Abstract Background: Prostate specific antigen (PSA)–α1-protease inhibitor complex (PSA-API) is a minor form of PSA in serum. It may be useful for prostate cancer (PCa) diagnosis, but its specific detection is hampered by nonspecific background. To avoid this, we developed an immunoassay for PSA-API based on proximity ligation. Methods: We used a monoclonal antibody (mAb) to total PSA (tPSA) to capture PSA, while using another anti-tPSA mAb together with an anti-API mAb as probes. We measured PSA-API by quantification of amplified DNA strands conjugated to the probes. We measured serum PSA-API in 84 controls and 55 men with PCa who had PSA concentrations of 4.0–10 μg/L. Results: The detection limit of the assay was 6.6 ng/L. The proportion of PSA-API to tPSA (%PSA-API) tended to be lower in men with PCa (2.8%) than without cancer (3.3%) but was not statistically significant (P = 0.363). When used alone, %PSA-API [area under the curve (AUC) 0.546] did not improve detection of PCa, whereas %fPSA (AUC 0.710) and the sum of %fPSA and %PSA-API (AUC 0.723) did. At 90% diagnostic sensitivity, the diagnostic specificity for cancer was not significantly better for %fPSA + %PSA-API than for %fPSA alone (36% vs 30%). Conclusions: Proximity ligation eliminated nonspecific background, enabling accurate measurement of PSA-API in serum specimens with moderately increased tPSA. The combined use of %PSA-API and %fPSA provided a modest improvement for PCa detection, but based on the current study cohort, it is uncertain whether the improvement has clinical utility. .