Summary The present study evaluated serial serum measurements of tissue polypeptide-specific antigen (TPS) in comparison with prostatespecific antigen (PSA) for assessment of tumour progression in patients with advanced prostate cancer receiving intermittent androgen suppression therapy (IAS). Twenty-three men were recruited into an IAS trial consisting of an initial 8 months of androgen suppression, followed by cycles of treatment cessation and resumption of therapy upon increases of PSA > 20 ng ml-' to prolong the hormone responsiveness of the tumour cells. Periods of androgen suppression resulted in reversible reduction in serum testosterone (< 1.8 nmol l-1) and PSA (< 4 ng ml-') and decreases in tumour volume (mean reduction for first cycle 24 ± 10%), indicating partial growth arrest and apoptotic regression of the tumours. In contrast to PSA values, non-specifically elevated TPS values were found in 8 of 23 patients. In 15 of 23 patients, TPS fell during periods of apoptotic tumour regression and increased simultaneously with testosterone and preceded the increases in PSA by 2 months during the period of treatment cessation. Although TPS represents a highly sensitive marker of tumour proliferation in this IAS clinical model of controlled tumour regression and regrowth, its low specificity compared with PSA limits its usefulness to monitoring of prostate cancer patients with proven absence of non-specific elevations of this marker.