Introduction

In 2012, an estimated 241,740 prostate cancer diagnoses will be made in the United States, accompanied by an estimated 28,170 prostate cancer deaths.¹ Beginning around 1994-1996, with widespread use of serum prostate-specific antigen (PSA) testing and digital rectal examination (DRE) for prostate cancer screening, and with increased treatment of clinically localized prostate cancer with surgery or radiation therapy, age-adjusted prostate cancer death rates have fallen steadily. Although this trend might indicate a beneficial impact of prostate cancer screening and/or early prostate cancer treatment on prostate cancer mortality, mass screening of the general population for prostate cancer remains controversial.² One challenge for prostate cancer screening is the prevalence of the disease in the United States: autopsy series have revealed small prostate cancers in as many as 29% of men between ages 30 and 40 and 64% of men between ages 60 and 70.³ Obviously, not all of these men are at risk for symptomatic or life-threatening prostate cancer progression. In fact, many such men, if diagnosed with prostate cancer, may be at greater risk for treatment-associated morbidity.

Currently, for U.S. men, the lifetime risk of a diagnosis of prostate cancer is about 1 in 6, whereas the lifetime risk of death from prostate cancer is on the order of 1 in 35.¹ Over the past two decades, treatment approaches for men with prostate cancer have changed dramatically, with improvement in established prostate cancer treatments and the introduction of new prostate cancer treatment approaches. Now, men diagnosed with prostate cancer often face a bewildering array of treatment choices. Clearly, the physicians that care for these men must weigh the risks of prostate cancer progression against the potential for side effects from treatment, in the context of other health risks and life choices, to best use the current collection of treatments for the greatest benefit.

To aid physicians who care for men with prostate cancer, this chapter will provide an overview of prostate cancer etiology, biology, screening, detection, diagnosis, prevention, and treatment.

Prostate Anatomy and Function

The prostate is a male sex accessory gland that surrounds the urethra and contributes secretions to the ejaculate (Figure 84-1). Located in the pelvis, the prostate sits adjacent to the bladder and rectum, is surrounded incompletely by a thin capsule composed of collagen, elastin, and smooth muscle, and at the apex of the gland, forms part of the urethral sphincter apparatus.⁴ Nerves to the corpora cavernosa of the penis, needed for penile erection, travel through fascia along the posterolateral surface of the prostate. These nerves can be recognized as a neurovascular bundle by urologists and preserved during radical prostatectomy to minimize sexual dysfunction postoperatively.^5,6 The prostate parenchyma has been divided into three zones that can be seen by transrectal ultrasonography, and recognized readily by surgical pathologists examining radical prostatectomy specimens: a central zone, surrounding the ejaculatory ducts and accounting for some 25% of the prostate; a transition zone, near the prostatic urethra with 10% of prostate tissue normally; and a peripheral zone, with the bulk of prostate tissue encompassing the posterolateral region of the prostate (Figure 84-2).^7,8

In addition to prostate cancer, the prostate also frequently manifests benign enlargement (benign prostatic hyperplasia or BPH) and chronic or recurrent inflammation (prostatitis). Like prostate cancer, each of these conditions can elevate the PSA, confounding the use of serum PSA testing for prostate cancer screening. When present, BPH tends to be located near the prostatic urethra (in the transition zone), whereas prostate cancer and the prostate cancer precursor lesions proliferative inflammatory atrophy (PIA) and prostatic intraepithelial neoplasia (PIN) usually arise in the periphery (the peripheral zone). Prostatic inflammation, though often prominent in the peripheral zone, can be seen throughout the prostate. Remarkably, although prostate cancer, BPH, and prostatitis all commonly afflict U.S. men, and can be simultaneously present in a single prostate gland, mechanistic associations of the three diseases have been difficult to demonstrate.

The prostate requires androgenic hormones, and an intact androgen receptor, for normal growth and development. In the prostate, the major circulating androgenic hormone, testosterone, produced by Leydig cells in the testes upon stimulation by luteinizing hormone (LH), is converted by 5α-reductase (nicotinamide-adenine dinucleotide phosphate-dependent δ⁴-3-ketosteroid 5α-oxidoreductase) to 5α-dihydrotestosterone (DHT).⁹ DHT, a more potent androgen than testosterone, binds to intracellular androgen receptors, alters androgen receptor conformation to promote dissociation from chaperone proteins, triggers androgen receptor dimerization and transport into the cell nucleus, and activates the expression of selected target genes.¹⁰ Stereotypically, androgen receptor target genes are characterized by the presence of androgen response element (ARE) DNA sequences within the transcriptional regulatory region, permitting direct binding and trans-activation by the androgen receptor.¹¹ For genes like KLK3 (encoding PSA), which are activated by the androgen receptor selectively in prostate cells, and not in cells of other tissues, the transcriptional regulatory region also contains additional DNA sequences (prostate-specific enhancer or PSE) conferring prostate-specific expression.¹²

The normal prostate epithelium is composed of basal epithelial cells, characterized by the expression of cytokeratins K5 and K14, and p63, columnar secretory epithelial cells, which express the androgen receptor, PSA, cytokeratins K8 and K18, prostate-specific membrane antigen (PSMA), and prostate-specific acid phosphatase (PAP); and rare neuroendocrine cells, that secrete chromogranin A, neuron-specific enolase, and synaptophysin (Figure 84-3).¹³ The basal epithelial cell compartment likely contains pluripotent prostatic stem cells, capable of self-renewal proliferation and of differentiation. In contrast, columnar secretory cells, specialized to produce secretions for the ejaculate, are terminally differentiated, particularly under the influence of androgenic hormones. The prostate epithelium is in turn supported by a stroma containing fibroblasts, smooth muscle cells, nerves, and blood vessels. Stromal cells, which also express the androgen receptor, secrete polypeptide growth factors, such as keratinocyte growth factor (KGF), that contribute to the regulation of epithelial homeostasis via a paracrine signaling mechanism.^14,15 Abnormal stromal–epithelial interactions, with disordered regulation of epithelial cell proliferation and differentiation, may contribute to the pathogenesis of both prostate cancer and BPH.¹⁶

Prostate cancer cells, and PIN cells, arise from the prostatic epithelium. Even though transformed, such cells typically retain many of the phenotypic attributes characteristic of differentiated columnar secretory cells, including the expression of androgen receptor, PSA, PSMA, and PAP. Prostate cancers reminiscent of basal epithelial cells are exceptionally rare; prostate cancers with features of neuroendocrine cells are somewhat more common.¹⁷ However, unlike normal columnar epithelial cells, neoplastic prostate epithelial cells are capable of proliferation. This has led to the concept that the target cell for neoplastic transformation in the prostate may be an “intermediate” cell, in transit from a basal epithelial stem cell to a differentiated columnar secretory epithelial cell, with properties of both stem cells and differentiated cells.^18,19 Another feature of neoplastic prostate epithelial cells, as compared with normal basal or columnar secretory cells, is that the neoplastic cells appear to use androgen receptor signaling not only for differentiation, but also for proliferation, as most prostate cancer cells tend, at least initially, to display some dependence on androgens for maintenance of growth and survival. Somatic fusions between an androgen-regulated gene, TMPRSS2, at chromosome 21q22, and genes encoding members of the ETS family of transcription factors, commonly found in prostate cancers, may provide a mechanistic explanation by which androgen signaling can promote prostate cancer cell growth.^20,21 Ultimately, in life-threatening prostate cancer, prostate cancer cells escape from the prostate gland, proliferate in lymph nodes, in bones, and in other organs and become less and less dependent on androgenic hormones.

Familial clusters of prostate cancer have been recognized since at least 1956, when Morganti et al. reported that men with prostate cancer were more likely to have relatives with prostate cancer than men without a prostate cancer diagnosis.²² In a study conducted more than three decades later, when detailed family histories were collected from men with prostate cancer and their spouses, the men with prostate cancer were more likely to have a brother or father with prostate cancer.²³ Twin studies, comparing the tendency for concordant prostate cancer development between monozygotic twins, sharing all of their genes, and dizygotic twins, sharing half of their genes, have also hinted at a significant contribution of hereditary to prostate cancer: in a study of 44,788 pairs of twins in Sweden, Denmark, and Finland,²⁴ 42% of the prostate cancer cases (with a 95% confidence interval of 29% to 50%) were attributed to heredity. In principle, familial clustering of prostate cancer cases could be a result of inherited susceptibility genes, shared exposure to carcinogenic stresses, or to some sort of detection or diagnosis bias (e.g., the brother of a man diagnosed with prostate cancer may be more likely to pursue screening for prostate cancer). To discover the genetic contributions to prostate cancer in spite of these potential sources of bias, both linkage analyses of genetic loci in high-risk prostate cancer families and genomewide association studies (GWAS) of prostate cancer susceptibility in large populations have been conducted. Each approach has generated a number of candidate genes or gene regions (as many as 30 or more), with several in common, confirming the contribution of heredity to prostate cancer risk but underscoring the complexity of inherited prostate cancer susceptibility.²⁵

Among the growing number of prostate cancer susceptibility genes discovered by mapping studies are RNASEL and MSR1.26,27 RNASEL encodes a latent endoribonuclease component of an interferon-inducible 2′,5′-oligoadenylate-dependent RNA decay pathway that functions to degrade viral and cellular RNA upon viral infection.²⁸ MSR1 encodes subunits of a trimeric class A macrophage scavenger receptor capable of binding bacterial lipopolysaccharide and lipoteichoic acid, and oxidized high and low density serum lipoproteins (oxidized HDL and LDL).²⁹ For both RNASEL and MSR1, not only have mutations been linked to prostate cancer susceptibility in families but variant alleles have been predicted to account for as many as 13% and 3% of sporadic prostate cancer cases, respectively.^27,30,31 The identification of RNASEL and MSR1 as candidate prostate cancer susceptibility genes has intensified interest in the possibility that infection and/or inflammation might contribute to the pathogenesis of human prostate cancer. In mice, targeted disruption of RnaseL leads to diminished interferon-α activity and increased susceptibility to viral infection,³² whereas targeted disruption of Msr-A leads to increased vulnerability to infection with Listeria monocytogenes, Staphylococcus aureus, Escherichia coli, and Herpes simplex virus type 1.^29,33–35 Further genetic support for this etiologic mechanism has come from analyses of common variants of other genes encoding participants in host inflammatory responses, including TLR4 and other members of toll-like receptor signaling pathways, MIC-1, IL1-RN, and COX-2, have also been associated with prostate cancer risk.^36–40

An increased risk for prostate cancer has long been known for men in breast cancer families carrying BRCA2 mutations, characterized by disease with an aggressive natural history arising before age 55 years.⁴¹ Nonetheless, a role for BRCA2 genotyping in general prostate cancer practice has not been established. However, as evidence has accumulated that several additional germline DNA sequence variants may be associated with prostate cancer, the possibility that genetic testing might be used to aid in prostate cancer screening, detection, diagnosis, or risk stratification has emerged. In one analysis, five such sequence variants, three single nucleotide polymorphisms (SNPs) at 8q24 and one each at 17q12 and 17q24.3, were found to have a marked association with prostate cancer, especially for men with a family history of the disease, showing a 9.46-fold increased risk (with 95% confidence interval of 3.62 to 24.72) for a prostate cancer diagnosis.⁴²

Prostate Inflammation and Prostate Cancer

Chronic or recurrent inflammation is known to play a causative role in the development of many human cancers, including cancers of the liver, esophagus, stomach, large intestine, and bladder. Inflammatory changes have been recognized in prostate tissues for many years, leading to speculation that inflammation might contribute in some way to prostate cancer development.⁵⁸ However, over the past few years, evidence has accumulated in support of a more critical role for prostatic inflammation in the pathogenesis of prostate cancer. Inflammatory changes are present in almost all radical prostatectomy specimens from men with prostate cancer. Because inflammation in the prostate is not usually associated with symptoms, the prevalence of prostate inflammation is not known, and the association with prostate cancer has been difficult to test.^59,60 A syndrome of irritative voiding symptoms and pelvic pain, perhaps attributable to inflammation near the prostatic urethra, is reported by some 9% or more of men between 40 and 79 years of age, with as many as 50% of such men suffering more than one episode by age 80 years.⁶¹ Most episodes of symptomatic prostatitis are not clearly attributable to specific infectious agents. Even so, sexually transmitted infections do appear to increase prostate cancer risk.^62,63 Nonetheless, if prostate infection and inflammation lead to prostate cancer, the mechanism does not appear likely to involve direct transformation of prostate epithelial cells by microbial DNA. Instead, the production of microbicidal oxidants by inflammatory cells, such as superoxide, nitric oxide, and peroxynitrite, may promote prostate cancer development by triggering cell and genome damage.^64,65 Increased production of oxidants by inflammatory cells in the prostate may be why decreased prostate cancer risk has been associated with intake of a variety of antioxidants or of nonsteroidal antiinflammatory drugs, and why RNASEL and MSR1, two of the prostate cancer susceptibility genes identified thus far, encode proteins that function in host responses to infections.

Despite these provocative hints, the contribution of prostate inflammation to prostatic carcinogenesis has been difficult to assess. However, in 1999, De Marzo et al. provided the most compelling linkage of prostate inflammation to prostate cancer by proposing that a prostate lesion, termed proliferative inflammatory atrophy (PIA), might be a precursor to PIN and to prostate cancer (Figure 84-4).⁶⁶ Areas of the prostate, containing epithelial cells that do not fully differentiate into columnar secretory cells, have long been recognized as focal atrophy lesions by prostate pathologists.^58,67 The term PIA has been used to describe those focal atrophy lesions that contain proliferating epithelial cells, are associated with chronic inflammation, and are often located adjacent to PIN lesions and/or prostate cancers.⁶⁸ The epithelial cells in PIA lesions typically express high levels of stress-response polypeptides such as GSTP1, GSTA1, and cyclooxygenase 2 (COX-2). Loss of GSTP1 expression in rare PIA lesions, attributable to de novo GSTP1 CpG island hypermethylation, may be what leads to the development of PIN and prostate cancer.⁶⁹

The hypothesis that inflammation might promote prostate cancer development offers new challenges to prostate cancer epidemiology, to the search for prostate cancer susceptibility genes, and to the molecular pathogenesis of prostate cancer. Although prostatic inflammation is common in regions of the world with high prostate cancer risks, whether regions of the world with low prostate cancer risks have less prostatic inflammation has not been determined. Provocative new data hint that PIA lesions appear to be more common in the prostate peripheral zone in men from high-risk prostate cancer regions than in men from low-risk prostate cancer regions. To better test the association of prostate inflammation and prostate cancer, perhaps new biomarkers of prostate inflammation, assayable in blood, urine, or prostate fluid, can be developed for use in epidemiology studies. In addition, polymorphic genes encoding regulators of immune responses will likely need to be systematically evaluated for prostate cancer risk associations. Ultimately, if the hypothesis is correct, prostate cancer risks may be reduced by therapeutically attenuating prostate inflammation.

Prostate cancer cells typically contain a plethora of somatic genome alterations, including gene mutations, gene deletions, gene amplifications, chromosomal rearrangements, and changes in DNA methylation (Figure 84-5). In the United States, prostate cancer diagnoses are typically made in men between 60 and 70 years of age, whereas small prostate cancers have been detected at autopsy in nearly 30% of men between 30 and 40 years of age.³ Thus, the somatic genome changes present in prostate cancers have often accumulated over many decades. The acquisition of somatic genome changes in the prostate may be influenced by lifestyle as well: although small prostate cancers have been detected at autopsy in men from geographic regions with low prostate cancer mortality, these small prostate cancers are usually only present in much older men.^72–72 Also, in the United States, prostates removed at radical prostatectomy for prostate cancer usually contain more than one prostate cancer lesion (Figure 84-6).

Over the years, several techniques have been used to catalog genome accidents in prostate cancer cells, including karyotyping, fluorescence in situ hybridization (FISH), comparative genome hybridization, loss of heterozygosity analyses, and genomewide microarray and/or sequencing approaches. In one such study, each prostate cancer case exhibited a mean of 3866 base mutations (range 3192 to 5865), 20 nonsilent coding sequence mutations (range 13 to 43), and 108 rearrangements (range 43 to 213).⁷³ In another, DNA hypermethylation was found at 5408 regions of the genome, with 73% of the regions near genes (5′, 3′, or intron–exon junctions), and 27% of the regions at conserved intergenic sites.⁷⁴ Often, these analyses reveal different chromosomal abnormalities in different cancer cases, in different cancer lesions in the same cancer case, and in different areas within the same cancer lesion. The propensity to develop such a heterogeneous collection of somatic genome lesions over so many years, and in a manner so sensitive to environment and lifestyle, suggests strongly that prostate cancers likely arise as a consequence of either chronic or recurrent exposure to genome-damaging stresses, defective protection against genome damage, or some combination of both processes. The resultant genomic instability may be the reason some prostate cancers progress to threaten life.^75,76

One characteristic somatic genome alteration, a rearrangement, drives the production of fusion transcripts between an androgen-regulated gene, TMPRSS2, at chromosome 21q22, and members of the ETS family of transcription factors (Figure 84-7).²⁰ Fusion partners for TMPRSS2 include ERG (also at chromosome 21q22), ETV1 (at chromosome 7p21), and ETV4 (at chromosome 17q21).^20,77,78 The gene rearrangements may result from a mishap during the androgen receptor transcriptional trans-activation of TMPRSS2 in which tangling or untangling of DNA by TOP2B triggers DNA double-strand breaks that recombine with ETS family partner genes via nonhomologous end-joining (Figure 84-8).^79,80 The appearance of the resultant fusion transcripts provides a plausible mechanism for the dependence of prostate cancer cells on androgenic hormones for growth and survival, as the expression of ETS family transcription factors can be stimulated by androgen action. TMPRSS2–ERG fusions have been detected in ~60% of prostate cancers and in >20% of PIN lesions.⁸¹ ERG is highly expressed by many prostate cancers, and not at all by others, though neither ERG expression nor the presence of ERG fusion transcripts appears to have great prognostic significance.^81–84 Nonetheless, the presence of TMPRSS2-ERG fusion transcripts does seem to define a molecular subset of prostate cancer. Rarer somatic alterations, including SPOP mutations (6% to 15%) and SPINK1 overexpression (~10%), are restricted to cases devoid of TMPRSS2-ETS family rearrangement rearrangements.^85,86

Hypermethylation of CpG island sequences encompassing the regulatory region of GSTP1, encoding the π–class glutathione S-transferase (GST) is the most common somatic genome change yet reported for prostate cancer.87,88 GSTs catalyze the detoxification of carcinogens, and of other reactive chemical species, via conjugation with the intracellular scavenger glutathione. In mice, targeted disruption of π-class GST genes leads to increased skin tumors after treatment with the carcinogen 7,12-dimethylbenzanthracene (DMBA).⁹⁰ Similarly, human prostate cancer cells devoid of GSTP1 appear especially vulnerable to genome damage mediated by exposure to N-OH-PhIP, the charred meat carcinogen that causes prostate cancer when fed to rats, and by exposure to oxidant stresses.⁹¹ In the normal prostate epithelium, GSTP1 is present at high levels in basal cells, and in lower levels in columnar secretory cells, though the enzyme can be induced in columnar epithelial cells subjected to genome-damaging stresses. In contrast, the enzyme is almost never present in prostate cancer cells. In nearly all cases, the absence of GSTP1 expression in prostate cancer cells can be attributed to hypermethylation of GSTP1 CpG island sequences, a somatic genome change that prevents GSTP1 transcription. Absence of GSTP1 expression and GSTP1 CpG island hypermethylation may also be characteristic of cells comprising PIN lesions, thought to be precursors to prostate cancer.⁹² The mechanism by which hypermethylated GSTP1 CpG island alleles arise during prostatic carcinogenesis remains to be elucidated. Nonetheless, prostate cells carrying inactivated GSTP1 genes appear to enjoy some sort of selective growth advantage early during the development of prostate cancer.

NKX3.1 encodes a prostate-specific homeobox gene essential for normal prostate development that may be a target for somatic loss on chromosome 8p21.⁹³ NKX3.1 has been shown to bind DNA and to repress PSA expression via interactions with ETS transcription factors.^94,95 Mice carrying one or two disrupted Nkx3.1 alleles manifest prostatic epithelial hyperplasia and dysplasia.^96,97 In men, loss of 8p21 DNA sequences occurs early during prostatic carcinogenesis, with 63% of PIN lesions and >90% of prostate cancers, showing loss of heterozygosity at polymorphic 8p21 marker sequences in one report.⁹⁸ However, although mapping studies have indicated that NKX3.1 lies within a common region of deletion, encompassing 2 megabases at 8p21, molecular pathology analyses have not yet established NKX3.1 as a somatic target for inactivation during prostatic carcinogenesis because somatic NKX3.1 mutations have not been identified. Nonetheless, loss of NKX3.1 expression does appear to accompany prostate cancer progression.

PTEN, a tumor suppressor gene encoding a phosphatase active against both proteins and lipid substrates, appears to be a common target for somatic alteration during prostate cancer progression.99–106 PTEN is an inhibitor of the phosphatidylinositol 3′-kinase/protein kinase B (PI3K/Akt) signaling pathway needed for cell cycle progression and cell survival. Although PTEN is expressed by normal prostate epithelial cells, and by cells present in PIN lesions, the expression of PTEN is often diminished in prostate cancers, with many prostate cancers containing collections of neoplastic cells with no PTEN.¹⁰⁷ PTEN defects have been found in a wide variety of cancers and cancer cell lines.¹⁰¹ For prostate cancer, a number of somatic PTEN alterations have been reported, including homozygous deletions, loss of heterozygosity, mutations, and probable CpG island hypermethylation. However, despite common losses of 10q sequences near PTEN in prostate cancers, somatic mutations at the remaining PTEN alleles are not as frequent. In a study of prostate cancer metastases recovered at autopsy, somatic PTEN alterations were even more common than in primary prostate cancers, and a significant heterogeneity in PTEN defects in different metastatic deposits from the same patient was also evident.¹⁰⁶ Haploinsufficiency for PTEN may contribute to the phenotype of transformed cells in the prostate. Pten^+/– mice display prostatic hyperplasia and dysplasia, and crosses of Pten^+/– mice with Nkx3.1^+/– mice have revealed that Pten^+/–Nkx3.1^+/– mice and Pten^+/–Nkx3.1^–/– mice develop lesions reminiscent of human PIN.^110–110

Defective regulation of p27, a cyclin-dependent kinase inhibitor encoded by CDKN1B, may also accompany prostatic carcinogenesis.111,112 In PIN cells and prostate cancer cells, p27 levels are almost always diminished, though the mechanism(s) for the reduction in p27 levels appear complex: somatic loss of DNA CDKN1B sequences at 12p12-13 have been reported for only 23% of localized prostate cancers, 30% of prostate cancer lymph node metastases, and 47% of distant prostate cancer metastases.¹¹³ In place of CDKN1B gene alterations, p27 polypeptide levels may be lowered indirectly by inadequate PTEN repression of the PI3K/Akt signaling pathway.^116–116 In this way, low p27 levels may be as much a result of loss of PTEN function as of CDKN1B alterations. The critical contribution of PTEN to epithelial growth regulation in the prostate is evident in mice, where disruption of Cdkn1b alleles leads to prostatic hyperplasia, and Pten^+/–Cdkn1b^–/– mice develop prostate cancer by 3 months of age.^111,117

Metastatic prostate cancer is almost always treated with androgen deprivation, antiandrogens, or a combination of androgen deprivation and antiandrogens.118,119 However, despite such treatment, androgen-independent prostate cancer cells eventually emerge and progress to threaten life. Curiously, in these cells, androgen receptor expression and androgen receptor signaling remain intact despite the absence of androgens.^120,121 Somatic alterations of AR have been reported for many prostate cancers, especially for androgen-independent prostate cancers. AR amplification, accompanied by high-level expression of androgen receptors, may promote the growth of androgen-independent prostate cancer cells by increasing the sensitivity of the cells to low androgen levels.¹²² AR mutations, encoding androgen receptors with altered ligand specificity have also been detected; for some of the mutant androgen receptors, even antiandrogens can act as agonist ligands.^125–125 When 44 mutant androgen receptors from prostate cancers were evaluated for transcriptional regulatory capabilities, 16% of the receptors had lost transcriptional activation activity, 45% of the receptors had gained some transcriptional regulatory ability, 32% of the receptors maintained some partial transcriptional modulatory activity, and the remaining 7% behaved like wild-type receptors.¹²⁶ In addition to somatic AR gene changes, androgen-independent prostate cancer cells with wild-type androgen receptors may activate androgen receptor signaling even in the absence of androgens, via posttranslational modifications of the androgen receptor and/or androgen receptor coactivators in response to other growth factor signaling pathways.^{120,127–130}

With widespread opportunistic prostate-specific antigen (PSA)–based screening for prostate cancer, most cases are diagnosed at a stage when symptoms due to the disease are absent. Nevertheless, locally advanced disease can cause lower urinary tract symptoms (obstructive or irritative), hematospermia, a decrease in ejaculate volume, and erectile dysfunction if the cancer extends beyond the prostate to involve the bladder neck, urethra, ejaculatory ducts, seminal vesicles, or erectogenic nerves, respectively. Metastatic involvement of the skeleton, bone marrow, pelvic lymph nodes, or periureteral lymphatics can lead to bone pain, pancytopenia, lower extremity edema, or retroperitoneal fibrosis, respectively. Paraneoplastic syndromes from ectopic hormone production by small cell variants of adenocarcinoma, and disseminated intravascular coagulation (DIC) have been associated with metastatic prostate cancer.

Serum PSA and Prostate Cancer Detection

The serum PSA value, along with the findings at DRE, correlate directly with the risk of prostate cancer at biopsy (Table 84-1). Furthermore, the PSA level at a young age anticipates the future risk of being diagnosed with prostate cancer decades later (Figure 84-9). Gann et al. first showed that the risk of a prostate cancer diagnosis, including the diagnosis of life-threatening disease, incrementally and directly with PSA over the decade after a baseline measurement, even at low PSA levels (below 4.0 ng/mL); a finding that has now been confirmed by many others.^173–173 These observations may allow a more targeted approach to prostate cancer screening using a baseline PSA value to direct screening intensity.^174,175

Table 84-1

Positive Predictive Value of DRE and PSA in a Multicenter Screening Trial

DRE	PSA	PPV (%)
Abnormal	Any	21.4
Any	>4	31.5
	4–10	26.1
	>10	52.9
Normal	>4	24.4
Abnormal	<4	10.0
	4–10	40.8
	>10	69.1

DRE, Digital rectal examination; PPV, positive predictive value; PSA, prostate-specific antigen.

Data from Catalona WJ, Richie JP, Ahmann FR, et al. Comparison of digital rectal examination and serum prostate specific antigen in the early detection of prostate cancer: results of a multicenter clinical trial of 6,630 men. J Urol 1994;151:1283.

In the early years of PSA testing, most clinicians used PSA as a dichotomous test (i.e., the PSA was “elevated” or not) with elevations triggering a prostate biopsy. However, data from the placebo arm of the Prostate Cancer Prevention Trial (PCPT¹⁷⁶), which accrued men with a serum PSA value <3.0 ng/mL, has suggested that there is no PSA cut-off level with both high sensitivity and high specificity for prostate cancer, and that virtually no level is low enough to exclude the presence of prostate cancer. Instead, serum PSA tests may be more appropriately thought of as measures of a continuum of prostate cancer risk, with the risk of cancer (and of high-grade cancer) rising directly with the PSA level, an observation first made by Gann et al. using a cohort population study design.¹⁷¹

Buy Membership for Hematology, Oncology and Palliative Medicine Category to continue reading. Learn more here