Premium
Derivation and application of upper limits for prostate specific antigen in men aged 50–74 years with no clinical evidence of prostatic carcinoma
Author(s) -
DALKIN B.L.,
AHMANN F.R.,
KOPP J.B.,
CATALONA W.J.,
RATLIFF T.L.,
HUDSON M.A.,
RICHIE J.P.,
SCARDINO P.T.,
FLANIGAN R.C.,
DEKERNION J.B.,
WATERS W.B.,
KAVOUSSI L.R.,
MACFARLANE M.T.
Publication year - 1995
Publication title -
british journal of urology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.773
H-Index - 148
eISSN - 1464-410X
pISSN - 0007-1331
DOI - 10.1111/j.1464-410x.1995.tb07712.x
Subject(s) - medicine , rectal examination , urology , percentile , receiver operating characteristic , confidence interval , prostate cancer , transrectal ultrasonography , prostate specific antigen , prostate , carcinoma , population , gynecology , cancer , mathematics , statistics , environmental health
Objective To derive age‐specific upper limits for prostate specific antigen (PSA) level in men 50–74 years of age with no clinical evidence of prostatic carcinoma, and to test the sensitivity and specificity for cancer detection of these upper limits. Subjects and methods A total of 6166 men were recruited for a multicentre study of prostate cancer detection and underwent a serum PSA determination and digital rectal examination (DRE). Men considered to be clinically free of prostatic carcinoma were those with a normal DRE and a PSA level ≤ 4.0 ng/mL, and men with an abnormality in either parameter who underwent ultrasonography‐guided prostate biopsy that revealed no evidence of carcinoma. By these criteria, 5469 men had no evidence of prostatic carcinoma. Dividing the population into 5‐year age increments, three statistical methods were assessed to derive upper limits for serum PSA level by age; the mean +2 sd , the 99th percentile, and a 97.5% prediction interval based on linear regression. Results Newly‐derived upper limits calculated by each method in the 50–54 and the 70–74 age group were 3.9 ng/mL and 7.6 ng/mL (mean + 2 sd ), 5.2 ng/mL and 14.0 ng/mL (99th percentile), and 4.7 ng/mL and 8.2 ng/mL (97.5% prediction interval). The sensitivity of the newly‐derived upper limits was tested using receiver operating characteristic curves derived from men with no suspicious findings on DRE and a serum PSA concentration > 4.0 ng/mL. Although the specificity of the test increased with increasing PSA upper limits, no upper limits derived from these three methods yielded adequate sensitivity to detect cancer; sensitivities by age range were from 53 to 94%, using mean + 2 sd , from 25 to 50% with the 99th percentile, and from 47 to 64% with the 97.5% prediction interval. Conclusion We do not recommend age‐referenced adjustments in upper limits for serum PSA concentration, but recommend that an upper limit of 4.0 ng/mL be used in all men 50–74 years of age.