Iseki R, Ohori M, Piccorelli A, Yu C, Piccorelli A, Ohno Y, Tachibana M, Kattan MW. Development of a nomogram for predicting a positive repeat prostate biopsy. World J Clin Urol 2014; 3(1): 47-53 [DOI: 10.5410/wjcu.v3.i1.47]
Corresponding Author of This Article
Ryo Iseki, MD, Department of Urology, Tokyo Medical University, 6-7-1 Nishi-shinjuku, Shinjuku, Tokyo 160-0023, Japan. iseki@tokyo-med.ac.jp
Research Domain of This Article
Urology & Nephrology
Article-Type of This Article
Original Article
Open-Access Policy of This Article
This article is an open-access article which was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/
Ryo Iseki, Makoto Ohori, Yoshio Ohno, Masaaki Tachibana, Department of Urology, Tokyo Medical University, Shinjuku, Tokyo 160-0023, Japan
Alice Piccorelli, Changhong Yu, Annie Piccorelli, Michael W Kattan, Department of Quantitative Health Sciences, Cleveland Clinic, Cleveland Clinic Foundation, Cleveland, OH 44195, United States
ORCID number: $[AuthorORCIDs]
Author contributions: All the authors contributed to this work.
Correspondence to: Ryo Iseki, MD, Department of Urology, Tokyo Medical University, 6-7-1 Nishi-shinjuku, Shinjuku, Tokyo 160-0023, Japan. iseki@tokyo-med.ac.jp
Telephone: +81-3-33426111 Fax: +81-3-33444813
Received: September 27, 2013 Revised: November 22, 2013 Accepted: December 13, 2013 Published online: March 24, 2014 Processing time: 178 Days and 19.8 Hours
Abstract
AIM: To find risk factors of cancer in patients who had a repeat biopsy and to develop the nomogram using our cohort.
METHODS: Among 3500 patients who had a prostate biopsy over 11 years between 2000 and 2010 at our hospital, we studied a total of 807 repeat biopsy sessions in 459 patients who had at least 1 initial negative biopsy. At each biopsy session, we recorded patient age, number of previous biopsy sessions, number of biopsy cores, number of previously negative biopsy cores, months from the initial biopsy, months from the previous biopsy, serum PSA, PSA slope, digital rectal examination findings, hypoechoic lesions suspicious for a cancer on transrectal ultrasonography, total prostate volume, transitional zone (TZ) volume, PSA density, PSA TZ density and history of high grade prostatic intraepithelial neoplasia (HGPIN) or atypical small acinar proliferation (ASAP). Clinical and pathological variables were correlated with the outcome of repeat biopsies. A nomogram was developed based on logistic regression analyses and calibration was performed.
RESULTS: Overall, 17% of repeat biopsies had a cancer. With receiver operating characteristics analyses, the highest area under the curve (AUC) was obtained based on all available 13 variables, which were age, PSA, digital rectal examination, PSA density, prostate volume, TZ volume, PSA TZ density, cumulative number of biopsy cores, HGPIN, ASAP, months from previous negative biopsy, initial negative biopsy and number of biopsy cores. Based on multivariable logistic regression analysis, a nomogram was constructed with an AUC of 0.74, which was greater than that of any single risk factor. The calibration plot seemed to be good.
CONCLUSION: Our nomogram for predicting a positive repeat biopsy can provide probabilities for cancer and may help clinical judgment on whether to do a repeat prostate biopsy.
Core tip: Although prostate cancer is found in about 30% of patients at the initial biopsy session, there is a need to identify those with a negative result but who are at high risk. Although individual risk factors have been found to be associated with cancer, patient counseling requires the integration of multiple risk factors to obtain a prediction for the individual. We developed a nomogram that predicts a positive biopsy after a previous negative biopsy session. It provides a wide range of probabilities for cancer and may help clinical judgment of whether to do a repeat prostate biopsy.
Citation: Iseki R, Ohori M, Piccorelli A, Yu C, Piccorelli A, Ohno Y, Tachibana M, Kattan MW. Development of a nomogram for predicting a positive repeat prostate biopsy. World J Clin Urol 2014; 3(1): 47-53
With the advent of serum prostatic specific antigen (PSA) and transrectal ultrasonography in the late 1980s, prostate needle biopsy has markedly increased for a diagnosis of prostate cancer[1,2]. About 70000 cases with prostate biopsies in a year in Japan[3] have been estimated, so roughly 50000 cases should have no cancer in biopsy specimens if the detection rate of cancer is about 25%. It is obvious that no cancer in the first biopsy session does not give the perfect answer so a repeat biopsy will be done over and over[4,5]. The procedure of prostate biopsy itself is not difficult but the complications, such as hematuria, dysuria and urinary tract infection, do exist and we should also not underestimate discomfort during the biopsy procedure.
On the other hand, the indication for rebiopsy has remained unclear. Clinicians tend to recommend a repeat biopsy because of high PSA levels, an increasing PSA, abnormalities on digital rectal examination and/or imaging studies, and the presence of high grade prostatic intraepithelial neoplasia (HGPIN) and/or atypical small acinar proliferation (ASAP) in a previous biopsy. Age, race and family history are also sometimes considered. Realistically, however, it is extremely difficult to get the right answers to make a clinical decision at the time of whether to do a biopsy with these many factors and backgrounds. In the past, many investigators studied to find risk factors to predict a cancer in a repeat biopsy and to construct models to provide the accurate prediction of a cancer. Among them, a nomogram, the mathematical model based on a logistic analysis with multiple factors, seems to be the best model to provide an accurate prediction for an individual patient. Therefore, we tried to find risk factors of cancer in patients with a repeat biopsy and develop the nomogram using our cohort.
MATERIALS AND METHODS
We studied 459 patients who had at least 1 initial negative biopsy and underwent a total of 807 repeat biopsy sessions at our hospital for 11 years between 2000 and 2010. The indication of repeat biopsy and number of biopsy cores varied among a total of 10 attending physicians. However, the number of biopsy cores has changed; 8 cores in 2000 to 2005, 10 cores in 2005 to 2007, and 12 cores in 2008 to the present. The majority of patients had a transrectal biopsy under local anesthesia of 10 cc of 1% lidocaine under the guidance of transrectal ultrasonography (TRUS). Also, the additional cores were taken if there were hypoechoic lesions suspicious for a cancer on TRUS. On TRUS, the volume of the prostate gland and transition zone were estimated with an ellipsoid formula, as previously described. PSA density[6] was calculated as PSA divided by prostate volume and PSA transition zone density (PSATZD)[7] was calculated as PSA divided by transition zone volume. PSA slope in ng/mL yearly was estimated by the difference between the most current serum PSA measured minus serum PSA at the previous biopsy, divided by elapsed time in years[8].
Pathology
We obtained the pathology outcome from the official pathology reports. The majority of pathological reports indicated HGPIN as an indicator of cancer precursor and we neglected the reports of low grade PIN. We also recorded ASAP if the pathological reports indicated that there were a few atypical cells which were not enough to diagnose a cancer.
Statistical analysis
In analyses, the association of positive repeat biopsy with each clinical and biopsy feature was evaluated using multivariable logistic regression analysis. All the predictor variables included in the multiple logistic regression were based on a systematic literature review and clinical relevance with model selection. Restricted cubic splines were used for numeric or ordinal variables to accommodate the non-linear relationship. Generalized estimating equations (GEEs) implemented to adjust the clustering effect resulted from the repeat biopsies from the same patients. To evaluate the predictive accuracy of each factor in combination, we used the area under the receiver operating characteristic (ROC) curve (AUC) as the discrimination metric. Also, in the calibration plots, we compared the predictive values with the actual outcomes. Bootstrapping was used to correct for over-fitting bias. Based on the logistic analysis, we developed a nomogram to predict the probability of a positive repeat biopsy. Statistical significance was determined with a P value less than 0.05. All analyses were carried out using the commercially available software, STATA (Stata Corporation, College Station, TX, United States) and S-plus (Insightful Corporation, Seattle, WA, United States) or the open-source statistical software R-2.12.2 (R Development Core Team 2011) with Design and Zelig packages added.
RESULTS
Overall, 17% of patients had a prostate cancer on repeat biopsy. Table 1 lists the characteristics of the patients. A mean of 1.7 repeat biopsy sessions (or 2.7 all biopsy sessions including the initial biopsy) per patient were performed. Table 2 shows the rates of a positive repeat biopsy according to each variable. The cancer detection rates at biopsy 2 to 4 were consistent at 16%-18%. The number of patients was too small but the detection rates were high at 38%-50% if palpable nodules were identified. As prostate volume measured with ultrasonography increased, detection rates of cancer significantly decreased (P < 0.001). Of 38 repeat biopsies with HGPIN in previous biopsy specimens, 11 (32%) had cancer compared to 116 (15%) of 769 patients with no HGPIN (P = 0.032). Similarly, of 64 biopsies with ASAP, 16 (25%) had a cancer compared to 116 (16%) of 743 biopsies with no ASAP (P = 0.05).
Table 1 Clinical and pathological characteristics n (%).
Predictor variables
No cancer
Cancer
P value
n = 675
n = 132
Digital rectal ex- amination
Negative
642 (95.1)
122 (92.4)
0.209
Positive
33 (4.9)
10 (7.6)
TRUS
Negative
603 (89.3)
108 (81.8)
0.015
Positive
72 (10.7)
24 (18.2)
HGPIN in previous biopsy
Negative
648 (96)
121 (91.7)
0.032
Positive
27 (4)
11 (8.3)
ASAP in previous biopsy
Negative
627 (92.9)
116 (87.9)
0.051
Positive
48 (7.1)
16 (12.1)
Age, yr
Mean
67
68.8
0.008
SD
7.3
6.3
No. of previous biopsy sessions
Mean
1.7
1.7
0.894
SD
1
1.1
No. of biopsy cores
Mean
11.3
11.9
0.006
SD
2.4
2.5
Cumulative No. of cores
Mean
16.1
15.8
0.791
SD
12.3
11.4
Months from the initial biopsy
Mean
26.8
25.7
0.576
SD
21.0
22.7
Months from the previous biopsy
Mean
16.9
15.7
0.353
SD
13.7
13.6
PSA slope (ng/mL per year)
Mean
1.2
1.3
0.824
SD
5.7
13.2
PSA (ng/mL)
Mean
9.8
11.6
0.011
SD
6.7
9.8
Prostate volume (cc)
Mean
53.0
43.8
< 0.001
SD
25.8
27.2
TZ prostate volume (cc)
Mean
30.2
21.2
< 0.001
SD
20.4
17.9
PSA density (ng/mL per cc)
Mean
0.2
0.3
< 0.001
SD
0.2
0.4
PSA TZ density (ng/mL per cc)
Mean
0.4
0.8
< 0.001
SD
0.4
0.9
Table 2 Association between cancer detection rates and clinical variables according to the number of biopsy sessions.
Variable
Total No./No. Ca (%)
Biopsy 2
Biopsy 3
Biopsy 4
Biopsy 5
459/75(16.3)
208/34 (16.7)
82/15(18)
41/3(7)
Digital rectal examination
Negative
437/73 (16.7)
195/29(14.9)
76/12 (15.8)
40/3 (7.5)
Positive
22/2 (9.1)
13/5(38.5)
6/3 (50)
1/0 (0)
Prostate volume on TRUS, (cc)
Less than 20
15/7 (47.7)
6/4 (66.7)
0/0 (0)
0/0 (0)
20-40
181/39 (21.5)
69/12 (17.4)
24/6 (25.0)
14/1 (7.1)
40-60
145/19 (13.1)
75/12 (16.0)
35/5 (14.3)
16/2 (12.5)
60-80
71/7 (9.9)
32/6 (18.8)
13/2 (15.4)
5/0 (0)
80 or Greater
47/3 (6.3)
26/0 (0)
10/2 (20.0)
6/0 (0)
PSA density (ng/mL per cc)
Less than 0.15
196/16 (8.2)
82/11 (13.4)
29/6 (20.7)
10/0 (0)
0.15-0.30
185/33 (17.8)
84/10 (11.9)
36/6 (16.7)
22/3 (13.6)
0.30-0.45
46/14 (30.4)
20/3 (15.0)
10/1 (10.0)
7/0 (0)
Greater than 0.45
32/12 (38.5)
22/10 (45.5)
7/2 (28.6)
2/0 (0)
PSA TZ density (ng/mL per cc)
Less than 0.25
156/9 (5.8)
58/7 (12.1)
19/3 (18.8)
10/0 (0)
0.25-0.5
168/22 (13.1)
85/11 (12.9)
42/7 (16.7)
18/2 (22.2)
0.5-1.0
98/26 (26.5)
40/6 (15.0)
15/3 (20.0)
12/2 (18.3)
Greater than 1.0
37/18 (48.6)
25/10 (40.0)
6/2 (33.3)
1/0 (0)
PSA (ng/mL)
Less than 4
18/1 (5.6)
1/1 (100)
0/0 (0)
0/0 (0)
4-10
313/48 (15.3)
136/19 (14.0)
49/8 (16.3)
20/2 (10.0)
10-20
108/21 (19.4)
54/11 (20.3)
25/5 (20.0)
17/1 (5.9)
Greater than 20
20/5 (25.0)
17/3 (17.6)
8/2 (25.0)
4/0 (0)
PSA slope (ng/mL per year)
Less than 0.75
237/30 (12.7)
102/15 (14.7)
44/8 (18.2)
21/1 (4.8)
0.75 or greater
222/45 (20.3)
106/19 (17.9)
38/7 (18.4)
20/2 (10.0)
Previous HGPIN
Negative
437/67 (15.3)
200/32 (16.0)
79/14 (17.7)
40/3 (7.5)
Positive
22/8 (36.4)
8/2 (25.0)
3/1 (33.3)
1/0 (0)
Previous ASAP
Negative
427/68 (15.9)
190/30 (15.8)
75/13 (17.3)
38/2 (5.3)
Positive
32/7 (21.9)
18/4 (22.2)
7/2 (28.6)
3/1 (33.3)
Normal histology
406/61 (15.0)
184/29 (15.8)
72/12 (16.7)
37/2 (5.4)
Total
459/75 (16.3)
208/34 (16.4)
82/15 (18.3)
41/3 (7.3)
Table 3 shows the results of the multivariable logistic regression analysis in predicting a positive biopsy after adjusting the clustering effect with GEEs. Of all variables, age, number of previous negative biopsy sessions, number of biopsy cores, cumulative number of biopsy cores, PSA slope, PSA density and history of ASAP were statistically significantly associated with repeat biopsy findings (all P < 0.05).
Table 3 Multivariable logistic regression analysis in predicting positive repeat biopsy with generalized estimating equations adjustment for clustering repeat biopsies from the same patients.
Predictor variables
Q1
Q3
Odds ratio
Lower 0.95
Upper 0.95
P
Age, yr
62
73
1.6
1.1
2.5
0.002
No. of previous biopsy sessions
1
2
2.3
0.8
6.2
0.002
No. of biopsy cores
10
12
1.6
1.3
2.1
0.001
Cumulative No. of cores
8
20
0.3
0.1
0.8
0.001
Months from the initial biopsy
10.3
36
0.6
0.2
2.0
0.727
Months from the previous biopsy
7.1
22.7
1.5
0.6
3.8
0.651
PSA slope (ng/mL per year)
-0.3
2.4
0.9
0.9
1.0
0.007
PSA (ng/mL)
6
11.4
1.0
0.3
4.0
0.098
Prostate volume (cc)
33.5
60.5
3.3
0.5
20.8
0.081
TZ prostate volume (cc)
15
36.1
0.1
0.0
1.0
0.116
PSA density (ng/mL per cc)
0.1
0.3
2.5
0.3
24.7
0.045
PSA TZ density (ng/mL per cc)
0.2
0.6
0.8
0.1
6.9
0.328
DRE-positive: negative
NA
NA
0.7
0.3
1.7
0.393
TRUS-positive: negative
NA
NA
1.4
0.8
2.6
0.218
HGPIN-positive: negative
NA
NA
2.2
0.9
5.2
0.080
ASAP-positive: negative
NA
NA
2.2
1.1
4.7
0.035
A nomogram was constructed incorporating all predictors (Figure 1). Although some variables were not statistically significant predictors of a positive repeat biopsy, they were retained in the nomogram because omitting them from the model tended to decrease accuracy. Also, other factors become falsely better if clinically important factors are omitted. The nomogram is used by first locating the patient position on each predictor variable scale. Each scale position has corresponding prognostic points (top axis). Point values for all the predictor variables are determined consecutively and summed to arrive at a total point value. This value is located on the total point axis and directly below is the prediction for finding cancer. The nomogram was internally validated with 1000 bootstrap resamples and achieved an AUC of 0.74.
Figure 1 Nomogram (age, number of previous biopsy session, number of biopsy cores, number of previously negative biopsy cores, months from initial biopsy, months from previous biopsy, prostate specific antigen slope, prostate specific antigen, digital rectal exam, transrectal ultrasonography, prostate volume, transition zone volume, prostate specific antigen density, high grade prostatic intraepithelial neoplasia, atypical small acinar proliferation, prostate specific antigen transition zone density) for predicting the probability of a positive repeat biopsy.
Nomogram is used by first locating the patient’s position on each factor axis. Each factor has corresponding prognostic points (top axis). The points for each factor are added to achieve a total point that is subsequently plotted on the total point axis (the second scale from the bottom). The probability of prostate cancer is estimated by drawing a straight line downwards to the bottom line from the location of the total point .TRUS: Transrectal ultrasonography; HGPIN: High grade prostatic intraepithelial neoplasia; ASAP: Atypical small acinar proliferation; TZ: Transitional zone; PSA: Prostatic specific antigen; DRE: Digital rectal exam.
Figure 2 shows the calibration by plotting the nomogram predicted probability for positive repeat biopsy against the patient’s observed probability. In general, the performance of the nomogram appears to be reasonably accurate while the predicted probability seems to overestimate a positive repeat biopsy in the higher range of predicted probability.
Figure 2 A calibration plot of the nomogram predicting the probability of a positive repeat biopsy.
DISCUSSION
In 2003, Lopez-Corona et al[9] first reported the nomogram to predict a cancer in 661 patients with a previous negative biopsy using 8 factors, such as age, digital rectal examination, a total number of biopsy cores, HGPIN, ASAP, PSA, PSA slope and family history. Since then, to our knowledge there are a total of 8 published nomograms[9-18], including 2 with an extended biopsy strategy[10,18], to predict a positive biopsy in a setting of repeat biopsy. We have summarized these nomograms in Table 4. All nomograms, except 2, included age, PSA and DRE as standard predictive factors. Among variables in the previous nomograms, 4 nomograms included the results of %free PSA which seemed to be most valuable predictor since it had a large scale of points in the nomograms with a relatively higher AUC of 0.76 to 0.856[10,11,14,18]. Following %free PSA, prostate volume seemed to be important since 6 out of 9 nomograms included it as a significant predictor. Only Sakura et al[18] reported that age was the largest scale in the nomogram, while another 7 nomograms showed a short scale for age. There were many other factors and none of 9 nomograms used identical factors, which may represent the complexity of the situation of a repeat biopsy.
Table 4 Published nomograms to predict positive prostate biopsy in repeat biopsy.
Age, No. of biopsy core, PSA, PSA density, prostate volume, TZ volume, PSA TZ density, TRUS, DRE, HGPIN, ASAP, months from initial biopsy and previous biopsy, No. of previous negative biopsy, PSA slope
11.4
17
Early studies reported relatively high rates of cancer detection on repeat biopsy after an initial biopsy finding of HGPIN, but recent studies have reported substantially lower rates[19,20]. Park et al[21] analyzed 45 patients with ASAP and 43 with HGPIN on initial biopsy and found cancer in 51% of repeat biopsies in each group. Abouassaly et al[22] reported that the findings of ASAP were associated with a high likelihood of cancer on repeat biopsy regardless of the number of cores taken on initial biopsy. In our study, in 32 patients with ASAP and 22 patients with HGPIN in the previous biopsy session, cancer was found in 21.9% of repeat biopsies in the ASAP group and in 36.4% in the HGPIN group. Since the proportion of ASAP and/or HGPIN of the entire group is small, the impact on the detection rates of cancer may not be huge but we still think that they should be included in the nomogram since they are a different entity from the other markers and it would be important to include the findings by pathologists.
In studies by Walz et al[10] and ourselves, PSA density was the best factor to predict a cancer in a repeat biopsy. Benecchi et al[13] also included PSA density, which was the second predictor after %free PSA, while Walz et al[10] indicated that PSA density was a stronger predictor than %free PSA. Previously, Keetch et al[23] noted that PSA density and PSA slope were the best predictors of cancer on a repeat biopsy. In men with a PSA density of ≥ 0.15 ng/mL per milliliter and a PSA slope of 0.75 ng/mL per year, the rate of cancer detection was 46% compared to 13% in men with a PSA density of < 0.15 ng/mL per milliliter and a PSA slope of < 0.75 ng/mL per year (P < 0.001)[23]. In contrast, Fowler et al[24] found that percent free PSA was the best predictor of cancer compared with the other PSA derivatives. Eventually, the inclusion of these important factors enhances the detection rate of cancer in a repeat biopsy and a nomogram would be the best method to combine these factors to provide an accurate prediction for the individual patient.
There are several limitations in the present study. Firstly, we could not include the results of %free PSA since it was frequently not measured at our clinic. However, all variables in our nomogram are readily available so it is not necessary for patients to come to the clinic frequently to get the result of %free PSA. Also, in the present study, we showed a similar accuracy to predict a positive repeat biopsy without %free PSA. Thus, our nomogram can help many patients and their physicians after appropriate validations. Secondly, there are 16 variables in our nomogram to achieve the highest AUC. The accuracy, AUC 0.74, seems to be reasonable but some may feel that there are too many variables to calculate the probability at a busy clinic. We agreed that the calculation on paper may not be realistic so we have put our nomogram on the web. Thirdly, because the nature of a nomogram is based on multivariable analysis, the values of some variables were reversed. For example, the points on the nomogram increase with a decrease of the PSA transition zone density. This happened because we included similar variables, such as transition zone volume, PSA density and prostate volume. Therefore, it may slightly confuse clinicians when they put each value of variables into our nomogram. However, the inclusion of such variables is valuable to get accurate estimations. Fourthly, in the present study, we did not do model selection that would be accomplished on the modeling data alone. Instead, we built the model by including all predictors that were clinically relevant to the disease based on the literature whether they were statistically significant or not on the current data. With enough number of events in the modeling data, theory driven models normally have better generalizability than data driven models when applied to other patient populations and preserve the predictive performance that was evaluated based on the modeling data[25,26].
In conclusion, we developed a nomogram to predict a cancer in a repeat biopsy using 16 readily available clinical factors with reasonable calibration. This helps both patients and clinicians to decide whether to do a repeat prostate biopsy.
COMMENTS
Background
Prostate cancer is found in about 30% of patients at the initial biopsy session but there is a need to identify those with a negative result who are at high risk. Although individual risk factors have been found to be associated with cancer, patient counseling requires the integration of multiple risk factors to obtain a prediction for the individual.
Innovations and breakthroughs
The authors developed a nomogram that predicts a positive biopsy after a previous negative biopsy session.
Applications
It provides a wide range of probabilities for cancer and may help clinical judgment of whether to do a repeat prostate biopsy.
Peer review
This paper is well-written and the topic is timely and appropriate.
Footnotes
P- Reviewers: Capitanio U, Gofrit ON, Mazaris E, Zhang J, Vaishampayan U S- Editor: Song XX L- Editor: Roemmele A E- Editor: Wu HL
Ohno Y, Nakamura T, Murata K. Prediction of Cancer Incidence in Japan.Cancer Statistics Report 2004. Tokyo: Shinohara Shuppan 2004; 201-217.
[PubMed] [DOI][Cited in This Article: ]
Keetch DW, Catalona WJ, Smith DS. Serial prostatic biopsies in men with persistently elevated serum prostate specific antigen values.J Urol. 1994;151:1571-1574.
[PubMed] [DOI][Cited in This Article: ]
Lopez-Corona E, Ohori M, Wheeler TM, Reuter VE, Scardino PT, Kattan MW, Eastham JA. Prostate cancer diagnosed after repeat biopsies have a favorable pathological outcome but similar recurrence rate.J Urol. 2006;175:923-927; discussion 923-937.
[PubMed] [DOI][Cited in This Article: ][Cited by in Crossref: 26][Cited by in F6Publishing: 24][Article Influence: 1.3][Reference Citation Analysis (0)]
Benson MC, Whang IS, Pantuck A, Ring K, Kaplan SA, Olsson CA, Cooner WH. Prostate specific antigen density: a means of distinguishing benign prostatic hypertrophy and prostate cancer.J Urol. 1992;147:815-816.
[PubMed] [DOI][Cited in This Article: ]
Kikuchi E, Nakashima J, Ishibashi M, Ohigashi T, Asakura H, Tachibana M, Murai M. Prostate specific antigen adjusted for transition zone volume: the most powerful method for detecting prostate carcinoma.Cancer. 2000;89:842-849.
[PubMed] [DOI][Cited in This Article: ][Cited by in F6Publishing: 1][Reference Citation Analysis (0)]
Carter HB, Pearson JD. PSA velocity for the diagnosis of early prostate cancer. A new concept.Urol Clin North Am. 1993;20:665-670.
[PubMed] [DOI][Cited in This Article: ]
Chun FK, de la Taille A, van Poppel H, Marberger M, Stenzl A, Mulders PF, Huland H, Abbou CC, Stillebroer AB, van Gils MP. Prostate cancer gene 3 (PCA3): development and internal validation of a novel biopsy nomogram.Eur Urol. 2009;56:659-667.
[PubMed] [DOI][Cited in This Article: ][Cited by in Crossref: 141][Cited by in F6Publishing: 147][Article Influence: 9.8][Reference Citation Analysis (0)]
Moussa AS, Jones JS, Yu C, Fareed K, Kattan MW. Development and validation of a nomogram for predicting a positive repeat prostate biopsy in patients with a previous negative biopsy session in the era of extended prostate sampling.BJU Int. 2010;106:1309-1314.
[PubMed] [DOI][Cited in This Article: ][Cited by in Crossref: 25][Cited by in F6Publishing: 27][Article Influence: 1.9][Reference Citation Analysis (0)]
Keetch DW, McMurtry JM, Smith DS, Andriole GL, Catalona WJ. Prostate specific antigen density versus prostate specific antigen slope as predictors of prostate cancer in men with initially negative prostatic biopsies.J Urol. 1996;156:428-431.
[PubMed] [DOI][Cited in This Article: ][Cited by in Crossref: 55][Cited by in F6Publishing: 46][Article Influence: 1.6][Reference Citation Analysis (0)]
