Within this decade, several important updates in prostate cancer have been presented through expert international consensus conferences and influential publications of tumor classification and staging.
To present key updates in prostate carcinoma.
The study comprised a review of literature and our experience from routine and consultation practices.
Grade groups, a compression of the Gleason system into clinically meaningful groups relevant in this era of active surveillance and multidisciplinary care management for prostate cancer, have been introduced. Refinements in the Gleason patterns notably result in the contemporarily defined Gleason score 6 cancers having a virtually indolent behavior. Grading of tertiary and minor higher-grade patterns in radical prostatectomy has been clarified. A new classification for prostatic neuroendocrine tumors has been promulgated, and intraductal, microcystic, and pleomorphic giant cell carcinomas have been officially recognized. Reporting the percentage of Gleason pattern 4 in Gleason score 7 cancers has been recommended, and data on the enhanced risk for worse prognosis of cribriform pattern are emerging. In reporting biopsies for active surveillance criteria–based protocols, we outline approaches in special situations, including variances in sampling or submission. The 8th American Joint Commission on Cancer TNM staging for prostate cancer has eliminated pT2 subcategorization and stresses the importance of nonanatomic factors in stage groupings and outcome prediction. As the clinical and pathology practices for prostate cancer continue to evolve, it is of utmost importance that surgical pathologists become fully aware of the new changes and challenges that impact their evaluation of prostatic specimens.
Several important events have transpired within the past few years that have resulted in some important changes in the pathologist's approach toward the diagnosis and reporting of prostate cancer. The International Society of Urological Pathology (ISUP) has held several consensus conferences for prostate cancer, including for Gleason grading (2014),1,2 for best practices recommendations in the application of immunohistochemistry including for prostate (2013),3 and for handling and staging of radical prostatectomy (RP) specimens (2010).4 In 2016, the 4th edition of the World Health Organization (WHO) “blue book”5 was released, incorporating new entities as well as new recommendations, mainly taken from the ISUP conferences. In 2018, the 8th edition of the American Joint Commission on Cancer (AJCC) TNM staging manual6 was implemented, with some updates in staging approaches and prognostic grouping of prostate cancer. These changes, particularly in grading, have been endorsed in recent urology and clinical oncology guidelines.7–9 This review summarizes the key updates important for the practicing general surgical pathologist.
A series of key developments in prostate cancer during the past 2 decades culminated in a major modification in its manner of grading, from traditional Gleason grading to the grade group (GG) system.1 Vital to these changes are our enhanced understanding, refinements, and rearrangement of the Gleason patterns (GPs); the better grouping and stratification of their outcomes, including the effective separation of indolent tumors; and the widespread acceptance of the innovative active surveillance (AS) management for low-risk prostate cancers.
First, there is now the almost complete abandonment of the lowest GPs, 1 and 2.10 It is now well recognized by experts, through the benefit of immunostaining for basal cells, that most of the originally described GP 1 cancers were likely adenosis (atypical adenomatous hyperplasia) and that there are virtually no lesions that meet the stringent criteria set out for GP 1. Although the 2005 ISUP consensus still described GPs 1 and 2 as vanishingly rare and uncommon, their reporting has totally disappeared in needle biopsy specimens because the overall architecture of the entire focus is not completely appreciable in contemporarily used 18-gauge needle biopsies.11 More importantly, even if exceptional examples of GP 1 or GP 2 are identified, their prognostic impact will not be different from that of contemporary GP 3, which already has a baseline favorable outcome not in need of further substratification.
Second, there has been a gradual purification of GP 3 from more adverse patterns through grade refinements (Figure 1, A through C). Over the years, all cribriform patterns were shifted from GP 3 to GP 4, so that by the 2014 ISUP consensus, GP 3 was left with well-formed gland pattern only. This resulted in the migration of the few potentially metastatic tumors from GP 3 to GP 4, leaving the contemporary GP 3 with almost purely indolent tumors. In the study by Ross et al12 of more than 14 000 Gleason score (GS) 6 prostate cancers, none of the cases had lymph node metastasis. Other studies have also shown that even extraprostatic extension in GS 6 prostate cancers is uncommon and seminal vesicle invasion is exceptional.13,14 In a study by Eggener et al15 looking at 15-year prostate cancer–specific mortality, only 3 of more than 9500 organ-confined GS 6 prostate cancer patients died of the disease. These findings confirmed the indolent nature of contemporary GS 6 tumors, different from the historic GS 6 cancers in Dr Gleason's era, which has become the foundation for AS management in prostate cancer.
Third, among the different grade sums/scores (GS 6–10), studies have shown that GS 7 tumors are much more heterogenous, with differing behavior between 3 + 4 and 4 + 3 cancers.16–18 Gleason score 4 + 3 cancers are 3 times more lethal than GS 3 + 4 cancers.16 There has been an inflation of GS 7 cancers as a result of the grade migration from GS 6,19 and this increase in GS 7 is enhanced in RP because more GS 6 cancers stay in AS protocol, not undergoing RP. This inflation of GS 7 cancers further justifies the need to split GS 7 cancers.
Fourth, studies on prostate cancer outcome have identified the optimal grouping of the different GSs; this includes splitting GS 7 cancers and lumping GS 9 and 10 cancers. A study by Pierorazio et al20 showed prostate-specific antigen (PSA) biochemical recurrence (BCR)–free survival in GS 6 or lower, 3 + 4. 4 + 3, 8 and 9, and 10 cancers of 94.6%, 82.7%, 65.1%, 63.1%, and 34.5%, respectively, and this study became the basis for the new GG grading. This same breakdown of the GSs was also incorporated into the Partin tables.21
The above developments culminated into the codification of this 5-tiered GS stratification as the new GGs 1 (3 + 3), 2 (3 + 4), 3 (4 + 3), 4 (8), and 5 (9 and 10) in the 2014 ISUP consensus conference.1 Thus, it is now recommended that the GG be reported concomitantly with the GS, for example, GS 3 + 4 = 7 (GG 2), with the potential that with time only the GG will be applied. Although it appears that the GG is a conversion of the GS, GG can be derived directly, with each GG having a specific definition (Table 1). One of the main advantages of GG is that the lowest baseline grade is now 1 instead of the 6 with GS. This facilitates counseling patients for AS who may have misperceived their GS 6 tumors as intermediate grade because they are in the middle of the overall GS scale of 2 to 10. Further, the GGs now represent the most clinically relevant cutoffs important for prostate cancer management. Separation of 3 + 4 and 4 + 3 can be important for radiation therapy, and prognostically distinct GS 8 is separated from GS 9 and 10, the highest end of the spectrum with clearly poor prognosis. In practical terms, the occasional and time-consuming challenge in classifying the aggressive 4 + 5, 5 + 4, and 5 + 5 cancers can be avoided with the use of GG, as these subdivisions became clinically irrelevant. There is, however, some controversy with GG 4, which is very heterogenous and includes GS 3 + 5, 4 + 4, and 5 + 3. At this point, there is limited BCR or survival analysis of 3 + 5 versus 5 + 3.
Several studies have validated the use of the GG system in prostate cancer in biopsy and RP, including in patients treated with other forms of therapy.22–28 The largest multi-institutional study of more than 20 800 men with prostate cancer that assessed PSA BCR showed in RP specimens hazard ratios relative to GG 1 of 1.9, 5.1, 8.0, and 11.7 for GG 2, GG 3, GG 4, and GG 5, respectively.22 The GG system has also been shown to be associated with risk of prostate cancer–specific mortality and bone metastasis progression.24 A recent study using a Surveillance, Epidemiology, and End Results database of more than 330 000 racially diverse prostate cancer patients showed that the new 5-tiered GG system is predictive of prostate cancer–specific mortality regardless of therapy received and clinical stage at diagnosis.23 The new GG system for grading prostate cancers has now been incorporated into the 2016 WHO blue book,5 the 8th edition of the AJCC cancer staging manual,6 the College of American Pathologists (CAP) prostate cancer protocol checklists,29 and urology and oncology treatment and screening guidelines, including the National Comprehensive Cancer Network guidelines.7–9
GP Definitions and Interpretations
Several refinements in the definition and interpretation of GPs 3, 4, and 5 have been addressed in the 2014 ISUP consensus conference (Table 2).1 As described above, contemporary GP 3 is now exclusively composed of well-formed glands, characterized by single, discrete glands with complete circumference of cells forming lumina. Classification as well-formed gland is regardless of size and shape, as some may be small (microacinar or atrophic), large (pseudohyperplastic, microcystic, prostatic intraepithelial-like), or falsely elaborate (tunneling, tangential sectioning). The 2014 ISUP consensus further recognized that branched glands should now be allowed in GP 3 (Figure 2).1
Distinction between some GP 3 and GP 4 glands can be challenging, particularly between small and tangentially sectioned GP 3 glands versus poorly formed GP 4 glands and between tight clusters of GP 3 glands versus fused GP 4 glands.30–32 Interobserver studies have shown that pathologists usually have problems classifying poorly formed glands as GP 4, particularly when seen admixed with intervening GP 3 glands versus when seen as a discrete cluster of GP 4 glands.31,32 Although a practical approach is to ensure that tangential sectioning is not the cause of glands that appear poorly formed, some authors have proposed that interobserver reproducibility may be improved in the interpretation of poorly formed glands as GP 4 when a minimum number (>5) of poorly formed glands are in a cluster and not intermixed with well-formed glands.31 This approach has limitations in applications while dealing with extremely small foci of cancer. The 2014 ISUP consensus provides some guidance on this issue.1 First, a diagnosis of GP 4 should be made at ×10 magnification to enhance consistency. Second, in problematic cases with occasional or seemingly poorly formed or fused glands, the diagnosis should be deferred to GP 3. Third, likewise in borderline GP 4 versus GP 3 glands, because of artifactual distortion, the diagnosis should favor assigning a lower grade. Examining of multiple levels is of immense help to mentally reconstruct a 3-dimensional image of the glands. Alternatively, cutting deeper levels to see if the poorly formed glands will “open up” and declare themselves as GP 3 glands, although a sensible approach, is not strongly recommended because this scenario is common and such an approach would be impractical for the pathologist and the laboratory. In terms of identifying tight clusters of GP 3 glands, the lining of individual glands must be clearly delineable, unlike in fused GP 4 glands, where 2 or multiple lumina can be separated by a single lining of cells.
It is now recognized that GP 4 in conventional acinar adenocarcinoma has 4 basic patterns: cribriform, fused, poorly formed, and glomeruloid glands (Figure 3, A through D).1 Among these patterns, several recent studies have shown that cribriform has an outcome worse than noncribriform GP 4 patterns.33–36 Among GS 7 prostate cancers, cribriform architecture was shown to be an independent predictor of PSA BCR, metastasis, and cancer-specific survival.34,35 In the 2005 ISUP consensus, large cribriform glands were shifted from GP 3 into GP 4.11 However, small cribriform glands with strict criteria (including smooth punched-out glandular spaces) were kept as GP 3 at that time. Subsequent studies, however, showed that consensus diagnosis of small cribriform GP 3 based on the 2005 ISUP criteria is exceedingly rare among genitourinary experts and that small cribriform glands (not larger than concurrent benign glands) show similar predictive values for PSA biochemical failure to large cribriform glands.36,37 Thus, the 2014 ISUP consensus with a unanimous vote assigned all cribriform glands regardless of size as at least GP 4.1
Another pattern that was unanimously voted and assigned as GP 4 is the glomeruloid pattern.1 Unlike the cribriform pattern, only a few studies have specifically investigated glomeruloid glands, although the aggregate evidence to date supports classifying these glands as GP 4.38,39 The study by Lotan and Epstein38 showed that glomeruloid glands in biopsy are overwhelmingly associated with GP 4 (80%), including large irregular cribriform glands, with areas showing transition from small to large glomeruloid and cribriform pattern, and suggested glomeruloid to be a precursor lesion of the cribriform architecture. In a recent study by 2 of the authors (C.B. and G.P.),39 the recently described glomeruloid pattern in RP was shown to be a negative predictor for BCR-free survival among GS 7 cancers, with improved survival compared with nonglomeruloid GS 7 cancers. Further, GS 7 cancers with glomeruloid pattern only were shown to have significantly lower 5-year BCR-free survival than GS 6 cancers, validating their designation as a GP 4 pattern.
Hypernephromatoid GP 4 pattern, a historic term, which resembles renal cell carcinoma and was relegated in the 2005 ISUP consensus to an uncommon GP 4 pattern, is now recommended by the 2014 ISUP consensus to be no longer used.1,11
Several refinements in GP 5 were also made in the 2014 ISUP consensus.1 Gleason pattern 5 now includes small solid cylinders, medium to large nests with rosettelike spaces, and focal but unequivocal comedonecrosis (Figure 4, A and B). Discrete glands (otherwise GP 3) with necrotic debris (and not comedonecrosis) in the lumen do not equate to GP 5. Solid cords or cylinder and medium to large solid nest patterns have been shown to be associated with the other traditional high–Gleason-grade patterns.40
Rules in GS Including Tertiary Grade
The rules for Gleason grading prostate cancer with 1 or 2 GPs in biopsy and RP are straightforward, and the original 2005 ISUP consensus rules were retained in the 2014 ISUP consensus.1,2,11 When there are 2 GPs present, the primary and secondary grades remain as the most predominant and less predominant grades, and when there is only 1 GP present, the grade is then doubled. When there are 3 GPs present, the rule in prostate biopsy instituted in the 2005 ISUP consensus was retained in the 2014 ISUP consensus, in that the most predominant of the 3 GPs is regarded as the primary grade and the highest of the remaining 2 GPs will become the secondary grade. In RP, grading when there are 3 GPs present may be different in contemporary times, and is not universally practiced. The traditional way of reporting 3 GPs in RP is simply based on the predominance of the 3 GPs present, into the most (primary grade), second most (secondary grade) and least common (tertiary grade) patterns, with the least common pattern reported if it's GP 5 (tertiary grade 5). However, there is a growing acceptance among several genitourinary experts that a cutoff of 5% for tertiary grade 5 should be applied, and that GP 5 should be bumped into a secondary grade if it is 5% or higher.2,41,42 For example, cancer with 60% GP 4, 30% GP 3, and 20% GP 5 should be graded as 4 + 5 = 9 (or GG 5) and not as 4 + 3 = 7 with tertiary grade 5, the former approach in concept being similar to grading in the prostate biopsy. There is a valid concern by experts that because the tertiary pattern 5 is not incorporated in prognostic tables and nomograms, discounting more than 5% GP 5 as tertiary grade (and not as secondary grade) may lead to a significant undergrading of prostate cancer. The 2014 ISUP consensus did not vote specifically to have a cutoff for the amount of GP 5 to be kept as a tertiary grade 5 in RP.2 However, it was agreed that the term minor high-grade pattern should be preferred over tertiary pattern, in essence the same principle that the tertiary grade 5 should only be in a small amount. When using GG in the presence of less than 5% tertiary grade 5, the recommended approach of reporting should be GG2 with minor high-grade pattern, or potentially GG2+. For either approach, one helpful way of avoiding confusion for the urologists on the other side of the report is to simply enumerate the percentages of GP 4 and GP 5 in a comment when reporting the presence of a tertiary grade 5.
Intraductal carcinoma of the prostate (IDC-P) is an intraluminal proliferation of carcinoma cells within acini and/or ducts with histologic features exceeding and more ominous than those for high-grade prostatic intraepithelial neoplasia (PIN) (Figure 5, A and B).5,43–46 The most commonly used histologic criteria for the diagnosis of IDC-P are those proposed by Guo and Epstein,43 which with slight modification have been incorporated into the 2016 WHO blue book5 (Table 3). Most IDC-Ps are considered as retrograde colonization of invasive carcinoma into the ducts or acini. Presence of IDC-P in biopsy is associated with high-grade prostate cancer and adverse factors in RP, and is significantly associated with higher PSA BCR rates after RP.34,43 Further, IDC-P in biopsy is also an independent predictor of early PSA BCR and metastasis in prostate cancer treated with radiotherapy.47 When IDC-P is identified with GS 6 prostate cancers only in biopsy or TUR specimens, IDC-P remains having an adverse prognosis.48 Even in prostate biopsies of patients with metastatic disease at presentation, presence of IDC-P is significantly associated with worse cancer-specific survival.49 Intraductal carcinoma of the prostate may occur rarely in biopsy without a concomitant stromal-invasive carcinoma and remains associated with higher grade and stage at RP.50 Thus, it has been suggested by some authors50 that the presence of IDC-P in a biopsy regardless of concomitant invasive carcinoma warrants a definitive therapy. Others51 recommend a more conservative approach of recommending a repeat biopsy, because most IDC-P is associated with invasive high-grade carcinoma. The later careful approach is brought about by the challenge in distinguishing IDC-P from PIN; overdiagnosis of PIN as IDC-P may lead to overtreatment. Although the distinction between the two (IDC-P and PIN) is based primarily on morphology (extent of proliferation and degree of cytologic atypia), ancillary immunohistochemistry may be useful. Cytoplasmic PTEN is lost in approximately 85% of IDC-P, in contrast to PIN, where expression is usually preserved, and can be helpful for their distinction.52 Erythroblast transformation-specific–related gene (ERG) rearrangement also occurs with higher frequency in IDC-P than in PIN.53,54
When identified in RP, IDC-P is associated with higher GS and stage and is an independent risk factor for progression-free and cancer-specific survival.55,56 Rarely, extensive IDC-P may occur in RP without the presence of invasive carcinoma.57
The 2014 ISUP consensus recommended not grading IDC-P.1 Presence of IDC-P should be subtracted when grading the concurrent admixed invasive carcinoma, or if IDC-P is isolated (without accompanying invasive cancer) no grade should be reported. The presence of IDC-P, however, must be reported in biopsy, and it is recommended that a comment that IDC-P is invariably associated with aggressive prostate cancer be made. Although the clinical significance of IDC-P is well established in biopsies with intermediate-risk prostate cancer, the impact of its presence in high-risk (GS 8–10) prostate cancer is not clear. We do not recommend reflex immunostaining of any cribriform, solid nest, or glands with comedonecrosis with basal cell markers to screen for IDC-P in needle biopsies; distinction must be attempted first by morphologic examination.
A small subset of IDC-P may occur without an adjoining invasive carcinoma, or may occur adjoined with microscopic focus of carcinoma showing transition from PIN, and these lesions are regarded by some authors as precursor-like IDC-P.58 Compared with regular-type IDC-P (with invasive cancer), precursor-like IDC-P has significantly lower associated GS, less-advanced stage, and lower BCR-free rate. On the other hand, some authors59,60 have also suggested that some classic PIN-like areas may actually represent retrograde colonization of prostate cancer (or not a precursor lesion), and thus may represent a form of early or “low-grade” IDC-P. Currently, the term IDC-P is strictly reserved for an intraepithelial neoplasm with histology more ominous than PIN, fulfilling the morphologic diagnostic criteria for IDC-P that have implications in management.1,5,43 Cribriform lesions that have features borderline between IDC-P and PIN have been regarded by some authors45,61,62 as “atypical cribriform lesion” or “atypical cribriform intraductal lesion insufficient for the diagnosis of IDC-P,” and their identification in isolation in biopsy should warrant a repeat biopsy. Because of the stark disparity in management of IDC-P and PIN, some authors45 now label all cribriform PIN in biopsy as an “atypical intraductal cribriform lesion.”
Besides PIN, other differential diagnoses for IDC-P include intraductal spread by urothelial carcinoma and ductal adenocarcinoma.45 In challenging situations, confirmation of an intraductal urothelial carcinoma can be facilitated by immunohistochemical staining (p63, GATA3, and uroplakin 2–positive lesional cells). Ductal prostatic adenocarcinoma is characterized by tall columnar cells that are often stratified, has more expansile growth, may have more robust papillary structures, and lacks basal cells. There is some degree of confusion among urologists with regard to IDC-P and ductal adenocarcinoma because of their overlapping names. A helpful approach is to provide a descriptive comment when either of these entities is diagnosed.
The classification of the uncommon prostatic neuroendocrine tumors has been modified in the 2016 WHO blue book,5 based on the system proposed by experts gathered under the auspices of the Prostate Cancer Foundation (Table 4; Figure 6, A through D).63 This new classification now includes adenocarcinoma with Paneth cell–like neuroendocrine differentiation and large cell neuroendocrine carcinoma (LCNEC). The previous entity of focal neuroendocrine differentiation in prostatic adenocarcinoma was refined to be referred to as neuroendocrine differentiation in usual prostate adenocarcinoma, which is characterized by neuroendocrine marker (synaptophysin or chromogranin) expression in a usual adenocarcinoma morphology that also retains the expression of prostatic markers (PSA, prostate-specific acid phosphatase [PSAP], prostate-specific membrane antigen [PSMA], NKX3.1, or P501S). Aggregate data suggest that neuroendocrine differentiation in usual adenocarcinoma has no impact on clinical outcome. There are conflicting data on extent of neuroendocrine positivity and response to androgen deprivation therapy63 ; some suggest that those with diffuse expression be singled out because this positivity could be part of the spectrum of primary or metastatic prostate cancers after androgen deprivation therapy that exhibit overlapping features between small cell and acinar adenocarcinoma and could be more clinically aggressive.64
Adenocarcinoma with Paneth cell–like neuroendocrine differentiation is characterized by the presence of cells with bright eosinophilic granules that are positive for neuroendocrine markers and ultrastructurally contain neurosecretory granules.65–67 The Paneth cell–like cells usually intermingle with usual adenocarcinoma cells, but their amount varies, and this morphology may be more florid. Rarely, these cells may show amphophilic cytoplasm and form cords or solid nests of bland cells with amphophilic cytoplasm that on conventional Gleason grading would be assigned a GP 5. Based on observed outcome, this morphology does not portend aggressive behavior, and hence, when detected, areas with amphophilic cytoplasm and neuroendocrine differentiation should be discounted from GP assessment and grading.67
Within the spectrum of purer neuroendocrine tumors, the extremely rare prostatic carcinoid is now officially labeled as well-differentiated neuroendocrine tumor (carcinoid), the aggressive small cell neuroendocrine carcinoma (SCNEC) is retained, and the new group of LCNEC is added.5,63 These tumors are generally negative or have low expression of prostatic markers. For diagnosis, well-differentiated neuroendocrine tumor should not be closely associated with usual prostate adenocarcinoma and must lack the expression of prostatic markers to distinguish it from the carcinoid-like pattern that may be occasionally seen in usual prostate cancer. Also, the tumor should not be related to the urethra, to exclude a urethral well-differentiated neuroendocrine tumor. Based on limited data, well-differentiated neuroendocrine tumor may present with higher-stage disease, including lymph node metastasis, and yet portends a favorable outcome.63,64 Association, if any, with multiple neuroendocrine neoplasia syndrome stigmata should be explored.
Small cell neuroendocrine carcinoma is morphologically and immunohistochemically similar to its extraprostatic counterparts, including more than 50% showing aberrant TTF1 expression.63,68–71 Expression of prostatic markers is generally absent, or focal if present. Of note, expression of the basal cell markers high-molecular-weight keratin (HMWK) and p63 is encountered in about one-third and one-fourth of SCNECs, respectively.68 About half of SCNEC occurs in patients with prior prostatic adenocarcinoma, particularly those treated with androgen deprivation therapy; thus, it is felt that the neuroendocrine differentiation is induced by androgen suppression. About half of the mixed usual adenocarcinoma and SCNEC harbors the TMPRSS2-ERG fusion, and this molecular event may aid in the identification of prostatic SCNEC at a metastatic site; immunohistochemistry is less reliable.72,73 Small cell neuroendocrine carcinoma is an aggressive tumor, with more than half of patients presenting with metastatic disease, and has a median cancer-specific survival of only 19 months.74
The newly added LCNEC is extremely rare, with fewer than 10 bona fide cases reported.75,76 Large cell neuroendocrine carcinoma is characterized by solid sheets, ribbons, or nests of neuroendocrine cells with peripheral palisading and geographic necrosis. The tumor cells usually have enlarged nuclei with prominent nucleoli and brisk mitotic activity, resembling more closely a high-grade usual adenocarcinoma. Most reported LCNECs have developed in patients with prior adenocarcinoma treated with androgen deprivation therapy, similar to SCNEC, suggesting that androgen suppression leads to clonal expansion of malignancy with neuroendocrine phenotype. In most cases, a minor-component adenocarcinoma with hormonal treatment effect may be seen intermingled with the LCNEC. Limited data in LCNEC suggest an aggressive tumor, with dissemination and death from metastasis occurring in less than a year.75 Overall, neuroendocrine differentiation, although uncommon, usually increases with androgen deprivation, and its incidence is relatively higher in the setting of castration-resistant prostate cancer.77 The characterization of the entire spectrum of aggressive prostatic neuroendocrine tumors in the castration-resistant setting is still evolving, and thus some subtypes are categorized based solely on clinical criteria.77
The 2016 WHO blue book5 has expanded the previously recognized subtypes of adenocarcinoma to include microcystic adenocarcinoma and pleomorphic giant cell adenocarcinoma (Table 5). The previous categories of prostatic oncocytic carcinoma and lymphoepithelial-like carcinoma have been dropped. Microcystic adenocarcinoma is characterized by carcinoma with cystic architecture containing cystically dilated glands with rounded contour and flat luminal cell lining. This carcinoma pattern has mixed features of the more established pseudohyperplastic and atrophic adenocarcinoma patterns (Figure 7).78 The microcysts are generally subcentimeter in size and often have luminal wispy mucin and eosinophilic crystalloids. Closer inspection will allow for the appreciation of nuclear enlargement, nucleomegaly, and cytoplasmic atrophic changes. The immunoprofile is similar to that of conventional adenocarcinoma, with overexpression of α-methylacyl CoA racemase (AMACR) and absence of basal marker (eg, HMWK, p63, CK 5/6) positivity. The microcystic adenocarcinoma is mostly a GS 6 (3 + 3) tumor. Recognition of this pattern is important to avoid underdiagnosis of carcinoma. There is no other clinical or prognostic significance.
Pleomorphic giant cell carcinoma is a rare aggressive type of prostate carcinoma at the undifferentiated end of the spectrum characterized by the presence of bizarre, anaplastic, and giant tumor cells (Figure 8).79,80 This variant is often admixed with usual prostatic adenocarcinoma and other carcinoma types (atrophic, ductal, etc), and the component with extreme pleomorphism varies. Most of the pleomorphic giant cells typically do not exhibit expression of prostatic markers, in contrast to the admixed usual adenocarcinoma cells. Data on the few cases published thus far suggest that the tumor portends a poor prognosis with most patients developing metastasis or dying from the disease. It is important to rule out a bladder giant cell carcinoma with or without urothelial component while considering this rare carcinoma in the prostate.
One of us (G.P.) recently described a series of the unique cystadenocarcinoma, an emerging subtype of prostatic carcinoma.81 These cystic tumors can be massively enlarged (>10 cm), presenting as a pelvic mass with compressive symptoms (giant multilocular cystadenocarcinoma [GMCC]) or as smaller lesions of multiple cysts (individual cysts measuring ≤1 cm) diagnosed clinically in a similar manner to the usual adenocarcinoma (microscopic cystadenocarcinoma). Histologically, cystadenocarcinomas are usually lined by ductal-type adenocarcinoma cells and demonstrate intracystic papillae that can be exuberant (Figure 9, A through C). The cystic fluid content may show a markedly high PSA level. Noncystic concomitant usual prostatic adenocarcinoma may also be seen admixed with the cystic tumor. GMCC appears to have a more aggressive behavior than usual adenocarcinoma, with a higher proportion presenting as locally aggressive disease, including adherence to the perirectal tissue or peritoneum. Microscopic cystadenocarcinoma likely represents an earlier stage of GMCC. It is unclear if GMCC is related to the equally rare and benign giant multilocular cystadenoma. Some authors have reported cystadenoma transitioning to cystadenocarcinoma, suggesting that some GMCC may also arise from giant multilocular cystadenoma.82
NEW REPORTING APPROACHES AND RECOMMENDATIONS
Tumor Volume in Needle Biopsy
There is controversy as to the manner of quantifying separate discontinuous foci of cancer in a biopsy core, although how this finding is reported has critical ramifications when managing patients with low-risk prostate cancers. Some authors recommend measuring the separate tumor foci end to end, and that includes the intervening benign glandular tissue in a core, instead of adding or “collapsing” the separate tumor foci (Figure 10).83–87 The end-to-end measurement has been shown by some studies to be more predictive of stage and risk for margin positivity at RP. The discontinuous tumor foci in the biopsy core are often the result of a single tumor from the corresponding region of the prostate. Because the tumor follows the curve of the peripheral zone, the needle may sample only both the ends and different parts of the curving tumor with intervening benign glands. Although evidence suggests that end-to-end measurement is superior, this approach may potentially lead to overestimation of tumor volume when multiple small tumor foci are truly present and sampled concomitantly. More clinical guidelines for AS use the 50% cutoff for tumor quantification as one criterion for assessing eligibility, and overestimation or underestimation may result in suboptimal management.88 The current CAP protocol checklist for prostate cancer provides the option of reporting both approaches of quantification until more evidence settles this issue.29 Because large foci of prostate cancer may show heterogeneous morphologies within the same focus, the mere presence of different tumor histologies (eg, atrophic and foamy gland) should not automatically suggest different foci when discontinuously observed in prostate needle biopsies.
GP 4 Percentage and Pattern
Several studies have shown a linear correlation of an increasing percentage of GP 4 in GS 7 cancers in biopsy and RP specimens with increasing risk for PSA BCR.39,89,90 Reporting percentage of GP 4 in biopsy with GS 7 cancers thus allows prognostic stratification helpful in management decisions. The 3 + 4 and 4 + 3 categories are relatively broad: GS 3 + 4 cancer with 5% GP 4 may behave more like a GS 6 cancer, and the patient can be included in AS protocol, unlike a 3 + 4 cancer with 45% GP 4. Reporting of percentage of GP 4 in biopsy also provides transparency for cases that have discrepancy in diagnoses among pathologists and allows correlation with the final GS in RP. For example, a 4 + 3 cancer with 55% GP 4 in biopsy may be interpreted by others as 3 + 4 cancer or may have a final GS of 3 + 4 in RP. Similarly, in RP specimens, reporting percentage of GP 4 in GS 7 cancers provides similar benefits. As mentioned above, because of the tumor grade migration to GS 7, the GP 3/4 modifications, and a large number of RPs being scored as GS7, recording percentage of GP 4 allows further stratification of this large cohort. The 2014 ISUP consensus recommended reporting percentage of GP 4, and this recommendation is now included in the 2018 CAP protocol for prostate cancer reporting as an option.1,29 However, the manner of reporting percentage of GP 4 is not yet established. For practicality, reporting may be by less than 5%, 5%, and followed by quartile or decile percentage breakdowns. We do not recommend reporting percentage of GP 4 in GS 8 through 10 cancers, and likewise percentage of GP 5 in GS 9 and 10 cancers, as the clinical value of this level of reporting in GS 8 through 10 cancers has not been established. However, in biopsies with multiple cores and with GS 8 in 1 core, reporting the percentage of GP 4 in other cores with lower GS may provide a better assessment of the final global GS. For example, GS 4 + 3 = 7 with 90% GP 4 in one concomitant core with GS 4 + 4 = 8 will indicate a global GS closer to GS 8 in RP.
Reporting the specific patterns of GP 4 in GS 7 cancers, particularly the presence and amount of cribriform pattern, has also been suggested.33,39 Among GP 4 glands in GS 7 cancers, cribriform architecture has the highest risk for PSA biochemical failure and is a strong prognostic indicator for distant metastasis and cancer-specific death.34–36,39 Because of its clinical significance, it has been suggested that cribriform pattern may be incorporated as a separate variable in future grading practice refinements.2,33 Because of the emerging prognostic import of cribriform glands in a biopsy with GS 7 cancer, their finding should potentially exclude a patient from an AS approach, and this is most relevant particularly when the percentage of GP 4 is not high. In RP, reporting the presence of cribriform glands will also help further stratify the relatively large cohort of GS 7 cancers for more accurate prognostic stratification.39 For example, RP with organ-confined 3 + 4 cancer with 5% GP 4 and no cribriform glands may behave more like a GS 6 cancer and may be followed clinically in the same manner. Further data on the importance of recording and quantification of cribriform morphology are clearly warranted.
Cancer According to Biopsy Core Sampling, Submission, and Fragmentation
Biopsies of prostate cancer are now performed mostly by an extended biopsy protocol (12-site sampling) approach.91,92 This has replaced the traditional sextant (6 sites) by adding lateral biopsies (6) to the bilateral base, middle, and apical (6) biopsies. The extended biopsy protocol is more sensitive in detecting prostate cancer than the sextant method. In the United States, reporting the findings separately for each site (specimen level) is widely practiced, providing 12 individual findings. In other countries, including Canada, summarizing the report for all sites (case level) may be preferred, providing one summarized or condensed finding. The approach in reporting GS, number of cores involved, and percentage of tumor in the positive cores has been the subject of discussion, with recommendations being proposed. These variables are important for the patient's management, particularly in consideration for AS in patients with low-grade cancers.87,88
Reporting of GS for each site, total number of cores involved, and percentage involvement of each involved core should be provided.2,29,88,93 Most clinical practices in the United States consider the highest GS of the multiple involved cores with differing grades as the final GS, in contrast to taking the GS of all the positive cores altogether (global GS). This is supported by several studies showing that the highest GS in biopsies with multiple differing GSs correlates best with the corresponding grade in RP.94–97 Reporting the number of involved cores of the 12 sites and percentage quantification for each core as discussed is also important for triaging patients for AS protocols. Most guidelines require that not more than 2 or 3 of the 12 cores be positive for the patient to be considered eligible for AS.88
In an ideal setting, if one core is sampled from each of the 12 sites and each core is submitted in a separate container, then reporting of GS and percentage quantification for the core, as well as counting the total number of cores involved, is straightforward. However, occasionally the urologists will sample more than one core from one site, resulting in multiple cores in one container from a site. For that site with multiple cores, we recommend reporting the combined GS for the cores, because they likely represent sampling of the same tumor and provide percentage quantification for each involved core. For example, if 2 cores were sampled from the left medial apex with both being positive, reporting may be phrased as “prostatic adenocarcinoma, Gleason 3 + 4 = 7 (GG 2) involving 2/2 cores, with 5% and 10% involvement of the individual cores.”
It is not uncommon that urologists will submit the 12 cores in 6 containers representing right and left base, middle, and apical sites and that the medial and lateral biopsies from each side are submitted together in one container. This results in each container having at least 2 cores representing medial and lateral biopsies. Ideally in this setting, reporting of the GS and percentage quantification for individual cores should be done. However, because the tumor of the 2 positive sites/cores may potentially come from a single large tumor involving the 2 adjacent sites, using the above approach of reporting a combined GS for all cores in the same container and percentage quantification for individual cores is preferable. For example, if 2 cores (of a 12-core sampling) are submitted in one container as left apex with both being positive, reporting may be phrased as in the above example. These 2 cores likely represent the left lateral apex and left medial apex of a 12-core standard sampling. Inking of the 2 cores with 2 different colors by the urologist and proper identification as lateral and medial will help resolve this problem and will allow separate reporting of the 2 different sites.
Another problematic situation is when the 12 cores are submitted in only 2 containers, representing right and left, with 6 cores submitted in each container representing one side. In this case, to allow for assessment of the AS criteria, an attempt should be made to separately report the percentage positivity and accurate counting of involved cores, if possible. Because the highest GS is used most by clinicians for clinical management, reporting of one (the highest) GS for the container is appropriate. It is important to communicate the predicaments in dealing with this type of specimen submission to the submitting urologists to avoid having this problem regularly.
In situations with multiple cores submitted in one container in which there is core fragmentation and counting is not possible, our approach, whether for a 12-, 6-, or 2-container submission, is to report a combined GS for that container and provide global percentage quantification of tumor involvement of all the core fragments in the container (Figure 11, A through C). One exception to this is that if the involved core(s) happens to be intact amid benign fragmented cores, then a percentage quantification may be provided for the positive intact core. A single intact core is usually about 1.5 cm in length. For counting, we make a comment of how many fragments are involved in multiple fragmented cores to provide some estimate. Other pathologists prefer to comment that accurate counting of positive cores in this setting is not possible or applicable.
In situations where there is inherent suboptimal estimation of quantitation such as fragmentation and/or submission of multiple cores in one container, if the cancer happens to be high grade (GS 8–10), an approximate provision of the percentage quantification (overall percentage involvement) is acceptable because in this situation the patient is no longer a candidate for AS.
Grading Multifocal Tumors and Reporting Margins in RP
Prostate carcinoma is often multifocal, and in situations where a dominant nodule is present (highest stage or grade), the other, smaller low-grade tumors should be discounted when grading the dominant tumor.2 Although this approach was recommended in the 2010 ISUP consensus on handling and reporting of prostate cancer,98 anecdotally, we still encounter cases in consultation of multifocal prostate cancer with a dominant nodule wherein a (global) RP grade inclusive of the other, lower-grade foci is provided. For example, a distinct dominant nodule of 4 + 3 = 7 is present along with several 3 + 4 = 7 foci in other parts of the prostate. A clinically misleading global grade of 3 + 4 = 7 may be provided by the pathologist rather than 4 + 3 = 7 because of multiple other small foci with a predominant GP 3. The approach to assign the case as 4 + 3 = 7 will also show better correlation with the biopsy GS (which is based on the highest GS score in the case among all positive biopsies) because the corresponding site in the biopsy will have a GS similar to that of the dominant tumor nodule. In reporting, the GS of the dominant tumor should be presented in the final diagnosis and synoptic report, whereas the other foci of lower grades may be mentioned in the comment, to avoid confusion to the urologists seeing multiple GSs in the report. The CAP protocol checklist recommends reporting the size of the dominant tumor as an optional variable.29
Margin positivity in RP is recognized as one of the variables that is strongly predictive for BCR.93,99 A positive margin is considered when the tumor directly touches the ink or, unequivocally, when the cancer gland is transected at the ink. Several studies have shown that the positive margin may be stratified in predicting BCR, with the most-recognized cutoff of 3 mm.100–103 The CAP prostate cancer protocol checklist recommends measuring the linear extent of the margin and is divided into limited (<3 mm) and nonlimited (≥3 mm).29 Several studies103–105 have also showed that the Gleason grade of cancer at the margin is predictive of BCR. The CAP protocol also now recommends reporting the Gleason grade at the margin as an optional variable.29
An update summary for prostate cancer staging in the 8th AJCC manual is presented in Table 6.6,106 The 8th AJCC manual has removed the subcategorization of pT2 (formerly pT2a, pT2b, and pT2c). Subcategorization of pT2 was not considered relevant for the following reasons: (1) involvement of more than 50% of one side (formerly pT2b) is extremely uncommon; (2) it is more common for cancer to cross to the other side (pT2c) before involving more than 50% of one side; (3) separate bilateral minute tumor foci were staged as the highest pT2 subcategory (pT2c); and (4) finally, there are no large clinicopathologic studies that validate the prognostic effect of this subcategorization in RP specimens.98,107–110 In clinical staging, however, pT2 subcategorization (cT2a, cT2b, and cT2c) assessable by digital rectal examination remains important for clinical management and is thus retained in the 8th AJCC manual.
Nonanatomic Factors in Staging
An important goal of the 8th AJCC manual is to have a staging system that is not only relevant for cancer surveillance and registry purposes, but also pertinent for the individual patient at the point of care.6,111 This goal is most readily achievable by incorporating nonanatomic factors in staging, and is best exemplified by the changes in the 8th AJCC manual for prostate cancer staging.112 In the AJCC prognostic stage groups for prostate cancer, nonanatomic factors of grade (GGs) and serum PSA levels are incorporated. This approach may result in organ-confined prostate cancers in the presence of elevated PSA (≥20 ng/mL) and/or GG 5 being upstaged to stage group III. This example would have been designated as stage II cancer in the 2010 AJCC manual.113
The 8th AJCC manual has evaluated 15 prostate cancer prognostic models identified in the public domain for possible endorsement, using strict inclusionary and exclusionary criteria established by the AJCC Precision Medicine Core.114 Of 15 prognostic tools, only 2 met all the AJCC quality criteria and were endorsed.115,116 These 2 prognostic models, however, are only for metastatic castration-resistant prostate cancer patients, and no currently available model is approved for localized or locally aggressive prostate cancer, where staging and prognostication are more important. The AJCC will continually assess prognostic and predictive nonanatomic biomarkers (including new genomic technologies) and prognostic tools for future prostate cancer staging.
The authors have no relevant financial interest in the products or companies described in this article.
A portion of this manuscript was presented at the Houston Society of Clinical Pathologists Symposium on Genitourinary Pathology; April 8, 2017; Houston, Texas.