To construct a clinical nomogram instrument to estimate individual risk for having prostate cancer (PC) for patients undergoing prostate specific antigen (PSA) screening, using all risk factors known for PC.

We conducted a cross-sectional study of 3,108 men who underwent a prostate biopsy, including a subset of 408 volunteers with normal PSA levels. Factors including age, family history of PC (FHPC), ethnicity, urinary symptoms, PSA, free:total PSA ratio, and digital rectal examination (DRE) were incorporated in the model. A nomogram was constructed to assess risk for any and high-grade PC (Gleason score ≥ 7).

Of the 3,108 men, 1,304 (42.0%) were found to have PC. Among the 408 men with a normal PSA (< 4.0 ng/mL), 99 (24.3%) had PC. All risk factors were important predictors for PC by multivariate analysis (P, .01 to .0001). The area under the curve (AUC) for the nomogram in predicting cancer, which included age, ethnicity, FHPC, urinary symptoms, free:total PSA ratio, PSA, and DRE, was 0.74 (95% CI, 0.71 to 0.81) and 0.77 (95% CI, 0.74 to 0.81) for high-grade cancer. This was significantly greater than the AUC that considered using the conventional screening method of PSA and DRE only (0.62; 95% CI, 0.58 to 0.66 for any cancer; 0.69; 95% CI, 0.65 to 0.73 for high-grade cancer). From receiver operating characteristic analysis, risk factors including age, ethnicity, FHPC, symptoms, and free:total PSA ratio contributed significantly more predictive information than PSA and DRE.

In a PC screening program, it is important to consider age, family history of PC, ethnicity, urinary voiding symptoms, and free:total PSA ratio, in addition to PSA and DRE.

The practice of prostate cancer screening is widespread among primary care physicians and specialists in the United States and Canada.1,2 Approximately 75% of men age 50 or older undergo regular prostate specific antigen (PSA) screening for prostate cancer.2 When it was first approved in 1986 by the US Food and Drug Administration, abnormal PSA levels were accurate in detecting all prevalent cases of prostate cancer.3,4 However, recent studies have now demonstrated that PSA alone cannot accurately predict new instances of prostate cancer because of the low proportion of prostate cancer volume to the amount of normal prostate among incident cases.5,6 Abnormal PSA levels cannot reliably distinguish patients with prostate cancer from those with benign prostatic hyperplasia,3 and a high rate of prostate cancer is present among men considered to have healthy PSA levels.6 Further, a proportion of patients diagnosed with prostate cancer through a PSA screening program have nonaggressive forms that do not require any treatment.7,8

Thus, there is uncertainty in using PSA as a screening instrument for prostate cancer among decision makers, physicians, and patients. We have previously shown that incorporating established risk factors for prostate cancer in a multivariate model can significantly improve the positive predictive value of PSA.9,10 By combining a panel of predictive variables in a nomogram model, including age, ethnicity, family history of prostate cancer, and prostate volume, we were able to improve the positive predictive value of PSA to detect prostate cancer among 2,637 patients who underwent one or more prostate biopsies.9 However, in this study, all patients had an abnormal PSA level, some patients had multiple prostate biopsies, and free:total PSA ratio measurements were not available, which has been accepted as a standard adjunct PSA test.11 Further, because prostate volume was incorporated in the nomogram model, patients would have to undergo invasive transrectal ultrasonography for accurate prostate volume assessment, making these results not clinically useful for patients who undergo their first screening assessment.

To individualize prostate cancer risk using a simple, noninvasive tool at the time of first PSA and digital rectal examination (DRE) screening, we developed a predictive nomogram model from a cohort of 3,108 men who underwent a prostate biopsy for the first time to determine the presence of prostate cancer. To address our previous limitations, we systematically measured free:total PSA among all 3,108 patients from banked sera. Also, we included a subset of men consisting of volunteers with healthy PSA levels who underwent a prostate biopsy. We incorporated age, ethnicity, family history of prostate cancer, the presence of urinary symptoms, total PSA, free:total PSA, and DRE to construct nomograms to estimate an individual's risk for having prostate cancer at the time of first biopsy. We did not include prostate volume in the model, but rather used urinary obstructive symptoms as a surrogate measure. These factors can be easily and noninvasively determined at the time of screening. This clinical tool will aid physicians and patients in accurately assessing an individual's risk for prostate cancer at the time of PSA and DRE screening.

Study Subjects

Patients were drawn from a sample of 3,010 eligible men who were referred to the Prostate Centers of the University of Toronto (Sunnybrook Health Sciences Centre and University Health Network) between June 1999 and June 2005. Patients were included in the study if they had an abnormal PSA value (≥ 4.0 ng/mL) or DRE. All patients underwent transrectal ultrasonography guided needle core biopsies. Patients eligible for this study were unselected and were accrued consecutively. No patient had a past history of prostate cancer. Patients were excluded if their PSA was greater than 50 ng/mL, where the decision to biopsy would be considered unequivocal (n = 54) if patients were not capable of giving consent to participate in a research study (n = 46), or if patients could not provide sufficient baseline information including ethnic background or family history of prostate cancer (n = 53). Of the remaining 2,857 men, 2,700 (95%) agreed to participate.

These subjects were originally accrued for an ongoing prostate cancer genetics and biomarker study where DNA and serum were collected and banked. Many factors have been and continue to be evaluated.12-14 As a part of this study, volunteers who were willing to undergo a prostate biopsy with a PSA in the healthy range (0 to 4.0 ng/mL) were accrued to serve as potential healthy controls if no cancer was detected at biopsy for our biomarker studies. These patients were not part of the referral stream for consideration of a prostate biopsy for the 2,700 patients. Rather, these patients were assessed at the prostate centers where patients with healthy PSA levels were evaluated as part of a prostate cancer screening program offered to the general public. No incentives were given to the volunteers and no major complications from the biopsy occurred. Over the study period, 408 men agreed to undergo a prostate biopsy for a healthy PSA level. Of the 408 men, 160 (39.2%) had an abnormal DRE, similar to the 47.9% incidence rate of abnormal DREs among patients with healthy PSAs from the Prostate Cancer Prevention Trial.15 Thus, a total of 3,108 subjects were used as our study cohort. All research was conducted with informed consent and with the approval of the hospital research ethics board.

Baseline Data Information and Primary End Point

A urological voiding history (AUA Symptom Score16), DRE results, serum PSA level, family history of prostate cancer information, and ethnic background were obtained by research personnel through questionnaire administration and medical record review. All data were stored within a centralized database. From our ongoing biomarker study, stored serum (at −70°C) was obtained for free:total PSA ratio measurements for each patient using standardized commercial kits (Beckman-Coulter, Brea, CA).

Six to 15 ultrasound-guided needle core biopsies were performed (median, 8), using an 18-gauge spring-loaded biopsy device. Samples were obtained using a systematic pattern and additional targeted samples were obtained from suspicious areas. The primary end point was the histologic presence of adenocarcinoma of the prostate in the biopsy specimen. All grading was based on the Gleason scoring system.17

Data Analysis

Cases were defined as patients with prostate cancer from core biopsies and controls were defined as having no evidence of cancer. Potential factors associated with increased prostate cancer risk were compared between cases and controls, including age, ethnicity, family history of prostate cancer, the presence of lower urinary tract voiding symptoms, total PSA levels, free:total PSA ratio, and DRE. We did not include the total number of needle cores taken at the time of biopsy because the majority of patients (> 80%) underwent one systematic pattern and eight total number of needle cores. Also, from the current and past studies,18 we did not find a significant correlation between the number of needle cores and the probability for cancer, and therefore, was not considered in the model. Further, Presti et al19 compared a variety of needle core patterns from six to 10 needle cores and found that eight was the optimal strategy. Unconditional logistic regression analysis was used to examine how each of these factors, alone and in combination, would predict the presence of prostate cancer.

Creation of Nomograms

Ordinal logistic regression was used to model the probability of having low or high grade cancer. Three outcome levels were defined: no cancer; low grade cancer (Gleason score 6 or lower); and high grade cancer (Gleason score 7 or higher). Continuous variables were modeled with restricted cubic splines to avoid linearity assumptions. The logistic regression model was the basis for constructing a nomogram. Of 3,108 patients, 2,108 were used for model building, and the remaining one third (n = 1,000) were used for model assessment.

Nomogram validation comprised two activities. First, discrimination was quantified with the area under the receiver operating characteristic curve. Second, calibration was assessed. This was done by grouping patients into deciles (each of size 100) with respect to their nomogram-predicted probabilities and then comparing the mean of the group with the observed proportion of patients with any or high-grade cancer. All analyses were performed using S-Plus 2000 professional software version 2 (Statistical Sciences, Seattle, WA) with the Design and Hmisc libraries added.20

Of 3,108 men, 1,304 men (42%) were found to have adenocarcinoma of the prostate at biopsy (cases), and 1,804 patients (58%) had no evidence of cancer (controls). Among the 408 subset of men with a healthy PSA (< 4.0 ng/mL), 99 (24.3%) had cancer at biopsy. Of 1,304 patients with cancer, more than one half had a Gleason score of 7 or higher; 19 (1.5%) had a Gleason score of 4 to 5; 534 (40.9%) had a Gleason score of 6; 606 (46.5%) had a Gleason score of 7, and 145 (11.1%) had a Gleason score of 8 to 10 cancer. All studied risk factors were found to be significantly associated with prostate cancer detection from univariate and multivariate analysis (Tables 1 and 2).

Nomograms to Estimate Risk for Prostate Cancer and Aggressive Prostate Cancer

We constructed a nomogram to predict both the presence of prostate cancer and aggressive forms of prostate cancer (patients with Gleason score 7 or higher cancers; Fig 1). We evaluated the nomogram accuracies in predicting the presence of prostate cancer and aggressive prostate cancer by comparing the predicted and actual probabilities for prostate cancer and aggressive prostate cancer in the validation set (Fig 2). The AUC was high for predicting the probability of prostate cancer. The total AUC for all factors in predicting prostate cancer was 0.74 (95% CI, 0.71 to 0.81) based on the nomogram (Fig 3). The total AUC for predicting high grade cancer was 0.77 (95% CI, 0.74 to 0.81; Fig 3).

To determine whether the nomogram was better at predicting the presence of prostate cancer compared with using the conventional screening method of PSA and DRE alone, we constructed a nomogram that only considered PSA and DRE in the model. The total AUC for this model was significantly lower than the more comprehensive nomogram (AUC, 0.62; 95% CI, 0.58 to 0.66 for any cancer; AUC, 0.69; 95% CI, 0.65 to 0.72 for high-grade cancer). To examine which factors were important in predicting prostate cancer within the comprehensive nomogram model, we examined how the AUC of the predictive model would drop by removing age, family history of prostate cancer, ethnicity, urinary symptoms, and the free:total PSA ratio, and compared the AUC drop if PSA and DRE were removed. If PSA and DRE were removed from the nomogram model, the incremental AUC drop was only 0.010 (Table 3). However, if we removed all of the other risk factors the incremental AUC drop was 0.082 (Table 3). Further, each of the risk factors alone including age, ethnicity, family history of prostate cancer, and lower urinary tract voiding symptoms were the same or better than PSA when comparing the incremental drop in the AUC (Table 3).

Also, the nomogram performed significantly better than conventional screening with PSA and DRE for patients with a healthy PSA level. For patients with a PSA lower than 4.0 ng/mL, the AUC for the nomogram was 0.74 (95% CI, 0.63 to 0.86) for predicting any cancer and 0.77 (95% CI, 0.66 to 0.88) for predicting high-grade cancer. Again, when we compared this with a conventional screening model with only PSA and DRE, the AUC was significantly less (AUC, 0.60; 95% CI, 0.48 to 0.72 for any cancer; AUC, 0.63; 95% CI, 0.48 to 0.79 for high-grade cancer). The incremental drop in the AUC was 0.085 if the nomogram factors were removed from the model compared with a drop of only 0.003 if PSA and DRE were removed from the model (Table 3). Also, for this subgroup of patients, age (incremental drop in AUC, 0.063) was the most important predictor for prostate cancer, beyond that of PSA (incremental drop in AUC, 0.002) and free:total PSA (incremental drop in AUC, 0.019; Table 3).

For patients with a PSA ≥ 4.0 ng/mL, free:total PSA was the most important predictor (Table 3). In this group, the AUC for the nomogram model that considered all the factors was 0.73 (95% CI, 0.69 to 0.76), which was significantly better than the model with PSA and DRE alone (AUC, 0.59; 95% CI, 0.56 to 0.63) for predicting cancer.

Finally, to determine which, if any, of these factors were associated with predicting the presence of aggressive prostate cancer at diagnosis, we compared each of the factors between patients with nonaggressive cancer and aggressive cancer. Significant predictors for high-grade cancer were age, DRE, PSA, and free:total PSA ratio (Table 4).

To our knowledge, this is the first study that combines all established risk factors and tumor markers for prostate cancer into a comprehensive clinical instrument that determines an individual's risk for prostate cancer. Compared with the conventional screening method of PSA and DRE, our nomogram model that includes age, ethnicity, family history of prostate cancer, urinary symptoms, and free:total PSA ratio, was significantly better in predicting prostate cancer. These factors in the nomogram are easy to assess by any primary or tertiary physician at the time of PSA and DRE screening. Using all risk factors and tumor markers, we were able to improve the AUC from 0.62 for conventional screening to 0.74 in predicting prostate cancer. Although better accuracy would be desirable, this would provide an important platform and benchmark to evaluate novel biomarkers.

The main advantage of this nomogram instrument is that clinicians can assess prostate cancer risk on an individual basis and make management decisions. For example, for a patient with a PSA of lower than 4.0 ng/mL, but who has other risk factors for prostate cancer, a biopsy may be justified based on the nomogram. In contrast, if the nomogram predicts a low chance for having aggressive prostate cancer for an older patient with a high PSA level, then it would be reasonable for the patient to forego a biopsy. The exact probability cutoff for undergoing or foregoing a biopsy would be left with the treating physician and patient and should be individualized. A list of sensitivities and specificities has been provided based on various probability cutoffs for the nomogram (Fig 3).

An important finding from this analysis was when compared with other factors used in the nomogram, PSA and DRE alone, which are considered the conventional screening methods, did not contribute significantly to the AUC in the predictive model, particularly for patients with a healthy PSA level. Age, DRE, and free:total PSA ratio were the strongest predictors (Table 3). Although individually, the other predictors were not as strong as those three factors, when combined they contributed important information. The AUC of age, DRE, and free:total PSA alone in predicting prostate cancer was 0.71 (95%CI, 0.68 to 0.74) compared with 0.74 with all factors. Thus, these factors should always be considered in the context of a prostate cancer screening program, particularly when these factors are routinely available and easy to obtain from a simple clinical history and physical examination.

Others have examined some of these factors for initial21,22 and repeat23 biopsy, but none have been as comprehensive as this study. Finne et al21 studied 758 men with an abnormal PSA from a Finnish prostate cancer screening study and developed a nomogram for prostate cancer risk, but this model did not include information on family history of prostate cancer or ethnicity. Karakiewicz et al22 also developed a nomogram to evaluate prostate cancer risk, but only included age, PSA, free:total PSA ratio, and DRE in the model. They also did not examine patients with healthy PSA levels.

From the Prostate Cancer Prevention Trial (PCPT), Thompson et al24 used results from the placebo arm to assess prostate cancer risk and also considered age, ethnicity, and family history, but did not consider urinary symptoms. Also, free:total PSA ratios were not measured, which have been an established adjunct test for PSA.11 In addition, in this study, 89% of the 5,519 patients had a healthy PSA level and only 150 patients had a PSA higher than 6 ng/mL. One of the major limitations of PSA is its low predictive value at abnormal levels for which this study attempts to improve. The majority of our cohort (n = 2,700) were men with a PSA between 4 and 20 ng/mL. Also, our cancer detection rate for men with a healthy PSA (24%) was similar to men in the placebo arm of the PCPT with a healthy PSA level. We also showed that the free:total PSA ratio was an important predictor that the PCPT could not address.24

Another aspect of this study was the construction of a nomogram that can predict the presence of aggressive forms of prostate cancer (Gleason score 7 or higher). Albertsen et al25 in a large population-based survey showed that patients with low-grade cancer (Gleason score 6 or lower) have significantly fewer life years lost from prostate cancer compared with patients with high-grade cancers. However, many experts would agree that patients with prostate cancer with a Gleason score 7 or higher require aggressive treatment.25,26

One limitation of the nomogram is the potential for misclassification of cases and controls. We and others have also shown that patients with an initial negative biopsy can have cancer at repeat biopsies (up 25%).18,27 We did not assess repeat biopsies. Others have designed nomograms for patients to undergo repeat biopsy.28 Also, our grade distribution also had a high proportion of Gleason score 7 or higher. This may be due to, in part, a contemporary trend of upgrading previously reported lower grade tumors29 and the fact that we included patients with PSA levels of up to 50 ng/mL. Finally, referral bias may also have reduced the predictive ability of PSA and DRE. However, our results of the performance of PSA and DRE alone are similar to those of the placebo arm of the PCPT which showed an AUC for PSA and DRE of 0.67.30

Although this study reflects a prostate cancer screening population, it is important for other centers to confirm and validate our findings. A multi-institutional study is currently underway. Nevertheless, this instrument provides important information for physicians and patients in assessing an individual's risk for prostate cancer and we have made it available for the general public.31

Although all authors completed the disclosure declaration, the following authors or their immediate family members indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. 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: N/A Leadership: Michael W. Kattan, Oncovance Technologies Consultant: Michael W. Kattan, GlaxoSmithKline Stock: N/A Honoraria: N/A Research Funds: D. Andrew Loblaw, Astra Zeneca Testimony: N/A Other: D. Andrew Loblaw, AstraZeneca, Sanofi-Aventis

Conception and design: Robert K. Nam, Steven A. Narod

Financial support: Robert K. Nam

Administrative support: Robert K. Nam, Ants Toi

Provision of study materials or patients: Robert K. Nam, Ants Toi, John Trachtenberg, Steven A. Narod

Collection and assembly of data: Robert K. Nam, Ants Toi, John Trachtenberg, Steven A. Narod

Data analysis and interpretation: Robert K. Nam, Ants Toi, Laurence Klotz, John Trachtenberg, Michael A.S. Jewett, Sree Appu, D. Andrew Loblaw, Linda Sugar, Steven A. Narod, Michael W. Kattan

Manuscript writing: Robert K. Nam, Ants Toi, Laurence Klotz, John Trachtenberg, Michael A.S. Jewett, Sree Appu, D. Andrew Loblaw, Linda Sugar, Steven A. Narod, Michael W. Kattan

Final approval of manuscript: Robert K. Nam, Ants Toi, Laurence Klotz, John Trachtenberg, Michael A.S. Jewett, Sree Appu, D. Andrew Loblaw, Linda Sugar, Steven A. Narod, Michael W. Kattan

Table

Table 1. Comparison of Factors Associated With Prostate Cancer Between Cases and Controls

Table 1. Comparison of Factors Associated With Prostate Cancer Between Cases and Controls

FactorCancer
No Cancer
P
No.%No.%
No. of patients1,30442.01,90458.0< .0001
Median age, years66.063.6
    Mean65.963.5
    Range41.3-93.839.9-89.6
Family history of PC.004
    Absent1,09340.91,57759.1
    Present21148.222751.8
Ethnicity< .0001
    Asian4123.713276.3
    White1,09742.81,46457.2
    Black14049.714250.3
    Other2628.36671.7
LUTS (AUA symptom score)
    Median57< .0001
    Mean7.68.6
    Range0-350-35
DRE
    No nodule93137.81,53062.2< .0001
    Nodule37357.727442.3
Median PSA, ng/mL7.496.74< .0001
    Mean9.78.1
    Range0.3-48.70.05-45.2
Free:total PSA ratio0.120.17< .0001
    Mean0.140.19
    Range0.007-0.90.007-0.9

NOTE. Percentages reflect the probability of having prostate cancer for each categorical factor.

Abbreviations: PC, prostate cancer; LUTS, lower urinary tract voiding symptoms; DRE, digital rectal examination; PSA, prostate specific antigen.

Table

Table 2. Multivariate Analysis of Factors Associated With PC

Table 2. Multivariate Analysis of Factors Associated With PC

FactorRegression Coefficient*P
Age0.047< .0001
Family history of PC
    AbsentBaseline
    Present0.43
Ethnicity
    WhiteBaseline
    Asian−0.69< .0001
    Black0.33.01
LUTS−0.02.0001
DRE
    No noduleBaseline
    Nodule0.68< .0001
PSA0.019.003
Free:total PSA ratio−6.26< .0001

Abbreviations: PC, prostate cancer; LUTS, lower urinary tract voiding symptoms; DRE, digital rectal examination; PSA, prostate specific antigen.

*Multivariate model includes age at biopsy, family history of prostate cancer, ethnicity, presence of LUTS, DRE, PSA, and free:total PSA ratio.

†Age at biopsy (per year), LUTS, PSA (per ng/mL), and free:total PSA ratio considered as continuous variables within the multivariate model.

‡Baseline group for ethnic background defined as white and other.

Table

Table 3. ROC Analysis and Comparison of Incremental Drops in the AUC for Each Predictor Variable When Removed From the Nomogram Model

Table 3. ROC Analysis and Comparison of Incremental Drops in the AUC for Each Predictor Variable When Removed From the Nomogram Model

FactorIncremental Drop in AUC
All PatientsPSA < 4.0 ng/mLPSA ≥ 4.0 ng/mL
AUC for PC0.740.740.73
    95% CI0.71 to 0.810.63 to 0.860.69 to 0.76
No. of patients3,1084082,700
Conventional screening
    PSA0.0010.0020.001
    DRE0.0090.0010.018
Total drop in AUC0.0100.0030.019
Additional nomogram factors
    Age0.0160.0630.011
    Ethnic background0.0010.0010
    Family history0.0010.0010.001
    LUTS0.0010.0010.001
    Free:total PSA ratio0.0630.0190.073
Total drop in AUC0.0820.0850.086

Abbreviations: ROC, receiver operating characteristic; AUC, area under the curve; PC, prostate cancer; LUTS, lower urinary tract voiding symptoms; PSA, prostate specific antigen; DRE, digital rectal examination.

Table

Table 4. Comparison of Factors Between Patients With Aggressive Cancer and Nonaggressive Cancer

Table 4. Comparison of Factors Between Patients With Aggressive Cancer and Nonaggressive Cancer

FactorNonaggressive Prostate Cancer
Aggressive Prostate Cancer
P
No.%No.%
Gleason score≤ 6≥ 7
No. of patients553751
Median age, years64.367.2< .0001
    Mean64.367.0
    Range42.7-88.541.3-93.8
Family history of PC
    Absent45841.963558.1
    Present9545.011655.0.40
Ethnicity
    Asian1639.02561.0.48
    White47443.262356.8
    Black5539.38560.7
    Other830.81869.2
LUTS
    Absent31741.944058.1.65
    Present23643.131156.9
DRE
    No nodule44547.848652.2< .0001
    Nodule10829.026571.0
Median PSA, ng/mL6.508.35< .0001
    Mean7.711.2
    Range0.3-45.00.6-48.7
Free:total PSA ratio0.140.10< .0001
    Mean0.160.12
    Range0.007-0.90.02-0.9

Abbreviations: PC, prostate cancer; LUTS, lower urinary tract voiding symptoms; DRE, digital rectal examination; PSA, prostate specific antigen.

© 2007 by American Society of Clinical Oncology

Supported by National Cancer Institute of Canada and the Terry Fox Foundation, Grants No. 010284 and 015164.

Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.

1. Bunting PS, Goel V, Williams JI, et al: Prostate-specific antigen testing in Ontario: Reasons for testing patients without diagnosed prostate cancer. Cmaj 160::70,1999-75, MedlineGoogle Scholar
2. Sirovich BE, Schwartz LM, Woloshin S: Screening men for prostate and colorectal cancer in the United States: Does practice reflect the evidence? JAMA 289::1414,2003-1420, Crossref, MedlineGoogle Scholar
3. Stamey TA, Yang N, Hay R, et al: Prostate-specific antigen as a serum marker for adenocarcinoma of the prostate. N Engl J Med 317::909,1987-915, Crossref, MedlineGoogle Scholar
4. Gann PH, Hennekens CH, Stampfer MJ: A prospective evaluation of plasma prostate-specific antigen for detection of prostatic cancer. JAMA 273::289,1995-294, Crossref, MedlineGoogle Scholar
5. Stamey TA, Caldwell M, McNeal JE, et al: The prostate specific antigen era in the United States is over for prostate cancer: What happened in the last 20 years? J Urol 172::1297,2004-1301, Crossref, MedlineGoogle Scholar
6. Thompson IM, Pauler DK, Goodman PJ, et al: Prevalence of prostate cancer among men with a prostate-specific antigen level < or = 4.0 ng per milliliter. N Engl J Med 350::2239,2004-2246, Crossref, MedlineGoogle Scholar
7. Albertsen PC, Hanley JA, Gleason DF, et al: Competing risk analysis of men aged 55 to 74 years at diagnosis managed conservatively for clinically localized prostate cancer. JAMA 280::975,1998-980, Crossref, MedlineGoogle Scholar
8. Barry MJ, Albertsen PC, Bagshaw MA, et al: Outcomes for men with clinically nonmetastatic prostate carcinoma managed with radical prostatectomy, external beam radiotherapy, or expectant management. Cancer 91::2302,2001-2314, Crossref, MedlineGoogle Scholar
9. Nam RK, Toi A, Trachtenberg J, et al: Making sense of PSA: Improving its predictive value among patients undergoing prostate biopsy. J Urol 175::489,2006-494, Crossref, MedlineGoogle Scholar
10. Nam RK, Toi A, Klotz LH, et al: Nomogram prediction for prostate cancer and aggressive prostate cancer at time of biopsy: Utilizing all risk factors and tumour markers for prostate cancer. Canadian J Urol 13::2,2006-10, Google Scholar
11. Catalona WJ, Smith DS, Wolfert RL, et al: Evaluation of percentage of free serum prostate-specific antigen to improve specificity of prostate cancer screening. Am J Med 274::1214,1995-1220, Google Scholar
12. Nam RK, Reeves JR, Toi A, et al: A novel serum marker, total prostate secretory protein of 94 amino acids (PSP94), improves prostate cancer detection and helps identify high grade cancers at diagnosis. J Urol 175::1291,2006-1297, Crossref, MedlineGoogle Scholar
13. Nam RK, Zhang WW, Jewett MA, et al: The use of genetic markers to determine risk for prostate cancer at prostate biopsy. Clinical Cancer Res 11::8391,2005-8397, Crossref, MedlineGoogle Scholar
14. Nam RK, Zhang WW, Trachtenberg J, et al: A single nucleotide polymorphism of the human kallikrein-2 gene (KLK2) highly correlates with serum HK2 levels and their combination significantly enhances prostate cancer detection. J Clin Oncol 21::2312,2003-2319, LinkGoogle Scholar
15. Thompson IM, Goodman PJ, Tangen CM, et al: The influence of finasteride on the development of prostate cancer. N Engl J Med 349::215,2003-224, Crossref, MedlineGoogle Scholar
16. Barry MJ, Fowler FJJ, O'Leary MP, et al: The American Urological Association symptom index for benign prostatic hyperplasia: The Measurement Committee of the American Urological Association. J Urol 148::1549,1992-1557, Crossref, MedlineGoogle Scholar
17. Gleason DF, Mellinger GT: Prediction of prognosis for prostatic adenocarcinoma by combined histological grading and clinical staging. J Urol 111::58,1974-64, Crossref, MedlineGoogle Scholar
18. Nam RK, Toi A, Trachtenberg J, et al: Variation in patterns of practice in diagnosing screen-detected prostate cancer. BJU Int 94::1239,2004-1244, Crossref, MedlineGoogle Scholar
19. Presti Jr JC, Chang JJ, Bhargava V, et al: The optimal systematic prostate biopsy scheme should include 8 rather than 6 biopsies: Results of a prospective clinical trial. J Urol 163::163,2000-166, Crossref, MedlineGoogle Scholar
20. Harrell FE: Regression Modeling Strategies With Applications to Linear Models, Logistic Regression, and Survival Analysis . New York, Springer-Verlag, 2001 Google Scholar
21. Finne P, Auvinen A, Aro J, et al: Estimation of prostate cancer risk on the basis of total and free prostate-specific antigen, prostate volume and digital rectal examination. Eur Urol 41::619,2002-626, ; discussion 41:626-627, 2002 Crossref, MedlineGoogle Scholar
22. Karakiewicz PI, Benayoun S, Kattan MW, et al: Development and validation of a nomogram predicting the outcome of prostate biopsy based on patient age, digital rectal examination and serum prostate specific antigen. J Urol 173::1930,2005-1934, Crossref, MedlineGoogle Scholar
23. Lopez-Corona E, Ohori M, Scardino PT, et al: A nomogram for predicting a positive repeat prostate biopsy in patients with a previous negative biopsy session. J Urol 170::1184,2003-1188, ; discussion 170:1188, 2003 Crossref, MedlineGoogle Scholar
24. Thompson IM, Ankerst DP, Chi C, et al: Assessing prostate cancer risk: Results from the Prostate Cancer Prevention Trial. J Natl Cancer Inst 98::529,2006-534, Crossref, MedlineGoogle Scholar
25. Albertsen PC, Fryback DG, Storer BE, et al: Long-term survival among men with conservatively treated localized prostate cancer. JAMA 274::626,1995-631, Crossref, MedlineGoogle Scholar
26. Catalona WJ: Management of cancer of the prostate. N Engl J Med 331::996,1994-1004, Crossref, MedlineGoogle Scholar
27. Djavan B, Ravery V, Zlotta A, et al: Prospective evaluation of prostate cancer detected on biopsies 1, 2, 3 and 4: When should we stop? J Urol 166::1679,2001-1683, Crossref, MedlineGoogle Scholar
28. Yanke BV, Gonen M, Scardino PT, et al: Validation of a nomogram for predicting positive repeat biopsy for prostate cancer. J Urol 173::421,2005-424, Crossref, MedlineGoogle Scholar
29. Albertsen PC, Hanley JA, Barrows GH, et al: Prostate cancer and the Will Rogers phenomenon. J Natl Cancer Inst 97::1248,2005-1253, Crossref, MedlineGoogle Scholar
30. Thompson IM, Ankerst DP, Chi C, et al: Operating characteristics of prostate-specific antigen in men with an initial PSA level of 3.0 ng/ml or lower. JAMA 294::66,2005-70, Crossref, MedlineGoogle Scholar
31. ProstateRisk.ca: About us. http://prostaterisk.ca Google Scholar

COMPANION ARTICLES

No companion articles

ARTICLE CITATION

DOI: 10.1200/JCO.2007.10.6450 Journal of Clinical Oncology 25, no. 24 (August 20, 2007) 3582-3588.

Published online September 21, 2016.

PMID: 17704405

ASCO Career Center