To identify the zonal location of prostate cancers before surgery, by analysing the mapping of ultrasonography-guided systematic sextant biopsies for differences between cancers located in the transition zone (TZ) and peripheral zone (PZ); and to compare the correlation between Gleason scores of needle biopsies and those of radical prostatectomy (RP) specimens.
Patients and methods
In all, 186 patients with TZ (46) and PZ cancers (140) underwent ultrasonography-guided systematic sextant biopsy and RP at the same institution. The clinical and pathological characteristics, and the anatomical location of positive biopsies, were determined and compared using t-tests and chi-square tests. Differences between Gleason scores of needle biopsies and those of RP specimens were evaluated and compared by Cohen κ testing.
TZ cancers had a significantly lower rate of positive biopsies in the middle (63% vs 80%) and base (50% vs 80%) of the prostate than had PZ cancers. Positive biopsies were exclusively obtained from the apex in 19.6% of TZ and 5% of PZ cancers (P = 0.002). There was exact agreement between Gleason scores of needle biopsies and those of RP specimens in 15.2% of TZ (κ = 0.02) and 55% of PZ cancers (κ = 0.25), respectively.
Compared with PZ cancers, TZ cancers had a different anatomical pattern of positive biopsies, with lower rates in the middle and base of the prostate. The finding of positive biopsies exclusively in the apex favoured prostate cancer located in the TZ. Furthermore, the correlation between needle biopsy Gleason scores and those of the RP specimens was clearly lower in TZ cancers.
Traditionally, most transitional zone (TZ) cancers were diagnosed incidentally during TURP undertaken for clinical BPH . This has changed with the widespread use of PSA testing and consecutive prostatic biopsy in men with a pathologically elevated PSA value. Meanwhile, cancers located in the TZ can be assessed in about a quarter of patients undergoing contemporary radical prostatectomy (RP) for localized prostate cancer stage T1c . TZ cancers are of special interest, particularly as these tumours more frequently show organ-confined disease and lower Gleason scores, despite higher serum PSA levels and cancer volumes, than peripheral zone (PZ) cancers [3–5]. Hence, as TZ cancers also have a more favourable prognosis under these conditions, higher serum PSA levels of these men compared with PZ cancers may be misleading in planning treatment . At present, nomograms and algorithms which use the preoperative serum PSA level as a prognostic marker represent the mainstay in predicting pathological stage or prognosis, but they do not take the zonal location in account [7,8].
Thus additional preoperative information about the probable zonal location might be helpful to estimate the outcome more accurately, particularly when the information would be easy to obtain. However, there are few studies addressing whether TZ cancers could be predicted by the pattern of positive biopsies in the common systematic sextant biopsy scheme.
Notably, the Gleason score of needle biopsies, which is another important marker for all nomograms and algorithms, may differ considerably from that of the RP specimen [9–12]. However, the correlation between both scores for cancers of different zonal location has been little studied to date. Thus, we also assessed the correlation of Gleason score of the needle biopsies with those of the RP specimen separately for TZ and PZ cancers.
Patients and methods
Between January 1994 and December 1997 a consecutive series of 505 patients underwent retropubic RP for localized prostate cancer in our department. For comparison 46 patients were selected with TZ cancer and 140 with pure PZ cancer in whom prostate cancer was diagnosed by TRUS-guided sextant biopsies, which were taken exclusively at our outpatient clinic. Detailed data on cancer volume and zonal location were available for all of these men. Patients with previous TURP and those who had received any kind of preoperative neoadjuvant hormonal therapy or radiation therapy were not considered for this evaluation.
Prostate cancer was assessed by a routine staging evaluation, including serum PSA measurement (Axsym assay, Abbott Park, IL, USA, before the DRE) a DRE and TRUS of the prostate (7 MHz transrectal scanner with an end-firing probe). Patients underwent biopsy mapping of the prostate with six systematic sextant TRUS-guided biopsies, obtained using an 18 G spring-loaded biopsy gun, from the apex, middle and base of the right and left prostatic lobes in the parasagittal plane. Each biopsy core was embedded separately in paraffin wax, and step-sections stained with haematoxylin and eosin. The anatomical mapping of the biopsies was assessed and the percentage of prostate cancer determined in all six biopsy cores (length of cancer/per total tissue length × 100).
All prostatectomy specimens were completely inked on their surfaces and processed according to the Stanford protocol using serial transverse sections at 3 mm . Whenever possible, whole-mount sections of the prostate gland were cut (Fig. 1) and stained with haematoxylin and eosin. Tumour areas were marked on the slides with a water-resistant pen (Fig. 1). By identifying the TZ boundary, the location of the largest tumour area was determined. Briefly, TZ tumours were considered when more than half the cancer area was in the TZ. This criterion was adopted from previous studies [4,5]. Tumour volume was measured by computer-assisted planimetry . The clinical and pathological stages were assessed according to the Fifth edition of the TNM system  and graded according to the classification proposed by Gleason .
The significance level of differences was determined at P < 0.05 in all statistical tests; the t-test was used for independent samples to assess mean differences between TZ and PZ cancers in age, serum PSA level, prostate volume, cancer volume, number of positive biopsies and the percentage of cancer per sextant biopsy set. The chi-square test was used to compare the two tumour groups for clinical and pathological stage, Gleason scores of the needle biopsies and RP specimens, positive lymph node status, surgical margin status and nerve-sparing procedure. Furthermore, the chi-square test was used to test the biopsy status for the three anatomical locations and for the exclusive occurrence at one anatomical side. The odds-ratio coefficient was used as an estimator for the relative risk. Agreements between the Gleason scores of the needle biopsies and RP specimens were calculated using Cohen's κ, which is a measure of concordance between observations with values of −1 to 1; a κ of < 0.40 represents low, 0.40–0.75 good and> 0.75 very good agreement.
The mean (sd, range) age of all 186 men was 60.4 (6.4, 44–75) years; Table 1 shows the clinical and pathological variables of the two groups. In detail, 19 of 46 TZ cancers (41%) were completely within the TZ; Figure 1 shows an example of a carcinoma located exclusively in the TZ. The mean preoperative serum PSA levels (15.5 vs 10.9 ng/mL) and mean tumour volumes (8.7 vs 5.5 mL) were significantly higher in patients with TZ cancers. The prostate volumes of both groups did not differ statistically. There was no significant difference in clinical stages. Men with TZ cancers had a higher rate of organ-confined disease (74% vs 56%) and positive lymph nodes were only seen in PZ cancers (four). Of all patients, 23 (12%) had positive surgical margins, and this was significantly higher in TZ cancers (22% vs 9%). Nearly half of patients had undergone a nerve-sparing procedure in both groups. Table 1 also shows the Gleason scores of the needle biopsies and RP specimens, both of which were significantly more favourable in TZ cancers.
Table 1. The comparison of clinical and pathological variables, needle biopsy and RP specimen Gleason score, needle biopsy results with location, and incidence of positive needle biopsies at only one anatomical location, between prostate cancers located in the TZ and PZ
No. of men
prostate volume, mL
Clinical stage, n (%)
Mean (sd) cancer vol, mL
Pathological stage, n (%)
Lymph node involvement, n (%)
No. surgical margins, n (%)
Nerve sparing, n (%)
Needle biopsy, n (%)
Prostatectomy specimen, n(%)
Needle biopsy result with position, n (%)
with –ve biopsies
with 1 or 2 +ve biopsies
with –ve biopsies
with 1 or 2 +ve biopsies
with –ve biopsies
with 1 or 2 +ve biopsies
Positive needle biopsy at one location, n (%)
The mean (sd) number of biopsies bearing cancer foci was significantly (P = 0.026) lower in TZ, at 2.24 (1.34), than in PZ cancers, at 2.76 (1.38). In addition, the mean (sd) percentage of cancer per systematic sextant biopsy set was also significantly (P < 0.001) lower in TZ, at 12.3 (11.8)% than in PZ cancers, at 20.3 (15.9)%. Cancer was detected in both lobes in 39% of TZ and 41% of PZ tumours.
Table 1 also shows the differences between TZ and PZ cancers with the location of cancer foci-bearing positive and negative biopsies. TZ tumours were detected significantly less often when biopsies were obtained from the middle and base of the prostate, and there was a decrease in detected cancer foci from the apex to the base. In contrast, most cancer foci of PZ tumours were found in biopsies from the middle and base. At the apex the incidence of positive biopsies was not statistically different between TZ and PZ tumours.
Table 1 also shows the comparison of the incidence of positive needle biopsies at one anatomical location between TZ and PZ tumours. The rate of cases with positive biopsies at the apex only was significantly higher in TZ cancers. The odds ratio of exclusively positive apical biopsies was 4.62 (95% CI, 1.61–13.24), representing a nearly five-fold increase in the risk of TZ cancer. There was no association between the number of positive biopsies exclusively obtained from the middle and from the base of the prostate and the type of cancer.
Table 2 shows the correlation between the Gleason scores of needle biopsies and those of the RP specimens. Overall, there was exact agreement in 45% and under-grading of the RP specimens was more common than over-grading (30% vs 25%). Most of the false grading differed by one score only (Table 2). An inaccuracy of two Gleason scores was more common in TZ than PZ cancers. TZ cancers had a distinctly lower exact agreement between Gleason scores of needle biopsies and those of the RP specimens than had PZ cancers. Cohen's κ analysis yielded a value of 0.25 in PZ cancers, which represented a poor agreement beyond chance. In TZ cancers the κ was 0.02, i.e. no correlation. In addition, there was an obviously different pattern of false grading between TZ and PZ cancers, as Gleason scores of needle biopsies in TZ cancers were more frequently over-graded than those of PZ cancers (Table 2).
Table 2. Correlation between needle biopsy Gleason scores and RP specimen Gleason score, with the κ coefficient
Differed by 1 score
Differed by 2 score
κ coefficient (sem)
Prostate cancers located in the TZ show several histological and genetic differences from PZ cancers [16–18]; notably, a different biological and clinical behaviour was also described previously. In particular, an important clinical difference is that TZ cancers may have lower tumour stages than PZ cancers, despite obviously higher PSA values [4,5]. PSA is an important marker in preoperative staging and higher values in men with TZ cancers may lead to a more pessimistic prognosis and a probable change in treatment method . Considering this, knowing the type of cancer would be helpful in counselling and planning therapy. Unfortunately, imaging methods are unable to identify TZ cancer accurately  and thus we investigated the possibility of predicting the type of cancer by an easily obtainable finding, e.g. the pattern of positive biopsies of the common systematic sextant biopsy scheme.
In the present study, positive biopsies of TZ cancers were more common towards the apex of the prostate; this pattern could be explained by the typical location of TZ cancers. Nouguchi et al. assessed the predominant location as anterior in the middle to the apex, by examining 148 RP specimens with TZ cancers. Furthermore, Chen et al. showed by a computer-plotting model of biopsy cancer foci that TZ cancers were concentrated in the apical two-thirds and in the furthest anterior portion of the prostate.
In contrast to TZ cancers, PZ cancers had the highest rate of detected cancer foci in the middle and base of the prostate, but the incidence of positive apical biopsies was similar to that in TZ cancers (Table 1). The present findings confirm the multifocal character of prostate cancers and compare well with other studies . In particular, positive biopsies of PZ cancers tended to be dispersed over the whole prostate, as positive biopsies from one anatomical location were very rare (Table 1). Notably, this was even true for TZ cancers, but it represented a nearly five-fold increase in the risk of TZ cancer when cancer foci were exclusively found in apical biopsies. However, sole apical biopsies were detected in only a few patients. In any case, this additional information might be helpful in decision-making, especially when the serum PSA level was high with no suspicious findings on a DRE.
Besides the PSA level the Gleason score of needle biopsies is another important prognostic marker. A good correlation with that of the RP specimens is warranted in clinical practice. When using 18 G biopsy needles, complete agreement between Gleason scores of biopsies and surgical specimens is reported in 28–58% of cases [10–12,22,23]. The present study showed exact agreement in 45% of all cancers; only a minority (4.8%) differed by two scores. This finding compares well with that of Spires et al., who found a difference of two or more scores in only 6%. The missing of well-differentiated cancers in that and in the present study may explain the more favourable findings, as even these cancers had the highest disagreement in other studies [10,11]. However, using Cohen's κ a coefficient of 0.19 represented a poor correlation beyond chance between the scores. This is in line with other investigations, which reported a κ of 0.15–0.26 [9,10].
Interestingly, in previous studies the correlation between Gleason scores of needle biopsies and that of RP specimens has not been addressed separately for TZ cancers. Remarkably, with an exact agreement in only 15%, the Gleason scores of TZ cancers had virtually no correlation (κ 0.02). Furthermore, the Gleason scores of needle biopsies were higher than those in the RP specimens in about two-thirds of cases. In particular, Gleason score 6 instead of 5 was most frequently assessed in the needle biopsies. This stands in clear contrast to PZ cancers, where under-grading was more common. Additionally, under-grading was also the error most frequently reported by other investigations not distinguishing between TZ and PZ cancers [9,10,23]. The reasons for this inverse trend of false grading between TZ and PZ cancers are unclear.
In TZ cancers, the higher PSA values and the tendency to overestimate Gleason scores in the needle biopsies give a more pessimistic impression and might also be misleading in planning treatment. Thus knowing the tumour type would be helpful and underlines the importance of identifying the cancer type before treatment. Hence, in the common sextant biopsy the occurrence of positive biopsies at the apex only supplies valuable information about the presence of TZ cancer. In the present series, additional TZ biopsies were not routinely taken; the benefits of this in improving cancer detection were discussed controversially in recent reports [2,24]. Nevertheless, additional TZ biopsies should be taken if there is reasonable clinical suspicion of TZ cancer . This might help to identify TZ cancers in more patients than the sextant biopsy scheme alone; further studies would be warranted on this topic.
The authors thank Fedor Daghofer, PhD (Karl-Franzens-University, Graz, Austria) for his assistance with data analysis.