Indian Centre For Advanced Scientific And Technological Research
 
icastor + 91-8420396666
icastor maths@icastorindia.org // eng@icastorindia.org
 

--- View Details ---

Author Name H. H. Ahmed Hussien, K. E. Awadalla
Title APPLICATION OF LOGISTIC REGRESSION IN PREDICTING SPREAD OF CANCER
Category    Mathematics
Year :     2011
Month :   June
Abstract The treatment and prognosis of cancer depends on how much the disease has spread. One of the regions to which cancer may spread is the lymph nodes. If the lymph nodes are involved, the prognosis is generally poorer than if they are not. That is why it is desirable to establish as early as possible whether the lymph nodes are involved. For certain cancers, exploratory surgery just to determine whether the nodes are involved. The aim of this study is to fit a logistic regression model to predict whether the nodes are positive for cancer based on the values of the variables that cart be measured without surgery. Data: The data was collected from Radio- Isotope National Center for 148 Sudanese patients diagnosed with prostate cancer from 1999 to 2007. For each patient we report the age, serum acid phosphatase, the stage of the disease, the grade of the tumor and X-ray result as well as whether the cancer had spread to the regional lymph nodes at the time of surgery. Result: In the context of this study of prostate cancer in Sudanese patients, the model shows that age and SAP are major determinants of whether the nodes are involve (P-value <0.0l), but a variety of other factors, such as the stage of the disease, and X-ray result are also important in refining predictions of lymph node involvement. Conclusion: The model is the best fit to calculate the probability of lymph node metastasis at different ages and it is much easier to interpret for clinical use.
Download pdf