Karol Axcrona, Department of Urology, The Norwegian Radium Hospital, Oslo University Hospital, Ullernchausseen 70, 0310 Oslo, Norway. e-mail: email@example.com
Study Type – Diagnostic (exploratory cohort)
Level of Evidence 2b
What's known on the subject? and What does the study add?
Several authors have previously reported that transrectal prostate biopsy has a false-negative rate of 20–30%, and that a number of prostate cancers missed on transrectal biopsy can be detected by transperineal biopsy. It has also been shown that most of these tumours are located anteriorly in the prostate gland.
The present study showed a high rate of prostate cancer in patients with previous negative transrectal biopsies but elevated PSA levels, and that the cancers were located anteriorly in the prostate gland. Also, most of these cancers were clinically significant in patients that underwent RP, i.e. a high proportion of cancers were high-grade/high-stage tumours. We also showed that the transperineal biopsy technique can be applied successfully to patients with a closed anal orifice after previous surgery for rectal cancer. Transperineal biopsy can be done safely without routine antibiotic prophylaxis.
• To investigate the outcomes of transperineal prostate biopsies in patients with elevated prostate-specific antigen (PSA) levels and negative transrectal biopsies. The aim of this retrospective study was to evaluate the diagnostic yield of the transperineal biopsy approach in these patients, and to evaluate the pathology findings in subsequent radical prostatectomy (RP) specimens in patients undergoing RP.
PATIENTS AND METHODS
• In all, 69 consecutive patients with previous negative transrectal biopsies but elevated PSA levels investigated at urological units in Norway who had been referred to The Norwegian Radium Hospital were included.
• The patients had undergone a mean (median; range) of 2.42 (2; 0–7) transrectal biopsies. The mean (range) age was 63.1 (42–78) years. The median (range) PSA level was 12 (4.3–229) ng/mL.
• The patients were examined with transperineal biopsy of the prostate between July 2007 and February 2009. The results of the transperineal biopsies were reviewed for Gleason biopsy score, and these were compared with the histopathology results of the RP specimens, i.e. final Gleason scores.
• Pathological stage of the prostate specimens and tumour volume were also reviewed.
• Prostate cancer was found in the biopsies of 38 of 69 patients (55%).
• In all, 20 of 38 patients had a Gleason score estimated at ≥3 + 4 = 7.
• In all, 26 patients underwent RP. The surgical specimens revealed pathological stage pT2c in 65%, pT3a in 27% and pT3b in 8% of the cases.
• In all, 23 of the 26 RP specimens showed a final Gleason score of ≥7. The vast majority of cancers detected were situated in the anterior/ventral portion of the prostate.
• Transperineal biopsy of the prostate in patients with an elevated PSA level after negative transrectal prostate biopsies appears to be a feasible and important option for further investigation to detect prostate cancer.
• The present study shows that the transperineal biopsy allows good access of the anterior/ventral part of the prostate.
• Histopathology reports on the RP specimens obtained from the patients that underwent RP revealed significant cancer.
The use of serum PSA in the screening for prostate cancer has increased detection rates  and has caused a stage migration toward localised disease [2,3]. The diagnosis of prostate cancer is based on a positive finding in biopsies. Investigation of patients with an elevated PSA level has for the past two decades been performed using systematic biopsies guided by TRUS, first described by Hodge et al. . A negative systematic prostate biopsy does not exclude clinically significant cancer, and several authors have modified the technique in an attempt to increase the sensitivity [5,6]. Nonetheless, urologists still regularly face the problem with patients who have persistently elevated PSA levels after negative prostate biopsies.
Several studies investigating the detection rate obtained from prostate biopsies have shown that significant cancers are missed by the initial biopsies. Rabbani et al.  enrolled 118 patients in a study of repeated sextant biopsies before radical prostatectomy (RP) with or without neoadjuvant hormonal therapy and reported a 23% false-negative rate in the re-biopsies. However, even extended biopsy protocols fail to detect up to 20% of significant cancers detected by pathological evaluation of RP specimens [8,9].
Chen et al.  performed a computer-simulated study on 180 RP specimens, and found that sextant biopsies detected 73% of prostate cancers with a cancer volume of >0.5 mL. Missed cancers were located mainly in the anterior part of the prostate and in the anterior horn of the peripheral zone. Emiliozzi et al.  compared in vivo transrectal vs transperineal six-core biopsy, and found a cancer detection rate of 40%. In all, 95% of these cancers were detected by the transperineal approach compared with 79% by the transrectal approach. Merrick et al.  reported a 42% cancer-detection rate in transperineal template-guided saturation biopsy of 102 patients with a previous negative TRUS-guided biopsy. Cancers were detected in all areas, but the anterior and apical regions were involved most frequently. Moran and Braccioforte  reported a 39% cancer-detection rate in stereotactic transperineal biopsy of 747 patients with previous negative transrectal biopsies. The apical region, in particular the anterior apex, had a significantly higher incidence of cancer than the rest of the prostate gland in these cases.
We present a study of 69 patients who had undergone multiple series of negative transrectal prostate biopsies at urological units in Norway because of elevated PSA levels. Patients underwent re-examination with transperineal prostate biopsies at The Norwegian Radium Hospital. The results of these are presented.
PATIENTS AND METHODS
The Norwegian Radium Hospital is a national third-line cancer hospital, and most patients were initially investigated elsewhere before referral to our institution. From July 2007 to February 2009, 69 patients were re-examined. The patients' characteristics regarding clinical T stage, mean age, and mean PSA level are listed in Table 1. A mean of 2.42 transrectal biopsy series had been performed before admission to our institution for further investigation. Generally, transrectal prostate biopsies had been performed using sextant and octant biopsy schemes as described previously, but some patients also had extended biopsy schemes. Three patients had previously been operated on for rectal carcinoma with abdomino-perineal rectal amputation and had therefore not been subject to transrectal biopsies of the prostate. The Regional Ethical Committee of Norway (REK2010/2923) approved the study.
Table 1. Baseline patient characteristics
Patients with cancer
Patients without cancer
Three of the patients investigated had not been biopsied transrectally because of a closed anus due to previous abdomino-perineal rectal amputation for rectal carcinoma.
The biopsy procedure was conducted as a day-care procedure in an operating room. The patients were given i.v. sedation and placed in the lithotomy position. The perineum was prepared with a 5% chlorhexidine solution, and the scrotum draped anteriorly to access the perineum. Local anaesthetic was infiltrated in the perineal skin and s.c. tissue.
TRUS was performed using a BK Medical 3535 ultrasound scanner (BK Medical, Herlev, Denmark) with a 6–10 MHz transrectal probe (model 8551) with an external needle-holder device (A3510) mounted on the ultrasound probe. The biopsies were taken using an 18 G biopsy gun (TruPath, Boston Scientific, Natick, MA, USA or Bard Max Core, Bard Denmark AB, Helsingor, Denmark) with an 18-mm sample notch. Brachytherapy templates for guiding the transperineal biopsies were not used. Transperineal biopsies of the prostate can also be obtained from patients after abdomino-perineal amputation of the rectum using the same biopsy setup. Generally, five biopsies from the right ventral/basal, right ventral/apical, left ventral/basal and left ventral/apical lobes were taken (Fig. 1A,B). In some patients who had had MRI of the prostate and where findings indicated a tumour, directed biopsies were taken from the suspicious area in addition to biopsies from the whole gland. If the MRI was negative, uncertain, or if no MRI was done, saturation biopsies were taken. Of the 38 patients diagnosed with prostate cancer, 22 patients were biopsied with >17 biopsy cores; in patients without findings of prostate cancer 26 of 31 patients were biopsied with >17 biopsy cores (Table 1). Antibiotic prophylaxis was not given routinely but prophylaxis was given to patients with previous infections after transrectal biopsy. Appropriate endocarditis prophylaxis was given to patients with mechanical heart valves or other risk factors.
One patient, who had previous spontaneous urinary retention, had a permanent catheter inserted after the biopsy. The catheter was removed on the second day and he had immediate spontaneous micturition.
ROBOT SYSTEM AND OPERATIVE PROCEDURE
A robot-assisted laparoscopic RP was performed with a three-armed Da VinciTM surgical system (Intuitive Surgical, Sunnyvale, CA, USA). Our operative approach for RP was based mainly on the Vattikutti Institute procedure published by Tewari et al. . A nerve-sparing operation was performed in 12 of 26 patients, and 13 of 26 patients underwent pelvic lymph node dissection. Dissection of the prostate did not present any difficulties when patients were operated on using the robot-assisted RP after investigation with transperineal saturation biopsies. The vesico-urethral anastomosis was sewn with a synthetic absorbable suture (Biosyn 2-0) double-armed thread using the technique previously described by van Velthoven et al. .
Immediately after RP the RP specimens were sent to the Department of Pathology for processing as described previously . The prostate was fixed for 2–3 days in 10% buffered formalin, was weighed, measured, and inked by three different colours for orientation and for analysis of the surgical margins. The mid part of the prostate was paraffin embedded to produce large whole mount 5-mm thick slides. The apical region, the bladder neck portion of the prostate, and the seminal vesicles were also embedded completely and sectioned.
Final pTNM stage, Gleason grade and score, and tumour volume were obtained on 4-µm thick slides stained with haematoxylin and eosin. The final histopathological classification of prostate cancer was done according to the International Society of Urological Pathology (ISUP) Consensus on Gleason grading of prostate cancer  (Table 2).
Table 2. Summary of Gleason score of patients' transperineal prostate biopsies
Biopsy Gleason score
N (n= 38)
3 + 2
3 + 3
3 + 4
4 + 3
4 + 4
4 + 5
Biopsies and RP specimens were evaluated by pathologists with >15 years of experience in urological pathology.
The characteristics of patients diagnosed with cancer and those of patients in whom cancer was not detected after transperineal prostate biopsies are given in Table 1. Before referral to The Norwegian Radium Hospital, the patients were examined with an average of 2.42 biopsy series and an average of 20.39 transrectal prostate biopsy cores at urological units in Norway. The mean PSA level was 19.8 ng/mL. In all, 38 of the 69 patients (55%) examined with transperineal prostate biopsies were diagnosed with prostate cancer. The patients diagnosed with cancer had an average of 4.1 positive cores. The Gleason score for the biopsy specimens are presented in Table 2. A biopsy Gleason score of ≤6 was detected in 18 of 38 patients (47%), Gleason score 3 + 4 = 7a in eight of 38 (21%) patients and Gleason score ≥4 + 3 = 7b in 12 of 38 patients (32%) (Table 2). In all, 31 patients were considered to have a clinically significant cancer based on the Gleason grade, number of positive biopsies, PSA level, and the sum of cancer length in all biopsies. Seven patients were subjected to active surveillance . Four patients were referred for external beam radiation therapy (EBRT), and 26 patients underwent RP. One patient with negative transperineal biopsies was scheduled for repeated transperineal biopsy and then elected to have a TURP because of bothersome LUTS. The TURP histopathology revealed cancer of the prostate with a Gleason score of 7a.
One patient was diagnosed with recurrence of a locally advanced rectal carcinoma (cT4) and received antiandrogen monotherapy concomitant with EBRT to treat the rectal cancer recurrence and prostate cancer. He died from the rectal cancer. Another patient had a primary locally advanced rectal cancer stage cT4; this patient started androgen-deprivation therapy (ADT) and EBRT with a preoperative intent to treat his rectal cancer and with a boost for curative intent for his prostate cancer. One patient received ADT and EBRT as the primary treatment. One patient with haemophilia B started ADT and EBRT. Another patient had verified skeletal metastases and a PSA level of 229 ng/mL at the time of the transperineal prostate biopsies and had already started ADT at the time of the transperineal biopsy and remained on ADT.
Seven patients were considered to have clinically insignificant cancer according to the eligibility criteria  and were enrolled in a schedule for active surveillance. They will be examined regularly. One 78-year-old man with a PSA level of 27 ng/mL and clinical T2 prostate cancer with a Gleason score of 3 + 3 = 6 in 10 of 38 biopsies was recommended for watchful waiting.
RESULTS OF HISTOPATHOLOGICAL EXAMINATION OF THE RP SPECIMENS
Seven of 15 patients (47%) had the Gleason score upgraded from ≤7a to Gleason score ≥7b, when their biopsies were compared with the definite Gleason score obtained from the RP specimen (Table 3).
Table 3. Characterization of transperineal biopsy Gleason score compared with histopathological Gleason score in RP specimens
Core needle biopsies
Gleason score ≤ 6
Gleason score 7a
Gleason score ≥ 7b
RP specimen, n (%)
Gleason score ≤ 6
Gleason score 7a
Gleason score ≥ 7b
In all, 26 of the 69 patients studied (38%) were considered to have significant cancers prone to radical surgical treatment. Two of the 26 patients operated on had a retropubic RP at their respective local hospitals; the remaining 24 patients had a robot-assisted RP at our institute. The final pathology findings are listed in Table 4. Interestingly, the vast majority of the patients operated on for prostate cancer in the present study had an advanced cancer as judged by the histopathology report (pTNM). Three of 26 patients (11.5%) had a final Gleason score of 6 and 23 of 26 patients (88.5%) had a final Gleason score ≥7 (Tables 3 and 4). In all, 16 of 26 patients (61.5%) had a final Gleason score ≥7b (Table 3 and 4). Twelve of 26 patients (46%) had an upgraded Gleason score, and two patients had a downgraded Gleason score when the final histopathology report was compared with the results of the biopsies. Eight patients (31%) had positive surgical margins. The tumour volumes were calculated, and these also showed significant cancers (Table 4) . In all, 22 of 24 prostate cancers had a volume of >0.5 mL. The postoperative PSA level was <0.2 ng/mL in all patients operated on.
Table 4. Final histopathological stage and Gleason grade in RP specimens
Negative surgical margins
Positive surgical margins
Mean (range) tumour volume, mL
Two patients with positive surgical margins/pT2c Gleason 4 + 3 underwent radical RP elsewhere. The mean tumour volume of the largest lesion is shown. Tumour volume was calculated according to Chen et al. . In all, 22 of 26 prostate cancers were located anteriorly/basally. Two of 26 tumours had a tumour volume of <0.5 mL (both pT2c Gleason 3 + 3).
pT2a Gleason 4 + 4
pT2c Gleason 3 + 3
pT2c Gleason 3 + 4
pT2c Gleason 4 + 3
pT3a Gleason 4 + 3
pT3a Gleason 4 + 4
pT3a Gleason 4 + 5
pT3b Gleason 4 + 3
pT3b Gleason 4 + 4
Thus, there was a marked upgrading in Gleason score from the prostate biopsy to the final histopathological evaluation of the RP specimens (Tables 2 and 3).
In all, 22 of 26 RP specimens (85%) showed that the cancer was located in the anterior/ventral portion of the prostate.
Figure 2 shows an example of a patient with an elevated PSA level who was examined and treated at our institute.
COMPLICATIONS RELATED TO TRANSPERINEAL BIOPSY
Few complications were recorded after the biopsy procedure. One patient was admitted with fever and bacteraemia, and he was treated successfully with oral ciprofloxacin. One patient had acute urinary retention on the evening after biopsy; his prostate was 200 mL. He was treated with a permanent catheter, which was retained until he was operated on after a few weeks later. One patient had a lower urinary tract infection and urinary retention, and was treated at his local hospital with antibiotics and a catheter, which was removed within a few days.
Different prostate biopsy strategies have been published. The original sextant biopsy described by Hodge et al. [6,7] had a false-negative rate of 20–33%. Increasing the number of cores or areas of the prostate sampled has been shown to increase the rate of cancer detected [5,19,20].
The optimal technique for re-biopsy is as controversial as the initial biopsy strategy. Studies performing re-biopsies in the same manner as the initial biopsies have shown that a substantial percentage of patients are diagnosed with cancer in the second or subsequent biopsy set [20,21]. Studies of the anatomical localisation of cancer missed in the initial biopsy have shown the anterior and apical regions of the prostate to have the highest density of cancer [22,23].
Several studies that included a larger sample of patients have shown that about one in five prostate cancers is located in the anterior region of the prostate [24,25]. A transperineal biopsy approach provides easy access to these regions of the prostate gland, and several authors have reported a higher rate of prostate cancer found using this approach [9,11,26,27]. Two prospective randomised studies on transperineal vs transrectal 12-core prostate biopsies found a similar yield of positive biopsy cores in both approaches [28,29]. One retrospective study found no difference in cancer detection in the re-biopsy setting between transrectal and transperineal biopsies when investigated with 24-cores biopsies . Taken together, these findings suggest that it is necessary to investigate patients with high PSA levels using repeat biopsies and that the amount of biopsied material is important; i.e. repeat biopsies should include ≥12 cores.
The present results show that the transperineal biopsy can be performed using an ultrasound probe without using a brachytherapy template, which makes the procedure easier. As noted by Takenaka et al. , the complications of transperineal biopsy have received little attention. In the present study, this approach had few complications, and none were critical (e.g. bleeding). The transperineal biopsy procedure was performed in a day-care unit, and the patients were anaesthetised with i.v. sedation, which makes the procedure more complex.
Comparison between the biopsy Gleason score and the definite Gleason score obtained from the histopathology report showed a marked upgrading. Upgrading of the Gleason score from the biopsy to a definite histopathology report on RP specimens is well documented [8,31]. However, in the present study, the percentage of patients whose score were upgraded was very high; i.e. 85% of the low Gleason grade biopsies lower or equal to Gleason score 7a were upgraded.
Considering that up to 20% of all prostate cancer is located in the anterior/ventral prostate, one could speculate that these cancers are contributing to an increase in prostate cancer death rate. These cancers might not cause obstructive voiding symptoms, and they are likely to be missed in patients with an elevated PSA level investigated by transrectal biopsy.
A recent publication by Ploussard et al.  showed that a considerable proportion of patients considered for an active surveillance programme based on preoperative parameters (i.e. Gleason score ≤6), had a significant cancer based on the histopathology report. Although the sample size in the present study is small, our results suggest that most patients with previous negative transrectal biopsies but with positive transperineal biopsies should undergo active radical treatment for prostate cancer because the proportion of significant cancers is high. The present results also suggest that patients could be offered treatment even if the Gleason score for the biopsies is low because there was a substantial upgrading of the final Gleason score for the RP specimens. Further histopathological analysis of RP specimens included in the present study will reveal whether the cancers in this specific patient sample have special properties, e.g. lower Gleason score in the periphery of the cancers where the positive biopsy might have come from. Long-term follow-up will be needed to determine if these cancers have a worse prognosis compared with cancers of similar pT stage and Gleason score detected in the transrectal biopsy.
In conclusion, the results of the present study and previous published studies suggest that patients with elevated PSA levels could be offered an examination with transperineal biopsies after two negative sets of transrectal biopsies. Ours is the first study to compare directly the outcomes of transperineal prostate biopsies with the final histopathology. We found a high proportion of high-grade/high-stage prostate cancers in patients with previous negative transrectal biopsies and elevated PSA levels. Further investigations will show whether these anterior/ventral high-grade prostate cancers are a separate entity amongst prostate cancers compared with cancers derived from the basal/dorsal peripheral zone or if they behave similarly biologically. Because of the high proportion of high-grade/high-stage cancers finally detected, investigation of patients with previous negative transrectal biopsies and elevated PSA levels using transperineal prostate biopsies might be recommended.
We express our gratitude to Jostein Eikanger at the Institute for Medical Informatics, Oslo University Hospital and Øystein Horgmo at the Photographic and Video Services, University of Oslo, for providing the illustrations. We thank for financial support from the Norwegian Radium Hospital Legacy.