast cancer recurrence among Malaysian women. Out of 1149 patients who were diagnosed and undergo treatment at Department of Surgery, Hospital Kuala Lumpur and only 454 cleaned data were used in this study. Data obtain retrospectively cohort from year 2006 until 2011 . The outcome was dichotomized into recurrence (code 1) and remission (code 0) with thit1een categorical predictors categorized into patients' background and medical factors. Breast Cancer Recurrence scorecard (BCR-scorecard) model and Logistic Regression model were developed and compared using SAS Enterprise Miner 7 .1. BCR-scorecard model is better predictive model with lower misclassification rate (18%) compared to Logistic Regression model (23%). Five important risk factors were identified: histological type, race, stage, tumour size and vascular invasion in predicting recurrence status.

Item Type: Thesis (Masters)
Creators:
CreatorsEmail
Jamian, Nurul HusnaUNSPECIFIED
Subjects: Q Science > QA Mathematics > Mathematical statistics. Probabilities
R Medicine > RC Internal Medicine > RC0254 Neoplasms. Tumors. Oncology (including Cancer)
Divisions: Faculty of Computer and Mathematical Sciences
Item ID: 12031
Last Modified: 14 May 2016 07:50
Depositing User: Admin Pendigitan PTAR
URI: http://ir.uitm.edu.my/id/eprint/12031

Actions (login required)

View Item View Item

Downloads

Downloads per month over past year