Direct comparison between transrectal and transperineal extended prostate biopsy for the detection of cancer

Authors


Satoru Kawakami md phd, Department of Urology, Graduate School, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-Ku, Tokyo 113-8519, Japan. Email: s-kawakami@tmd.ac.jp

Abstract

Aim:  To establish whether extended transrectal (TR) and extended transperineal (TP) biopsies are equivalent in detecting prostate cancer.

Methods:  Due to an elevated prostate-specific antigen (PSA) greater than 2.5 ng/mL or abnormal digital rectal examination findings, 783 men underwent a transrectal ultrasound-guided three-dimensional 26-core biopsy, a combination of TR 12-core and TP 14-core biopsies. Using recursive partitioning, the best combination of sampling sites that gave the highest cancer detection rate at a given number of biopsy cores was selected either with a TR or a TP approach. The cancer detection rate and characteristics of detected cancers were compared between the TP 14-core and the TR 12-core biopsies and between selected subset biopsy schemes.

Results:  Prostate cancer was detected in 283 of the 783 men (36%). There was no statistical difference in cancer detection rate or in the characteristics of detected cancers between TP 14-core and TR 12-core biopsies. As far as the best combination of sampling sites was selected, there was no statistical difference in cancer detection rates or in the characteristics of detected cancers between the TP and the TR subset biopsy schemes up to 12 cores. TP and TR biopsies performed equally, regardless of a history of negative biopsy, a digital rectal examination finding, the PSA level or the prostate volume.

Conclusions:  We demonstrated for the first time that extended TP biopsy is as effective as its TR counterpart in detecting cancer and the characteristics of detected cancers, as far as sampling sites are selected to maximize the cancer detection rate.

Introduction

Despite a global trend from conventional sextant to extended biopsy,1,2 the optimal biopsy strategy for prostate cancer detection still remains to be defined. Even the number of biopsy cores required for the optimal detection of prostate cancer in the transrectal (TR) approach is controversial.3 When compared to TR biopsy, studies on extended transperineal (TP) biopsy have been relatively limited and there has been no consensus on standard extended TP biopsy protocol.4–13 The lack of standardization at the moment is related to both TP and TR approaches.

We have demonstrated that, as far as the cancer detection rate is concerned, the three-dimensional combination of TP and TR approaches outperforms the single TP or TR approach at a given number of biopsy cores either in the initial14 or in the repeated biopsy setting.15 These results suggest that it might be better to take both the TP and the TR approaches into consideration in the exploration of the optimal biopsy protocol. To our best knowledge, whether the TP biopsy performs equally with the TR biopsy has not been fully elucidated to date. In the current study, we aimed to clarify whether the extended TP and the extended TR biopsies are equivalent in detecting prostate cancer by analysing the result of the three-dimensional 26-core (3D26) systematic biopsy in which a direct comparison of both approaches is possible in an identical cohort.

Methods

Patient population

The current study cohort was 783 Japanese men who agreed to undergo the 3D26 biopsy at our institutions between June 2001 and May 2006. Biopsy criteria included an elevated prostate-specific antigen (PSA) level above 2.5 ng/mL and/or abnormal digital rectal examination (DRE) findings. Patients were excluded if they presented with a previous history of prostate cancer, acute prostatitis or proven urinary tract infection. Men with PSA levels greater than 40 ng/mL and palpable tumors were also excluded. Written informed consent was obtained from all the men before the procedure. Before biopsy, all men underwent an evaluation of PSA (Tandem R or AxSYM), urinalysis and DRE.

Three-dimensional 26-core biopsy

As shown in Figure 1, the 3D26 biopsy protocol was designed to achieve adequate sampling of the entire peripheral zone by combining a TP 14 and a TR 12 sampling sites.14,15 In the TP 14-core biopsy, 12 cores were taken from the peripheral zone (A1, A2, P1, P2, AL and PL sites) and 2 from the transition zone (TZ) as reported elsewhere.7 The TR 12-core biopsy is composed of the conventional parasagittal sextant 6-core (pa, pm and pb sites) biopsy16 and far-lateral sextant 6-core (la, lm and lb sites) biopsy as described previously.1,14,15 The biopsy was performed using an 18-gauge biopsy needle driven by an automatic Magnum Biopsy gun (C R Bard, Inc., Covington, GA) under local, spinal or general anesthesia in 3 (0.4%), 541 (69%), and 239 (31%) men, respectively. With the patient in the lithotomy position, the prostate gland was imaged with an ultrasound machine (ALOKA SSD-2000, Aloka, Tokyo, Japan) equipped with both a biplanar 5/7.5 MHz transrectal probe (ALOKA UST-672) and an end-fire 5 MHz transrectal probe (ALOKA UST-670P-5). Prostate volume was calculated using the prolate ellipsoid formula (width × length × height × π/6). Patients were discharged the following morning after confirmation of hemostasis and smooth urination without pyrexia. Biopsy-related morbidity was evaluated as described elsewhere.7 Any complications requiring prolonged hospitalization or re-hospitalization during the four-week postbiopsy period were considered major. All other complications were considered minor. Biopsy tissue cores were processed and evaluated as described previously.7,14,15

Figure 1.

Transverse, sagittal and coronal projections of the three-dimensional 26-core biopsy, a combination of (A) transperineal 14-core and (B) transrectal 12-core biopsies. The site-specific cancer positive rates (%) were superimposed on the right side of the panels.

Data analysis

Cancer detection rates were compared between the TP 14-core and the TR 12-core biopsy schemes in the entire study cohort as well as in patient subgroups divided by the biopsy setting (initial vs repeat); DRE findings (normal vs abnormal); PSA (≤10 ng/mL vs >10 ng/mL); and prostate volume (≤37 mL vs >37 mL). Cut-off value for prostate volume was based on the median value of the current cohort.

Cancer characteristics including the biopsy Gleason score, percent positive core (percentage of cancer positive cores in the 26 cores) and maximum cancer length (the longest cancer length in biopsy tissue core in the patient) were compared between cancers with positive cores in the TP 14 sampling sites (TP + cancers) and those with positive cores in the TR 12 sampling sites (TR + cancers) as well as between cancers with positive cores solely in the TP 14 sampling sites (TP + TR- cancers) and those with positive cores solely in the TR 12 sampling sites (TP-TR + cancers).

We also compared the cancer detection rates of the subset biopsy schemes selected from the TP 14 sampling sites or the TR 12 sampling sites in the entire study cohort as well as in each patient subgroup. Using recursive partitioning, we selected the best combination of sampling sites providing the highest cancer detection rate at a given number of cores (2–12 cores) either with the TP or the TR approach.17 At first, the sampling site with the highest cancer detection rate was selected as a 2-core biopsy scheme in each approach. The next sampling site was recruited to achieve the highest detection rate at a 4-core biopsy setting. This step was repeated up to 12-core in the TR approach or up to 14-core in the TP approach. We considered the cancer positive rate of the 3D26 biopsy as 100%. Continuous variables were expressed as the median (interquartile range). Association was evaluated based on Fisher's exact test for nominal variables and a one-way analysis of variance (anova) for continuous variables. All calculated P-values were 2-sided and P < 0.05 was considered statistically significant. All analyses were performed using JMP version 6.0 (SAS Institute Inc. Cary, NC).

Results

Performance of the 3D26 biopsy

The median (interquartile range; range) values of age, PSA and prostate volume of the 783 men were 66 (61–71; 40–81) years, 6.6 (4.8–10.0; 0.9–52.8) ng/mL and 37 (28–50; 10–122) mL, respectively. The biopsy setting was initial in 518 (66%) and repeat in 265 (34%) men. Of the 783 men, 131 (17%) had abnormal DRE findings. Using the 3D26 biopsy, adenocarcinoma was detected in 283 (36.1%) of the 783 men. The biopsy Gleason score was 5–6 in 116 (41%) cancers, 7 in 137 (48%) cancers and 8–10 in 30 (11%) cancers. Sampling site-specific cancer positive rates were superimposed on the biopsy scheme in Figure 1.

Minor complications were observed in 21% of the subjects, gross hematuria being most prevalent (18%), followed by minor rectal bleeding in 11%. Major complications were far less frequent, consisting of acute bacterial prostatitis in four patients (0.5%) and rectal bleeding in three (0.4%), all managed conservatively. The complication rate did not differ according to the method of anesthesia.

Comparison between the TP 14-core and the TR 12-core biopsies

As shown in Table 1, among the 283 cancers detected by the 3D26 biopsy, 243 cancers had positive core in the TP 14 sampling sites (TP + cancers) and 231 had positive core in the TR 12 sampling sites (TR + cancers). There was no statistical difference in cancer detectability between the TP 14-core biopsy (86%; 243/783) and the TR 12-core biopsy (82%; 231/783) (P = 0.51). As we reported in the preliminary studies [14, 15], cancer detection rates of the TP 14-core biopsy and the TR 12-core biopsy were significantly lower than that of the 3D26 biopsy (P = 0.032 and P = 0.005, respectively). Among the 283 cancers, 40 (14%) cancers had positive cores solely in the TR 12 sampling sites (TP-TR + cancers), 52 (19%) solely in the TP 14 sampling sites (TP + TR- cancers) and the remaining 191 (67%) in both TP and TR sites (TP + TR + cancers). Characteristics of cancers did not differ significantly between 243 TP + cancers and 231 TR + cancers or between 52 TP + TR- cancers and 40 TP-TR + cancers (Table 2).

Table 1.  Number of men with a positive biopsy result in the transperineal 14 and transrectal 12 sampling sites
 Transperineal 14 sites
Cancer positiveCancer negativeTotal
Transrectal 12 sites
 Cancer positiveTP + TR + 191TP − TR + 40TR + 231
 Cancer negativeTP + TR − 52TP − TR − 500TR − 552
 TotalTP + 243TP − 540Total 783
Table 2.  Characteristics of patients and cancers according to the results of the three-dimensional 26-core biopsy
Transperineal 14 sites
Transrectal 12 sites
TP +TR +P-valueTP + TR –TP − TR +P-value
Positive
Any
Any
Positive
 Positive
Negative
Negative Positive 
  • Findings on the transperineal 14-core biopsy in TP + and TP + TR- groups and those on the transrectal 12-core biopsy in TR + and TP-TR + groups. Values are expressed as median (interquartile range) or number of patients. DRE, digital rectal examination; PSA, prostate-specific antigen.

Baseline characteristics
 Number of patients243231 5240 
 Age [year]68 (64–72)68 (64–73)0.8968 (63–72)68 (62–73)0.81
 Biopsy setting (initial/repeat)169/74156/750.8734/1821/190.28
 DRE (normal/abnormal)176/67166/650.8743/933/70.98
 PSA [ng/mL]7.6 (5.3–12)7.7 (5.3–12)0.787.2 (4.9–11)7.5 (4.6–10)0.44
 Prostate volume [mL]29 (22–40)29 (23–42)0.8131 (24–42)36 (30–51)0.06
Cancer characteristics
 Biopsy Gleason score (5–6/7/8–10)96/119/28100/106/250.7030/18/429/11/00.12
 % positive cores14 (7–29)17 (8–33)0.478 (4–8)4 (4–8)0.06
 Maximum cancer length [mm]5 (3–8)5 (3–8)0.862 (1–3)2 (1–4)0.90

Comparison between selected subset biopsy schemes

Next we compared the cancer detectability of each subset biopsy scheme selected from the TP 14 or from the TR 12 sampling sites at 2–12-core biopsy settings. Figure 2 shows the best combinations of either the TP or the TR approach determined by the recursive partitioning method. We observed that the best combinations of sampling sites are not identical to the order of site-specific cancer detection rates shown in Figure 1. As far as the best combination of sampling sites was selected, there was no statistical difference in cancer detection rates between TR and TP approaches up to 12-core biopsy setting (Fig. 2). For example, TP 12-core biopsy, a set of TP sampling sites omitting the P1 site from the TP 14-core biopsy scheme, is equipotent to the TR 12-core biopsy scheme. There was no significant difference in the characteristics of cancers detected by the selected TP and the selected TR biopsies at 2–12-core biopsy settings (data not shown).

Figure 2.

Maximum cancer detection rate achieved by the best combination of sampling sites in transrectal biopsy (black circles) and transperineal biopsy (colored circles) according to the number of biopsy cores. The order in which transperineal or transrectal sampling sites are recruited into the best combination was shown in the upper and lower panels, respectively.

Patient subgroup analyses

The cancer detection rates of the 3D26 biopsy in patient subgroups are compared in Table 3. Cancer detection rates were higher in men with an abnormal DRE, a PSA of ≥10 ng/mL or a prostate volume <37 mL, than in men with a normal DRE, a PSA of <10 ng/mL or a prostate volume of ≥37 mL, respectively. The cancer detection rate on the initial biopsy setting was not different from that on the repeat biopsy setting. The cancer detection rates of the TP 14-core biopsy and the TR 12-core biopsy as well as selected subset biopsy schemes of the TP and the TR approaches were compared in each patient subgroup. As shown in Figure 3, the TP and the TR approaches performed equally, irrespective of the biopsy setting, the DRE finding, the PSA level or the prostate volume. In each patient subgroup, there was no significant difference in the characteristics of cancers between those detected by the selected TP and those detected by the selected TR biopsies at 2–12-core biopsy settings (data not shown).

Table 3.  Cancer detection rates of the three-dimensional 26-core biopsy according to patient subgroups
VariablesCategories% cancer detectionP-value
  1. DRE, digital rectal examination; PSA, prostate-specific antigen.

Entire cohort 36.1 (283/783)
Biopsy settingInitial36.7 (190/518)0.6946
Repeat35.1 (93/265) 
DRENormal32.2 (210/652)<0.0001
Abnormal55.7 (73/131) 
PSA≤10 ng/mL32.4 (194/599)<0.0001
>10 ng/mL48.4 (89/184) 
Prostate volume≥37 mL23.1 (89/385)<0.0001
<37 mL48.7 (194/398) 
Figure 3.

Comparison of cancer detection rate of the best combination of sampling sites in transrectal biopsy (black circles) and transperineal biopsy (colored circles) according to biopsy setting (A), DRE finding (B), PSA (C) and prostate volume (D).

Discussion

In the extended biopsy schemes, the present study demonstrated for the first time that the TP approach is equipotent to the TR approach in detecting cancer as well as the characteristics of the cancers detected. In the published reports only one study has been reported comparing TR and TP approaches directly.18 Although they suggested a superiority of the TP 6-core biopsy to the TR 6-core biopsy in a relatively small study cohort, we found an equipotency of the TP 6-core and the TR 6-core biopsies in the current study cohort. This discrepancy may be due to different patient backgrounds, a different study size and, most importantly, our selection of the best sampling sites from the 26-core biopsy scheme.

The current study suggests how to choose the sampling sites to maximize the cancer detection rate with a minimum number of biopsy cores using either the TP or the TR approach as has been reported previously.14,15 The importance of tissue sampling from the anterior portion of the prostatic apex (the A1 and A2 sites in the TP biopsy scheme in Fig. 1) and from the far-lateral peripheral zone (the PL and AL sites in the TP biopsy scheme and the la and lm sites in the TR biopsy scheme in Fig. 1) was reconfirmed in the current study. Our results, however, do not necessarily exclude the possibility of a better TP or TR biopsy scheme that targets sampling sites other than those included in the 3D26 biopsy scheme. For example, recently reported deeply angled apical sampling through the TR route might improve apical cancer detection.19 Whether such a TR sampling strategy makes transperineal sampling unnecessary remains to be seen.

In addition cancer detectability we found that the characteristics of cancers are also equivalent between cancers detected by TP approach and those detected by the TR approach. Further study is warranted to examine whether cancers detected by the TP approach and those detected by the TR approach are pathologically equivalent in radical prostatectomy specimens.

Another issue to be clarified is whether the TP and TR biopsies are indicated differentially to a specific subgroup of patients. We found no difference in the cancer detection rate between the TR and TP biopsy schemes in men with abnormal DRE. As far as the cancer detection rate is concerned, we have already shown that the three-dimensional combination of TP and TR sampling sites with a total of 14–16 cores outperforms the extended TR or extended TP biopsies.14,15 The current results suggest that, if only a single TP or TR approach is allowed, we can adopt either of these approaches depending on the patient situation, regardless of the biopsy setting, the DRE finding, the PSA level or the prostate size.

In conclusion, the present study demonstrated for the first time that an extended TP biopsy is equipotent to its TR counterpart in regard to cancer detectability and the characteristics of cancers detected, as far as sampling sites are selected to maximize the cancer detection rate.

Ancillary