The patient population with a rising prostate specific antigen (PSA) post-therapy with no evidence of disease on standard imaging studies currently represents the second largest group of prostate cancer patients. Little information is still available regarding the specificity and sensitivity of positron emission tomography (PET) tracers in the assessment of early biochemical recurrence. Ideally, PET imaging would allow one to accurately discriminate between local vs nodal vs distant relapse, thus enabling appropriate selection of patients for salvage local therapy. The vast majority of studies show a relatively poor yield of positive scans with PSA values <4 ng ml−1. So far, no tracer has been shown to be able to detect local recurrence within the clinically useful 1 ng ml−1 PSA threshold, clearly limiting the use of PET imaging in the post-surgical setting. Preliminary evidence, however, suggests that 11C-choline PET may be useful in selecting out patients with early biochemical relapse (PSA < 2 ng ml−1) who have pelvic nodal oligometastasis potentially amenable to local treatment. The role of PET imaging in prostate cancer is gradually evolving but still remains within the experimental realm. Well-conducted studies comparing the merits of different tracers are needed.