European Urology

Specimen submission, gross description, and site designation

Concurrent with the rise of PSA screening and increasingly sensitive imaging techniques, the average number of prostate needle biopsy cores has risen from 2 to 6 to 12 over the past 20 yr [5], [6], and [7]. With this expansion, the primary purpose of needle biopsy has shifted from the targeting of specific areas of concern on rectal examination to the systematic mapping of the gland for cancer involvement and quantity [7]. In modern practice this information is routinely used to determine (1) the indication or lack of indication for any form of therapy or follow-up, (2) the type of therapeutic options offered to the patient, (3) the extent of resection (ie, nerve sparing or not) for patients opting for surgery, and (4) the nature and dosing of radiation therapy.

Given the import of these results, the submission, handling, and description of biopsy cores assume clinical significance. Whether needle cores are submitted in two containers (right and left sides) or in separate containers with specific site designations (eg, right lateral apex, right lateral mid, right lateral base) is not uniform among urologists or institutions. However, the potential importance of knowing the specific location of the biopsy, and therefore the location of cancer, is well recognized; this information allows for avoidance of diagnostic pitfalls (eg, normal anatomic structures, such as the prostatic central zone and seminal vesicles [base biopsy], that may mimic prostatic intraepithelial neoplasia [PIN] or cancer, respectively) [8], detailed correlation with clinical and imaging studies, and effective treatment planning [7] and [9]. In addition, a number of studies have correlated the presence and amount of cancer in different regions with risk of higher pathologic stage and margin positivity [10]. On a practical level, processing and pathologic assessment of needle biopsies are greatly facilitated if biopsies are separated. Less material is lost when cutting single biopsies; reading biopsies one by one is easier and facilitates identification of minimal foci of cancer [11]. Therefore, when cores are submitted separately or assigned a clear site designation by container, the pathology report should reflect this labeling. Some urologists place multiple cores into each container and attempt site designation based on inking of each core. However, this practice may result in fragmentation and/or nonevident ink on the cores, so it is not recommended for site-specific designation.

Gleason grading: background and historical context

The modern system for grading prostatic adenocarcinoma emerged from work by Donald F. Gleason in the 1960s based on a specimen cohort from the Veterans Administration Cooperative Research Group [12]. Nearly 50 yr later, the Gleason grading system remains novel in that it is based on the architectural pattern of the tumor alone (Gleason patterns 1–5); the sum of the two most common patterns—that is, primary Gleason pattern plus secondary Gleason pattern equals Gleason score (GS)—conveys the most clinical meaning. While additional morphologic descriptors were added to patterns 3, 4, and 5 in subsequent 1974 and 1977 publications [13] and [14], all these observations emanate from an era in which PSA screening was nonexistent, most patients presented with palpable and/or advanced disease, and prostatic tissue was typically obtained from transurethral resection (TUR) specimens or other large specimens.

With the introduction of PSA screening, as well as the advent of thin-needle biopsy techniques and expanded sampling over the last two decades, it has become necessary for pathologists to diagnose and grade PCa on smaller and better characterized samples. As a result of increasing volumes coupled with the importance assigned to Gleason grading in modern predictive models [3] and [4], pathologists have garnered much experience in the application of the Gleason grading system. Not surprisingly, this has led to gradual evolution in practice.

A well-documented example of this phenomenon is the group of lesions formerly diagnosed as GS 1 + 1 = 2. It is now well known that many such cases would have benefited from modern immunohistochemical staining with basal cell markers and today might be classified as adenosis (adenomatous hyperplasia), a benign entity [15]. Additional difficult and/or underrecognized areas in Gleason grading, such as what pattern to assign to small to medium-sized cribriform glands, poorly formed glands, and variants of carcinoma, have also been encountered.

In 2005, the International Society of Urologic Pathology (ISUP) convened a conference on Gleason grading to address emerging issues in the field based on existing data, as well as the personal and institutional experience of a large international group of urologic pathologists. The resulting manuscript, “The 2005 International Society of Urological Pathology (ISUP) Consensus Conference on Gleason Grading of Prostatic Carcinoma,” serves as a provisional diagnostic guide to modern Gleason grading [16]. Importantly, the modifications to Gleason grading codified in the 2005 ISUP paper represent collective changes introduced over the course of the 1990s and early 2000s based on much-expanded experience with assessment of prostatic needle biopsy and RP specimens. A few publications since 2005 have addressed morphologic findings for which limited to no literature existed [17], [18], and [19].

Needle biopsy Gleason grading: usual scenarios

Nearly all prostatic carcinomas seen in needle biopsy specimens are of the usual (acinar or conventional) type, to which the Gleason grading system may be applied. Gleason patterns 1–2 (GS 2–4), which require nodular circumscription as a diagnostic criterion, are not easily evaluable in the limited tissue of needle biopsy. In light of poor correlation with prostatectomy grade and reproducibility among experts, GS 2–4 are for practical purposes not diagnosed in these specimens [15] and [16]. Conversely, Gleason pattern 5, including single cells, sheet of cells, and comedocarcinoma, is essentially unchanged from its original descriptions [12], [13], and [14]. Overall, the 2005 ISUP recommendations convey a significant contraction of Gleason pattern 3 and a consequent expansion of Gleason pattern 4, with Gleason pattern 3 typically the lowest assigned grade. The most profound impact of these changes has been on grading of prostatic needle biopsies, with GS 7 now being the most commonly assigned score in many settings [20] and [21].

In modern terms, discrete and well-formed, infiltrative glands—even when small—have been retained within Gleason pattern 3 (Fig. 1). In contrast, practice patterns have diverged with regard to cribriform glands with rounded contours, as well as ill-defined glands with poorly formed lumina, originally considered Gleason pattern 3 [16]. A percentage of small to medium-sized cribriform lesions label with basal cell markers and are better recognized today as cribriform high-grade PIN [22]. Of many images presented to ISUP experts in 2005, only rounded, well-circumscribed glands having the same size as normal glands, as well as evenly spaced lumina and cellular bridges of uniform thickness, were diagnosed as pattern 3. A subsequent study in which 10 well-known uropathologists were asked to grade a highly selected set of images thought to be representative of cribriform Gleason pattern 3 found that nearly all cases were considered Gleason grade 4 [18] (Fig. 2). In routine practice, therefore, cribriform glands, regardless of size, are nearly always diagnosed as pattern 4.

Fig. 1 Gleason pattern 3: small to medium-sized discrete acini with focal tangential sectioning.

Fig. 2 Medium-sized cribriform gland with somewhat irregular luminal spaces (on left) that would be assigned Gleason pattern 4.

A related feature of PCa is glomerulations or glomeruloid structures, characterized by dilated glands with intraluminal cribriform structures, a morphology not accounted for in the original Gleason system. While the 2005 ISUP group did not reach consensus on this histology, a recent study reported that 45 biopsies with glomerulations showed an association with Gleason pattern 4 cancers in the same biopsy in >80% of cases [19]. This evidence, along with the significant morphologic overlap with, and occasionally observed transitions to, cribriform Gleason pattern 4 carcinoma, favors classifying glomerulations as pattern 4 (Fig. 3).

Fig. 3 Glomerulations demonstrating significant morphologic overlap with and transition to cribriform Gleason pattern 4 carcinoma.

The 2005 ISUP conference also highlighted the controversy surrounding classification of “ill-defined glands with poorly-formed glandular lumina” (Fig. 4). While there is some consensus that such foci should be graded as pattern 4, this morphology represents a significant challenge for the Gleason grading system, with few instructive images in the existing literature. The ISUP panel cautioned that a “cluster of ill-defined glands in which a tangential section of pattern 3 glands cannot account for the histology” would be diagnosable as Gleason pattern 4 [16], a determination that in many cases necessitates evaluation of multiple levels and sections of such glands.

Fig. 4 Gleason score 3 + 4 = 7 carcinoma. Note multiple poorly formed glands with ill-defined lumina and/or incomplete nuclear complement.

Needle biopsy Gleason grading: special scenarios

Although Gleason grading is and always has been fundamentally based upon a sum of the first and second most common patterns, uropathologists have evolved reporting strategies for some specific scenarios in which (1) morphologic patterns are not well addressed within the original Gleason system, (2) the classic grading might not be clinically precise, and (3) the patient has received prior therapy. While some of these recommendations are consonant with the original Gleason system, the method of applying these rules has been clarified over time. Table 2 and Table 3 summarize these recommendations.

Needle biopsy Gleason grading: prostate cancer variants

The Gleason grading of a number of variants has been modified from the original system, as reflected in Table 2. The group of mucin-related tumors is a controversial and evolving area within Gleason grading. Carcinomas associated with extravasated mucin (either focal or abundant) and/or mucinous fibroplasia present a diagnostic challenge because of significant distortion of tumor architecture [23]. In the biopsy context, it is difficult to evaluate true mucinous (colloid) carcinoma, which requires the presence of >25% mucin pools for its diagnosis [24]. However, carcinomas with mucinous features, typically composed of irregular cribriform glands in a mucinous background, may be diagnosed. Such cases may also show individual glands in the same background or simulated “gland within gland” patterns representing single distorted acini and caused by encroachment of acellular mucin in and adjacent to neoplastic glands (Fig. 5). A similar finding is carcinoma with mucinous fibroplasia (collagenous micronodules), indicating the delicate ingrowth of fibrous tissue in and among glands, which may result in fused- or cribriform-appearing glands.

Fig. 5 Carcinoma with mucinous features. Note that although some truly fused glands (pattern 4) are present, much of the cancer consists of discrete glands (pattern 3) with varying degrees of distortion by extravasated mucin.

While colloid carcinoma with cribriform glands is routinely graded as 4 + 4 = 8 in RP specimens, there is no consensus regarding cases with discrete glands in a background of extravasated mucin or mucinous fibroplasia. At the 2005 ISUP conference, some attendees suggested that the mucin or mucinous fibroplasia be extracted and the underlying architecture graded [16]. As such, a percentage of these cases would be graded as GS 3 + 3 = 6. While early studies of mucinous carcinoma from the pre-PSA era showed associations with Gleason patterns 4–5 nonmucinous PCa and adverse outcomes [24] and [25], more recent studies have recognized the variability in the epithelial component of mucin-containing carcinomas and have reported no death from disease and limited biochemical recurrence without clinical evidence of local or distant recurrence in patients who had mucinous carcinoma treated by RP [26] and [27]. In theory, the latter results support grading based on the architectural configuration rather than by the presence of extracellular mucin in these tumors, a finding that should be confirmed in larger series.

Increasing clinical precision with Gleason grading on needle biopsy

There are certain circumstances in which reporting primary plus secondary Gleason grades may be inexact, as the traditional GS is unlikely to be representative of cancer in the gland (Table 3). For instance, in the context of abundant high-grade cancer, lower-grade patterns should not be incorporated in the GS. If the pathologist encounters a 15-mm core with 13 mm of cancer in which 0.5 mm displays Gleason pattern 3 and the remainder is Gleason pattern 4, the pathologist should diagnose GS 4 + 4 = 8 [16]. Conversely, any amount of high-grade tumor should be included in the GS, as it very often reflects more significant high-grade tumor in the gland. Hence, a 15-mm core with 13 mm of cancer in which 0.5 mm displays Gleason pattern 4 and the remainder is Gleason pattern 3 should be diagnosed as GS 3 + 4 = 7 [16]. To apply the second rule correctly, the possibility of tangential cuts of pattern 3 glands masquerading as fused or poorly formed glands must be excluded.

Although Gleason himself noted the presence of more than two patterns in approximately 50% of RP specimens, quantitating this phenomenon in needle biopsy has been more difficult, with the few existing reports suggesting an incidence of 1.5–4% [28] and [29]. How to address tertiary Gleason patterns in the biopsy context is controversial, as any incorporation of the third most common pattern is by definition contrary to Gleason's original approach [12]. Nonetheless, in some cases, the pathologist encounters a core with three patterns of cancer—most typically, patterns 3, 4, and 5 (eg, 3 + 4 = 7 or 4 + 3 = 7 with a minor component of 5). As an example of a pathology recommendation based on empirical experience and heavily influenced by clinical practice, the ISUP group recommended that such cases be classified overall as high grade (primary grade plus highest grade) because of the possibility that the highest grade is a more significant component in the gland. Consequently, the highest grade would be used by clinicians when assessing risk using a variety of predictive models, which only allow for two grades. For example, a core with 10 mm of cancer that is composed of 65% pattern 3, 25% pattern 4, and 10% pattern 5 would be diagnosed as GS 3 + 5 = 8 [16]. A subsequent study has supported this “first plus worst” approach, finding earlier time to, and higher percentage of patients with, biochemical recurrence in patients with GS 7 with tertiary pattern 5 compared with GS 7 alone [29].

When cores are submitted in separate containers and/or have a clearly designated location, it is presumed that the urologist–oncologist will be interested in this information for treatment planning. Therefore, the pathologist should assign a separate GS to each sampled core, rather than an overall or averaged score for the entire biopsy session [16], [30], and [31]. Such practice avoids weakening the predictive power of the highest GS (eg, one core showing 4 + 4 = 8 and multiple cores showing 3 + 3 = 6; overall GS would assign 3 + 4 = 7) and is buoyed by studies demonstrating that the highest GS in a specimen correlates with grade and stage at RP [32] and [33]. There is no uniform manner of grading cores of differing GS in the following contexts: (1) multiple cores in one container without site designations and (2) multiple fragmented cores in one container even with site designation. These settings are problematic, as the relationship of each core or fragment to another is unclear, and the potential for overgrading is increased. So as not to impose a seemingly precise assessment in an inherently imprecise scenario, logic dictates that the pathologist would assign an overall GS in these cases.

Grading irradiated cancer

Radiation therapy (external beam and/or brachytherapy [“seeds”]) is commonly used to treat clinically localized or locally advanced PCa. In the setting of a rising PSA after radiation therapy, a biopsy is typically performed in an attempt to distinguish local (in the prostate) recurrence from metastatic disease and for histologic confirmation if a salvage RP is to be attempted. Occasionally, the pathologist will not be informed of a history of radiation therapy, so it is essential to recognize the changes in benign and malignant glands that occur with this intervention, which have been well described elsewhere [34].

While in the past, benign tissue with marked therapeutic effect may have been diagnosed as atypical, increased recognition that these changes are therapy related has aided pathologists in their correct identification. Cancerous foci exhibiting profound treatment effect secondary to radiation typically display infiltrative, poorly formed glands or single cells with moderate to abundant vacuolated clear cytoplasm and prominent nucleoli [34] (Fig. 6). When only irradiated cancer is seen, the case may be signed out as “adenocarcinoma with profound treatment effect” and not graded. When usual-type adenocarcinoma is solely present after therapy, such that the observed cancer is indistinguishable from that of a patient who had not received radiation, the cancer is graded. Although not codified in the literature, in cases in which both gradable cancer and cancer with treatment effect are seen, a reasonable approach is to assign a GS and add a note stating, “The assigned Gleason score reflects the gradable portion of the carcinoma (%); the remaining cancer shows profound treatment effect.”

Fig. 6 Adenocarcinoma with profound radiation treatment effect. Note poorly formed glands and single cells with vacuolated clear cytoplasm.

Determining whether gradable cancer is present is crucial for clinical management, as studies of postradiation biopsies with 10-yr follow-up indicate that the biochemical recurrence-free and distant failure rates for patients having only cancer with profound treatment effect are similar to the rates for patients with benign biopsies, as opposed to patients with gradable cancer [35]. Said another way, the presence or absence of gradable cancer in a biopsy after radiation therapy is a major indicator of clinical outcome.

Needle biopsy Gleason grade as a measure of risk

Independent of these recent modifications, accumulated evidence from >40 yr of application has shown the biopsy GS to be the most significant predictor of pathologic outcomes at RP, as well as one of the key predictors of clinical outcomes after RP and radiation therapy [4], [21], [36], [37], [38], [39], [40], and [41]. Furthermore, GS on needle biopsy may be used to determine therapeutic choices, the extent of neurovascular bundle resection, or performance of a pelvic lymph node (LN) dissection. Consequent to the evolution previously described, the value of grouping GS (ie, GS ≤6, 7, 8–10), such that each group behaved worse than the group below it, was recognized. Further substratification of GS 7 based on primary grade (ie, GS 3 + 4 = 7 vs GS 4 + 3 = 7) has also been shown to influence pathologic and clinical outcomes [42], [43], and [44] and is routinely reported. It has also been demonstrated that tumors with biopsy GS 9–10 are associated with a much worse prognosis than tumors with GS 8, such that GS 8–10 might not be considered a homogeneous group [45]. As Gleason grading is incorporated into every predictive tool that has been designed for PCa [4], [39], and [40], the accurate and reproducible application of this system has clinical meaning. In the past two decades, major educational and media efforts, including Web sites, publications, and courses, have resulted in significantly better correlation for GS on biopsy among community and academic pathologists, such that >80% of cases showed exact agreement in a recent large series [37].

Few formal studies have evaluated the impact of the 2005 ISUP conference [20], [21], [46], [47], [48], and [49], and the studies that have done so were small cohorts with limited follow-up. As many of the “changes” or “consensus” positions represent modifications by groups of pathologists over time, such an exercise may also not be fruitful, as using 2005 as a dividing line between an “old” and a “new” system may be biased and inaccurate. While these studies have generally documented a higher percentage of biopsy specimens with GS ≥7 in cohorts after 2005 and a somewhat improved biopsy–prostatectomy GS correlation and prediction of biochemical-free progression after RP, more robust data with long-term follow-up addressing these questions are awaited.

Extent of involvement

For core biopsy specimens, the absolute number of cores examined and involved is routinely reported. In cases with one core submitted per container, this assessment is simple. In the event of multiple cores per container, the degree to which tissue fragmentation has taken place will affect this determination.

Once a diagnosis of cancer has been rendered for a given core, there are multiple measures of tumor quantification that have been reported to correlate with pathologic grade and stage, as well as predict biochemical recurrence [50], [51], [52], and [53]. Many of these evaluations are tedious, are not routine in contemporary practice, and may add little to the predictive accuracy of more simple measurements. However, there is overwhelming consensus that in addition to the number of cores involved, some quantitation of tumor extent on a per-core basis should occur, whether by visual estimation of linear extent in millimeters, percentage of the core involved, or both [54]. The latter may be reasonable, given that a variety of clinical nomograms and protocols use different measures. When multiple cores are submitted in the same cassette, there is a higher likelihood of fragmentation [11], such that it may be most prudent to report the percentage of the overall fragmented specimen involved by cancer in these cases.

Within a given core, foci of cancer may be present continuously or discontinuously along the length of the specimen. In the former case, length in millimeters, percentage of core involvement, or both are readily assessed. When multiple foci of carcinoma are separated by intervening benign prostatic glands and stroma, some pathologists will “collapse” the tumor by disregarding the intervening tissue [50], while others will measure the farthest distance between the outermost foci and report the entire length or percentage as if there were one unbroken focus (eg, three small foci of carcinoma discontinuously involving 80% of the core) [55]. This method may result in vastly differing tumor quantitation, which may affect nomogram predictions or eligibility for active surveillance. Two contemporary studies of this specific issue convey different findings. The first study showed that in cores with discontinuous foci of cancer separated by ≤5 mm of intervening benign stroma, different methods used to estimate cancer length have equal prognostic significance [50]. In contrast, a recent report has suggested that for cancer-bearing cores in which the needle biopsy GS is reflective of the entire tumor in the RP specimen, quantitating discontinuous foci as one unbroken focus correlates better with pathologic outcomes [55]). Given the limited evidence, it is not possible to draw a definitive conclusion at this time.

Perineural invasion

Perineural invasion, defined as cancer tracking along or circumferentially around a nerve, is a relatively ubiquitous finding in RP specimens. In needle biopsies, an incidence between 11% and 38% has been reported in large cohort series [56]. There appears to be functional bidirectional communication between nerves and prostatic carcinoma cells accounting, at least in part, for perineural growth, which is a major route of extraprostatic extension [57]. However, there has been significant controversy over the past two decades as to whether this finding on needle biopsy predicts extraprostatic spread at RP and/or biochemical recurrence after therapy (surgical or radiation) [58], [59], [60], [61], [62], [63], [64], and [65]. Reviews by Bismar et al. and Harnden et al. reveal that while most studies find perineural invasion to be predictive of extraprostatic extension in univariate analysis, its importance is not retained once PSA, clinical stage, and biopsy GS (common preoperative parameters) are considered in multivariate analysis [51] and [56]. Similarly, there are conflicting data as to whether perineural invasion predicts recurrence after surgery or radiation therapy. Importantly, the meta-analysis by Harnden et al. has shown that studies that analyzed perineural invasion in specific patient groups stratified by PSA levels, clinical stage, GS, and/or biopsy tumor extent have found it to be an independent prognostic factor [56].

It is also clear from these collective studies that surgeons react in different ways to a report of perineural invasion on needle biopsy. While some groups initially considered this finding an indication to abandon nerve-sparing surgery, recent data suggest that bilateral nerve sparing may be performed without compromising oncologic efficacy in the majority of patients [60]. Taking into account the relative ease of identifying perineural invasion, its proposed significance in at least some patient groups, and the variation in therapeutic choices among centers and clinicians, many pathologists routinely report this finding.

High-risk lesions and putative precursors: small foci of atypical glands, suspicious for carcinoma

Atypical small acinar proliferation (ASAP) and small focus of atypical glands, suspicious for carcinoma are terms that refer to a gland or focus of glands that is suspicious for cancer but lacks sufficient architectural and/or cytologic atypia for a definitive diagnosis. This diagnosis is rendered in <5% of specimens in large cohorts [66] and [67]. If used correctly, these terms reflect the pathologist's uncertainty as to whether a given glandular focus can be assigned a cancer diagnosis. It is therefore important that ASAP/small focus of atypical glands, suspicious for cancer not become a “wastebasket” diagnosis subsuming a large spectrum of lesions. Rather, this diagnosis should be a last resort—one in which the pathologist, after careful consideration using hematoxylin and eosin stain criteria and ancillary immunohistochemical studies as appropriate, is unable to arrive at a definitive benign or cancer diagnosis.

There are many reasons for a finding of ASAP. Common among them is the inability to identify architectural features (clustered, infiltrative glands) and/or cytologic (nuclear, nucleolar, cytoplasmic) features to establish a definitive cancer diagnosis. Some of the more common struggles include atypical glands that are few in number, foci with procedural-related crush or fragmentation artifact, crowded glands with minimal cytologic atypia, glandular foci associated with significant inflammation, and small acinar foci in which outpouching/tangential sectioning of high-grade PIN cannot be distinguished from limited cancer adjacent to PIN [68]. All speak to the need for examining multiple levels and sections of the biopsy.

It is important to recognize atypical foci suspicious for cancer in prostatic needle biopsies because of their association with cancer on repeat biopsies [66], [69], [70], [71], [72], and [73]. In this sense, ASAP may be seen as a risk factor for the subsequent finding of cancer, with the existing literature reporting an average 40.2% (median: 38.5%) risk of cancer following this diagnosis, a rate that has been stable for nearly two decades [66]. In some cases, a focus of atypical glands is closely associated with a focus of high-grade PIN, a phenomenon that seems to carry a risk of cancer on repeat biopsy similar to ASAP [74]. Regardless of personal or group practice regarding terminology, it is incumbent upon pathologists to communicate to their colleagues the clinical import of these findings so that appropriate follow-up, in the form of early repeat needle biopsy, may be performed. Some experts have recommended that repeat biopsy concentrate more sampling in the area of the initial atypical site, with relatively less sampling elsewhere in the gland [66]—a policy that is more easily carried out when the initial biopsy contains specific site designations.

High-risk lesions and putative precursors: high-grade prostatic intraepithelial neoplasia

Although PIN was first described by McNeal in the 1960s, formal characterization did not occur until the late 1980s, when it was first termed intraductal dysplasia and quickly evolved to prostatic intraepithelial neoplasia[75]. Current evidence from a variety of sources has rendered high-grade PIN the only well-established precursor to prostatic adenocarcinoma [66], [68], and [76].

Morphologically, PIN describes architecturally benign prostatic glands lined by atypical cells. After initially being divided into three grades, PIN was more concisely classified as either low grade (approximating grade 1) or high grade (approximating grades 2–3), with prominent nucleoli being the primary distinguishing factor [75]. In the past two decades, however, it has become evident that (1) there is low interobserver reproducibility for a diagnosis of low-grade PIN, with even urologic pathologists having difficulty separating this entity from slight variations of normal prostatic glandular architecture, and (2) low-grade PIN does not convey a significantly increased risk of cancer in follow-up biopsy when compared with an initial benign diagnosis [66]. As a result, the diagnosis of low-grade PIN has largely faded from the pathology reporting spectrum such that a diagnosis of PIN today is understood to refer to high-grade PIN.

Recent reviews reveal a large range of incidence, from 0% to 24.6% (mean: 7.7%; median: 5.2%) on initial biopsy, with no apparent relationship between PIN detection and number of cores sampled, year of sampling, or academic compared with community practice settings [66] and [68]. This wide variation may be partially explained by the subjective nature of evaluating “cytologic atypia”—specifically, the presence of prominent nucleoli (how prominent? how many?)—as well as multiple histologic artifacts (thick sections, fixatives that enhance nucleolar detail). The difficulty in defining atypia is highlighted in the responses to a survey of 64 urologic pathologists that inquired as to how prominent/how many nucleoli are required for a PIN diagnosis. The answers were as follows: “any visible at 40× magnification,”“any visible at 20× magnification,”“any visible regardless of magnification,”“in >10% of secretory cells at 40× magnification,”“in >10% of secretory cells at 20× magnification,” and “in >10% of secretory cells regardless of magnification.” These answers represented 16%, 17%, 19%, 11%, 9%, and 13% of replies, respectively, clearly demonstrating the great variability in the application of this diagnosis [67]. These findings indicate the need for more specific diagnostic criteria to increase the reproducibility of a PIN diagnosis.

While the incidence of high-grade PIN does not appear to be dependent on the number of cores sampled, with studies in the 6- and 12-core eras showing similar variability in PIN detection [66], a significantly decreased incidence of cancer detection following a high-grade PIN diagnosis has been observed [77], [78], and [79]. While the literature reveals a huge span of cancer incidence after high-grade PIN diagnosis, ranging from 2.3% to 100%, a more careful look reveals an incidence of approximately 50% in studies from the 1990s, which dropped to approximately 20% after 2000 [66]. This change approximates the shift toward more extended biopsy schema, which is now the norm. Furthermore, recent studies that examined the risk of cancer on rebiopsy following a diagnosis of high-grade PIN compared with that following a benign diagnosis have shown no statistically significant differences [66] and [80]. This finding has led some practitioners to propose that early repeat needle biopsy is not required for men within 1 yr of a PIN diagnosis, especially if there is only one core with high-grade PIN [79]. When the initial biopsy is multifocal (more than one core with PIN), the risk of cancer on immediate repeat biopsy is approximately 40% [81] and [82] and justifies repeat biopsy within the first year. However, the long-term risk of cancer with unifocal high-grade PIN on initial biopsy remains unknown. Until data exist, it may be reasonable to perform a repeat biopsy between 1 and 3 yr later, as suggested by some groups [83] and [84].

High-risk lesions and putative precursors: intraductal carcinoma

Intraductal carcinoma is characterized by malignant epithelial masses conforming to the contours of often expanded native ducts and/or acini displaying basal cells. Early descriptions from RP specimens drew attention to the fact that in contradistinction to high-grade PIN, intraductal carcinoma is rare in areas away from carcinoma [85]. This dichotomy is also reflected in needle biopsies, in which intraductal carcinoma is rarely seen in the absence of invasive cancer [86]. Further studies revealed associations with high GS and tumor volume, as well as increased rates of extraprostatic extension, seminal vesicle invasion, and recurrence after prostatectomy [86], [87], and [88]. Based on this evidence, most experts have argued that intraductal carcinoma is part of the evolution of PCa (a late event) or, alternatively, an aggressive precursor (which may or may not arise from PIN) [87] and [89]. Recent follow-up series of needle biopsies containing exclusively intraductal carcinoma have shown that the overwhelming number have invasive cancer with GS >7 and pT3 disease at subsequent RP [86] and [90]. These associations reveal the critical importance of separating high-grade PIN from intraductal carcinoma on needle biopsy.

Diagnosing intraductal carcinoma may be difficult, as the description of this condition may overlap with that of high-grade PIN in a given case. A number of authors have proposed more specific architectural and cytologic features that they feel to be beyond what is acceptable for a PIN diagnosis [86] and [90]. The most commonly agreed-on criteria are intraductal foci with dense cribriform (more solid than cribriform) to solid masses with or without comedonecrosis. Although not always present, the other feature that has been repeatedly associated with intraductal carcinoma is marked nuclear atypia in the form of striking nucleomegaly, hyperchromasia, and/or overt pleomorphism (should be well beyond the increased nuclear size, hyperchromasia, and prominent nucleoli seen in high-grade PIN) (Fig. 7).

Fig. 7 Intraductal carcinoma: solid growth of malignant cells with marked nuclear atypia. Note the evident basal cells at multiple points in the periphery of the duct.

However, in the absence of these collective findings, the threshold for diagnosing this pattern of carcinoma is more blurred. For instance, some experts would not consider an intraductal carcinoma diagnosis in the presence of loosely cribriform or micropapillary intraductal lesions, always classifying them as high-grade PIN, while others would accept these architectures within the intraductal carcinoma spectrum only when accompanied by marked nuclear atypia [88] and [90]. In practice, the identification of rounded or circumscribed masses of malignant cells with complex architecture and/or obvious nuclear atypia and a preserved basal cell layer should raise the diagnostic possibility of intraductal carcinoma. Given the well-established correlation with high-grade, high-stage disease at RP, when detected, the presence of intraductal carcinoma should be noted in needle biopsy reports; some experts recommend definitive therapy when intraductal carcinoma is diagnosed on biopsy [86] and [90].

Specimen handling and sectioning

Most pathology laboratories receive RP specimens in formalin. However, with expanded emphasis on tissue procurement and snap-freezing of fresh tissues for molecular and genomic studies, pathologists increasingly receive prostatectomies without fixative [107]. This condition raises the possibility of altered protein, DNA, RNA, or gene expression, depending on the length of ischemia, a presumption that remains controversial [112], [113], and [114].

Once received, RP specimens are measured in vertical (apex-to-base), transverse (right-to-left). and sagittal (anterior-to-posterior) dimensions. A specimen weight is also determined, which most commonly conveys the weight of the prostate with seminal vesicles attached. Specimens are routinely inked to enable (1) accurate assessment of surgical margins and (2) accurate identification of laterality—when more than one ink color is used. Most pathologists use more than one color to facilitate orientation of the prostate [107].

While prostates from which fresh tissue will be harvested may be sectioned in the fresh state, sectioning is clearly facilitated when the gland is fixed, allowing for more uniform slicing. At the prostatic apex, evaluation of prognostically significant features, including the presence of tumor at the margin, necessitates evaluation of the entire convexity of the apical surface. The most effective method is taking a section approximately 3–5 mm from the most apical portion of the gland and then “coning” the resultant disk to submit the entire apex. A recent ISUP conference on RP reporting found that most urologic pathologists preferred sagittal coning (to ensure blocks of uniform thickness) over the radial cone method used in the cervix [107]. In this way, each coned fragment has one inked surface that reflects the true apical margin.

Assessment of the bladder-neck margin has clinical import, yet the optimal method for evaluation is much less clear. To report tumor in “bladder-neck” tissue at RP, one must see cancer glands in thick muscle bundles (detrusor muscle–like or muscularis propria–like) outside the prostate [95]. However, the degree of bladder-neck tissue resected with the specimen by the urologic surgeon may vary and is not easily visualized because of tissue contraction upon removal from the patient. Moreover, it has been demonstrated that detrusor-like muscle bundles continue over the anterior and lateral aspects of the prostate from base to midgland [115]. Hence, while most laboratories use a similar sectioning and coning method as in the apex [107], this protocol may not truly reflect the bladder-neck margin. These factors should be considered when reviewing the literature regarding bladder-neck margin positivity, as will be discussed further shortly. Although limited data are available, a viable alternative may be a thin, 1-mm shave or removal of the roughened muscular tissue immediately surrounding the urethra at the base to best approximate the true bladder-neck margin.

Sampling of the seminal vesicles is likewise vital to PCa staging. While there is general agreement that sections should be taken at the junction of the prostate and seminal vesicles bilaterally [110] to exclude the possibility of tumor invasion by direct extension, there is wide variability in how much, if any, of the remainder of the seminal vesicle should be sampled. Although no evidence exists for additional sampling, other routes of tumor spread to seminal vesicles by lymphovascular invasion or in conjunction with extraprostatic extension have been reported [116]. It may be reasonable, therefore, to consider some degree of enhanced sampling, for example, one additional section from the mid–seminal vesicle in addition to the junctional section.

Whether one partially or totally embeds the prostate is largely dependent on the nature of the institution and the institution's investment in research, tissue harvesting, and/or correlation with imaging studies such as magnetic resonance imaging. Regardless of the approach, the most diagnostically sound and clinically useful method is one that provides maximal information on grade, stage, and margin status. While there are many approaches for subtotal sampling, a decade ago Sehdev et al. compared 10 sampling techniques in patients with cT1c tumors with one or more adverse pathologic findings (eg, GS ≥7, extraprostatic extension, margin positivity) and described a method with comparable results to whole-gland submission. This method entailed embedding every posterior section and one midanterior section from both right and left. If either anterior section had a potentially dominant (by size) tumor, all anterior sections were submitted. This method detected >95% of adverse features [117] and represents a practical alternative for institutions not wishing to submit the entire gland. Centers opting for any subtotal submission of the gland should balance its benefits against the additional effort expended in keeping track of remaining tissue, subsequent embedding of additional blocks, dictating amended reports, and/or a delayed final diagnosis [107].

Multifocality

The tendency of PCa to develop in a multifocal fashion is well established, with reported rates between 60% and 90% in surgically removed glands [118]. Although the biologic basis for multifocality still requires clarification [119] and [120], this aspect of PCa has a significant impact on RP reporting, especially in assigning zonal origin, identifying the index or dominant tumor nodule, and grading and staging.

Zonal origin

Numerous studies have claimed that transition-zone tumors should be considered and reported separately from peripheral-zone tumors [121] and [122]. In part, these observations were based on a series of studies from the late 1980s and early 1990s in which investigators argued that transition-zone tumors could be identified using distinctive histology, including well-differentiated glands of variable size and contour, composed of tall cuboidal to columnar cells with clear to pale pink cytoplasm, basally oriented nuclei, and occasional eosinophilic luminal secretions [121]. These studies concluded that this “clear cell” appearance was a marker of transition-zone tumors, which were associated with a more indolent course, higher cure rate, and overall more favorable prognosis [122]. Paradoxically, while transition-zone tumors may be of larger volume and associated with higher serum PSA values than peripheral tumors, most reports have maintained that transition-zone tumors show lower GSs [123] and [124]. Few studies, however, compared tumors arising in the transition zone with tumors arising in the anterior peripheral zone, the predominant glandular tissue of the apical prostate.

A recent large-scale histopathologic analysis of 197 anterior dominant tumors, in which zone of origin was determined using an anatomy-sensitive approach emphasizing the variability in anterior prostatic anatomy from apex through base, showed that the majority of dominant anterior tumors in the prostate are actually of anterior peripheral–zone origin. In comparing 97 cases of anterior peripheral–zone origin and 70 cases of transition-zone origin, no significant differences in GS, incidence of extraprostatic extension, or overall surgical margin positivity rate were observed [125]. Other groups have also reported no significant differences in GS between transition- and peripheral-zone tumors [126] and [127]. Therefore, while it is important to recognize anterior prostatic tumors, which represent an increasing percentage of dominant lesions [125], there is less definitive evidence at this time to specify zone of origin in the pathology report.

Defining the index, or dominant, tumor

The notion of an index, or dominant, tumor was originally proposed at Stanford University by McNeal and Stamey, who measured the volume of the largest tumor nodule and correlated this volume with outcome [128].While empirical experience and logic dictate that the dominant nodule by size will most often be associated with the highest GS and will be the stage-determining lesion, up to one-third of cases may not conform to this rule [118]. While it is relatively easy to report a dominant nodule location in the former case, some studies have questioned the prognostic significance of this data element. Furthermore, in the one-third of cases in which size, grade, and stage do not converge in a single tumor nodule, there is no consensus as to how the index lesion should be designated [108]. This situation, of necessity, will lead to diagnostic challenges in Gleason grading and staging, as will be described.

Grading of specimens with separate tumor nodules

While the general principles, historical background, and recent modifications in morphology in regard to Gleason grading are equally applicable to biopsy and RP specimens, a number of reporting elements specific to prostatectomy remain. One such element is the grading of cases with separate tumor nodules. This phenomenon is best illustrated with two examples. The first example is a gland with multiple tumor nodules in which the largest nodule has GS 4 + 4 = 8, while multifocal smaller nodules with GS 3 + 3 = 6 are also present. Assigning an overall GS in such a case may result in a diagnosis of 4 + 3 = 7 or even 3 + 4 = 7, depending on the extent of the multifocal disease. In light of limited data, the 2005 ISUP conference [16] recommended assigning a separate GS to each dominant tumor nodule. In this case, the reported GS would only reflect the dominant nodule by size—that is, GS 4 + 4 = 8—without the need to record smaller foci of lower-grade tumor.

A second example of a challenging scenario is a prostate gland with multiple tumor nodules in which the largest nodule has GS 3 + 3 = 6, while a smaller nodule shows GS 3 + 4 = 7. Here, grading on the basis of the dominant nodule by size alone may underestimate the biologic potential of the tumor. Hence, the ISUP group recommended reporting two GSs—one for the largest nodule (ie, GS 3 + 3 = 6) and one for the nodule with the highest grade (ie, GS 3 + 4 = 7). This approach would lead to separate GSs for at most two nodules in the overwhelming majority of cases [16]. However, given the lack of evidence in the literature, we may posit that another reasonable approach in this case may be to assign one GS of 3 + 4 = 7, as this approach may be used in a more straightforward fashion by clinicians in prognostic nomograms. A similar strategy may be used when no dominant nodule is present and scattered small foci of GS 3 + 3 = 6 and 3 + 4 = 7 make up the tumor.

Tertiary Gleason grades

The definition of tertiary Gleason grade in RP specimens is not analogous to that of needle biopsy because (1) the entire tumor is available for examination and (2) the multifocal nature of PCa affects its assessment [16]. Technically, the extent of a tertiary component can vary from <1% to approximately 30%. While there is no consensus definition, a number of authors have used <5% higher-grade tumor (usually pattern 5), choosing to regard the highest pattern as the secondary pattern if it is more abundant than 5% [129], [130], [131], and [132]. A significant difficulty is imposed by the routine omission of tertiary grades in clinical management because of the presence of only two grades in existing nomograms. Nonetheless, recognition and assignment of tertiary grades in RP specimens are widely practiced, with data suggesting that GS 3 + 4 = 7 tumors with a minor component of pattern 5 have similar stage and risk of biochemical progression to GS 8 tumors, for example [130]. Interestingly, while RP with GS 4 + 3 = 7 tumors and a minor component of pattern 5 fare worse than GS 4 + 3 = 7 tumors, they are not akin to GS 4 + 5 = 9 tumors [130], underscoring the impropriety of adopting a “first plus worst” approach, as in needle biopsy specimens. More controversial is whether it is appropriate to assign tertiary patterns in cases with, for example, overwhelming Gleason pattern 3 and < 5% pattern 4. Although some practitioners have suggested that such tumors have a prognosis intermediate between GS 3 + 3 = 6 and 3 + 4 = 7 tumors [129], future large-scale cohorts will be needed to fully address these issues.

Grading after androgen ablation (hormonal) therapy

Since the prostate is an androgen-responsive organ and the androgen pathway plays a key role in the development of function of the gland, androgen-related molecules/enzymes are molecular targets for hormonal ablation therapy, especially in patients with advanced disease. Grading of treated cancers is included here because in some clinical environments, limited hormonal therapy may also be administered prior to an RP to reduce gland size. Therefore, as in biopsies after radiation therapy, it is important to recognize changes in benign and malignant glands introduced by hormonal ablation.

While benign tissue may exhibit glandular atrophy in the form of cytoplasmic diminution resulting in glandular lining cells that appear cuboidal and flat and relative basal cell prominence, as well as, occasionally, squamous metaplasia, stromal edema, and/or fibrosis [133], the profound treatment effect on cancer glands is more pronounced. Glands may have little cytoplasm and hyperchromatic, yet pyknotic, nuclei in which only their infiltrative growth is indicative of cancer (Fig. 8). Aggregates of cells with pyknotic nuclei and abundant xanthomatous cytoplasm, resembling histiocytes, as well as largely acellular mucin pools with rare floating single cells, may also be seen [133], such that positive immunohistochemical labeling for pancytokeratin and PSA, as well as negative basal cell markers/positive racemase, may be required to establish the diagnosis. Limited cancer with treatment effect may be a cause of understaging in such patients, and careful evaluation of the margins and extraprostatic tissue should be undertaken [133]. Reporting recommendations for grading are similar to those for radiation therapy, though whether a tumor shows profound treatment effect or not is of less clinical consequence, as hormonal therapy does not negate adverse outcomes in the long term.

Fig. 8 Adenocarcinoma with profound hormonal treatment effect. Note small glands and single cells with little cytoplasm and pyknotic nuclei in the lower half of the figure; benign glands show loss of cytoplasm, cuboidal appearance of secretory cells, and relative prominence of the basal cell layer.

Organ-confined disease: pT2 substaging

A controversial area of RP reporting that is still in evolution regards substratification of organ-confined disease [108]. With the advent of TNM staging in 1992, pathologic stage T2 PCas were assigned to one of three categories to parallel the clinical staging system: (1) pT2a for tumors occupying less than one-half of one lobe, (2) pT2b for tumors occupying greater than one-half of one lobe, or (3) pT2c for tumors involving both lobes. However, differences between the staging systems (digital rectal examination compared with pathologic evaluation of RP specimens) were evident, and using this substaging, few pT2b tumors were identified [134]. The pT system was simplified in 1997 to include pT2a (tumors confined to one lobe) and pT2b (bilateral disease) [135]. This modification created the illogical circumstance in which bilateral small foci of disease could receive a higher stage than a unilateral large lesion [108]. Limited clinical utility and correlation with clinical staging led to reversal of the 1997 TNM in its 2002 and 2009 iterations, such that three subcategories are now listed [136]. However, the past decade has yielded a number of studies showing that pathologic substaging of organ-confined disease by any of the previously mentioned systems lacks prognostic import [108], [137], and [138]. While many practitioners still report pT2 substages at this time, a consensus at the recent ISUP conference was that this practice was optional and should be modified in the future [108].

Extraprostatic extension

In pathologic terms, extraprostatic extension refers to the presence of tumor beyond the borders of the gland. While this terminology may convey ease in application, the reality of determining extraprostatic extension in practice is highly dependent on anatomic location and the presence of desmoplastic reaction to tumor and/or biopsy-related changes. The basic boundary of the prostate is a condensed fibromuscular layer of prostatic stroma rather than a true, epithelium-lined capsule. Early observations showed that although the boundary was usually intact in the posterior and posterolateral aspects of the gland, this was not the case in the apex, anterior, or bladder-neck regions [139]. Moreover, even in regions with a well-defined edge to the prostate, tumor- or biopsy-related fibrous change may cause difficulty in evaluation, making assessment of extraprostatic tumor spread quite challenging. Not surprisingly, interobserver variability studies among pathologists targeting extraprostatic extension report the most variation in areas and cases without clear anatomic landmarks [140].

The most easily recognizable sign of extraprostatic extension is tumor admixed with periprostatic fat. In the posterolateral prostate, a pT3a stage may also be assigned to tumor identified within loose connective tissue and/or perineural spaces of the neurovascular bundles and, when present, to distinct tumor nodules within desmoplastic stroma that bulges beyond the prostatic contour [109]. The latter may be diagnosed after visually tracking along the edge of the prostatic stroma to confirm interruption of the edge of the prostate. In the absence of clear histologic boundaries in the apex, anterior, and bladder-neck regions, such evaluations are imprecise. There is debate as to whether extraprostatic extension can be diagnosed at the apex and how to separate this finding from apical margin positivity [139] and [140]. The current convention is to designate a tumor as organ-confined at the apex as long as tumor is not present at the inked margin. If tumor is present at the apical margin, it is staged as pT2+ because of the vague boundaries of the prostate in this area such that distinguishing a positive margin in an area of intraprostatic incision versus an area of extraprostatic extension cannot be done. The presence of skeletal muscle (apex) and blood vessels (apex through base) in both the anterior stroma and the anterior extraprostatic space, coupled with blending of the prostatic stroma with extraprostatic smooth muscle bundles (mid-gland to base), leaves invasion into or at the level of adipose tissue as the most reasonable diagnostic feature of extraprostatic extension in the anterior prostate [141].

Once the presence or absence of extraprostatic extension has been established, some method of quantitation is routine [109]. The two most common approaches are those of Epstein et al. [142] and Wheeler et al. [143], both of which distinguish “focal” from “established” extraprostatic extension. The former defined focal as a few neoplastic glands just outside the prostate, with any more glands being established; the latter defined focal as extraprostatic tumor occupying less than one high-power field in no more than two sections and established as any degree more than that [142] and [143]. Using these subjective, yet readily applicable, criteria, clinically meaningful separation of pT3a patients can be achieved. Extraprostatic extension is a significant parameter in nearly all postoperative predictive tools in use today [4], [39], and [144]. Although pathologists typically report the location or locations of extraprostatic extension, this parameter has no known prognostic significance in the absence of a positive margin at the site.

Bladder-neck invasion

In a significant change from prior versions, the 2009 TNM classification categorizes microscopic bladder-neck invasion as pT3a, rather than together with gross invasion (pT4) [136]. This change represents the culmination of a decade of work in which the clinical significance of microscopic bladder-neck invasion was challenged [109]. The overwhelming number of studies found that usual grading and staging parameters, but not microscopic bladder-neck invasion alone, were independent predictors of progression or that patients with this finding have a greater likelihood of 3- and 5-yr progression-free survival than patients with seminal vesicle invasion [145], [146], and [147].

While this finding is now the consensus in the urologic pathologic community, there exists some variability in defining microscopic bladder-neck invasion based on the specimen-handling considerations previously highlighted. All prior studies have called microscopic bladder-neck invasion when malignant cells or glands invade thick smooth-muscle bundles in the bladder-neck section [109]. However, equating this finding with bladder-neck margin positivity will depend on whether specimens are coned (not equated) or shaved (equated). Importantly, microscopic bladder-neck invasion should be distinguished from tumor intermixed with benign prostatic glands in the bladder-neck section; this finding may represent either a false-positive margin due to the pathologist's obtaining a shave margin that is too thick or a true-positive margin in an area of intraprostatic incision by the surgeon.

Seminal vesicle invasion

Tumor infiltration of the muscular wall of the seminal vesicle is a well-established adverse prognostic feature in PCa [110] and [148]. Nearly two decades ago, Ohori et al. studied a cohort of patients with seminal vesicle invasion and described three routes of spread from the prostate: (1) direct spread along the ejaculatory ducts at the base, (2) extraprostatic extension into periseminal vesicle soft tissue with ensuing seminal vesicle invasion, and (3) discontinuous spread (in cases in which no prostatic base tumor was identified) [116]. While it is possible that the latter may reflect lymphovascular invasion, the distinction of seminal vesicle invasion types is not routine reporting practice [105] and [110].

There are three significant caveats regarding assessment of seminal vesicle invasion. The first is in assessment and staging of tumor invading periseminal vesicle soft tissue. While early studies designated these tumors within the rubric of “seminal vesicle invasion,” this finding is currently staged as pT3a (extraprostatic extension). Second, in the unusual case in which tumor is present in endothelial-lined lymphovascular spaces within the seminal vesicle wall alone, without overt muscular wall invasion, there is no consensus as to whether pT3a (tumor beyond the prostate) or pT3b (seminal vesicle invasion akin to muscular wall invasion) should be assigned. Similarly, whether to diagnose pT3b disease when tumor is seen at the ejaculatory duct–seminal vesicle junction in prostatic sections is controversial. Two approaches follow: (1) Only diagnose pT3b when tumor is seen in the extraprostatic seminal vesicle or (2) allow for the diagnosis of tumor invading the “base of the seminal vesicle,” but require that the seminal vesicle have a well-formed muscular coat and be topographically separate from prostatic glandular tissue [110]. Future studies in appropriately selected cohorts will be necessary to clarify these findings.

Lymph node metastasis

Pelvic LN dissection is the standard means for detecting LN metastasis in PCa [149], and LN metastasis is overwhelmingly associated with high-grade, high-stage, and large-volume disease [150]. Over the past three decades, the manner in which urologists and oncologists view the finding of LN positivity in relation to patient management has evolved. While finding LN metastasis on frozen section analysis was once an absolute contraindication to RP, more recent studies have suggested the curative potential of LN dissection [151], [152], and [153]. Coinciding with the advent of risk stratification tools that aid in selecting and avoiding intervention and the increasing popularity of minimally invasive surgery, evidence from groups such as CaPSURE suggests a steady decline in the performance of LN dissection, especially for patients in low- and intermediate-risk groups [149]. However, great debate still exists in the determination of which risk categories warrant LN dissection and the extent of LN sampling that should be routinely performed [149]. Complicating matters, two recent studies in large cohorts of LN-positive patients have shown that the typical limited sampling (external iliac LN only) detects only one-third of positive LNs and that 30–40% of LN metastases may occur contralateral to the dominant tumor in the prostate [150], suggesting that current trends in LN sampling may need reassessment.

As a result, some urologic surgeons still send frozen-section LN samples, though studies have revealed a relatively high false-negative rate [154], and pathologists may be asked to evaluate either no LN, a limited sampling (usually bilateral external iliac nodes) submitted as “right and left pelvic LNs,” or occasionally, more extensive sampling by LN packet (external iliac, obturator, and hypogastric). This situation leads to extreme variability in the average number of LNs identified [110]. While pathologists routinely report the number of LNs and the number involved by tumor, there is significantly more variation in reporting the diameter of the largest LN, the diameter of the largest metastatic focus, and the presence of extranodal extension [110], the independent prognostic value of which is not well defined.

Surgical margins

Like other reported parameters in this paper, surgical margin status is a known prognosticator for PSA recurrence and disease progression in PCa [111]. A positive surgical margin is defined as tumor cells at the inked margin of the prostatectomy specimen, with an incidence of 11–38% in large series [155]. This finding may occur in a region of extraprostatic extension (pT3 R1 in the 2009 TNM classification) or by intraprostatic incision into an otherwise organ-confined tumor (pT2 R1 in the 2009 TNM classification or pT2+ in many institutions) [136]. While most investigations have shown that it is predominantly positive surgical margins in pT3 disease that are relevant in terms of recurrence risk [156], the value of surgical margin positivity in otherwise organ-confined disease (pT2+) has more recently been elucidated. Chuang et al. reported a worse 5-yr actuarial freedom from biochemical recurrence for pT2+ patients than for patients with organ-confined/margin-negative disease [157], and Stephenson et al. detailed a 7-yr progression-free probability of 94% (95% confidence interval [CI], 93–95) for organ-confined/margin-negative disease compared with 76% (95% CI, 72–80) for pT2+ patients [158]. Hence, reporting of overall margin status as positive or negative should be uniform in pathology practice.

In many institutions, a report of a positive margin is cause for initiating radiation therapy, as there is some evidence to suggest that doing so reduces the rate of rising PSA after prostatectomy [159]. However, recognizing that a significant percentage of patients with positive margins never experience PSA recurrence, other clinicians opt for a more selective application of this adjuvant therapy, choosing instead to follow patients closely and treat them only if biochemical recurrence occurs or there is clinical/radiologic evidence of progression. The rationale for the latter approach is based on the estimated number of patients one would have to treat to prevent one recurrence. A wealth of somewhat conflicting evidence has been reported for a host of parameters that attempt to substratify positive surgical margins by number (ie, multiple vs solitary or absolute number), site of margin positivity (eg, apical vs other, bladder neck vs other, in an area of vs away from an area of pT3 disease), extent of margin positivity (ie, extensive vs focal or number of millimeters), or even GS at a site of margin positivity [147], [156], [158], [160], [161], [162], and [163] and to find associations between one or more of these features and biochemical recurrence. Yet a recent meta-analysis demonstrated that while many parameters were independently prognostic, no single parameter improved the predictive accuracy of a standard nomogram in which surgical margin status was modeled as positive or negative [158]. Even in highly selected cohorts, therefore, further evaluation of the range of features of positive margins may be necessary to determine which, if any, is the most robust predictor of outcome, warrants routine reporting, and helps select patients most likely to benefit from adjuvant radiation therapy [164].

Tumor quantitation

Some measure of tumor quantitation is a cornerstone of tumor reporting in most organ-based cancers. It has been long known that PCa tumor volume correlates well with common adverse features such as high GS, extraprostatic extension, seminal vesicle invasion, and clinical outcomes [165] and [166]. Although a large number of institutions have studied the independent prognostic value of tumor volume in their respective cohorts, wide differences exist; approximately half of large series each find an association with biochemical recurrence or no association [108], [166], [167], and [168]. These results can be somewhat explained by differences in tumor quantification method, composition of the cohort (pre-PSA screening compared with later), and failure to demonstrate significance when added to commonly reported PSA and pathologic findings. Hence, whether overall tumor volume was calculated by estimation of tumor percentage, number of blocks with tumors, ratio of involved to uninvolved blocks, greatest length times greatest thickness, or by one of these measures for the dominant nodule alone [166] significantly affects the ability to compare one study with another. Similarly, the intimate correlation with other prognostic factors leaves the independent value of tumor volume uncertain at this time. Though in practice many pathologists do report some measure of tumor volume/size and this method is widely advocated by a variety of professional groups, including a recent ISUP conference [108], given the lack of uniformity and definitive evidence, there is no one recommended method for doing so.

Lymphovascular invasion

Identifying lymphovascular invasion requires the presence of tumor cells within endothelial-lined spaces conforming to the contour of the space and, when possible, attachment to the endothelium [109]. Care must be taken to exclude common artifacts, including retraction around cancer glands or mechanical displacement of tumor cells [169]. The reported incidence of lymphovascular invasion ranges from 5% to 53%, with most studies finding strong associations with high GS, positive surgical margins, extraprostatic extension, LN metastasis, and in univariate analysis, biochemical progression [170]. Akin to the situation with tumor volume, however, the independent prognostic value of this marker is debatable because of differences among studies in specimen handling, definition of lymphovascular invasion, number of and follow-up between cohorts, inclusion of patients with LN involvement, and whether specimens were re-reviewed or garnered from the report [169]. While few investigators have stratified patients by stage and other features, Herman et al. and Yamamoto et al. have studied pT3aN0 patients, finding lymphovascular invasion in 35% and 28% of cases and independent predictive value for PSA failure and clinical progression in multivariable analysis [170] and [171]. While further studies with standardized definitions and pathologic examination are needed, the relative ease of identification and reporting warrants the inclusion of lymphovascular invasion in routine pathology reports.