Predictors of Tumor Status among Breast Cancer Patients

Name of dataset: wisbcdata.xlsx

The data is from the Wisconsin Breast Cancer which consists of 683 cases of potentially cancerous tumors. Traditionally whether a tumor is malignant or benign is determined with an invasive surgical biopsy procedure. An alternative less invasive technique called "fine needle aspiration" allows examination of a small amount of tissue from the tumor (FNA). For the Wisconsin data, FNA provided nine-cell features for each case; a biopsy was then used to determine the tumor status as malignant or benign.

Can the Fine Needle Aspiration (FNA) Technique be used as a Substitute for Biopsy?

The project has the following outcome (Y) and predictors (Xi) variables:

Outcome: Y = tumor status

Predictors: X1 - X9

1) Do the cell features allow us to predict tumor status? That is, can we use FNA as an alternative to the biopsy procedure for future patients?

(2) What are the sensitivity and specificity of the FNA based on the model?

(3) What features are the predictors of tumor status?

Develop your hypotheses based on the predictor variables you are interested to test and investigate.

Make a decision on which variables to retain in the final model based on the results of a detailed analysis.

1. Using the wisbcdata.xlsx data, fit the logistic regression model for

= probability of tumor status, using x19 as the predictors:

2. give a linear approximation for the estimated effect of a 1 unit increase in each independent variable.

3. Construct a 95% confidence interval to describe the effect of each variable on the odds of tumor status. Interpret.

4. test the hypothesis that predictors have no effect on tumor status. Interpret.

Outcome: Y = tumor status Predictors: X1 X9 Fit logistic regession model of tumor status (Y) using Xl-X9. Description of variables Vaable Definition Tumor status ( 0 = benign, 1 = malignant) Clump thickness Cell size uniformity Y X1 X2 X3 Cell shape uniformity X4 Marginal adhesion X5 Single epithelial cell size X6 Bare nuclei X7 Bland chromatin X3 Normal nucleoli X9 Mites es