Jun Nakashima md, Department of Urology, Keio University School of Medicine, 35 Shinanomachi, Shinjuku, Tokyo, Japan. Email: email@example.com
Objectives: To investigate the clinical value of prostate specific antigen velocity (PSAV) in predicting the extraprostatic extension of clinically localized prostate cancer.
Methods: One hundred and three patients who underwent radical prostatectomy for clinically localized prostate cancer were included in the analysis. The correlation between preoperative parameters, including PSA-based parameters, clinical stage, and histological biopsy findings, and the pathological findings were analyzed. Logistic regression analysis was performed to identify a significant set of independent predictors for the local extent of the disease.
Results: Sixty-four (60.2%) patients had organ confined prostate cancer and 39 (39.8%) patients had extraprostatic cancer. The biopsy Gleason score, PSA, PSA density, PSA density of the transition zone, and PSAV were significantly higher in the patients with extraprostatic cancer than in those with organ confined cancer. Multivariate logistic regression analysis indicated that the biopsy Gleason score, endorectal magnetic resonance imaging findings, and PSAV were significant predictors of extraprostatic cancer (P < 0.01). Probability curves for extraprostatic cancer were generated using these three preoperative parameters.
Conclusions: The combination of PSAV, endorectal magnetic resonance imaging findings, and biopsy Gleason score can provide additional information for selecting appropriate candidates for radical prostatectomy.
Radical prostatectomy is most effective when the disease is organ confined at the time of surgery.1 However, a significant number of patients undergoing radical prostatectomy for clinically localized prostate cancer are found to have extraprostatic extension in the final pathological analysis.2–4 Extraprostatic extension of prostate cancer in pathological specimens has been clearly shown to be an unfavorable prognostic finding in patients undergoing radical prostatectomy.1 Thus, an accurate preoperative assessment of disease extent is essential for the selection of appropriate therapy in patients with prostate cancer. Unfortunately, current clinical methods used preoperatively to predict the disease extent are still of limited value.
Many investigators have tried to combine various preoperative features, such as PSA, biopsy Gleason score, and clinical stage to predict the final pathological stage of radical prostatectomy specimens.5 Partin et al. have reported that the combination of clinical stage from digital rectal examination (DRE), PSA, and Gleason score predicted the pathological stage more accurately than any indicator alone (Partin Tables).2–4 However, it is generally agreed that DRE is limited by a significant level of clinical understaging, and that the incidence of unsuspected periprostatic soft tissue invasion after prostatectomy in DRE-based staging series is high.6 Therefore, using other more accurate imaging studies might provide additional information to that reported by Partin et al.
PSA is a valuable tumor marker for detecting, staging, and monitoring prostate cancer.7 In addition to cancer detection, various PSA-based parameters have been shown to be valuable to some extent for the prediction of extraprostatic disease in radical prostatectomy.8–17 Combinations of PSA-based parameters and other staging methods are required for accurate prediction of the pathological stage. However, the predictive values of the combinations of the PSA-based parameters and other staging modalities have yet to be determined. It has been previously reported that the combination of PSA density (PSAD), endorectal MRI findings, and biopsy Gleason score can provide useful information concerning the pathological stage.18 In that study, PSAD was the most valuable predictor among the PSA-based parameters for detecting extraprostatic disease in patients with clinically localized prostate cancer. However PSA velocity (PSAV) was not included in the previous study in spite of PSAV having been widely watched after the report to be a strong and independent predictor of tumor stage,19 biochemical progression and death from prostate cancer.20 The present study was undertaken to investigate the clinical value of PSAV, in comparison with other parameters including PSA and PSAD, for the prediction of extraprostatic disease.
The present study consisted of 103 patients (mean age, 66.4 ± 5.3 years, range 51–74 years) who underwent radical prostatectomy and pelvic lymphadenectomy for clinically localized prostate cancer at our institution. The mean length of PSA follow up before surgery was 27.9 ± 24.1 (range, 6.3–106.2) months. Patients who had received endocrine therapy prior to radical prostatectomy and for whom there were no PSAV data were excluded. Prostate cancer was histologically confirmed preoperatively by transrectal ultrasound-guided needle biopsy using a 10-MHz endorectal transducer. The staging procedures included DRE, computed tomography, endorectal magnetic resonance imaging (MRI), and bone scanning. All endorectal MRI studies were performed more than 4 weeks after prostate biopsy to reduce the possibility of postbiopsy artifacts. Tumors which attached to the prostate capsule more than 1 cm in width on endorectal MRI and tumors with a localized bulge were defined as a subgroup of suspected microscopic extension, while tumors with a contact length of less than 1 cm were defined as a subgroup of no extension.18,21 Patients who had definite findings of extracapsular extension and/or seminal vesicle involvement on endorectal MRI were excluded from this study. The prostate volume was calculated based on a prolate ellipsoid and the length was measured using transrectal ultrasound. Serum samples were obtained before DRE. PSA concentrations were determined by an enzyme immunoassay (AIA-PACK PA; Tosoh Company, Foster City, CA). PSAD and the PSA density of the transition zone (PSATZD) were calculated by dividing the PSA value by prostate volume and transition zone volume (TZ volume). The surgical specimens were step sectioned at 4 mm intervals perpendicular to the long axis of the gland and each section examined. Whole mount preparations were used to assess extraprostatic cancer. Extraprostatic cancer was defined as seminal vesicle involvement, malignant cells outside the prostatic capsule, or lymph node metastasis. The three consecutive PSA data were recorded as the value at first visit to our office as PSA1, the closest value in time before diagnosis as PSA3, and the value at the mid-point in time between PSA1 and PSA3 as PSA2. Using these three consecutive PSA values, PSAV were calculated as shown below using the method that was previously reported by Carter and Pearson.22 All PSA data were measured at our institution.
The groups were compared using the Mann–Whitney U-test. The independence of fit of categorized data was analyzed with the χ2 test. Receiver operating characteristic (ROC) curves were plotted with the sensitivity (true-positive fraction) on the y-axis versus 1 minus the specificity (false-positive fraction) on the x-axis. Independent factors for the prediction of extraprostatic cancer were identified using stepwise logistic regression analysis. The predicted probability of extraprostatic cancer was estimated as P = 1/(1 + exp−k). Logistic regression gives a score (k), where k is a +b1 × 1 + b2 × 2 + . . . , which is a linear combination of the predictors (×1, ×2 . . . ) in the model. The model coefficients (a, b1, b2 . . . ) were chosen to optimize the model's ability to predict the probability of extraprostatic cancer. Differences were considered statistically significant at P < 0.05.
A total of 64 (60.2%) patients had organ confined cancer in the pathological specimen while 39 (39.8%) patients had extraprostatic cancer including 37 (94.9%) patients with only extracapsular extension, 2 (5.1%) with seminal vesicle invasion, and no lymph node metastasis. Preoperative parameters in both groups are summarized in Table 1. Significant differences were noted in prostate volume (P = 0.0170), TZ volume (P = 0.0072), Gleason score (P = 0.0285) in the biopsy specimens, preoperative PSA (P = 0.0041), PSAD (P = 0.0001), PSATZD (P = 0.0007), and PSAV (P = 0.0002) between patients with organ confined cancer and those with extraprostatic cancer.
Table 1. Comparison of preoperative parameters in patients with organ confined prostate cancer and those with extraprostatic cancer
Organ confined (mean ± SE)
Extraprostatic (mean ± SE)
PSA, prostate specific antigen; PSAD, PSA density; PSATZD, PSA density of the transition zone; PSAV, PSA velocity; TZ, transition zone.
No. of patients
5.719 ± 0.129
6.231 ± 0.206
P volume (cc)
39.317 ± 2.070
31.973 ± 1.859
TZ volume (cc)
19.579 ± 1.560
13.551 ± 1.166
8.375 ± 0.448
11.758 ± 1.284
0.239 ± 0.016
0.415 ± 0.051
0.603 ± 0.070
1.105 ± 0.144
0.962 ± 0.213
3.197 ± 0.655
Based on DRE findings, 82 (79.6%) patients were classified as stage T1c and 21 (20.4%) were classified as either stage T2a or T2b. Fifty-two patients with a T1c tumor (63.4%) and 12 patients with T2a or T2b tumor (57.1%) had pathologically organ confined cancer. No significant correlation was found between the DRE findings and the pathological findings.
On the other hands, endorectal MRI demonstrated no extension in 84 patients, while it demonstrated suspected microscopic extension in 19. Among the no extension on endorectal MRI subgroup, 71.4% (60/84) had pathologically organ confined cancer, while 78.9% (15/19) of the suspected microscopic extension subgroup had pathologically extraprostatic cancer. A significant correlation was found between the endorectal MRI findings and the pathological findings (P = 0.0001).
The area under the ROC curve (AUC) was calculated in order to estimate the predictive value of the preoperative parameters for the prediction of extraprostatic cancer. The AUC was 0.612 for PSA, 0.626 for prostate volume, 0.647 for TZ volume, 0.612 for PSA, 0.685 for PSAD, 0.702 for PSATZD, and 0.706 for PSAV. The P-value between PSAV and PSA, between PSAV and PSAD and between PSAV and PSATZD were 0.44, 0.86, and 0.97, respectively, which were not significant. Although the P-values were not significant among these AUC, PSAV had the largest AUC value among the PSA-based parameters.
Multivariate stepwise logistic regression analysis was performed to assess the significant independent preoperative parameters to predict extraprostatic cancer. The analysis indicated that biopsy Gleason score (≤6 vs 7≤), endorectal MRI findings (no extension vs suspected microscopic extension), and PSAV were significant predictors of the extraprostatic extension of prostate cancer (P = 0.0009, 0.0002, and <0.0001, respectively, Table 2).
Table 2. Multivariate stepwise logistic regression analysis to predict extraprostatic prostate cancer
MRI, magnetic resonance imaging; PSAV, prostate specific antigen velocity.
Gleason score (≤6 vs≥7)
Probability curves for extraprostatic cancer were generated using the three preoperative parameters (PSAV, endorectal MRI findings, and biopsy Gleason score). Figure 1 shows plots representing the probability of extraprostatic cancer by a combination of the preoperative parameters. For example, a patient with a PSAV of 2 ng/mL/year, no extraprostatic extension on endorectal MRI, and a biopsy Gleason score 7≤ has a 57.5% probability of pathological extraprostatic cancer (Fig. 1).
Although staging evaluation of prostate cancer with various preoperative examinations may increase the percentage of cases being correctly diagnosed as organ confined cancer, preoperative staging tends to understage a significant number of patients, which often results in additional treatment. Therefore, accurate prediction of the local extent of prostate cancer by means of preoperative variables is crucial with respect to making a decision concerning treatment strategy. Identification of the preoperative parameters predictive of the pathological stage at radical prostatectomy has become a topic of major concern. In the present study, patients with clinically localized prostate cancer who were preoperatively evaluated by endorectal MRI and had undergone radical prostatectomy were studied to determine the preoperative variables that were significantly associated with extraprostatic extension. The results demonstrated that preoperative PSAV, biopsy Gleason score, and endorectal MRI findings are the best combination of predictors for the extraprostatic extension of prostate cancer. Partin et al. contributed greatly to the clinical assessment of prostate cancer through their nomograms for predicting the pathological stage using preoperative PSA, biopsy Gleason score, and clinical stage using DRE.2–4 Various nomograms are currently available for predicting the final pathological stage and biochemical recurrence free survival based on preoperative parameters.5 Although these nomograms appear to be cost effective and good predictors of the final pathological stage, there is still room for improvement. The introduction of endorectal MRI has improved the ability to correctly define local tumor staging. However, it seems to have limitations with respect to the diagnosis of microscopic extraprostatic cancer.23,24 To overcome the limitations of endorectal MRI, other staging modalities, including PSA-based parameters and biopsy findings, should be combined with endorectal MRI staging. Although PSA is widely used for the early detection, staging, and monitoring of patients with prostate cancer, the use of PSA alone as a tumor marker is not sufficiently sensitive or specific for staging.2–4,7 In recent years, several PSA-based parameters, such as percent free PSA, PSAD, PSATZD, and PSA-ACT have been evaluated and compared to clarify whether they could be used as superior preoperative predictors of extracapsular tumor invasion.8–18 In a previous study, multivariate logistic regression analysis demonstrated that PSAD was a significant predictor of the postoperative pathological outcome among these PSA-based parameters, as were endorectal MRI findings and the biopsy Gleason score.18 However, PSAV was not included as one of the PSA-based parameters.
In the present study, we investigated the clinical value of PSAV with PSAD. Previous studies have reported that the rate of change of PSA, such as PSAV and PSA doubling time before the diagnosis of prostate cancer, can predict tumor stage, grade, and the time to recurrence after radical prostatectomy.19,22,25 Moreover, D'Amico et al.20 have recently reported that men whose PSA level increases by more than 2.0 ng per milliliter during the year before the diagnosis of prostate cancer may have a relatively high risk of death from prostate cancer despite undergoing radical prostatectomy. In the present study, PSA, PSAD, PSATZD, and PSAV were significantly higher in patients with extraprostatic cancer than in those with organ confined disease. In addition, the area under the ROC curve for PSAV was slightly larger than that for PSAD or PSATZD. Furthermore, multivariate stepwise logistic regression analysis demonstrated that PSAV was the best predictor of extraprostatic disease among these possible PSA-based parameters. Therefore, PSAV is considered to be equal or superior to PSAD or PSATZD for predicting the local extent of the disease. Although PSAV is not always available for all patients, the clinical value of PSAV is thought to be significant in patients who have had their PSA determined more than two times. In addition, since PSAV was retrospectively calculated, there is a limitation that the intervals of the PSA measurements are not uniform among respective PSAV calculation in this study.
In conclusion, the present study has demonstrated that preoperative PSAV, like endorectal MRI findings and biopsy Gleason score, can be a good predictor of extraprostatic disease in patients with clinically localized prostate cancer. Although it is anticipated that our result for predicting extraprostatic cancer will be validated by other investigators in the future, the combination of PSAV, endorectal MRI findings, and biopsy Gleason score may provide additional information that will assist in selecting appropriate candidates for radical prostatectomy.