Multiple Imputation as a Means to Assess Mammographic vs. Ultrasound Technology in Determining Breast Cancer Recurrence
Irene Helenowski, Hakan Demirtaş, Seema Ahsan Khan, Firas G. Eladoumikdachi, Ali Shidfar · Turkiye Klinikleri Journal of Biostatistics · 2014
ABS TRACT Ultrasound and mammogram are two imaging tools used to determine breast cancer diagnosis. One prognostic factor measured by these technologies is tumor size and also can be used in the prediction of recurrence. Which technology offers better determination of diagnosis and gives better measurements of prognostic factors is an ongoing debate among clinicians. Examining the performance of these tools by the association between tumor size and recurrence could depend on the amount and nature of missing data, however. The purpose of this work is two-fold. The first purpose is to determine any relationship between recurrence and tumor size via ultrasound and mammogram by employing complete case analysis. The second purpose involves applying multiple imputation to determine the significance of any indicative associations found in the complete case analyses. In taking these approaches, we aim to show how multiple imputation can aid investigators in further understanding how to discern associations in their data relevant in clinical applications as predicting recurrence.