A Study That Provides Long-term Safety Follow-up and Examines Long-term Exposure to Abiraterone Acetate

This is a nonrandomized (individuals will not be assigned by chance to study treatments),
open-label (individuals will know the identity of study treatments), long-term safety
follow-up study of abiraterone acetate in approximately 300 patients from other completed
abiraterone acetate clinical studies. Patients must have received at least 3 months of
treatment with abiraterone acetate and a low-dose corticosteroid and, based on investigator
assessment, may benefit from continued treatment. This study will consist of a screening
period followed by open-label treatment of continued abiraterone acetate access. The patients
will continue with the same abiraterone acetate and low-dose corticosteroid dosing regimen
they were receiving in the previous abiraterone acetate clinical study until the investigator
determines that the patient is no longer receiving benefit or the sponsor terminates the
study. Patients can be withdrawn from the study if an alternative access (eg,
patient-assistance program or commercial source of abiraterone acetate) is available and
feasible. Each cycle of treatment will be 28 days. Investigators will monitor and assess the
patients for response to treatment or progression according to routine practice or as
clinically indicated to determine whether continued treatment with abiraterone acetate is
warranted. No efficacy data are being collected. Safety will be monitored throughout the
study for a maximum duration of 6 years. End-of-study assessments will be performed at least
30 days after the last dose of abiraterone or upon early withdrawal.

Inclusion Criteria:

- Currently participating in an abiraterone acetate clinical study considered complete
and had received at least 3 months of treatment with abiraterone acetate tablets.

Exclusion Criteria:

- Medical conditions that require hospitalization.

- Any condition or situation which, in the opinion of the investigator, may put the
patient at significant risk, may confound the study results, or may interfere
significantly with the patient's participation in the study.