|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 26, No 6 (February 20), 2008: pp. 835-841 © 2008 American Society of Clinical Oncology. DOI: 10.1200/JCO.2007.13.1490 Long-Term Prediction of Prostate Cancer: Prostate-Specific Antigen (PSA) Velocity Is Predictive but Does Not Improve the Predictive Accuracy of a Single PSA Measurement 15 Years or More Before Cancer Diagnosis in a Large, Representative, Unscreened Population
From the Departments of Surgery (Urology), Clinical Laboratories, Epidemiology and Biostatistics, Memorial Sloan-Kettering Cancer Center, New York, NY; and the Departments of Laboratory Medicine and Clinical Sciences in Malmö, Lund University, University Hospital UMAS, Malmö, Sweden Corresponding author: Hans Lilja, MD, PhD, Department of Clinical Laboratories, Surgery (Urology), and Medicine (GU-Oncology), 1275 York Ave, Box 213, Memorial Sloan-Kettering Cancer Center, New York, NY 10021; e-mail: liljah{at}mskcc.org
Purpose We tested whether total prostate-specific antigen velocity (tPSAv) improves accuracy of a model using PSA level to predict long-term risk of prostate cancer diagnosis.
Methods During 1974 to 1986 in a preventive medicine study in Sweden, 5,722 men aged Results The median time from second blood draw to cancer diagnosis was 16 years; median follow-up for men without prostate cancer was 21 years. In univariate models, tPSA level at second assessment and tPSAv between first and second assessments were associated with prostate cancer (both P < .001). tPSAv was highly correlated with tPSA level (r = 0.93). Twenty-year probabilities of cancer for men at 50th, 90th, and 95th percentile of tPSA and tPSAv were 10.6%, 17.1%, and 21.2% for tPSA, and 9.1%, 11.8%, and 14.1% for tPSAv, respectively. The concordance index for tPSA level was 0.771. Adding tPSAv, fPSA, %fPSA or velocities of fPSA and %fPSA did not importantly increase accuracy of tPSA to predict prostate cancer. Results were unchanged if the analysis was restricted to patients with advanced cancer at diagnosis. Conclusion Although PSA velocity is significantly increased in men with prostate cancer up to two decades before diagnosis, it does not aid long-term prediction of prostate cancer.
Prostate-specific antigen (PSA) is used to aid detection of prostate cancer and to monitor treatment effects.1-3 Its value for detection is compromised because certain nonmalignant conditions (inflammation and benign prostatic hyperplasia [BPH]) increase PSA to levels above current biopsy thresholds.4 To improve detection, total PSA velocity (tPSAv), the rate of change of tPSA over time has been suggested as a predictor. Several studies confirmed that greater tPSA velocities indicate higher risk of subsequent prostate cancer.5-8 Numerous reports suggested that a sharper rise in PSA before treatment is correlated with aggressive disease,9-11 early recurrence after treatment,12 and prostate-cancer–related death.13,14 Recent evidence suggests that tPSAv could be used to predict life-threatening prostate cancer 10 to 15 years before diagnosis.15 Data from the Malmö Preventive Medicine (MPM) study suggest that a single PSA measurement at age 44 to 50 years predict long-term risk of prostate cancer.16 This study consisted of a large, representative, population-based cohort of men in Sweden, where prostate cancer screening is not recommended and where PSA testing is uncommon. Because most cases were diagnosed clinically, rather than following PSA screening, any association of PSA with prostate cancer in this cohort will be free of verification bias. In MPM, a second blood sample was obtained from a third of participants 6 years from baseline, enabling us to calculate tPSAv between the first and second measurements. This allowed us to test whether tPSAv predicts subsequent clinical diagnosis of prostate cancer in an unscreened population. However, for a marker to be useful, it must contribute predictive information beyond that provided by established predictors. The key question is therefore not whether tPSAv predicts future prostate cancer diagnosis, but whether a clinician is better able to predict if he or she accesses prior tPSA data and calculates velocity, as opposed to just using the most recent tPSA level. We therefore compared the predictive accuracy of a model including tPSA with one that included both tPSA and tPSAv.
Patients The MPM study16-18 was designed to investigate risk factors for cardiovascular and metabolic diseases among middle-aged men and women. Between 1974 and 1986, 21,277 males aged 33 to 50 years participated at baseline (participation rate, 74%). Six years after baseline venipuncture, all men born during 1926 to 1931 and in 1938 were reinvited for a second screening. Figure 1 shows the flow of participants through the study.
Our current study cohort consisted of 5,722 men participating in the second screening (participation rate, 72%). EDTA-anticoagulated blood was collected at each occasion and archived at –20°C. Of the 5,722 participants, 812 (14%) had samples that were lost or could not be analyzed, and three were diagnosed with prostate cancer before their second blood draw, leaving 4,907 participants for analysis. Information regarding occurrence and date of prostate cancer diagnosis was obtained from the Swedish National Cancer Registry through December 31, 2003. Prostate cancer was identified in 443 of 4,907 men. We investigated whether our subset of 4,907 participants was representative of the entire MPM cohort, in which 1,045 of 21,277 men were diagnosed with prostate cancer through December 31, 2003. We compared baseline age and tPSA between the subset and the entire MPM cohort following methods used in our prior report.16 The median tPSA was similar among cases (1.07 ng/mL in entire cohort and in this subset) and among controls (0.55 v 0.53 ng/mL). Median age at baseline was 47 and 48 years, respectively. Moreover, the strength of the association between tPSA and prostate cancer was similar (area under receiver operating characteristic curve [AUC] of tPSA at baseline 0.762 in prior publication16 v 0.758 in this subset). Our subset therefore is highly representative of the entire MPM cohort. Patient clinical records were reviewed for 75% of the 241 cancer cases diagnosed by December 31, 1999. Data were extracted from the National Prostate Cancer Registry for cases diagnosed from 1997. Clinical data were obtained for 410 of 443 (93%) of cases. Median delay from second blood sample to diagnosis was 16.1 year (range, 1.7 to 22.0 years). Most patients had sextant biopsies performed and evaluated according to the Gleason system, but some were graded according to WHO criteria. Therefore, we translated all gradings into the WHO system. Of 396 patients, 136 (34%) were WHO-I, 163 (41%) patients were WHO-II, and 97 (25%) patients were WHO-III, with biopsy grade unavailable for 14 patients. Clinical stage was T1 for 139 of 407 (34%) patients, T2 for 146 (36%) patients, and T3-4 for 122 (30%) patients, with 3 patients lacking stage information. Bone scans were performed on 319 of 410 (78%) participants at diagnosis; 49 of 319 (15%) were positive. The initial cause for consultation was noted in 50 of 57 patient records available for men with T1 disease. Of these, 43 of 50 (86%) patients had previously consulted the Department of Urology (University Hospital UMAS, Malmö, Sweden) for lower urinary tract symptoms. Any men who had a PSA test as part of the prostate cancer diagnostic process were subject to the PSA threshold recommendations of the Southern Swedish Regional Prostate Cancer Advisory Panel. The recommended tPSA threshold was initially 4.0 ng/mL, but in late 1990s, it was lowered to 3.0 ng/mL, irrespective of transrectal ultrasound or digital rectal examination.
PSA Measurements
Statistical Methods
Baseline characteristics of our cohort are presented in Table 1. Median age of participants at second venipuncture was 53 years. There were 443 documented cases of prostate cancer; median follow-up time for men without prostate cancer was 21 years. The 20-year cancer-free probability was 90.5% (95% CI, 89.5% to 91.3%; Fig 2). On univariate analysis, the level of each PSA form at second assessment and their velocities were all associated with prostate cancer (all P < .0005). Higher levels of tPSA, tPSAv, fPSA, and fPSAv, and lower levels of %fPSA and %fPSAv were associated with increased risk of prostate cancer (Table 2). tPSA level at second assessment showed greater spread of risk than the velocities. For example, the 20-year probability of prostate cancer diagnosis at median tPSA (10.6% risk) was half the risk at the 95th percentile of tPSA (21.2% risk), whereas risk for tPSAv increased from 9.1% at median tPSAv to 14.1% risk at the 95th percentile.
Table 3 presents c-indices of all PSA forms at second assessment and c-indices when other PSA forms and velocities are added to a model based on tPSA at second assessment. On univariate analysis, tPSA had the highest predictive accuracy among all PSA forms (c-index of 0.771); the difference between the c-index of tPSA and that of tPSAv (0.712) was highly significant (P < .0005). Adjusting for tPSA, tPSAv remained a significant predictor of prostate cancer (P < .0005). However, addition of tPSAv to a model including tPSA level did not enhance predictive accuracy for prostate cancer (c-index, 0.771 for both models). This is likely due to the high correlation between tPSA and tPSAv (0.93). Results were similar with the addition of fPSA, fPSAv, %fPSA, and %fPSAv: c-indices from these models ranged from 0.770 to 0.777.
Although the overall accuracy of tPSAv plus tPSA was similar to that of tPSA alone, it remains possible that tPSAv is superior for decisions in which either high specificity or high sensitivity is required. We compared the two models by decision curve analysis,22 using cancer at 20 years as a binary end point (Fig 3). The x-axis is the threshold for action (ie, probability of subsequent cancer diagnosis when a man would opt for intervention such as intensive screening). The y-axis represents clinical "net benefit": the probability of true-positive minus probability of false-positive, where the latter is weighted by the threshold probability. If 100 men are tested, a difference in net benefit of 0.01 is equivalent to one additional man who will develop prostate cancer receiving a risk prediction above threshold level and therefore opting for an intervention, with no increase in number of false-positive predictions. Both tPSA alone and tPSA plus tPSAv are of clinical net benefit for most thresholds compared with intervening in all men (Fig 3 gray line) or no men (thick black line). However, clinical benefit of tPSA plus tPSAv is never much greater than that for tPSA alone, regardless of threshold probability.
We performed additional analyses for men whose tPSA at second assessment was in the upper 50%, 25%, and 10% of men in our study cohort (tPSA above 0.7, 1.2, and 2.0 ng/mL, respectively). In these analyses, fPSAv and levels of fPSA and %fPSA improved predictive accuracy by the largest amount; tPSAv and %fPSAv were less informative (Table 3). Since tPSAv did not improve prediction of any prostate cancer, we explored whether tPSAv was of benefit in the prediction of advanced prostate cancer. For this analysis, we focused on the 82 participants diagnosed with advanced prostate cancer by December 31, 1999, when our database was current for clinical chart review. We failed to find any enhanced predictive value contributed by PSAv irrespective of the definition of advanced cancer used. For clinical stage T3-4 or metastases at diagnosis, the c-index was 0.849 for tPSA plus tPSAv versus 0.848 for tPSA alone; for WHO grade 3 cancer at diagnosis, addition of tPSAv did not importantly increase the c-index (0.805 for tPSA plus PSAv v 0.800 for tPSA alone). The c-index of tPSA alone was significantly higher than for tPSAv alone at P less than .0005 for both definitions.
Interest in tPSAv results from the observation that tPSA levels are increased by BPH, a common condition in the age group subject to PSA screening.16,23 Using a large population-based cohort of men in early middle age who were likely to have a low incidence of BPH, we investigated whether tPSAv and other PSA forms can aid tPSA in long-term prostate cancer prediction. The study should be free of verification bias, as there was no recommendation for PSA testing in Sweden during the study period. The majority of MPM participants had no PSA tests before diagnosis. The substantial proportion of cancers staged as T1 at diagnosis (34%) could appear to be at odds with these cancers having been diagnosed clinically. However, in the chart review of the men staged T1 at diagnosis, a large majority (86%) had lower urinary tract symptoms as the initial cause for consultation. Our results showed that men subsequently diagnosed with prostate cancer have increased tPSA and fPSA levels, decreased %fPSA, and increased velocities of tPSA and its forms up to 20 years before cancer diagnosis. This long-term predictive value of tPSA levels reinforces our earlier findings,16 derived from a case-control study within the MPM cohort. Long-term predictive value for tPSAv was reported from the Baltimore Longitudinal Study of Aging, where a tPSAv of more than 0.75 ng/mL/yr was associated with increased risk of prostate cancer diagnosis within 5 years.5 A tPSAv more than 0.1 ng/mL/yr was associated with increased risk of cancer diagnosis within 10 years.6 However, to demonstrate that PSAv is of clinical value, a model incorporating both tPSA and tPSAv must have predictive accuracy superior to tPSA alone. We found no additional predictive value contributed by tPSAv or the velocity of any other PSA form. Similarly, tPSAv was associated with advanced prostate cancer diagnosis, but contributed no additional value for advanced cancer prediction when added to tPSA. Of note, the predictive value of tPSAv alone was 0.712, which was improved to 0.771 by adding tPSA. However, as the predictive value of tPSA alone was 0.771, we see no reason for a clinician to retrieve a previous PSA value to calculate tPSAv, as the tPSAv does not provide any additional information compared with the most recent PSA value alone. Carter et al15 stated that the addition of tPSA as a categoric variable to a model including age, date of diagnosis, and tPSAv did not significantly improve the performance of this model. However, they did not comment on whether the addition of tPSAv to a model including tPSA improved model performance. We think this is the critical question because of the additional data required and effort involved in calculating PSAv. To our knowledge, this study includes the largest number of prostate cancer cases (total, 443 cases) with the longest time delay to diagnosis (median of 16 years from second assessment) ever studied for evaluating long-term prostate cancer prediction. In comparison, the recent report on PSA velocity from the Baltimore Longitudinal Study of Aging focused on PSAv at 10 to 15 years before diagnosis and included 124 prostate cancer patients.15 Our study is subject to several limitations. PSA measurement is susceptible to sample degradation and to errors in the assay itself. We minimized these problems by using previously unthawed anticoagulated EDTA plasma instead of serum19 and the same, research-standard, dual-label assay for all analytes.20 The MPM cohort limited us to two PSA measurements for the calculation of PSA velocities. Though certain authors recommend that PSA velocity should be calculated from three separate measurements to reduce variability,24 Riffenburgh et al demonstrated that two PSA values are sufficient if samples were obtained more than a year apart.25 The interval of 6 years between samples is longer than any previously investigated,5-7,15,25,26 but the longer interval should provide a more robust estimate of velocity by reducing the influence of random intraindividual variation. Indeed, the median velocities measured (eg, 0.17 ng/mL throughout 6 years for tPSA) were above the expected level of random intraindividual variation (coefficient of variation 12% for tPSA, corresponding to 0.06 ng/mL for the median tPSA level in our study).27 However, it remains possible that PSAv suddenly accelerated shortly before the second blood draw in men subsequently diagnosed with prostate cancer, but not those remaining cancer-free. Pearson et al demonstrated that the exponential portion of the PSA curve begins on average 7 to 9 years before diagnosis.28 With the median time from the second assessment to cancer diagnosis 16 years in our study, likelihood is low for a sudden acceleration of tPSAv shortly before the second blood draw (Fig 4).
There are two related explanations for the finding that PSAv does not aid prediction of cancer in early middle-aged men. First, tPSAv was highly correlated with tPSA level at second assessment (correlation, 0.93). If two variables are highly correlated, using both provides little additional information over using only one. This finding is consistent with the recent report by Carter et al15 that tPSA and tPSAv were correlated (r = 0.7), and that tPSA did not improve the predictive value of a model that included tPSAv. Second, the growth curve for prostate cancer is likely to be exponential (Fig 4). Accordingly, the growth curve is nearly flat many years before diagnosis and starts to rise more steeply only as disease progresses. Therefore, it is possible that tPSAv gains greater value to predict prostate cancer later in the disease process when men are older, time to diagnosis is shorter, incidence of BPH is higher, and predictive value of tPSA is lower.29 That said, data from the ERSPC study in Rotterdam suggested that tPSAv measured close to the time of diagnosis did not improve the diagnostic accuracy of tPSA.26,30 Furthermore, using tPSA doubling time (tPSADT) rather than velocity is unlikely to be of benefit. Riffenburgh et al25 demonstrated by simple analytic geometry that tPSADT contains the same information as tPSAv and depends on both tPSA level and tPSAv. tPSADT is therefore another mathematical manipulation of a highly predictive variable. Thompson et al31 used log values to calculate tPSAv for 55- to 91-year-old men biopsied in the control arm of the Prostate Cancer Prevention Trial. They concluded that tPSAv alone was significantly associated with increased risk of prostate cancer but did not add predictive value when added to a prostate cancer risk model containing tPSA level. fPSA, fPSAv, %fPSA, and %fPSAv did not importantly aid prediction in our analysis in the current study, probably due to the long delay between the blood sample and cancer diagnosis.29 fPSA is a stronger marker of malignancy than tPSA,32,33 and patients in this cohort were most likely in an early, preclinical, or possibly premalignant state at the time of venipuncture. Our results have implications for understanding PSA dynamics two decades before prostate cancer diagnosis. Benefit for tPSAv as an additional predictor would be expected if tPSA increased suddenly and steeply in cancer cases. However, tPSAv did not improve predictive accuracy in our highly representative cohort, suggesting that the typical pattern of tPSA dynamics long before prostate cancer diagnosis is a gradual increase during many years. This also appears to be true for patients with advanced cancer at diagnosis.
Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a "U" are those for which no compensation was received; those relationships marked with a "C" were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors. Employment or Leadership Position: None Consultant or Advisory Role: None Stock Ownership: Hans Lilja, Patent holder for free PSA and hK2 blood assays Honoraria: None Research Funding: None Expert Testimony: None Other Remuneration: None
Conception and design: David Ulmert, Angel M. Serio, Matthew F. O'Brien, Charlotte Becker, James A. Eastham, Peter T. Scardino, Thomas Björk, Göran Berglund, Andrew J. Vickers, Hans Lilja Administrative support: Hans Lilja Provision of study materials or patients: Göran Berglund, Hans Lilja Collection and assembly of data: David Ulmert, Angel M. Serio, Charlotte Becker, Göran Berglund, Hans Lilja Data analysis and interpretation: David Ulmert, Angel M. Serio, Matthew F. O'Brien, Charlotte Becker, James A. Eastham, Peter T. Scardino, Thomas Björk, Göran Berglund, Andrew J. Vickers, Hans Lilja Manuscript writing: David Ulmert, Angel M. Serio, Matthew F. O'Brien, Charlotte Becker, James A. Eastham, Peter T. Scardino, Thomas Björk, Göran Berglund, Andrew J. Vickers, Hans Lilja Final approval of manuscript: David Ulmert, Angel M. Serio, Matthew F. O'Brien, Charlotte Becker, James A. Eastham, Peter T. Scardino, Thomas Björk, Göran Berglund, Andrew J. Vickers, Hans Lilja
We thank Janet Novak, PhD, and Susan Aiello, DVM, ELS, for editing the manuscript, and thank Gun-Britt Eriksson and Kerstin Håkansson for expert assistance with immunoassays.
This study was funded by P50-CA92629 SPORE from the National Cancer Institute, Swedish Cancer Society project No. 3555, Swedish Cancer Society Project No. 4715 European Union 6th Framework contract LSHC-CT-2004-503011 (P-Mark), and Fundación Federico SA. Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
1. Farkas A, Schneider D, Perrotti M, et al: National trends in the epidemiology of prostate cancer, 1973 to 1994: Evidence for the effectiveness of prostate-specific antigen screening. Urology 52:444-448, 1998[CrossRef][Medline] 2. Newcomer LM, Stanford JL, Blumenstein BA, et al: Temporal trends in rates of prostate cancer: Declining incidence of advanced stage disease, 1974 to 1994. J Urol 158:1427-1430, 1997[CrossRef][Medline] 3. Stephenson RA, Stanford JL: Population-based prostate cancer trends in the United States: Patterns of change in the era of prostate-specific antigen. World J Urol 15:331-335, 1997[CrossRef][Medline] 4. Ruckle HC, Klee GG, Oesterling JE: Prostate-specific antigen: Critical issues for the practicing physician. Mayo Clin Proc 69:59-68, 1994[Medline] 5. Carter HB, Pearson JD, Metter EJ, et al: Longitudinal evaluation of prostate-specific antigen levels in men with and without prostate disease. JAMA 267:2215-2220, 1992 6. Fang J, Metter EJ, Landis P, et al: PSA velocity for assessing prostate cancer risk in men with PSA levels between 2.0 and 4.0 ng/ml. Urology 59:889-893, 2002[CrossRef][Medline] 7. Ito K, Yamamoto T, Ohi M, et al: Usefulness of prostate-specific antigen velocity in screening for prostate cancer. Int J Urol 9:316-321, 2002[CrossRef][Medline] 8. Lynn NN, Collins GN, O'Reilly PH: The short-term prostate-specific antigen velocity before biopsy can be used to predict prostatic histology. BJU Int 85:847-850, 2000[CrossRef][Medline] 9. Goluboff ET, Heitjan DF, DeVries GM, et al: Pretreatment prostate specific antigen doubling times: Use in patients before radical prostatectomy. J Urol 158:1876-1878, 1997[CrossRef][Medline] 10. Hanks GE, Hanlon AL, Lee WR, et al: Pretreatment prostate-specific antigen doubling times: Clinical utility of this predictor of prostate cancer behavior. Int J Radiat Oncol Biol Phys 34:549-553, 1996[CrossRef][Medline] 11. Lee WR, Hanks GE, Corn BW, et al: Observations of pretreatment prostate-specific antigen doubling time in 107 patients referred for definitive radiotherapy. Int J Radiat Oncol Biol Phys 31:21-24, 1995[Medline] 12. Egawa S, Arai Y, Tobisu K, et al: Use of pretreatment prostate-specific antigen doubling time to predict outcome after radical prostatectomy. Prostate Cancer Prostatic Dis 3:269-274, 2000[CrossRef][Medline] 13. D'Amico AV, Chen MH, Roehl KA, et al: Preoperative PSA velocity and the risk of death from prostate cancer after radical prostatectomy. N Engl J Med 351:125-135, 2004 14. D'Amico AV, Renshaw AA, Sussman B, et al: Pretreatment PSA velocity and risk of death from prostate cancer following external beam radiation therapy. JAMA 294:440-447, 2005 15. Carter HB, Ferrucci L, Kettermann A, et al: Detection of life-threatening prostate cancer with prostate-specific antigen velocity during a window of curability. J Natl Cancer Inst 98:1521-1527, 2006 16. Lilja H, Ulmert D, Björk T, et al: Long-term prediction of prostate cancer in a large, representative Swedish cohort: Prostate kallikreins measured at age 44-50 predict prostate cancer up to 25 years before diagnosis. J Clin Oncol 25:431-436, 2007 17. Berglund G, Eriksson KF, Israelsson B, et al: Cardiovascular risk groups and mortality in an urban Swedish male population: The Malmö Preventive Project. J Intern Med 239:489-497, 1996[CrossRef][Medline] 18. Berglund G, Nilsson P, Eriksson KF, et al: Long-term outcome of the Malmö preventive project: Mortality and cardiovascular morbidity. J Intern Med 247:19-29, 2000[CrossRef][Medline] 19. Ulmert D, Becker C, Nilsson JA, et al: Reproducibility and accuracy of measurements of free and total prostate-specific antigen in serum vs plasma after long-term storage at -20 degrees C. Clin Chem 52:235-239, 2006 20. Mitrunen K, Pettersson K, Piironen T, et al: Dual-label one-step immunoassay for simultaneous measurement of free and total prostate-specific antigen concentrations and ratios in serum. Clin Chem 41:1115-1120, 1995 21. Rafferty B, Rigsby P, Rose M, et al: Reference reagents for prostate-specific antigen (PSA): Establishment of the first international standards for free PSA and PSA (90:10). Clin Chem 46:1291-1292, 2000 22. Vickers AJ, Elkin EB: Decision curve analysis: A novel method for evaluating prediction models. Med Decis Making 26:565-574, 2006 23. Thompson IM, Chi C, Ankerst DP, et al: Effect of finasteride on the sensitivity of PSA for detecting prostate cancer. J Natl Cancer Inst 98:1128-1133, 2006 24. Carter HB, Pearson JD, Waclawiw Z, et al: Prostate-specific antigen variability in men without prostate cancer: Effect of sampling interval on prostate-specific antigen velocity. Urology 45:591-596, 1995[CrossRef][Medline] 25. Riffenburgh RH, Amling CL: Use of early PSA velocity to predict eventual abnormal PSA values in men at risk for prostate cancer. Prostate Cancer Prostatic Dis 6:39-44, 2003[CrossRef][Medline] 26. Roobol MJ, Kranje R, de Koning HJ, et al: Prostate-specific antigen velocity at low prostate-specific antigen levels as screening tool for prostate cancer: Results of second screening round of ERSPC (Rotterdam). Urology 63:309-313, 2004[CrossRef][Medline] 27. Bruun L, Becher C, Hugosson J, et al: Assessment of intra-individual variation in prostate-specific antigen levels in a biennial randomised prostate cancer screening program in Sweden. Prostate 65:216-221, 2005[CrossRef][Medline] 28. Pearson JD, Carter HB: Natural history of changes in prostate specific antigen in early stage prostate cancer. J Urol 152:1743-1748, 1994[Medline] 29. Vickers AJ, Ulmert D, Serio AM, et al: The predictive value of prostate cancer biomarkers depends on age and time to diagnosis: Towards a biologically-based screening strategy. Int J Cancer 121:2212-2217, 2007[CrossRef][Medline] 30. Schröder FH, Roobol MJ, van der Kwast TH, et al: Does PSA velocity predict prostate cancer in pre-screened populations? Eur Urol 49:460-465, 2006[CrossRef][Medline] 31. Thompson IM, Ankerst DP, Chi C, et al: Assessing prostate cancer risk: Results from the Prostate Cancer Prevention Trial. J Natl Cancer Inst 98:529-534, 2006 32. Steuber T, Niemela P, Haese A, et al: Association of free-prostate specific antigen subfractions and human glandular kallikrein 2 with volume of benign and malignant prostatic tissue. Prostate 63:13-18, 2005[CrossRef][Medline] 33. Haese A, Graefen M, Steuber T, et al: Total and Gleason grade 4/5 cancer volumes are major contributors of human kallikrein 2, whereas free prostate specific antigen is largely contributed by benign gland volume in serum from patients with prostate cancer or benign prostatic biopsies. J Urol 170:2269-2273, 2003[CrossRef][Medline] Submitted June 20, 2007; accepted November 8, 2007.
Related Editorial
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2008 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
50 gave two blood samples about 6 years apart. We measured free (fPSA) and total PSA (tPSA) in archived plasma samples from 4,907 participants. Prostate cancer was subsequently diagnosed in 443 (9%) men. Cox proportional hazards regression was used to evaluate tPSA and tPSAv as predictors of prostate cancer. Predictive accuracy was assessed by the concordance index. 
