Is interval from an initial biopsy a significant predictor of prostate cancer at repeat biopsies?

Authors


Akinori Satoh md, Department of Urology, Saiseikai Central Hospital, 1-4-17 Mita, Minato-ku, Tokyo 108-0073, Japan. Email: zenpaku@ndmc.ac.jp

Abstract

Background:  Currently, there are no clear criteria for indicating repeat biopsies in patients with negative results at an initial biopsy of the prostate. The aim of the present study is to determine the clinical and pathological parameters which predict prostate cancer at repeat biopsies with special attention to the interval between biopsies in addition to prostate specific antigen (PSA) and its derivatives.

Methods:  We reviewed 100 patients who underwent an initial biopsy that proved negative for prostate cancer and required repeat biopsies between November 1996 and November 2003. Clinical parameters such as age, PSA and its derivatives, interval between biopsies, number of cores taken and initial biopsy histology were analyzed.

Results:  In total, 31 patients (31.0%) were found to have prostate cancer, 18 (25.7%) of 70 patients by the second biopsy and 13 (46.4%) of 28 patients by the third biopsy. Two patients underwent the fourth biopsy, which revealed no prostate cancer. The patients with a positive biopsy had a significantly longer interval between the biopsies than the patients with a negative biopsy (P = 0.0036). Furthermore, in both univariate and multivariate logistic regression analysis, only the interval between the biopsies proved to be an independent predictor of positive results at repeat biopsies (P = 0.0094 and 0.0019).

Conclusions:  Only the biopsy interval was a significant predictor of prostate cancer at repeat biopsies in both univariate and multivariate analysis.

Introduction

Currently, there are no clear criteria for indicating repeat biopsies in patients with negative results at an initial biopsy of the prostate, which annoys both patients and urologists in a clinical setting. Patients with a negative biopsy often go on to many repeat biopsies which may be in part unnecessary. Prostate-specific antigen (PSA) has been widely utilized in routine screening for prostate cancer. Although numerous previous studies have attempted to identify predictors of positive biopsies focusing on PSA and PSA derivatives, their results have been controversial. The aim of this study is to determine the clinical and pathological parameters which predict prostate cancer at repeat biopsies with special attention to the interval between biopsies in addition to PSA and its derivatives.

Methods

We reviewed our practice records of 100 patients who underwent an initial biopsy that proved negative for prostate cancer and were indicated repeat biopsies between November 1996 and November 2003.

All the patients underwent transrectal ultrasound (TRUS)-guided biopsy. The whole prostate and transition zone were measured in three dimensions, and volume was estimated with the prolate ellipsoid formula. PSA determinations were assessed by Markit-M PA assay (Dainippon Pharmaceutical, Osaka, Japan) before December 1998 and by E-test TOSOH-II PA (TOSOH, Tokyo, Japan) after January 1999. The PSA values measured by Markit-M PA were converted using the formula (TOSOH) = 2.775 × (Markit) + 2.117 (linear regression equation, r = 0.981) to match with the values measured by TOSOH-II PA.

Variables studied in each case were as follows: age, prostatic volume, PSA, PSA density (PSAD), PSA transition zone density (PSA-TZD), PSA velocity (PSAV), coefficient of variation (CV) of PSA, interval between biopsies, number of cores taken and prior biopsy histology. PSAD and PSA-TZD were calculated by dividing total PSA by whole prostatic volume and transition zone volume, respectively. PSAV was calculated with the least square regression analysis. CV of PSA was calculated by dividing standard deviation by mean of PSA values.

A descriptive study of all the variables included in the study was conducted. Comparisons of the parameters between the positive result group and the negative result group were performed using the Mann–Whitney U-test. Logistic regression analyses were used for identification of predictors of the positive result at repeat biopsy. Probability values of less than 0.05 were considered to be of statistical significance.

Results

One hundred patients with an initially negative biopsy underwent repeat biopsies. The age of the patients ranged from 51–82 years with a median of 67.5 years (Fig. 1). Of those, 30 patients (30.0%) underwent a third biopsy and 2 patients (2.0%) a fourth biopsy. PSA determinations were performed 1–29 times with a median of 4 times between the biopsies. The median PSA at the initial biopsy and the last biopsy were 7.9 ng/mL and 10.4 ng/mL, respectively. The median PSA velocity was 1.43 ng/mL per year and the median CV of PSA 16.8%. PSAD and PSA-TZD were calculable in 70 and 66 of the patients with a median of 0.21 ng/mL per mL and 0.41 ng/mL per mL. The interval between the initial biopsy and the last biopsy ranged from 2.9 to 85.3 months with a median of 17.3 months. The median number of biopsy cores taken was seven for the initial biopsy and eight for the last biopsy. Twenty patients (20.0%) were found to have atypical cells in one of the previous biopsies before the last biopsy.

Figure 1.

Age distributions of patients.

Of the 100 patients, 31 (31.0%) had prostate cancer detected at repeat biopsies; 18 of 100 patients (18.0%) at the second biopsy and 13 of 30 patients (43.3%) at the third biopsy. No cancer was detected at the fourth biopsy. Table 1 shows the parameters compared between the positive result group and the negative result group. The patients in the positive result group had a statistically longer interval between the initial biopsy and the last biopsy than those in the negative result group (P = 0.004). Figure 2 demonstrates the distribution of the interval between the biopsies. However, there were no statistically significant differences in other parameters between the groups.

Table 1.  Clinical and pathological parameters compared between negative result group and positive result group
ParameterNegative result group (n = 69)Positive result group (n = 31)P-value
  1. Data in parentheses are ranges (minimum to maximum), unless noted otherwise. †Prostate volume and PSAD were not obtained in 23 cases. ‡Prostate volume and PSAD were not obtained in seven cases. §Transition zone volume and PSA-TZD were not obtained in 25 cases. ¶Transition zone volume and PSA-TZD were not obtained in nine cases. CV, coefficient of variation; PSA, prostate-specific antigen; PSAD, PSA density; PSA-TZD, PSA transition zone density; PSAV, PSA velocity.

Median age67 (51–82)68 (59–82)0.07
Median prostate volume (mL)49.9 (12.1–141.7)43.8 (18.4–79.2)0.09
Median transition zone volume (mL)23.2 (2.7–81.0)§18.3 (7.2–47.7)0.12
Median PSA at initial biopsy (ng/mL) 7.9 (2.7–32.6) 8.0 (2.9–119.0)0.93
Median PSA at last biopsy (ng/mL)10.3 (4.2–58.4)10.5 (4.5–92.4)0.60
Median PSAD (ng/mL per mL) 0.20 (0.07–1.89) 0.24 (0.11–2.00)0.05
Median PSA-TZD (ng/mL per mL) 0.39 (0.11–2.64)§ 0.53 (0.17–4.72)0.10
Median PSAV (ng/mL per year) 1.78 (−21.4–24.1) 1.15 (−18.5–13.0)0.64
Median CV of PSA (%)17.3 (2.0–65.4)14.9 (4.6–43.7)0.50
Median interval between biopsies (month)14.5 (2.9–77.3)26.3 (4.3–85.3)0.004
Median numbers of core 8 (6–13)10 (6–13)0.24
Atypical cell at previous biopsies (%)16 (23) 4 (13)0.24
Figure 2.

Box and whisker plot of interval between initial biopsy and last biopsy in the negative result group and positive result group.

Univariate and multivariate logistic regression analysis was performed for identification of predictors of prostate cancer at repeat biopsies (Table 2). Only the interval between the initial biopsy and the last biopsy was a significant predictor of prostate cancer in both univariate and multivariate analysis. PSA and its derivatives were not associated with prostate cancer detected at the repeat biopsies.

Table 2.  Univariate and multivariate logistic regression analyses for identification of predictors of prostate cancer at repeat biopsy
VariableUnivariateMultivariate
Hazards ratioP-value95% CIHazards ratioP-value95% CI
  • Prostate volume and PSAD were not obtained in 30 cases.

  • ‡Transition zone volume and PSA-TZD were not obtained in 25 cases. Atypical cell, presence of atypical cell at previous biopsy; Interval, interval between initial biopsy and last biopsy; PSA, prostate-specific antigen; PSAD, PSA density; PSA-TZD, PSA TZ density; PSAV, PSA velocity; TZ, transition zone.

Age1.0610.06460.996–1.1301.0340.5491 0.927–1.152
Prostate volume0.9780.06970.954–1.0021.0410.5214 0.920–1.178
TZ volume0.9740.16360.939–1.0110.9380.3638 0.818–1.076
PSA at initial biopsy1.0240.18360.989–1.0610.9490.6350 0.764–1.178
PSA at last biopsy1.0220.14520.992–1.0530.9230.6546 0.651–1.310
PSAD3.6210.10530.763–17.187120920.24852.820 × 10−4–5.186 × 1013
PSA-TZD2.0560.06730.950–4.4490.1490.4250 0.001–16.006
PSAV0.9550.23550.885–1.0300.9470.5882 0.777–1.153
%CV of PSA0.9840.38130.948–1.0210.9810.6044 0.912–1.055
Interval1.0320.00941.008–1.0581.0680.0055 1.020–1.119
Numbers of core1.2610.09560.960–1.6551.3320.3845 0.698–2.541
Atypical cell0.4910.24090.149–1.6130.2970.2428 0.039–2.277

Discussion

Repeat biopsies of the prostate are often indicated in patients with a negative initial biopsy because of persistent abnormal PSA and/or digital rectal examination findings. However, to date, no established criteria seem to exist indicating repeat biopsies in patients with negative result at an initial biopsy.

Several studies reported that 20–40% of patients with an initially negative biopsy will be found to have cancer at a repeat biopsy.1–3 In our study, the detection rate was 31.0% at repeat biopsies, which concurs with the previous reports.

Considering the significant initial false-negative biopsy rate, it is indispensable to determine predictors of prostate cancer at repeat biopsies to avoid numerous unnecessary biopsies. Many clinical and pathological parameters have been studied as predictors of cancer at repeat biopsies. Among PSA and its derivatives, percent free PSA and PSAD of the transition zone were found to be the most accurate predictors.4 Kronz et al. demonstrated that the presence of high-grade prostatic intraepithelial neoplasia (PIN) at an initial biopsy is a predictor of cancer at a repeat biopsy.5 Some investigators reported that detection rate of cancer at repeat biopsies increases as the prostate volume increases.6,7 Percent free PSA was not evaluated in this study as this test was not available in all the cases. PSA derivatives including PSAD and PSA-TZD were not the predictor of cancer at repeat biopsies in the present study. But this may not be conclusive because the record of prostatic volume was obtained in the relatively small numbers of the patients (66/100). The presence of atypical gland was not a predictor of cancer. This would be because our findings of atypical gland did not include high-grade PIN and is compatible with the report that the presence of low-grade PIN has little clinical significance.8

The interval between biopsies is of significant interest. Urologists are often faced with the dilemma that too many biopsies in a short period of time would not be tolerated by patients both physically and mentally, whereas a too long course observation without repeat biopsies may lead to progression of cancer. In the present study, the patients in the positive result group had a statistically longer interval between the initial biopsy and the last biopsy than those in the negative result group (P = 0.004). Moreover, both univariate and multivariate analysis demonstrated that only the interval between the biopsies was the significant predictor of prostate cancer at repeat biopsies (P = 0.0094 and P = 0.0055). Furthermore, the cancer detection rate of subsequent biopsies was higher; 18.0% at the second biopsy and 43.3% at the third biopsy. According to the analysis of patients distribution of these two groups, this difference in cancer detection rate would be due to difference of the biopsy interval between the groups (median interval, 13.1 vs 27.6 months, P = 0.0004). As with the fourth biopsy, the detection rate (0%) would be unreliable, as only two cases were evaluated. It is rather difficult to compare our results with previous reports because there are few existing studies that specifically analyse this issue. In general, the detection rate of subsequent biopsies is lower. Lopez-Corona et al. showed that the rate decreased from 19.5% at the second biopsy to 13.5% after five or more biopsy sessions.9 Roehl et al. reported that the rate decreased from 30% at the first biopsy to 7% after six biopsy sessions.10 Our findings seemingly do not correspond with these reports. However, many previous reports mainly focused on the number of repeat biopsies and the interval between the first biopsy and the last biopsy were not discussed. Although the number of biopsies performed is same, it is possible that the interval between the first biopsy and the last biopsy is considerably different. Few studies to date performed multivariate analysis using the number of biopsies and the interval between the biopsies as covariates. Lopez-Corona et al.9 incorporated biopsy interval in their nomogram for predicting a positive repeat biopsy. But they did not show that biopsy interval was a statistically significant predictor in their analysis. Therefore, our finding that cancer detection rate of subsequent biopsy increases can not directly be compared with the findings of the previous reports.

The reason for increasing cancer detection rate at subsequent biopsies is unknown from this study. One possible mechanism is the increase in tumor volume as time elapses. Further study must be done to prove the hypothesis, for example, tumor volumetry using radical prostatectomy specimens.

To our knowledge, our study is the first to show that biopsy interval is a significant predictor of positive result at repeat biopsies. Unfortunately, because of the lack of enough data regarding prostatic volume and free PSA, we could not fully assess the significance of the known most important parameters such as percent free PSA and PSA-TZD for prediction of cancer at repeat biopsies. From our results, however, too early repeat biopsy is not recommended.

Conclusions

The patients with a positive biopsy had a significantly longer interval between the biopsies than the patients with a negative biopsy. Only the biopsy interval was a significant predictor of prostate cancer at repeat biopsies in both univariate and multivariate analysis.

Ancillary