Development of a nomogram model using a very large sample of patients (pts) to predict the risk of 99mTc-bone scan (BS) positivity in pts receiving androgen deprivation therapy (ADT) for prostate cancer (PCa).

4552 Background: Decisions to start imaging pts with PCa can be difficult due to the heterogeneous nature of the disease. Nomograms can combine information on multiple disease factors to improve predictive accuracy and are not influenced by subjective confounders that may affect clinical judgment. However, currently available nomograms for PCa predict current or future risk of positive BS in ADT-naive pts only. We have developed a new nomogram to predict current BS positivity in ADT-treated pts with PCa using a very large pt sample. Methods: All BS records from 1650 ADT-treated pts who received treatment at the Memorial Sloan-Kettering Cancer Center from 2000 to 2011 were analyzed. Pts were followed up from the start of ADT until the first positive BS or last hospital visit. Multivariate logistic regression analysis was used to model the variables that could be used for predicting likelihood of metastases and the variables were incorporated into the nomogram model. The current probability of bone metastas...