Anatomic, morphologic and genetic heterogeneity of prostate cancer: implications for clinical practice

Keywords::

• carcinogenesis
• field effect
• focal therapy
• Gleason grading
• multifocal
• precursor lesions
• prostate
• prostatic intraepithelial neoplasia
• staging and classification
• tumor heterogeneity

The tendency of prostate cancer to develop multifocally in the majority of cases is well documented. The likelihood of finding two or more separate tumors in a single prostate ranges from 60 to 90% [1–4]. The fact that prostate cancer often arises as multiple tumor foci and that these separate foci are independent clones, which may arise synchronously or metachronously, increases the complexity of establishing accurate grading and developing optimal treatment strategies [5]. Furthermore, the separate foci have varying degrees of biological aggressiveness based upon their unique genetic mutations, and can progress at different rates. It is likely that both environmental factors and inherited genetic predispositions influence prostate carcinogenesis. If this is indeed the case, an existing predisposition globally present in a patient’s prostatic epithelium could put that epithelium at risk of acquiring malignant properties after exposure to a variety of chemical or biological insults, such as xenotropic murine leukemia virus-related virus and inflammation, thereby forming multiple separate precursor lesions that could evolve into spatially separate and unique cancers within the same gland [6].

Prostate cancers most commonly arise either in the peripheral zone (70%) or the transition zone (25%) [7]. About one half of transition zone-predominant cancers exhibit distinctive morphology, being composed of tall columnar cells lining glands of variable size and contour, in contrast to peripheral zone cancers, which typically lack the clear cell histologic pattern, are more poorly differentiated and tend to be more infiltrative, whereas transition zone cancers tend to remain, at least initially, within the transition zone boundaries. Peripheral zone cancers, in addition to being the most prevalent type of cancer, are more amenable to transrectal palpation and biopsy, whereas transition zone tumors are probably more likely to be missed on biopsy and more likely to be encountered as transurethral resection specimens. Transition zone tumors present with higher prostate-specific antigen levels and higher tumor volumes, but more frequently show organ-confined disease and lower Gleason scores. Transition zone tumors typically lack TMPRSS2–ERG gene fusion, commonly seen in peripheral zone tumors [8].

The microscopic morphology of prostate cancer was elegantly analyzed by Donald Gleason, who devised a prostate cancer grading system according to the degree of glandular differentiation and the architectural distribution of the malignancy within the existing parenchyma. Gleason recognized that grade heterogeneity was common in multiple foci of adenocarcinoma in the same gland, and incorporated this hetero­geneity into his now-classic grading and scoring system, an observation confirmed by a study of 115 completely embedded and serially sectioned whole-mount prostatectomy specimens, which demonstrated an 87% prevalence of multifocality with extensive interfoci histologic heterogeneity within an individual specimen [3]. Analysis of 100 specimens in this study showed 290 separate tumor foci. Only nine of the specimens exhibited the same Gleason grade in all foci. The index tumor Gleason grade and the overall grade assigned to the case were the same in only two-thirds of the cases [3]. These data confirm and further emphasize the problem of multifocality and grade heterogeneity in routine cases of prostate cancer.

It is unclear whether the above-noted multifocality and grade heterogeneity are secondary to the development of multiple separate clones of cancer, or whether a single cancer arises initially and then gives rise to future generations of cancer with unique cytogenetic abnormalities. The first hypothesis posits that multiple independently growing clonal neoplasms gradually enlarge and eventually form a large histologically heterogeneic tumor, accompanied by a few small independent foci elsewhere in the gland. This notion is supported by studies showing decreased grade heterogeneity in small tumors compared with larger tumors in prostates with multifocal disease. In the second scenario, multi­focal tumors of different grades may represent the evolution of poorly differentiated areas from more well-differentiated areas as a result of progressive accumulation of additional chromosomal anomalies and progressively worsening genetic instability. It seems probable that the evolution of prostate cancer represents a combination of these two processes occurring concomitantly [9].

Cheng et al. analyzed allelic loss at chromosome 8p12–21, the location of a tumor suppressor gene involved in early prostatic carcinogenesis, and at the BRCA1 locus on chromosome 17q21 in separate tumors within the same prostate to determine whether or not the tumors had arisen independently. They found a random discordant pattern of allelic deletion in distantly separate tumors, whereas the same allele was consistently lost from different regions of the same tumor, consistent with an independent tumor origin in 15 out of 18 cases [5].

Although accumulated data strongly support the concept of an independent origin for many, if not most, prostatic tumors within an individual gland, it remains conceivable that spatially separate tumors may arise through intraglandular dissemination. In Cheng’s study of DNA alterations in spatially separate cancers within a prostate, cited above, the pattern of allelic loss in three out of their 18 patients was compatible with a mechanism of intra­glandular dissemination. In this scenario, a highly aggressive prostate cancer begins to spread early in its biologic course, forming satellite lesions via intraglandular dissemination and thereby rapidly involving large portions of the gland. In an analysis of 94 separate tumor samples from 30 patients who died from metastatic prostate cancer, Liu et al. demonstrated that multiple ­separate metastases may arise from a single precursor cancer cell [10].

The importance of tumor multifocality in prognosis and staging of prostate cancer is yet to be determined [5,9]. In our pathologic review of 1274 patients undergoing radical prostatectomy for clinically localized prostate cancer using whole-mount sectioning and tumor mapping, tumor focality failed to predict for biochemical recurrence both in univariate and multivariate models [11]. No differences were seen in 5-year biochemical recurrence-free survival rates for unifocal (68%) and multifocal tumors (69%) [11]. Conversely, unifocal cancer is not always associated with a favorable prognosis. As a matter of fact, half of unifocal prostate cancers were associated with intermediate- or high-risk ­factors [11,12].

The well-documented multifocal and multiclonal nature of prostatic carcinogenesis has been a source of longstanding uncertainty regarding the validity of pathologic staging with respect to its true impact on long-term prognosis. The most controversial aspect of the 2010 American Joint Committee on Cancer tumor–node–metastasis (TNM) staging system is the subclassification of pT2 prostate cancer. T2 tumor is defined as organ-confined prostate cancer, with pT2a representing involvement of <50% of one lobe, ppT2b more than 50% of one lobe and pT2c representing involvement of both lobes. Questions have arisen regarding the true validity of stage pT2b tumors. In the first study to investigate the pathologic prevalence of pT2b tumors through examination of 369 totally embedded and serially sectioned whole-mount radical prostatectomy specimens, not a single unilateral tumor occupying greater than one half of a single lobe (pT2b) could be identified in the large whole-mount prostate cohort from the Indiana University [4]. Prostate cancers were multifocal in 312 cases (85%). The majority of the specimens were pathologic stage T2 (n = 276; 75%). Taking into consideration the average prostate weight (35 g) as well as the predominance of tumor multifocality, it would be unusual to identify a tumor involving more than half of one lobe (approximately an 8 cm³ tumor) without involving the other lobe [4]. These data bring into question the validity of stage pT2b as it is currently defined – that is, a tumor occupying more than one half of a single lobe, but not involving both lobes.

A pT2 subclassification based on tumor volume (or tumor size) may be superior to the current 2010 pT2 staging classification based on the clinical impression of midline crossing and proportional occupation of a lobe [4]. Tumor size has been widely used as an important TNM staging parameter in numerous organ systems. There was an excellent correlation between tumor size (maximum tumor diameter) and prostate cancer volume [13]. We propose that maximum tumor diameter may be used for pT2 substaging [14]. The majority of cases with a maximum tumor diameter of ≤0.5 cm have a tumor volume of ≤0.5 cm³. It is well documented that patients with a tumor volume of ≤0.5 cm³ and Gleason score of <7 (so called ‘insignificant prostate cancer’) have excellent prognosis. It is interesting to note that each centimeter increase in maximum tumor diameter was associated with a 70% increase in the risk of recurrence in a previous analysis [13,15]. It has been documented that the median of maximum tumor diameter was 1.6 cm [13,15].Therefore, it appears to be reasonable to use 0.5 and 1.6 cm as cutoffs for substaging of pT2 prostate cancer. We have proposed a new subclassification of pT2 prostate cancers based on tumor size [14]. In this proposal, pT2a tumors are organ-confined cancers with a largest tumor dimension of ≤0.5 cm; pT2b tumors are organ-confined cancers with a largest tumor dimension of >0.5 cm, but ≤1.6 cm; and pT2c tumors are organ-confined cancers with a largest tumor dimension of >1.6 cm [14]. Further investigation is imperative to define the optimal cutoffs for pT2 substaging and to examine alternative approaches for pT2 subclassification.

Controversy has also revolved around the optimal application of Gleason grading in predicting outcome, since it is well documented that multiple small foci of prostate cancer with higher Gleason grades than that of the index tumor are often present in cancerous glands, raising concern that the established Gleason grading system may not accurately predict the metastatic potential of all carcinoma foci in a prostatic gland. In short, it is unclear whether a single focus on the index tumor is a valid predictor of the true malignant potential of prostate cancer composed of multi­focal tumors. A complete picture of the clinical and biological potential of prostatic carcinoma in a gland can only be achieved by evaluating all tumors within a gland, an aim that is beyond the capabilities of current imaging and biopsy techniques. As noted previously, a study by Arora et al. of 290 separate tumor foci in 100 prostatectomy specimens showed the weakness of focusing only on the Gleason profile of the primary index tumor: only nine of these specimens showed the same Gleason grade in all foci, and in many cases, the overall Gleason score did not correlate with the score of the index tumor [3].

Accumulated data suggest the worst histologic grade dictates the biological behavior of prostate cancer [16–19]. In light of these findings, it is currently recommended that needle biopsy grading should reflect the sum of the primary and worst grades observed in the biopsy material, regardless of the relative extents of a tertiary high-grade lesion, if present. In grading radical prostatectomies, summing of the grades of the primary and secondary patterns is recommmended, as well as reporting the presence of any high-grade tertiary elements present in the specimen, regardless of volume, in recognition of the potentially significant clinical impact of higher grade but low-volume tumors as well as the imperfect accuracy of needle biopsy in gauging the full extent of such disparate lesions within a prostate. In recent studies, it was found that the combined percentage of Gleason patterns 4 and 5 is one of the most powerful predictors of patient outcome, and appears superior to conventional Gleason score in identifying patients at increased risk of disease progression [17,20]. We recommend that the amount of high-grade cancer in a prostatectomy specimen should be taken into account in therapeutic decision-making and assessment of patient prognosis.

Increasingly, the past decade has seen attempts to apply an array of novel techniques to achieve selective ablation of perceived clinically significant prostate cancers [21,22]. This has been driven by the rationale of trying to avoid the complications of radical prostatectomy by selectively eradicating localized lesions, while at the same time achieving a good oncologic outcome. Patients with apparent locally confined, small volume and unilateral cancers at initial diagnosis, and some with recurrent cancer after failed primary therapy, have been managed by procedures such as cryosurgery, brachytherapy, high-intensity focused ultrasound and radiofrequency interstitial ablation, with varying degrees of success. A comprehensive analysis of these procedures and their outcomes is beyond the scope of this article.

In light of the well-documented multifocality of most prostate carcinomas and concomitant difficulty in the preoperative delineation of all clinically significant tumors, and the difficulties involved in accurately identifying and analyzing all lesions before treatment, particularly small lesions of high grade, it becomes questionable whether focused treatments such as those described in the last paragraph are appropriate treatment options in current practice. Even extended saturation biopsy techniques may fail to identify small but significant cancers. The experimental use of 3D computed tomography prostate reconstruction has shown some potential in increasing the reliability of biopsies, but this technique is still in its early stages. Localized treatments based on imperfect biopsy and imaging data may result in a failure to ablate all pathologically significant lesions in a prostate, which raises concern regarding long-term oncologic control and progression of untreated cancers elsewhere in the gland. Additional reliable data are needed before such treatments become standard of care. It is encouraging that a novel molecular approach, such as urinary PCA3 gene test, may more reliably predict multifocality and identify patients eligible for focal therapy [23].

In summary, the majority of prostates with adenocarcinoma contain multiple, topographically separate and clonally distinct foci of cancer. Intrafocal and interfocal genetic heterogeneity demonstrate the clonal development and evolution of morphologically distinct tumors. Accurate anatomic and histologic delineation of the full extent of disease is crucial in accurately predicting outcome but is not always reliably achieved by current biopsy techniques. The imperfect association between tumor volume, grade and aggressiveness has emphasized the importance of cancer multifocality and the determination of the pathologic characteristics of all of the independent foci in the prostate in accurately estimating prognosis and determining treatment plans, including focal ablative therapy.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.