Prostate cancer gleason score

Pathologists glean a wealth of clinical detail from the smallest piece of tissue

Prostate cancer is a leading cause of morbidity and mortality in men, accounting for about 30% of all new cases of cancer and 14% of deaths from cancer. Despite considerable advances in our ability to detect and treat prostate cancer, there have been no significant corresponding decreases in morbidity and mortality.[1] The two main issues for clinicians and pathologists involved in prostate cancer are early detection of the cancer and identifying the prognostic factors that predict outcome in individual patients.[2]

Early detection of prostate cancer, preferably in the preinvasive phase (in lesions such as high grade prostatic intraepithelial neoplasia), is important if a treatment can be found that will arrest development of the cancer. Although a relatively new concept, chemoprevention is a promising strategy for preventing or arresting the development of prostate cancer and is most effective in the early stages of cancer formation, when reversibility may be feasible.[3]

Much research effort has also gone into the prognostic factors that can predict outcome in individual patients with prostate cancer, and these were the subject of two recent international consensus conferences.[4 5] The goal is to tailor the therapeutic approach to the clinical, morphological, and molecular ligatures of each patient. Many of the clinically important predictive factors in prostate cancer are still derived from a pathologist's examination of tissue specimens using light microscopy, but the challenge of assembling the information is such that the use of artificial neural networks is expected to improve accuracy in diagnosis, staging, and treatment outcomes for prostate cancer.[4 5]

In the first of the consensus conferences, organised by the College of American Pathologists,[4] a multidisciplinary group of clinicians, pathologists, and statisticians analysed the existing predictive factors and stratified them into categories reflecting the strength of published evidence and taking into account the opinions of the prostate working group members of the College of American Pathologists. Factors were ranked as: category I--those proved to be of prognostic importance and useful in clinical patient management; category II--those that have been extensively studied biologically and clinically but whose importance remains to be validated in statistically robust studies; and category III--all other factors not well enough studied to show their prognostic value.

This ranking was endorsed by the World Health Organization's second international consultation on prostate cancer,[5] whose emphasis was mainly on biopsy derived predictive factors. In particular, this meeting recommended the adoption in clinical practice of all the pathology factors in category I; stated that those in category II may be included, based on local discretion; and did not recommend the use of any of the factors in category III.[5]

Factors in category I include preoperative serum prostate specific antigen level, TNM stage grouping, histological grade, and surgical margin status. In particular, the Gleason system is recommended as the standard for histological grading of prostate cancer. The Gleason score should be reported as the composite score and its component patterns--for example, Gleason 7 = 4 + 3, and it is recommended that the first reported pattern should be the most common and the second reported pattern the second most common. The highest grade pattern should also be reported, regardless of frequency. If the sample is taken at needle biopsy the Gleason score of the entire sample should be reported as a composite score. The WHO conference additionally recommended the use of WHO nuclear grade, pathological effects of treatments, and location of cancer within the prostate as prognostic markers.

Category II factors include tumour volume, histological type, and DNA ploidy. A fair to good correlation exists between the amount of cancer reported in biopsy specimens and that subsequently found in specimens taken at radical prostatectomy.[4 5] This correlation is greatest for large cancers. For this reason pathologists should include in their reports the total percentage of cancer in the total number of needle biopsy segments. Although the volume of cancer in specimens taken at radical prostatectomy has been shown to have predictive value in only some studies, also including this determination may be of value until further data are available.

Factors in category III include perineural invasion, neuroendocrine differentiation, microvessel density (angiogenesis), nuclear roundness, chromatin texture, other karyometric factors, proliferation markers, prostate specific antigen derivatives, and other factors (oncogenes, tumour suppressor genes, apoptosis genes, etc). Although the evidence of these is not yet good enough for their use in clinical practice, two factors in category III show particular promise--neuroendocrine differentiation and angiogenesis.[6 7]

Prostate cancer may show divergent differentiation towards a neuroendocrine phenotype in the form of neuroendocrine small cell carcinoma or carcinoid-like tumours.[6] Much more common, however, is focal neuroendocrine differentiation in prostate cancer, which may be pronounced in about 10% of carcinomas. The prognostic importance of local neuroendocrine differentiation in prostate cancer is controversial, but current evidence suggests that it has an influence on prognosis related to hormone resistant tumours or a role in the conversion to a hormone resistant phenotype.[5] Chromogranin A appears to be the best overall tissue and serum marker of neuroendocrine differentiation, and thus serum chromogranin A concentrations may be useful in assessing the emergence of or progression of hormone resistant cancer.[6]

Recent studies have shown that angiogenesis is a potent prognostic indicator for patients with prostate cancer and that the evaluation of endothelial growth factors is useful in assessing the angiogenic phenotype in prostate cancer.[7] At first, vascular density was assessed by staining endothelial cells in prostate tissue sections with antibodies specific to factor VIII related antigen. When normal prostate, nodules in benign prostatic hyperplasia, premalignant lesions (prostatic intraepithelial neoplasia), and prostate cancer were compared, an altered pattern and a progressively increasing number of microvessels were observed. Next to normal epithelium the vascular network appears orderly, but in cancer microvessels appear more randomly distributed.

It has been proposed that neovascularisation begins in benign prostatic hyperplasia and that it keeps progressing in a stepwise manner in the subsequent proliferative premalignant and malignant stages.[7] Furthermore, in prostate cancer the number of microvessels differs between clinically localised disease with a low Gleason score and that with a high Gleason score, reaching, in the latter, levels similar to those observed in prostate cancer with bone metastases. Significant tissue levels of vascular endothelial growth factor are present in prostate cancer and in prostatic intraepithelial neoplasia lesions, the expression being highest in association with neuroendocrine cells and correlated with an altered pattern of vascularisation. The vascular endothelial growth factor expression is downregulated by hormonal manipulation, except in the population of neuroendocrine cells. All this indicates that vascular endothelial growth factor may contribute to the establishment, progression, and regression of prostate neoplasia.

Oncogenes such as c-myc, c-erb-[B.sub.2] and bcl-2 are potential candidate markers, as are various tumour suppressor genes such as p27(Kip1), pp32r1/2, and PTEN, following findings which implicate their involvement in the control of development and progression of prostate cancer. Several metastasis suppressor, or anti-invasion, genes are also implicated in the progression of prostate cancer and are considered as potential future prognostic markers.[1]

Thus a wealth of clinical information is available in even the smallest tissue specimens of the prostate, including those taken at biopsy and during transurethral resections. Substantial effort has been expended recently years in describing the available factors and determining their predictive value for staging, cancer recurrence, and patient survival.

Rodolfo Montironi professor of pathology

Institute of Pathological Anatomy, University of Ancona School of Medicine, Regional Hospital, 60020 Torrette di Ancona, Italy (r.montironi@popcsi.unian.it)

RM is involved in two Italian trials evaluating the effect of androgen ablation on prostate cancer. These trials are funded by Takeda Italy and AstraZeneca Italy.