Prostate Cancer Research Paper

ANATOMY, PHYSIOLOGY, AND PATHOLOGY

The prostate is a walnut-sized gland located at the base of the pelvis beneath the urinary
bladder, surrounding the urethra (Figure 1). Its initial development is stimulated by
testosterone from the fetal testes but it remains incompletely developed until puberty, when it
grows to its adult size of about 20 g around age 20 years. The prostate develops 30–50
branched ducts that are lined with glandular epithelium and open into the prostatic urethra. Its
function is to contribute up to 30% of the volume of semen with slightly alkaline secretions
including prostaglandins, proteolytic enzymes, acid phosphatase, zinc, and citric acid, which
help maintain sperm viability.

McNeal (1981) first identified the prostate’s zonal anatomy. The peripheral zone, where 60–
70% of tumors occur, contains about 70% of the total volume and is accessible via digital
rectal examination (DRE). The central zone, containing 25% of the volume and the
ejaculatory ducts, is where inflammatory processes such as prostatitis arise. The transitional
zone, where 25% of tumors occur, contains only 5% of the volume. The anterior zone is
fibromuscular.

As the hormonal milieu changes with middle age, the prostate grows, a condition termed
benign prostatic hyperplasia (BPH). Due to the inelasticity of the prostate’s fibrous outer
capsule, increasing pressure is placed on the prostatic urethra, producing urinary symptoms.
Depending on severity, these problems come to medical attention and in this context cancer is
often diagnosed. BPH was once considered a risk factor for prostate cancer, but this view is
no longer held; both conditions simply occur commonly in the aging prostate gland. A small
proportion of cancers are found in tissue fragments from transurethral resections performed to
relieve obstructive symptoms of BPH.

Virtually all prostatic tumors are adenocarcinomas arising from the glandular epithelium and
most men will develop them if they live long enough. Although much research has sought
tumor markers that would separate the tigers from the pussy cats, this is still beyond reach. A
tumor’s potential lethality (aggressiveness) is currently based on its degree of spread beyond
the capsule and microscopic inspection by a pathologist, who identifies the two largest tumor
foci (primary and secondary patterns) and grades each from 1 to 5 according to the histologic
features described by Gleason (Gleason, 1977) (see Figure 2).
Gleason grades for the two foci, which are rarely more than one grade apart, are added to
give a sum from 2 to 10. Tumors with Gleason sums below 5 are not considered aggressive.
Historically, tumors with Gleason sums 5–7 were considered low risk but tumors with a
Gleason score of 7 have been added to the high-grade aggressive category (formerly Gleason
scores 8–10). The significance of Gleason sum 7 tumors remains controversial, especially
with respect to decisions about treatment. Approximately 30% of Gleason sum 7 tumors have
a primary focus of grade 4 and these are considered to be more aggressive than those having
a primary focus of grade 3. Some clinicians consider that any focus of Gleason grade 4
warrants clinical suspicion.

DIAGNOSIS, SCREENING, AND TREATMENT

Prostate cancer has no specific symptoms until advanced. Its size and the presence of any
lumps are assessed by DRE and ultrasound; the diagnosis of cancer is aided by measurement
of the serum levels of prostate specific antigen (PSA), a protease enzyme contributed by the
prostate to seminal fluid, some of which permeates into the bloodstream. Originally used
clinically to monitor cancer progression after treatment, PSA is now used for early detection.
High serum PSA levels almost certainly indicate the presence of malignancy, but its use also
leads to overdiagnosis, that is, the diagnosis of tumors that would never have progressed to
clinically significant tumors during life, the so-called pussycats.

Suspicious PSA levels are followed up with ultrasoundguided needle biopsies to establish a
histopathologic diagnosis. Prostate tumors are commonly small and multifocal, and there is
an element of chance in whether a tumor can be adequately identified by biopsy. The number
of biopsies used to establish a diagnosis has increased over time. The PSA threshold level
used to prompt clinical investigation has also fallen, especially in the United States.

PSA testing for early prostate cancer is already being performed on an ad hoc or
opportunistic basis in many countries. Randomized clinical trials (RCTs) are being conducted
to test the efficacy of systematic PSA screening in reducing prostate cancer mortality in the
whole population (Schroder 1994). Because many controls (men who were randomly
assigned not to have a PSA test) in these RCTs have in fact received PSA tests as part of their
normal community care, the power of the trials to provide definitive evidence of efficacy has
been reduced, and this has lengthened the duration of the trials by some years (Gohagen et
al., 1994).

Screening, to be effective, requires a cheap, sensitive, and specific test, a good understanding
of the natural history to identify a curable early form of the disease, and good evidence that
treatment of this form of disease reduces mortality by comparison with treatment of disease
diagnosed in the ordinary way at the onset of symptoms. Prostate cancer screening by PSA
testing currently fails to satisfy these criteria.

Given the uncertainties that surround its biological heterogeneity and limited evidence of
treatment benefits, prostate cancer management is complex. Men with low-grade, localized
cancer and a low PSA may opt for watchful waiting, with regular repeat PSA tests to monitor
changes, termed PSA velocity. Once PSA velocity reaches a certain point, decisions are made
with respect to treatment. Radical treatments for localized disease are surgery (prostatectomy)
or implantation with radioactive seeds to deliver localized radiation (brachytherapy).
Advanced disease is treated by androgen deprivation/blockade, which can involve surgical or
medical castration, and external beam radiation. Principal concerns are side effects, such as
impotence, incontinence, and damage to the rectum and bladder neck, and the uncertainty
posed by the limited evidence for substantial benefit of treatment.

VARIATION IN INCIDENCE AND MORTALITY

Autopsy studies carried out by Yatani and others (1982) showed that the prevalence of
microscopic, nonaggressive tumors (historically termed latent cancers) increased with age,
affecting a majority of prostates by age 60 years, but varied little between populations, any
variation in total incidence being due to differences in levels of invasive disease. Because of
their high prevalence, the detection of nonaggressive cancers depends mainly on the intensity
with which they are sought. The total incidence of prostate cancer is, therefore, highly
sensitive to the medical-care setting.

Prior to PSA testing, a 30-fold international variation in prostate cancer incidence was
observed, with high rates in Western countries, particularly the United States, and low rates
in Asia. Although age-standardized incidence rates differed between countries, trends were
stable for decades prior to PSA testing (see Figure 3). After PSA testing became widespread
in the late 1980s, incidence more than doubled in some populations. This happened because
nonaggressive tumors were in effect transferred from the pool of previously undiagnosed
disease to tumors requiring clinical attention. The dramatic increase occurred earlier and rose
higher in the United States than elsewhere. After peaking in 1992, incidence fell, but it has
remained at a higher level than before. Countries where PSA testing is not common do not
show this huge increase.

Figure 4 shows age-standardized prostate cancer incidence and mortality rates for a selection
of countries centered on 1985 and 1995, before and after PSA testing. The remarkable
features are the huge differences in incidence and minor differences in mortality. PSA testing
was slow to emerge in the UK and is rare in Denmark, and this is reflected in the rates.
Against a long-term trend of slowly increasing prostate cancer mortality, some countries have
experienced a small decrease that has sparked debate about the contribution of PSA testing. It
is impossible, however, to distinguish possible early detection effects from simultaneous
improvements in treating advanced cancers. Mortality has also fallen in countries without
PSA testing and this argues against the impact of PSA testing on mortality. Such
opportunistic testing is unlikely to have reduced mortality so quickly unless the lead time
(advancement of diagnosis by the test) were extremely short, which is improbable. We will
have to wait for randomized controlled trials (RCTs) to resolve this question.

Total prostate cancer survival has improved since the introduction of PSA testing, due mainly
to the diagnosis of an increased proportion of nonaggressive tumors. Relative survival
estimates at 5 years are virtually 100% for localized cancer and are around 33% for cancer
with distant metastases.

Prostate cancer shows strong ethnic variation, with African-Americans, especially those of
Jamaican descent, having the highest reported rates of incidence and mortality in the United
States. It had long been considered that native Africans were at low prostate cancer risk but
recent surveys suggest that the incidence for thoroughly screened African populations may be
as high as for African-Americans.

Japanese men have historically experienced a low incidence of prostate cancer, but this
changed among Japanese migrants to the United States. Incidence rates for first-generation
Hawaiian Japanese rose above those for Japan and the rates for the second generation rose
even further. It is impossible to ascribe this to specific changes in environment because little
is known about the environmental causes of prostate cancer. The fact that mortality rates also
increased in the migrants argues that at least some of the change might be due to
environmental factors.

Risk Factors For Prostate Cancer

Apart from age, family history, and race, few risk factors have been established for prostate
cancer, especially for its lethal forms. It is thus not currently possible to prevent prostate
cancer. Given the amount of research conducted to identify modifiable risk factors, this is
disappointing. One problem has been the biological heterogeneity of prostate tumors.
Another has been the dominance of reports in the literature from small case-control studies
that, due to their retrospective design, inherent biases and poor statistical power have
produced inconsistent findings concerning a large number of possible risk factors.

In the late twentieth century, the first reports began to appear from large cohort studies.
Unfortunately, the prostate cancers in these studies have been diagnosed during the PSA era,
and the majority of tumors are nonaggressive. Without rigorous efforts to identify risk factors
for aggressive cancer and distinguish tigers from pussycats, progress in our understanding of
prostate cancer will continue to be slow.

FAMILY HISTORY AND GENETIC FACTORS

Having a first-degree relative with prostate cancer incurs a twofold to threefold risk and this
increases with the number of relatives affected and earlier age at diagnosis. These
observations are consistent with an inherited genetic predisposition. Analyses of families with
multiple cases of prostate cancer have inferred dominant patterns of inheritance, as well as
recessive and X-linked patterns. It is therefore considered that mutations in more than one
gene might increase susceptibility to prostate cancer. Several loci have been identified, for
example, on chromosomes 1, 17, 20, and X, but reports have been difficult to replicate.
Putative prostate cancer susceptibility genes have also been reported, for example, ELAC2,
RNASEL, and MSR1, but none has been generally accepted. The BRCA2 gene associated
with increased risk of breast cancer for women is a possible exception, as male BRCA2
mutation carriers have increased risk of early-onset prostate cancer.

A region on chromosome 8q found in Icelandic families has been replicated in two other
studies and may contribute to increased risk for African-Americans (Amundadottir et al.,
2006). The risk associated with this region appears regardless of severity, so the putative gene
is likely to be associated with early events in carcinogenesis.

The search for prostate cancer genes may have been thwarted by tumor heterogeneity, since
different genes may be involved in susceptibility to aggressive and nonaggressive tumors.
There is evidence that men with a family history are more likely to have PSA testing and this
will lead to increased diagnosis of nonaggressive tumors in families and make the search for
susceptibility genes more difficult.

Association studies have been carried out for common polymorphisms in candidate genes
known to be in pathways important to prostate development and function. Candidates have
been selected from pathways such as steroid hormone metabolism, DNA repair, insulin, and
insulin-like growth factors (IGFs) and response to infection and inflammation. For example,
much has been published concerning genetic variants in the androgen receptor, 5 alpha
reductase type 2, the vitamin D receptor, IGF-1, interleukins, toll-like receptors, and a
number of cytochrome P450 enzymes. This approach has had a similar lack of success to the
search for susceptibility genes. Seldom have initial reports generated from small, poorly
designed studies been replicated by studies with greater statistical power. It is increasingly
recognized that genetic association studies need to be substantially larger than previously
thought. Some very large studies are under way, but an equivalent emphasis on tumor
aggressiveness is also required if knowledge is to be advanced.
HORMONES
Androgens are required for normal growth and function of the prostate. Testosterone and its
active form dihydrotestosterone (DHT) bind to the androgen receptor and result in increased
transcription and cell division. Some argue that androgens are permissive factors for prostate
cancer, in that their presence is essential for other carcinogenic factors to play their part.
Androgen deprivation in almost all forms leads to involution of the prostate, a fall in PSA
levels, and apoptosis of prostate cancer and epithelial cells. Prostate cancer is termed
hormone-dependent because when advanced it is initially responsive to androgens and can be
controlled by surgical or chemical castration. Men with congenital 5-alpha reductase
deficiency cannot convert testosterone to DHT and do not develop a normal prostate. It is
often said that eunuchs do not develop prostate cancer but the evidence for this observation is
slight.

It has long been reasoned that the hormone dependency of advanced disease might also
influence incidence, the androgen hypothesis. Many case-control studies tested this
hypothesis, with contradictory results. Early cohort study analyses supported the androgen
hypothesis, because circulating levels of testosterone were associated with prostate cancer
risk, while those of sex-hormone binding globulin (SHBG) were protective.

Early analyses failed to take into account tumor heterogeneity, combining all cancers as if
they were the same, the majority of those being nonaggressive. Later studies that have
stratified the analysis on Gleason score or tumor grade have reported that circulating levels of
androgens are associated with increased risk of nonaggressive tumors but a reduced risk of
aggressive tumors. This is consistent with the known function of androgens in maintaining
prostate cellular differentiation.

Clearly, the relationship between androgens and prostate carcinogenesis is complex. Putting
their role in the genesis of nonaggressive tumors to one side, the question remains of how and
at what stage androgens might lose their protective influence on cellular differentiation and
promote aberrant proliferation. A pertinent observation is that prostate cancer increases
dramatically with age from the fifth decade, a time linked with the androcline and changes in
hormonal milieu toward falling androgen and rising estrogen levels.

There is evidence that IGF-1 and its binding protein, IGFBP-3, are associated with risk, but
findings from cohort studies are contradictory. A review of early studies reported a positive
association with IGF-1 and a negative association with IGFBP-3 levels. Larger studies later
found little evidence of association with IGF-1 and some have found positive associations
with IGFBP-3, in direct contrast with the earlier observations. The latter, if real, may help
explain a link with obesity. IGFBP-3 levels are reported to decrease with increasing levels of
physical activity and are potentially modifiable.

SEXUAL ACTIVITY
The androgen hypothesis has also influenced research about sexual behaviors and markers of
androgen influence, for example, age at starting shaving, sexual activity and marriage, the
number of sexual partners including prostitutes, and the frequency of sexually transmissible
infections (STIs). With the exception of STIs, these reports have lacked consistency. Some
studies also report risk associated with diminished opportunity for sexual activity, for
example, increased prostate cancer mortality for Roman Catholic priests, somewhat
analogous to breast cancer in nuns. The case-control literature on these topics has its usual
flaws but additionally lacks a standard approach to measuring sexual activity.

A large case-control study and a prospective cohort study using a standard questionnaire on
ejaculatory frequency by decade of life reported similar findings, that men having high
ejaculatory frequency, especially in early adulthood, reduced their risk by about one-third.
The significance of the observation is not clear; hypotheses range from considerations of
prostatic duct hygiene to more subtle hormonal feedback relationships between the prostate,
testes, and the pituitary gland. The contrasting age-related relationships between androgen
levels, ejaculatory frequency, and prostate cancer incidence may offer some insight.

INFECTION AND INFLAMMATION

There is general support for the hypothesis that infection might be causally related to prostate
cancer. Self-reports of STIs are unreliable, however, and few studies have attempted to
measure evidence of infection directly. Further, many common STIs in men are
asymptomatic or insufficiently bothersome to come to medical attention and it is not possible
to identify culpable organisms for a substantial proportion of those that do.

The findings concerning human papilloma virus infection are essentially null, an earlier
positive association failing to be replicated by later studies. Studies have also failed to
demonstrate any consistent association with herpes simplex viruses. There is limited
serologic evidence that syphilis infection is associated with risk. On the other hand, a
protective effect of serum positivity to Chlamydia trachomatis infection has been reported,
consistent across different serotypes and demonstrating dose–response relationships. This
raises the possibility that some STIs might be able to increase the prostate’s immune
response.

The idea that one consequence of infection, chronic inflammation, might contribute to
prostate carcinogenesis is gaining ground. Inflammation in prostate specimens is common
and foci of proliferative inflammatory atrophy (PIA) are commonly observed in the
peripheral zone. Some consider PIA to be a precancerous lesion or a marker of an
intraprostatic environment that is favorable to cancer development, because it is often found
adjacent to areas of both intraepithelial neoplasia and cancer. Although not fully consistent, a
number of studies have shown a protective effect of aspirin and nonsteroidal
antiinflammatory drugs on prostate cancer risk, particularly for advanced disease. Further
evidence for the possible causal significance of inflammation arises from reports that variants
in genes involved in the inflammatory response to infection, for example, the interleukin and
toll-like receptor signaling pathways, are associated with risk. Intriguingly, some putative
susceptibility genes, for example, MSR1 and RNASEL, have roles in this pathway.

DIETARY AND NUTRITIONAL FACTORS

A large number of inconsistent dietary associations have been reported from ecological
correlation studies and case control studies. The measurement error associated with dietary
assessment coupled with the biases inherent in case-control studies provides a strong
rationale for focusing attention on reports from well-conducted cohort studies. Cohort studies
(mostly North American) have reported positive associations with saturated fat and/or meat,
dairy products (calcium), and zinc and negative associations with tomato-based foods
(lycopene), vitamin D, vitamin E, and selenium, fish (omega-3 fatty acids), soy, isoflavones,
and polyphenols. Cohort study findings are also not fully consistent, possibly because the
majority of cancers in these studies are PSA-detected and advanced tumors are rare. Although
there is a trend toward analysis by tumor grade or stage, statistical power is usually a limiting
factor.

It is clear from cohort studies that fruit and vegetables are not associated with risk (Key et al.,
2004). Although a protective role for phytoestrogens has been promoted, there is little
evidence that men who consume foods containing significant amounts are at lower risk. The
evidence regarding cruciferous vegetable consumption is also poor. No effect of cruciferous
vegetables was observed in either the European Prospective Investigation of Cancer (EPIC)
or the Health Professionals Follow-Up Study (HPFS). EPIC had no information on tumor
stage but in the HPFS there was evidence of a protective effect for disease confined to the
prostate in men aged less than 65 years (Giovannucci et al., 2003).

Although the data from different studies are not entirely consistent, tomato-based foods
(especially tomato sauce), and their principal carotenoid, lycopene, may be protective and,
more so, against advanced disease. Lycopene is concentrated in the prostate and has
antioxidant ability twice that of beta-carotene and ten times that of vitamin E. Associations
with other carotenoids are weaker than for lycopene but are similar in direction.

Micronutrients of interest for chemoprevention have been identified incidentally from RCTs
having other principal outcomes. The Alpha Tocopherol Beta Carotene (ATBC) trial in male
smokers showed an unexpected deficit in prostate cancer in the alpha tocopherol (vitamin E)
arm but no effect of beta carotene (Albanes et al., 1996). There is little evidence that vitamin
E within the normal dietary range is protective. Similarly, a trial of selenium for skin cancer
prevention also showed a deficit of prostate cancer. These agents, both of which have
antioxidant properties, are now subject to RCTs. A concern for the RCTs is the outcome,
which is unlikely to be aggressive cancer but a screen-detected endpoint such as elevated
PSA or microscopic disease. The ATBC trial’s other unexpected outcome was the increased
risk of lung cancer in the beta carotene arm, a salutary example to those who would promote
the supplementary consumption of micronutrients in high doses. It is matched by the finding
that men who take high doses of zinc supplements put themselves at increased risk of prostate
cancer.

Meat, particularly red meat, remains under suspicion, but it has been difficult to separate an
effect of meat from that of fat, for which meat-based foods provide a major source. Some
hold the view that the association with red meat might be related to carcinogenic by-products
of cooking methods, particularly char grilling. There is little available evidence, and similar
associations have not been observed with other grilled meats such as chicken or fish. Fish
consumption appears to be protective against advanced prostate cancer, possibly as a rich
source of marine fatty acids with known anti-inflammatory properties that might be relevant
to prostate cancer prevention.

Cohort studies have yielded mixed results concerning fat, but those that do report associations
have indicated them more consistently with advanced prostate cancer. Analyses by type of fat
and for specific fatty acids have produced similarly inconsistent results, the strongest
association being reported with alpha-linolenic acid. Although the evidence is inconclusive,
several hypotheses have been advanced to explain an association with fat, including an effect
on serum androgen levels and the possibility that certain types of fatty acids or their
metabolites may initiate or promote carcinogenesis.

Dairy foods have been suspected for their fat and calcium content. It has been hypothesized
that high calcium intake, possibly by lowering 1,25(OH)2 vitamin D levels, is associated with
poorer differentiation in prostate cancer and thereby with fatal prostate cancer. Rigorous
reviews and meta analyses have concluded that the risk associated with dairy foods, if any, is
very small.

Vitamin D is also manufactured by the skin when exposed to sunlight. Vitamin D is known to
promote differentiation and impede proliferation of prostate cells and its deficiency is,
therefore, hypothesized to decrease risk, but epidemiological studies are inconclusive.

Overweight and obesity are products of imbalance between energy intake and expenditure by
way of physical activity. The evidence that body size and composition are associated with the
risk of prostate cancer is inconsistent. Some inconsistencies are due to variations in design,
power, and physical measurements, many studies relying on self-report rather than direct
measurement. A comprehensive meta-analysis of the associations reported from cohort
studies concluded that body mass index was weakly associated with prostate cancer risk
(particularly for advanced tumors). Few studies have examined central obesity and prostate
cancer risk, though there is growing recognition of its possible importance. One cohort study
that made direct body measurements has shown waist circumference to be positively
associated with the risk of aggressive disease.

Several prospective cohort studies have found that obese men are more likely to die from
prostate cancer. Also, among men with prostate cancer, obese men are more likely to
experience biochemical progression (rising PSA levels). It has been postulated that obesity
might differentially influence the development of disease, reducing the risk of nonaggressive
disease, while increasing the risk of aggressive disease.

Given the positive associations observed between measures of obesity and prostate cancer, it
would be expected that physical activity would be protective against prostate cancer. Reviews
have concluded that physical activity may be inversely associated with prostate cancer risk
but the evidence is inconsistent, the magnitude of the risk reduction observed is small, and
these could be attributed to the studies’ methodological limitations.

OTHER FACTORS

ALCOHOL
Reviews generally conclude that there is no association between low to moderate alcohol
consumption and prostate cancer, but cannot exclude the possibility of an association with
heavy drinking. There is a single case-control study report of a protective effect of red wine
consumption, which, if real, may be related to the antioxidant properties of red wine pigments
such as resveratrol.
TOBACCO
Overall, epidemiological studies do not support an association between smoking and the risk
of developing prostate cancer. There is evidence, though, that smoking is associated with
increased prostate cancer mortality.

RADIATION
The prostate is one of the few organs in which cancer is not associated with ionizing
radiation, even at high doses. There is weak evidence of a negative association with
ultraviolet light exposure, possibly via vitamin D synthesis.

OCCUPATION
Although a large literature exists on occupational associations with prostate cancer, very little
has been firmly established. The risks that have been observed may be the result of
uncontrolled confounding with social class. In the PSA era, the degree of detection bias in
occupational studies has increased further, because men who believe that they have been
exposed to a possible carcinogen tend to seek PSA tests more frequently and are more likely
to have a diagnosis made than colleagues.

Early studies focused on occupational exposures to cadmium. As the prostate concentrates

zinc, it was considered that cadmium could replace zinc and be procarcinogenic, but follow-
up studies of cadmium smelter and battery production workers failed to confirm early
suggestions of increased risk.

Dioxin, a contaminant of a herbicide used in Vietnam, is similar to many components of

herbicides used in farming. A review of the linkage between dioxin and prostate cancer risk
by the National Academy of Sciences reported that the available data on the association
between dioxin exposure and prostate cancer was inconclusive. A number of studies have
looked at the risks associated with farming, particularly related pesticide and herbicide
exposures, but most have been ecologic and have failed to measure exposures at the
individual level or to control for confounders.

VASECTOMY
Studies have been inconsistent with respect to possible risks of prostate cancer following
vasectomy. Reviews have concluded that virtually all studies had been deficient in their
handling of bias that could account for the modest risks observed in the positive studies.

FUTURE PROSPECTS
Prostate cancer has changed irrevocably since the widespread use of the PSA test for early
detection. In many countries, incidence continues to increase, while in others early detection
has yet to begin. Although many men now diagnosed with prostate cancer would have died
with the disease rather than as a result of it, their diagnosis is just as meaningful in terms of
public health services as would have been a diagnosis based on symptoms. The manmade
epidemic of prostate cancer will continue to impact hugely on the use of health services not
only with respect to initial diagnosis and treatment, but also in relation to the management of
the long-term side effects of treatment, with their enormous potential to diminish quality of
life.

There is little prospect of change until advances are made in the molecular taxonomy of these
tumors, so that the tigers can be discriminated from the pussycats. Without such progress, the
epidemic of prostate cancer will continue unabated.