Selectively identifying aggressive prostate cancer will reduce unnecessary testing and overdiagnosis. Multivariate models can be of assistance but require proper validation and users must be aware of the setting from which the models were derived. See also pages 000-000.