The treatment options currently available for patients with clinically localized prostate cancer are radical prostatectomy, irradiation and deferred symptom-guided treatment. Today we do not have any controlled data indicating that any of these treatments is better than the other. Management policies that have evolved are, at best, based on structured comparisons of data from uncontrolled treatment series. In such comparisons the differences between the various treatment options with respect to disease-specific survival up to 10 to 15 years after diagnosis are modest. These comparisons are based on patient series diagnosed 15 to 20 years ago, i.e. before prostate specific antigen (PSA) was available as a marker for prostate cancer. The utilization of PSA has dramatically changed the conditions for diagnosing prostate cancer, in that with PSA we find more men with prostate cancer with smaller tumors at diagnosis. Today we find many tumors because of an elevated PSA followed by systematic biopsies in men without palpable prostatic lesions. Thus it is difficult to translate long-term outcome data from older patient series to the current situation in which tumors are found under changed clinical conditions. Consequently, a patient can only be informed about the different treatment options and their side effects. In the treatment decision process he has to balance a possible benefit in survival against side effects, according to his own preferences. We need randomized studies comparing the different treatment options. Such research is under way in Scandinavia, but due to the long course of the disease we will probably have to wait many years before we can answer the questions which sparked these studies.