The implementation of PSA in the everyday clinical practice has revolutionized prostate cancer screening leading to early detection of the disease and improved survival outcomes [
The optimal therapy for localized prostate cancer is still controversial with radical prostatectomy to be the main therapeutic option. The main tools used to estimate the preoperative risk for adverse pathological findings after surgery and the possibility for PSA recurrence after radical treatment are PSA, Gleason score, and clinical stage [
Although PSA density has a significant and established role in prostate cancer screening by increasing the diagnostic value of PSA [
The aim of the present study was to evaluate PSA density as a predictor of biochemical recurrence in patients undergoing a radical prostatectomy for clinically localized disease. Furthermore, a comparison was made between the latter, PSA, and Gleason summary to identify their predictive strength.
After we obtained an approval from the Ethics committee of our institution, we conducted a prospective study of 256 consecutive patients who underwent an open retropubic radical prostatectomy. All patients suffered from clinically localized prostate cancer, as it had been defined from the findings of preoperative digital rectal examination, transrectal ultrasound, and in some cases (29 patients) of magnetic resonance imaging, and they had >10 years of life expectance. An abdomen and pelvic computer tomography was asked from patients with PSA >20 ng/mL and/or Gleason score ≥7, and/or positive digital rectal examination while a bone scintigraphy was performed in cases with PSA >20 ng/mL, Gleason score >7 and/or clinical stage ≥T2c. The diagnosis of the disease was established by positive biopsy results. Patients with any clinical suspicion of locally advanced prostate cancer were excluded from the study. None of the patients had received neoadjuvant or adjuvant hormone therapy and/or radiotherapy. The study period was from February 2007 to April 2011.
All surgical procedures were made by a single surgeon. In all cases, a standard surgical procedure was followed as it has been described by Walsh and Donker [
The surgical specimen was examined of our institution pathologists and a histological report concerning the presence of organ confined disease, the presence of positive surgical margins, and the pathological grade and stage was obtained. Any extension of tumor outside of the prostatic capsule in the periprostatic fat was considered as advanced disease while the infiltration of the capsule without penetration was considered as localized disease. The 2009 TNM (tumour node metastasis) classification was used to define the pathological stage.
Follow-up protocol was consisted by visits every 3 months in order to evaluate the levels of serum PSA. A digital rectal examination was performed in all patients in every visit. Biochemical relapse was defined as 2 consecutive values >0.2 ng/mL, 2 weeks apart. We have to point that no adjuvant therapeutic mode was applied in all cases, even in the cases with advanced disease, until the time of biochemical failure. The median followup was 36 months.
The main endpoint of the study was to evaluate the role of PSA density in predicting biochemical recurrence in patients who were operated for clinically localized prostate cancer. Secondary endpoints were the correlation of PSA density with findings of pathological adverse events, in terms of positive surgical margins, extracapsular disease, seminal vesicles invasion and lymph nodes invasion, and to compare PSA density, PSA and Gleason score potential in predicting biochemical recurrence.
PSA density was calculated preoperatively during the biopsy procedure by using transrectal ultrasound by dividing the maximum preoperative PSA value and prostate volume. The latter was calculated based on the ellipse dimension theory formula [
Statistical analysis was performed by using SPSS version 17 (SPSS Inc., Chicago, IL, USA). Descriptive statistics are presented as the mean ± standard deviation and interquartile range for continuous variables and as the absolute and percent frequency for categorical variables. The numerical variables normality condition was studied by using Kolmogorov-Smirnov test. None of the variables had a normal distribution and thus Mann-Whitney
Twelve patients were excluded due to insufficient data (lost during the followup). Finally, 244 patients entered the analysis providing a margin error of the study estimated to 6.02%. The characteristics of the study cohort are summarized in Table
Study cohort characteristics.
Characteristics | Kolmogorov-Smirnov | |
---|---|---|
Number of patients, |
244 | |
Age (years) | ||
Mean ± std, IQR | 66.2 ± 6.49, 10 |
|
PSA (ng/mL) | ||
Mean ± std, IQR | 8.85 ± 3.66, 4.79 |
|
PSA density (ng/mL2) | ||
Mean ± std, IQR | 0.34 ± 0.38, 0.19 |
|
Preoperative Gleason score, |
||
|
87 (35.7) | |
|
118 (48.4) | |
|
39 (16.0) | |
Pathological Gleason score, |
||
|
86 (35.2) | |
|
112 (45.9) | |
|
46 (18.9) | |
Clinical stage, |
||
|
181 (74.2) | |
|
45 (18.4) | |
|
16 (6.6) | |
|
2 (0.8) | |
Pathological stage, |
||
|
25 (10.2) | |
|
8 (3.3) | |
|
142 (58.2) | |
|
47 (19.3) | |
|
22 (9.0) | |
Extracapsular disease, |
||
|
175 (71.7) | |
|
69 (28.3) | |
Surgical margins, |
||
|
167 (68.4) | |
|
77 (31.6) | |
Seminal vesicles invasion, |
||
|
222 (91.0) | |
|
22 (9.0) | |
Lymph nodes invasion, |
||
|
91 (37.3) | |
|
122 (50.0) | |
|
31 (12.7) | |
Biochemical failure, |
||
|
173 (70.9) | |
|
71 (29.1) |
Std: standard deviation; IQR: interquartile range.
Clinical and pathological characteristics of study cohort depending on the presence of biochemical recurrence.
No recurrence | Recurrence |
|
|
---|---|---|---|
Number of patients, (%) | 173 (70.9) | 71 (29.1) | |
Age (years), mean ± std, IQR | 66.3 ± 6.56, 9 | 65.9 ± 6.36, 11 | 0.527† |
PSA (ng/mL), mean ± std, IQR | 8.37 ± 3.23, 4.29 | 9.99 ± 4.35, 5.20 | 0.005†* |
PSA density (ng/mL2), mean ± std, IQR | 0.29 ± 0.25, 0.18 | 0.46 ± 0.57, 0.27 | 0.002†* |
Preoperative Gleason score, |
0.015‡* | ||
|
71 (41.0) | 16 (22.5) | |
|
79 (45.7) | 39 (54.9) | |
|
23 (13.3) | 16 (22.5) | |
Pathological Gleason score, |
0.423‡ | ||
|
65 (37.6) | 21 (29.6) | |
|
78 (45.1) | 34 (47.9) | |
|
30 (17.3) | 16 (22.5) | |
Clinical stage, |
0.201‡ | ||
|
134 (77.5) | 47 (66.2) | |
|
29 (16.8) | 16 (22.5) | |
|
8 (4.6) | 8 (11.3) | |
|
2 (1.2) | 0 (0) | |
Pathological stage, |
0.001‡* | ||
|
19 (11.0) | 6 (8.45) | |
|
7 (4.05) | 1 (1.41) | |
|
104 (60.1) | 38 (53.5) | |
|
36 (20.8) | 11 (15.5) | |
|
7 (4.05) | 15 (21.1) | |
Extracapsular disease, |
0.064‡ | ||
|
130 (75.1) | 45 (63.4) | |
|
43 (24.9) | 26 (36.6) | |
Surgical margins, |
0.021‡* | ||
|
126 (72.8) | 41 (57.7) | |
|
47 (27.2) | 30 (42.3) | |
Seminal vesicles invasion, |
<0.001‡* | ||
|
166 (96.0) | 56 (78.9) | |
|
7 (4.05) | 15 (21.1) | |
Lymph nodes invasion, |
<0.001‡* | ||
|
83 (100.0) | 39 (55.7) | |
|
0 (0) | 31 (44.3) |
Correlation of PSA density to adverse pathological findings and biochemical recurrence.
PSM | ECD | SVI | LNI | BCR | |
---|---|---|---|---|---|
Mean ± std | 0.32 ± 0.47 | 0.28 ± 0.45 | 0.09 ± 0.29 | 0.40 ± 0.45 | 0.29 ± 0.46 |
|
0.239 | <0.001* | 0.009* | 0.080 | 0.040* |
Statistics were made by using Pearson’s correlation (significance <0.05), *statistically significant, std: standard deviation, PSM: positive surgical margins, ECD: extracapsular disease, SVI: seminal vesicles invasion, LNI: lymph nodes invasion, and BCR: biochemical recurrence.
A univariate and multivariate analysis was conducted to explore and compare the ability of PSA, PSA density and Gleason score in prediction of biochemical failure (Table
Univariate and multivariate analysis of preoperative parameters regarding the predictive potential of biochemical recurrence.
|
Odds ratio | 95% CI | |
---|---|---|---|
Univariate analysis | |||
| |||
PSA | 0.002* | 1.126 | 1.045–1.213 |
PSA density | 0.005* | 3.408 | 1.454–7.987 |
Preoperative Gleason summary | 0.233 | 1.214 | 0.882–1.671 |
| |||
Multivariate analysis | |||
| |||
PSA | 0.006* | 1.112 | 1.030–1.200 |
PSA density | 0.009* | 2.861 | 1.293–6.330 |
CI: confidence interval, *statistically significant.
The incidence of prostate cancer has tremendously increased after the introduction of PSA in disease screening and a greater number of patients will be operated in order to radically excise the tumor. Given that a significant proportion of these patients will relapse after surgery, preoperative parameters predictive of tumor status and the potential for biochemical failure are required.
PSA density is mainly used to distinguish “benign” increases of PSA from “malignant” ones especially in cases belonging in the grey zone, meaning PSA rises between 4 and 10 ng/mL. Additionally, previous reports have revealed the valuable role of PSA density in estimating the possibility of adverse pathological findings after radical prostatectomy [
Kundu et al. [
In the present study we showed that PSA density significantly correlates with adverse pathological features after radical prostatectomy and with the possibility of PSA relapse, as well. In addition, it was found that PSA density was a strong preoperative predictor, though slightly inferior to PSA, for biochemical failure after radical therapy. Given that PSA relapse after radical surgery is seriously influencing the prognosis and quality of life due to additional therapies, preoperative estimation of patients with increased potential for such a failure is of great clinical importance since the suggested treatment protocols and patient information may be positively altered. Our findings suggest that PSA density deserves a place in preoperative nomograms since it may increase the prognostic ability of the current ones. In the continuous effort of urological community and researchers to better define prostate cancer nature and biological aggressiveness, PSA density seems to be significant for estimating the potential outcome after radical prostatectomy.
Prediction of biochemical failure after radical prostatectomy is of great importance since it represents an adverse factor for prognosis and survival. The most commonly used nomograms are based upon preoperative stage, PSA and Gleason score. Based on our results, PSA density is a significant predictor and its implementation in nomograms may increase their accuracy in estimating postoperative PSA relapse.
The authors declare that they have no conflict of interests.