We previously demonstrated that IgG responses to a panel of 126 prostate tissue-associated antigens are common in patients with prostate cancer. In the current report we questioned whether changes in IgG responses to this panel might be used as a measure of immune response, and potentially antigen spread, following prostate cancer-directed immune-active therapies. Sera were obtained from prostate cancer patients prior to and three months following treatment with androgen deprivation therapy (n = 34), a poxviral vaccine (n = 31), and a DNA vaccine (n = 21). Changes in IgG responses to individual antigens were identified by phage immunoblot. Patterns of IgG recognition following three months of treatment were evaluated using a machine-learned Bayesian Belief Network (ML-BBN). We found that different antigens were recognized following androgen deprivation compared with vaccine therapies. While the number of clinical responders was low in the vaccine-treated populations, we demonstrate that ML-BBN can be used to develop potentially predictive models.