We evaluated whether the prostate-specific antigen (PSA) mass or free PSA (fPSA) mass (i.e.,absolute amount of total circulating PSA or fPSA protein,respectively),versus serum PSA or fPSA concentration,improves the accuracy of predicting the total prostate volume (TPV) in relation to obesity.Among men whose multicore (≥12) transrectal prostate biopsy was negative,586 who had a PSA of ＜10 ng ml-1 and underwent the fPSA test prior to biopsy were enrolled.The PSA mass or fPSA mass (pg) was calculated by multiplying the serum level by plasma volume.At each TPV cut-off point (30 ml,40 ml,and 50 ml),the areas under the receiver operating characteristics curve (AUCs) of each variable were compared in obesity-based subgroups.AUCs of fPSA and fPSA mass for predicting TPV were significantly larger than those for PSA and PSA mass by 8.7％-12.1％ at all cut-off points.Subgroup analyses based on obesity showed that,although PSA mass and fPSA mass enhanced accuracy by 4％ (P =0.031) and 1.8％ (P =0.003),respectively,for determining TPVs of ≥30 ml and ≥50 ml in obese and overweight men,they did not improve the accuracy in most other combinations of the degrees of obesity with TPV cut-off points.Thus,compared with serum PSA or fPSA,the absolute amount of PSA or fPSA protein mass improved the accuracy of predicting TPV in obese men very minimally and only for certain TPV cut-off points.Hence,these indicators may not provide clinically meaningful improvement in predicting TPV in obese men.