- Top of page
- CONFLICT OF INTEREST
What's known on the subject? and What does the study add?
In the era of extended biopsy sampling of the prostate, multifocal high-grade prostatic intraepithelial neoplasia (HGPIN) is associated with a significantly higher rate of cancer diagnosis than unifocal HGPIN or a benign diagnosis. In addition, the cancers that are subsequently diagnosed in men with HGPIN on their initial biopsy tend to be smaller, lower grade and more commonly organ-confined. This has led to a reappraisal of the need and timing of repeat biopsies.
The present paper provides a series of recommendations on the optimal timing of repeat biopsies in men with HGPIN on biopsy, based on the current available evidence.
This is the first of a two part series reviewing the nature and clinical significance of in situ cellular proliferations in the prostate gland.
This first part examines prostatic intraepithelial neoplasia (PIN), while the second part in the next supplement discusses intraductal carcinoma and ductal adenocarcinoma of the prostate.
PIN is a precursor lesion in the development of some forms of adenocarcinoma of the prostate. In the 1990s, high-grade PIN (HGPIN) on biopsy was a significant predictor of carcinoma, but this was due to incomplete sampling with sextant biopsies. With more extensive sampling in the last decade, the likelihood of identifying cancer after a diagnosis of HGPIN is not significantly different from a benign diagnosis.
In several recent studies, it is now recognised that multifocal HGPIN is a better predictor of cancer than unifocal HGPIN. Most cases of cancer will be detected in the vicinity of the HGPIN, but up to 40% of cancers will occur in different sextants.
In assessing potential markers for carcinoma in men with HGPIN on biopsy, α-methylacyl coenzyme-A racemase (AMACR) has emerged as a promising diagnostic tool.
HGPIN with strong staining for AMACR is associated with a higher rate of cancer detection in subsequent biopsies compared with AMACR-negative HGPIN. Also, AMACR positivity in HGPIN is more commonly seen adjacent to carcinoma, and this may provide guidance as to the site of future biopsies.
- Top of page
- CONFLICT OF INTEREST
PIN is a neoplastic cellular proliferation confined to pre-existing prostatic ducts and acini with no evidence of stromal invasion. When the term was first proposed, PIN was divided into three grades from 1 to 3, similar to the grading of cervical intraepithelial neoplasia. Due to poor diagnostic reproducibility among pathologists, these categories were subsequently combined into low- and high-grade (HG), with the HG component encompassing grades previously referred to as PIN2 and PIN3. The separation of low-grade and HGPIN is made on morphology, and the presence of nucleoli within nuclei is the defining feature of HGPIN (Fig. 2). Unfortunately, this definition lacks precision, and is subject to variation for several reasons, including the diagnostic criteria of the reporting pathologist and technical factors, e.g. the quality of the sections and the type of fixative used. While the reported rates of HGPIN range from 0 to 25% [1,2], in contemporary practice it is expected that the rate of diagnosis of HGPIN in the absence of carcinoma is 4–8% .
Figure 2. High grade PIN on the left showing prominent nucleoli in the nuclei (arrowed), with benign acini on the right for comparison. (H&E, original magnification 400×).
Download figure to PowerPoint
The clinical significance of HGPIN itself is difficult to determine. HGPIN is a precursor lesion in the development of carcinoma in some but not all cases of carcinoma, with considerable epidemiological and morphological evidence summarised by Bostwick and Qian . Evidence supporting HGPIN as a precursor lesion includes: the appearance of HGPIN up to 10 years before the carcinoma; both HGPIN and cancer increase in incidence with age; HGPIN is more common, more extensive and more often multifocal in prostates in which there is adenocarcinoma; HGPIN and carcinoma tend to occur in the same sites in the prostate gland; and in several cases, cancer has been shown budding off glands with HGPIN.
There is a considerable body of molecular evidence linking HGPIN and carcinoma, summarised by Montironi et al.  and Mosquera et al. . HGPIN and carcinoma show similar chromosomal anomalies, telomere shortening, and over-expression of p16, p53, bcl-2, c-myc and α-methylacyl coenzyme-A racemase (AMACR). The TMPRSS2-ERG fusion protein, which is present in 50% of prostate cancers, is also present in ≈20% of cases of HGPIN and is not identified in BPH or normal prostate tissue [4–6]. In a study using comparative genomic hybridisation, Ribeiro et al.  elucidated a pathway of prostate carcinogenesis and proposed that there are two separate pathways: the first involves loss of 8p with the development of HGPIN and subsequent carcinoma, while the second pathway occurs through loss of 13q, leading to carcinoma without preceding HGPIN.
In keeping with this association of HGPIN with carcinoma, early studies in the 1990s showed a high predictive rate of up to 50% for the subsequent diagnosis of carcinoma in men with biopsy identification of isolated HGPIN in the absence of carcinoma . However, subsequent studies have shown this was most likely due to inadequate sampling of the prostate in the sextant biopsy era, rather than the predictive value of PIN [8–10]. Review of the results of more contemporary studies between 2000 and 2006 gives a median rate of 24% for the diagnosis of carcinoma on repeat biopsy after a diagnosis of HGPIN, compared with 19% for an original benign diagnosis, with no statistical significance between the two groups in seven of nine studies . In a more recent study, De Nunzio et al.  reported a rate of subsequent cancer diagnosis of 18% at 6 months in 12-core sampling.
At present, there are no clinical features to indicate which men with HGPIN on core biopsy are at increased risk of carcinoma on subsequent biopsy. Of the histological features, only multifocal HGPIN is of predictive value. While many earlier studies showed that the number of cores with HGPIN was not predictive of a subsequent diagnosis of cancer , many of these studies were probably based on inadequate sampling. More recent studies with a ≥10-core sampling showed that multifocal HGPIN (two or more cores) was a significant predictor of carcinoma compared with unifocal HGPIN (one core only) or a benign diagnosis. Merrimen et al.  reported a rate of cancer detection of 9.4% for unifocal HGPIN, compared with 29.9% for multifocal HGPIN, while Roscigno et al.  reported higher rates of 23.1% for unifocal HGPIN vs 37.6% for multifocal HGPIN. Lee et al.  reported results similar to Merrimen et al. , and gave a hazard ratio of 1.19 for unifocal HGPIN and 2.56 for multifocal HGPIN. Lee et al. extrapolated these findings further, and constructed modified Kaplan–Meier curves based on Cox regression models, which estimated the probability of detecting carcinoma at 1, 3 and 5 years in men with a benign diagnosis is 3.6%, 12.5% and 22.4% respectively, unifocal HGPIN was 4.4%, 14.7% and 26.1% and multifocal HGPIN was 9%, 29% and 47.8% respectively (Table 1). With bilateral HGPIN on biopsy, the likelihood of detection of cancer at these intervals is even greater at 12%, 37% and 57.8% respectively. These figures show that the likelihood of diagnosing cancer increases with time regardless of the diagnosis on the initial biopsy, but the diagnosis of cancer occurs more frequently and earlier in men who have multifocal HGPIN, especially if it is present bilaterally.
Table 1. Estimated probability of cancer detection at 1, 3, and 5 year intervals depending on HGPIN status of original biopsy (probability expressed as %, adapted from Lee et al., 2010)
| ||1 year, %||3 years, %||5 years, %|
|Unilateral multifocal HGPIN||9.4||29.6||48.1|
|Bilateral multifocal HGPIN||12.1||37.0||57.8|
Describing HGPIN in four or more cores as ‘widespread HGPIN’, Netto and Epstein  reported an incidence of carcinoma of 39% on repeat biopsy in these men. Also, in that study, the identification of carcinoma in men with widespread HGPIN increased with age, with a rate of 55% in men aged ≥70 years, compared with 33% for men aged <70 years. Roscigno et al.  developed a nomogram to predict the likelihood of identifying cancer on repeat biopsy after diagnosis of isolated HGPIN, based on age, PSA concentration, number of cores taken at the initial biopsy and multifocality of HGPIN, with 72% reported accuracy.
As well as controversy about the predictive value of a diagnosis of HGPIN on biopsy, there is similar controversy about the timing of repeat biopsy in these men. In contemporary practice with more extended sampling, the diagnosis of carcinoma on repeat biopsy at 12 months is low . Lefkowitz et al.  reported a rate of cancer diagnosis of 2.3% at 12 months after a diagnosis of HGPIN, but more importantly, in a small cohort of 31 men, identified cancer in 25.8% with a delayed repeat biopsy after 3 years . In unpublished data, these authors have shown a rate of cancer diagnosis of 23.2% at 3 years in 101 men and 28.6% at 6 years in 28 men . The authors postulated that the longer interval between biopsies allowed small foci of cancer time to grow sufficiently to be detected, or alternatively, allowed enough time for cancer to develop from HGPIN. They also noted that there was no significant change in the PSA concentrations in those men with cancer diagnosed on the subsequent delayed biopsy, suggesting that one cannot rely on PSA concentration alone in monitoring men with HGPIN on biopsy.
The diagnosis of cancer after a biopsy diagnosis of HGPIN can occur anywhere in the prostate gland. Given HGPIN is a likely precursor of carcinoma, it is not surprising the cancer will most commonly be diagnosed in the same or adjacent sextants as the HGPIN, but it can be found in the contralateral lobe in 10–40% of cases [2,17]. A recent interesting finding, is the presence of higher expression of AMACR in HGPIN adjacent to carcinoma . AMACR is up regulated in both PIN and carcinoma, but this is particularly seen in HGPIN adjacent to carcinoma (56%) vs HGPIN that is located away from the carcinoma (14%). If validated, AMACR may provide a guide as to the most appropriate sites for repeat biopsies (Fig. 3).
Figure 3. AMACR positive HGPIN, with AMACR positive cells staining red and basal cells brown. (Antibodies p504S (red), p63 (brown), 34BE12 (brown), haematoxylin counterstain, ×200 original magnification).
Download figure to PowerPoint
The first repeat biopsy after a diagnosis of isolated HGPIN will detect 71–100% of cancers, although there are up to 10% of men with a diagnosis of HGPIN who will not have a diagnosis of cancer until the third or fourth biopsy . All of these men had a diagnosis of HGPIN in at least one of the preceding biopsies before the diagnosis of cancer. The cancers that are detected on repeat biopsy are generally of small volume and organ confined [2,19] and, in keeping with this, only one core is positive in 71% of men, and the cancer is most commonly in <5% of the core.
There are few studies describing the findings at radical prostatectomy in men with carcinoma after an initial diagnosis of HGPIN. In the largest review of 45 men published recently , the carcinomas were organ confined in 84.4%, with mean (median, range) tumour volumes of 0.3 (0.12, 0.003–1.46) mL. Most of the tumours were Gleason score 6 (80%). No seminal vesicle invasion or nodal metastases were identified. Most of the cases of cancer in the Al-Hussain and Epstein  study were diagnosed ≤12 months of the diagnosis of HGPIN, suggesting they were present initially but were not identified because they were too small. These findings correlate with similar findings of a pooled analysis of 54 cases described earlier .
At present, there are no immunohistochemical or molecular markers in routine use that will identify those men with HGPIN who are at increased risk of cancer and who require re-biopsy at a shorter interval. As noted above, AMACR is a potential candidate, as men with at least one gland with HGPIN with positive staining for AMACR are 5.2-times more likely to have a subsequent biopsy diagnosis of prostate cancer compared with AMACR-negative HGPIN . Other potential candidates include prostate tumour overexpressed gene 1 (PTOV1)  and adenomatous polyposis coli (APC) methylation status .
From these studies, there are several observations and recommendations for the further management of men in whom a diagnosis of HGPIN is made on TRUS biopsy in the absence of carcinoma [2,16]:
Low-grade PIN is usually not reported due to a lack of reproducibility, and these lesions have no higher risk of carcinoma on re-biopsy compared with a benign diagnosis [2,23]
The finding of HGPIN in one core is generally not associated with a higher rate of detection of carcinoma compared with a benign biopsy in extended biopsy protocols [12,13]
. Lefkowitz et al
showed the rate of detection of carcinoma at a 3-year interval is higher than the rate of detection at 1 year in this group. Until further large prospective studies are completed, these findings suggest repeat biopsy after a diagnosis of HGPIN in one core should be conducted at 3 years rather than 1 year in the absence of other clinical indications.
The detection of multifocal HGPIN (two or more cores) is generally associated with a higher rate of cancer detection of at least 30%, and consideration of re-biopsy is recommended in 12 months. For widespread HGPIN (four or more cores with HGPIN) the rate of cancer detection is even higher at 39% at 3 years, and re-biopsy at a shorter interval of 3–6 months may be more appropriate.
In some cases of HGPIN, there are a few adjacent atypical cells in which carcinoma cannot be excluded. This is a difficult situation and may represent either tangential sectioning of HGPIN or alternatively represent an early stage in the development of carcinoma. It is recommended that these men should be managed in the same manner as men with an atypical gland focus, with consideration of re-biopsy ≤6 months, specifically targeting the site of glandular atypia. In this situation, it is expected that a subsequent diagnosis of carcinoma will be made in >40% of cases [2,24]
Finally, while cancer will most commonly be diagnosed in the same or adjacent sextants as the HGPIN, it can be found in the contralateral lobe in 10–40% of cases [2,17]
. Therefore, while targeting the site of the previous diagnosed HGPIN on re-biopsy is recommended, the entire prostate gland should be sampled.
HGPIN is rarely diagnosed in TUR prostate specimens, with an incidence of 2.8% of isolated HGPIN, or 4.2% including those cases with associated cancer . The risk of developing cancer in these men was 21% over 7 years. Therefore, when HGPIN is identified, the entire TUR specimen is usually all processed and examined histologically to exclude associated carcinoma. Further, it is recommended that younger men in whom HGPIN is identified at TUR prostate, should have TRUS biopsies of the peripheral zone to exclude associated carcinoma, while older men with a normal PSA concentration require clinical follow-up only.