Nomograms for prostate cancer diagnosis

AUA 2007 POSTER 1752, Anaheim, CA, May 23, 2007

NOMOGRAM WITH PSA KINETICS FOR PREVISION OF PROSTATE BIOPSY

Luigi Benecchi, Michele Potenzoni, Andrea Prati, Roberto Arnaudi, Carmelo Destro Pastizzaro, Antonio Savino, Nicoletta Uliano, Anna Maria Pieri, Daniel Martens, Domenico Potenzoni

Department of Urology, Fidenza Hospital (Parma), Italy

www.urologiaparma.com

urologiaparma@libero.it

INTRODUCTION

Since the discovery of prostate-specific antigen (PSA), detection and treatment of prostate cancer has changed dramatically.  The results of large screening programs have demostrated that diagnostic evalation of elevated serum PSA improves early detection and the likelihood of identifying organ-confined disease (1,2)

However, the conventional strategy for PSA screening, which calls for biopsies in all men with total PSA greater than 4 ng/ml, leads to many false positive results and is thus associated with a high cost in terms of unnecessary biopsies (4, 5).

A lot of methods have been used to enhance the specificity of PSA: PSA velocity, PSA density, PSA transition-zone density, Age -specific PSA level, ratio of free PSA to total PSA, level of alfa1-antichymotrypsin complex PSA (3) and artificial neural network (4).

Previous studies focused on risk factors for positive prostate biopsy. However, PSA ratio, velocity and density have not been tested or included together in these previous models and nomograms. We hypothesized that PSA ratio (%fPSA), PSA velocity and PSA density represent the foremost determinant of positive prostate biopsy.

MATERIALS AND METHODS

From January 2001 all men who underwent a repeated prostate biopsy with 6 or more cores entered the study. Men with PSA interference such as 5-alfa reductase therapy (finasteride or dutasteride) were excluded.

All patients were scheduled for transrectal sonography with biopsy because of abnormal digital examination findings and/or PSA levels of 4 micro/L or greater. 688 men entered the study.

Serum was obtained before any diagnostic procedure. Both total immunoreactive and free PSA were assayed using the chemiluminescent immunoassay Immulite (Diagnostic Products Corporation, Los Angeles , CA), in accordance with the manufacturer's instructions.

The PSA slope was obtained fitting the line of least squares (PSA versus time) for patients with 3 or more PSA assayed in 18 months or more before the last biopsy. Specifically, we fit the equation: y=a+bx to the data of each patient. Here y symbolises PSA and parameter a is the intercept. Parameter b is the slope and reflects the increase of PSA in one year and use the same unit (ng/ml/year) as PSA velocity (9) .

The factors we evaluated for the risk of a positive biopsy included age, DRE findings, total PSA level, free to total PSA ratio, PSA density, PSA slope, previous high grade PIN.

Multivariate logistic regression analysis, determined wich factors were indipendent predictors of prostate carcinoma in the model building set. Relative risk and 95% confidence intervals were also derived. A nomogram for a positive biopsy was developed from the final logistic model findings

 

RESULTS

We reported our results updated from that presented in the abstract.

prostate biopsy

benign

prostate cancer

n

688

458

230

%

100%

66.6%

33.4%

age

69.4 (49-90)

68.6 (49-86)

70.9 (51 - 90)

dre

35.7%

29%

50.4%*

PSA

8.8 (0.4-93.4)

8.45 (0.4 -43)

9.37 (2.12 - 93.4)*

percent free PSA

16,1 (2.2-65)

18 (3.8-65)

10.92 (2.2 - 44.5)*

PSA velocity

0.5 (-5-15)

0.4(-5-9)

1.1 (-3 - 15)*

PSA slope

0.5 (-7-15)

0.3 (-7 -7)

1.2 (-3 - 15)*

prostate volume

51 (10-330)

58 (20 - 330)

40 (10 - 150)*

PSA density

0.16 (0.01 -2.2)

0.14 (0.014 - 0.87)

0.23 (0.04 - 2.2)*

transition zone volume

33 (4-200)

35.7 (4 - 200)

24 (6 - 85)*

transition zone PSA density

0.26 (0.01 - 2.5)

0.23 (0.01 - 1.35)

0.45 (0.04 - 2.2)*

*p< 0.05 Mann-Whitney U test

Table 1 The descriptive characteristics of the entire cohort of 688 men.

In multivariate analyses, all the predictors except for age (p=0.7) were significant (p values<0.05). T he combined effect of PSA, %fPSA, DRE, prostate volume, PSA slope resulted in cumulative accuracy of 80.9%.

 

The nomogram consists of eight rows. The first row (points) is the point assignment for each variable. Rows 2 to 6 represent the variables included in the model. For an individual patient, each variable is assigned a point value (uppermost scale [points]) based on the clinical characteristics. To determine the point assignment , a vetical line is made between the appropriate variable and the points line. The assigned points for all six variables are summered, and the total is found in row 7 (Total Points) . Once the total is locaded in row 7, a vertical line is made between it and the corresponding value in the final row, row 8 (predicted prostate cancer probability).

 

DISCUSSION

Our nomogram utilize readily available clinical information and allow quick calculation This approach may allow identification of extremely low-risk individuals for whom the risks associated with prostate biopsy are judged to outweigh the benefits. Conversely, our nomogram may allow identification of men at sufficient risk of prostate cancer that they urologists elect to proceed with prostate biopsy even though clinical "guesstimates" would suggest that they are at low risk. The nomogram provides risk estimates that will have to be judged on an individual basis. A man with a negative DRE, PSA 5 ng/ml, PSA ratio 35%, PSA density 1, PSA slope 0 might be considered to be al relative low risk. Our nomogram suggests that he has a 29% risk of having prostate cancer. Should he undergo prostate biopsy? Given this scenario, some will judge that a 29% risk of prostate cancer justifies prostate biopsy, others will not.

We think that prostate biopsy is a surgical procedure with intrinsic risk (schok, hemorrhage, infection..) so should done if there is a 33-50% or more of risk of positive prostate cancer. If not is better to repeat after 6 or 12 months the nomogram evaluation.

CONCLUSION

We successfully developed an accurate model to predict the outcome of prostate biopsy. Addition of PSA, %fPSA, DRE, PSA density and PSA velocity sharply improves accuracy of our model.

Poster (AM07-0245) abstract code: 394 EAU Berlin 2007

 

Repeat prostate biopsy and PSA kinetic in men with previous 3 or more PSA measurements

Luigi Benecchi, Michele Potenzoni, Andrea Prati, Roberto Arnaudi, Carmelo Destro Pastizzaro, Antonio Savino, Nicoletta Uliano, Anna Maria Pieri, Daniel Martens, Domenico Potenzoni

Department of Urology Fidenza Hospital (Parma) Italy

www.urologiaparma.com

Introduction:

Previous studies focused on risk factors for prostate cancer on repeat biopsy. However, PSA ratio, velocity and transition zone density have not been tested or included together in these previous models and nomograms. We hypothesized that PSA ratio, velocity and transition zone density represent the foremost determinant of positive repeat prostate biopsy.

The aim of our study is to develop a predictive model that incorporates clinical data and PSA kinetic from general practice to detect prostate cancer in patients with a previous negative prostate biopsy.

The significant independent covariates for detecting carcinoma in the present study are age, DRE findings, total PSA level, free to total PSA ratio, PSA density, PSA slope, previous high grade PIN.

This predictive model might aid physicians in electing patients to repeat prostate biopsy.

Methods:

From January 2001, all men who underwent a repeated prostate biopsy with 12 or more cores entered the study.

Men with PSA interference such as 5-alfa reductase therapy (finasteride or dutasteride) were excluded.

All patients were scheduled for transrectal sonography with biopsy because of abnormal digital examination findings and/or PSA levels of 4 micro/L or greater.

The PSA slope was obtained fitting the line of least squares (PSA versus time) for patients with 3 or more PSA assayed in 18 months or more before the last biopsy.

Specifically, we fit the equation: y=a+bx to the data of each patient. Here y symbolises PSA and parameter a is the intercept. Parameter b is the slope and reflects the increase of PSA in one year and use the same unit (ng/ml/year) as PSA velocity

688 consecutive patients (mean age 68, range 50 to 84) underwent a 12 or more cores prostate biopsy

A repeat prostate biopsy was conducted in 419 men

A total of 130 cancers (31%) were found at the ultrasound guided repeat prostate biopsies.

 

Results:

The factors we evaluated for the risk of a positive biopsy included age, DRE findings, total PSA level, free to total PSA ratio, PSA density, PSA slope, previous high grade PIN.

Multivariate logistic regression analysis determined which factors were independent predictors of prostate carcinoma in the model building set.

Of the 419 patients reviewed 246 presented all clinical data for multivariate logistic regression.

Accuracy 80.9%

Conclusions:

We successfully developed an accurate model to predict the outcome of repeat prostate biopsy. Addition of %fPSA, DRE, PSA density, PSA slope and history of HGPION velocity sharply improves accuracy of our model.