Screening tests for Prostate cancer

The two main screening modalities for prostate cancer include digital rectal examination (DRE) and serum prostate-specific antigen (PSA; concentration more than 4ng/dl) and endorectal (transrectal) ultrasonography (TRUS).  The most widely used and oldest technique for detection of prostate cancer is the DRE ranges in estimates of sensitivity and specificity are reported, Estimates suggest a PPV of 15% to 30% and sensitivity of approximately 60%. Ultimately, only one in three patients with a positive DRE has prostate cancer.  With the development and application of intraluminal (rectal) probes with high resolution, studies have shown that small, nonpalpable malignant lesions of the prostate could be detected.  TRUS has fallen short of expectations, with large variation in reports of sensitivity and specificity, both ranging from 41% to 79%.  Despite this, TRUS is considered an excellent ancillary modality to increase accuracy of biopsy over the digital guidance alone.

PSA is a blood test that allows for earlier detection of many prostate cancers, with sensitivity of up to 80% to 85%.  Unfortunately, PSA tests have low specificity, resulting in high rates of false-positives.  Interest in the PSA grew in the late 1980s as PSA levels were shouwn to drop to undetectable levels after prostatectomy.  However, normal PSA values are found in approximately one-third of men with localized cancers, and PSA levels are often elevated in men with noncancerous conditions, such as benign prostatic hyperplasia.  Only approximately 7% of prostate cancers detected by screening are microfocal and low grade.

Some investigators have argued that integration of DRE with determination of PSA levels and the use of TRUS in selected cases should improve prostate cancer detection.  Using age-specific PSA ranges may be a promising strategy for increasing PSA sensitivity.  Prostate cancer screening is more controversial in older men; concerns about quality of life outweigh potential benefit of screening.  Several reports indicate that men over age 70 years are unlikely to benefit from PSA testing.

Nonrandomized Studies
Two nonrandomized studies are ongoing to evaluate screening tests for prostate cancer [the ACS National Prostate Cancer Detection Project (ACS-NPCDP) and a multi-center study headquartered at Washington University involving 6 university medical centers].  Results of the ACS-NPCDP trial indicate that a combined modality approach (DRE, TRUS and PSA) to prostate cancer detection yields high levels of early detection.  Data from the Washington University study also indicate that screening with PSA or PSA in combination with DRE, or both, is reasonable for detecting carcinomas at an early stage.  Continued follow-up to evaluate longer-term morbidity and mortality data from these studies will not be available until the end of the first decade of the twenty-first century.  Thus, clinicians and patients must make decisions in the absence of RCT data. 

What is Cancer Screening?

Cancer Screening
The goal of cancer screening is a very practical one to detect cancer at an early state when it is treatable and curable.  How ever, the reality is quite complex. For a screening test to be useful, the test or procedure should detect cancer earlier than would occur otherwise, and there should be evidence that ear lier diagnosis results in improved outcomes.  Advances in genetics and molecular biology undoubtedly will make it possible to detect cancer at earlier and earlier stages along the car cinogenesis pathway.  Thus, the lien between prevention and screening may narrow further, as it has for colorectal and cervical cancers.  In spite of the potential of molecular diagnostics, screening must be evaluated according to its present reality, which is to detect asymptomatic disease when it is potentially curable and can reduce death from cancers. The potential of screening still has not been fully realized.  A report by the National Cancer Policy Board and the Institute of Medicine estimated that a 19% decline in the rate at which new cancers occur and a 29% decline in the rate of cancer deaths could be achieved by 2015 through changes in behavior and greater dissemination of proven technologies, including cancer screening.  The National Cancer Policy Board also estimated that appropriate use of screening among person aged 50 and older could reduce the mortality from colorectal cancer by 30% to 80%; screening among women aged 50 and older could reduce mortality from breast cancer by 25% to 30%, and screening women age 18 and older could reduce the rate of cervical cancer mortality by 20% to 60%.

This is the overview of cancer screening what it is, key terms, measures of effectiveness, and consequences. We also review briefly the status of screening for several prevalent cancers (breast, cervical, skin, prostate, colorectal, and lung).  The recommendations are for the general population, and not for people with identified mutations in cancer susceptibility genes. High-risk populations are the subjects of other chapters.  Generally, it is recommended that people with cancer susceptibility mutations be screened at earlier ages and more frequently for the cancers to which they are predisposed.

This overview cannot provide comprehensive information on any of the cancer sites reviewed, because the literature for each is vast. Rather, the chapter provides a succinct summary of the field and a perspective for clinicians to use in determining the efficacy of particular screening tests.  Several texts and on line resources provide good overviews of the issues related to screening for specific kinds of cancer.  The National Cancer Institute Physician Data Query is an excellent source for the latest data on the specific screening tests discussed here.  Similarly, the American Cancer Society (ACS) provides current information on cancer rates and ACS screening guidelines.  Many other good sites are available as well.

What is Cancer Screening?
Appropriate cancer screening should lead to early detection of asymptomatic or unrecognized disease by the application of acceptable, inexpensive tests or examinations in a large number of persons. The results of a screening test then should be applied expeditiously to separate apparently well persons who probably have disease from those who probably do not.  The main objective of cancer screening is to reduce morbidity and mortality from a particular cancer among persons screened.

Several characteristics make particular cancers suitable for screening.  These include substantial morbidity and mortality, high prevalence in a detectable pre-clinical state, possibility of effective and improved treatment because of early detection, and availability of a good screening test with high sensitivity and specificity, low cost, and little inconvenience and discomfort.  Only breast, cervical, and colorectal screening have met the rigorous criteria of the U.S. Preventive Services Task Force and have sufficient high-quality evidence to justify population based screening.

Evaluation of a screening Test
In evaluating a screening test, it is essential to answer questions concerning its ability to accurately predict a presence of absence of disease.  If the test result is abnormal, what are the chances that disease is present? If the test result is normal, what are the chances that disease is absent? The validity of a screening test is measured by whether it correctly classifies those persons who have disease as test-positive and those without it as test-negative.
Sensitivity and specificity address the validity of screening tests.  Sensitivity is the probability of testing positive if the disease is truly present.  As sensitivity increases, the number of persons with the disease who are classified as test-negative (false-negative) decreases.  Specificity is the probability of screening negative if the disease is truly absent.  A highly specific test rarely is positive in the absence of disease and therefore results in a lower proportion of persons without disease who are incorrectly classified as test-positive (false-positive).  PV+ is an estimate of test accuracy in predicting presence of disease; PV – is an estimate of the accuracy of the test in predicting absence of disease.  Predictive value is a function of sensitivity, specificity, and prevalence of disease.  Accuracy is a measure of the percent of all results that are true results, whether positive or negative, that is total correct test results.  A screening test that results in many false-positive findings is inefficient and potentially dangerous, because it subjects people to follow-up procedures that are often costly, with a range of risks. Similarly, false-negative results can be life threatening, because they miss cancers that could possibly be identified and treated.  Every organized screening program must balance the potential for false positive and false-negatives when establishing criteria for follow-up.
Sources of bias are of particular importance in evaluating screening programs.  People who choose to participate in screening programs (volunteers) are likely to be different from the general population in ways that affect survival; thus, volunteer bias is a concern.  Lead-time bias is defined as the internal between diagnosis of disease at screening and when it would have been detected due to the development of symptoms.  If lead-time bias is not taken into account, survival erroneously may appear to be increased among, screen-detected cases as compared to unscreened cases.  Finally, length bias is the over-representation among screen-detected cases of those with a long pre-clinical period (thus, less rapidly fatal), leading to the incorrect conclusion that screening was beneficial.

Measures of Effectiveness

Several measures have been used to judge the effectiveness of screening.  As Goldie and Kuntz discussed, biologic characteristics of cancers and their precursor lesions may affect test efficacy.  In addition, we make assumptions about the heterogeneity of risk factors in screened and unscreened populations; measures of effectiveness tend to minimize true differences within population.  The most definitive measure of efficacy and least subject to bias is breast cancer mortality as determined by the comparison of screened and unscreened groups in a randomized clinical trial (RCT).  Other outcome measures have been proposed, including case finding and survival. Each has limitations.
Another potential outcome is improved quality of life.  Unfortunately, few trials have collected such data.  These data are being collected as part of the NCI’s Prostate, Lung, Colorectal, Ovarian (PLCO) cancer screening trial and the newer NCI-funded trial to compare special computed tomography (CT) for lung cance3r screening versus chest x-rays.
Increasingly, reports include the number needed to treat to convey the size of the population that do and do not benefit.  The number is usually very small for cancer screening-often in the range of five or fewer people benefiting per 5000 people screened.
In assessing effectiveness of screening technologies, the RCT has been the gold standard.  It is the most powerful methodology for demonstrating the value of screening in comparison to an unscreened group.  RCTs minimize biases inherent in other designs; especially lead time, length bias, selection bias, and over diagnosis.  Case-control studies also can provide useful information, and at less cost than RCTs.  In addition, increasingly sophisticated statistical modeling techniques may be helpful in assessing the impact of screening, especially in situations in which large RCTs cannot be conducted.  The U.S. Preventive Services Task Force PD! And others have rated levels of evidence; the RCT is uniformly regarded as the highest level of evidence.

Breast Cancer Screening & Clinical Trials

In 2003, approximately 211,300 new invasive cases of breast cancer (and 55,700 new ductal carcinoma in situ cases) and approximately 39,800 deaths due to the disease were estimated to have occurred in the United States.  For a woman of average risk, lifetime breast cancer incidence is 7.8%, and mortality is 2.3%.  Widely accepted techniques, for breast cancer screening, but with differing levels of evidence, include mammography, clinical breast examination (CBE), and breast self-examination (BSE), No cancer screening test has been studied more than mammography (with or without CBE).

Randomized Clinical Trials
Eight randomized trials have been conducted over more than 40 years to assess the impact of mammography.  Together, these trials have included more than 5,00,00 women, with 180,000 women aged 40 to 49.  The eight international RCTs have varied greatly.  Most trials have included women in their 40s, although two trials began accrual at age 45. One of the Canadian trials (the first National Breast Cancer Screening study (NBSS1) was designed to examine mammography and CBE versus usual care for women in their 40s, with a separate study (NBSS2) to assess mammography plus CBE versus CBE alone women aged 50 to 59.  The international studies have varied in key ways.

Positive and Negative Consequences of Screening

Every medical activity has positive and negative consequences, and screening is no exception.  Potential benefits include improved prognosis for those with screen-detected cancers, the possibility of less radical treatment, reassurance for those with negative test results, and resource savings if treatment costs are reduced because of less radical treatments.  The optimal outcome is a reduction in cancer mortality.  Potential negative effects of screening include physical, economic, and psychological consequences of false-positive and false-negatives, the potential for over diagnosis, the potential carcinogenic effects of screening, and the labeling phenomenon.  The last refers to the fact that telling individuals that they have cancer may change how they see themselves or how others see them.  In addition, Baines has noted the largely unexplained increase in cancer among the unscreened group as a negative consequence of screening.
Physicians should engage patients in discussions of the risk and benefits of cancer screening.  Because most people overestimate the risks for certain type of cancers (eg. Breast), they may inflate the need for screening and the potential benefits.  For some cancers, such as colorectal cancer, people may under estimate their personal susceptibility and may need encouragement to consider screening, with positive as well as negative consequences.  In the case of prostate cancer, for which the evidence is still equivocal and population-based screening is not recommended, it is especially important that men understand the limitations of screening and make informed decisions with full understanding of the potential downstream effects of positive and negative test results.  Informed decision making increasingly is becoming a paradigm with the potential to be achieved in contemporary clinical practice.  Informed decision making occurs when an individual understands the disease or condition being addressed and also comprehends what the clinical service involves, including its benefits, risks, limitations, alternatives, and uncertainties; has considered his or her own preferences, as appropriate; believes he or she has participated in decision making at a level that he or she desires; and makes a decision consistent with his or her preferences.  Decision aids are tools used to help patients examine the nature of screening tests and their benefits and limitations.
Ausoker outlined topics that should be included when helping patients to make informed decisions about cancer screening.  These include the purpose of screening; the likelihood of positive and negative findings and the possibility of false-positive or false-negative results; the uncertainties and risks involved; any significant medical, social, or financial implications of screening; and follow-up plans.

Mammography

It was not until 1963 that the first RCT and the only RCT conducted within the United States commenced at the Health Insurance Plan of New York, Women were aged 40 to 64 at entry; nearly 62,000women were randomized to the study (two-view mammography and CBE) or control group (usual care).  Two-view mammography and CBE) or ontrol group (usual care). Two-view mammography and physical examination were offered every 12 months for 3 years, and follow-up was continued for 18 years.

The Swedish Two-country Trial in Kopparberg and Ostergotland began in 1977 to 1978; 135,000 women were randomized to one-view mammography every 24 (younger than age 50 years) or 33 months (older than age 50 years).  Within those geographic areas, all women were invited to enroll by letters of invitation, using the population registry list.  Screening continued for four rounds for younger women and there rounds for older women.

The Malmo Trial was begun in 1976 in one city in Sweden. It was one of only two trials that began screening at age 45 years, and it sopped entry at age 69 years.  It used tow-view mammography every 18 to 24 months for five rounds; randomization was by cluster based on birth cohort.  Approximately 68,000 women were enrolled in this trial.

The Stockholm Study began in 1981 and enrolled approximately 59,000 women aged 40 to 49 years was by individual, whereas clustered randomization was used for women aged 50 to 59 years, who received tow-view mammography every 18 months.  Randomization of women aged 40 to 49 years was by individual, whereas clustered randomization was used for women aged 50 to 59 years. Verified results have not yet been published, but additional data were provided in 1997 for the Nationa Institutes of Health Consensus Conference on Breast Cancer Screening in women aged 40 to 49 years.

The Edinburgh trial began in 1976 as a randomized component of the larger, nonrandomized United Kingdom trial for the Early Detection of Breast Cancer.  Approximately 45,000 women aged 45 to 64 years were randomized to two-view mammography plus CBE on either a 12 or 24 month schedule.  The purpose was to assess the impact of mammography and CBE in reducing mortality from breast cancer.

The NBSSI was designed to examine the value of two-view mammography and CBE compared to usual care in women aged 40 to 49. Approximately 50,000 women were enrolled, starting in 1980, and received follow-up yearly for 5 years.  Unlike the other trials, the women were recruited as volunteers and then randomized.  As miller et al have reported, these women were different from the rest of the Canadian population in several ways – for example, they were less likely to smoke, and they had higher levels of education.

The second Canadian study, the NBSS2, also begun in 1980, enrolled nearly 40,000 women aged 50 to 59 years, and was designed to compare two-view mammography and CBE against CBE only.  In other words, the question was whether mammography has benefits over and above CBE in this age group.  This is the only trial planned to assess the additive impact of mammography in addition to CBE.  Critiques of this trial have appeared by several authors.  The criticisms are probably overstated, and the results should be discounted.

Non Randomized Clinical trials
A number of nonrandomized trials have been conducted around the world.  Much can be learned from these studies.  However, because of a number of design limitations, their results alone should not be used in establishing screening guidelines and policies.  The largest study of mammography and CBE was the US.  Breast Cancer Detection Demonstration Project (BCDDP): 280,000 women aged 35 and older were recruited and screened in 28 centers annually with mammograms and CBE during the 1970s.  The BCDDP was sponsored jointly by the NCI and the ACS.  Because BCDDP participants were not a random sample of the population, there were some important differences from women in the general population.  Most notably, the BCDDP population was at much higher risk, with a substantially higher incidence of breast cancer.   Subsets of women were followed as part of a case-control study conducted by Morrison et al. to examine case fatality rates within the BCDDP.  Breast cancer mortality was approximately 20% less than expected from national data.  A benefit was seen for younger women, but it was less than for older women.

EVIDENCE FOR MAMOGRAPHY

More than 3.5 million women-years of observation have been recorded for women of all ages, and more than 2.7 million women-years for women aged 40 to 49 years at entry from the breast cancer screening RCTs.30 One of the challenging aspects of tracking these trials is the constantly shifting nature of the data.22Thetrialists provide updates at different points in the time, and some reports use non-verified data. Thus, at any given point, review articles may vary substantially in the numbers they report (see Table 22-3). Now, several years later, with most of the international RCTs having more than 10 years of follow-up, trends are clearer: Six of the trials show reduction in mortality for women who where in their 40s at entry. However, a large variability remains in the relative risk of dying from breast cancer for women younger than 50 years. Also, significant controversy surrounds many of the reported numbers.

Although the randomized trials have included too few women older than age 70 years to offer guidance about screening for older women, the Forum on Brest Cancer Screening 37 recommended regular mammograms for women aged 70 years who are otherwise healthy a case-control analysis in the Nijmegen study38 confirmed the benefit of mammography for women older than 70 years.

Seven particularly important published metaanalyses, including the two mentioned earlier, provide assessments regarding the impact of breast cancer screening, especially mammography (Table 22-4). They generally have examined results separately for women aged 40 to 49 years and those aged 50 years and older. (Few women are in the older categories).  For women in their 50s and 60s, there is general agreement about the benefits of mammography . The most recent review done by the U.S.Preventive Services Task Force   shows a modest benefit for women aged 40 years and older 29. Thus, there is now general consistency in the evidence reviews.

BREAST CANCER SCRENING GUIDELINES

The question, ”At what age should women begin getting regular mammograms?,” has been one of the  most contentious in science and medicine .In 1993 , the International Workshop on Screening for Breast Cancer13 created considerable controversy when it concluded that for women aged 40 to 49 years ,randomized controlled trials  consistently demonstrated  no benefit from screening  in the first 5 to 7 years after  entry and  a marginal  benefit  after  that . The issue became even more   inflamed after a 1997   National Institute of Health Consensus Conference on Breast Cancer Screening for Women Ages 40 to 49.39 The report, contrary to expectations, found insufficient data to recommend that women in their 40s get regular mammograms. Disagreement still exits over whether the modest reduction in mortality warrants a recommendation that all women in there 40s get regular mammograms. Disagree warrants a recommendation that all women in their 40s be screened .40 The argument turns primarily on the small population benefit archived and the fact that most of the benefit occurs when screened women are in their 50s.  Only 1to2women’s lives would be extended per 1000 women of 40to 50 years of 10 years. However, in agreeing on a reduction in breast cancer mortality of 18%, the ACS and the NCI changed their screening recommendations, with the ACS now advising annual mammograms for women aged 40years and older. Annual screening for women in there 40s is based on the assumption of a shorter lead-time for younger women.

Controversy flared again in 2000, when Gotzsche and Olson published another metaanalysis of the world’s mammography trials and found most of the trials methodologically flawed.  They concluded that mammography did not show a statistically significant benefit for women in any group.

Over the past several years, there has been a convergence of opinion across most medical organizations, including those that are grounded in evidence-based decision making, that mammography benefits women aged 40 years and older.  Current controversies focus on the net benefits and optimal screening schedule.  The benefit is modest.  Because no trials have included sufficient numbers of older women to permit separate analyses by age, there are no good data on the upper limit of benefit.  Most informed sources (e.g., ACS) not recommend regular screening for women aged 40 years and over, with some organizations recommending annual and some biennial screening.

The evidence suggests that a 5% to 20% additional benefit in mortality reduction can be achieved by adding a high-quality CBE.  The US Preventive Services Task Force reviewed the data regarding CBE and concluded that they were insufficient to reach conclusions about efficacy or to issue population-based recommendations.  No evidence has shown that BSE reduces mortality from breast cancer.  Large trials conducted in China and Russia did not result in the hoped-for reduction of cancer mortality as a result of careful BSE instruction.  In its 2002 report, the US Preventive Services Task Force concluded that the data also are insufficient to reach conclusions about BSE.  Green and Taplin recommended that if women want to perform BSE, they are taught to do it correctly.  Given the small windows of time physicians have to counsel and teach, time probably is best spent giving strong messages about the importance of mammography and performing a thorough CBE.  Data from a large-scale service delivery program suggest that CBE may be most useful for women in their 40s where mammography may be somewhat less efficacious.  Although the recommendations of different medical organizations vary, most encourage women to have a CBE yearly.

Colorectal Cancer Screening

The importance of screening for colorectal cancer is based on the high incidence and mortality from this cancer and the availability of screening methods of demonstrated efficacy.   Cancers of the colon and rectum account for the third largest number of new cancer cases after lung cancer, with 105,500 colon and 42,000 rectum cases estimated, respectively, in 2003; approximately 57,100 deaths were predicted for colon and rectum cancers combined.  Whereas declining incidence rates have stabilized since 1995, the steady decline in death rates has accelerated since the mid-1980s, especially for whites.

Four procedures are currently in use for colon cancer screening: fecal occult blood test (FOBT), sigmoidoscopy, colonoscopy, and high-contrast barium enema.  All professional organizations that have published guidelines recommend screening for adults 50 years and older with some combination of these four modalities.  For the general population, age 50 years was chosen, because that is when the incidence of colorectal cancer begins to increase and where efficacy is supported by evidence.  For individuals at high risk, age 40 years is generally recommended as the starting age.

1.    Fecal occult blood test, Direct support for use of FOBT comes from three large randomized controlled trials, including a Minnesota trial of 46,501 participants between the ages of 50 and 80 years of age.  This study found that annual FOBT with rehydration of the samples decreased 13-year cumulative mortality from colorectal cancer by 33% and biennial screening by 21% Most (75% to 84%) of this reduction resulted from the test itself rather than from incidental discovery of cancers by follow-up colonoscopy.  Data from case-control studies are generally consistent with the conclusions of these trials.  The main limitation of the FOBT is its limited specificity.    
2.    Flexible sigmoidoscopy.  The advantage of sigmoidoscopy screening over FOBT is that it frequently includes the actual removal of cancer or a precancerous lesion in a biopsied polyp, thus combining screening and treatment in one step.  Another advantage is that it needs to be performed only infrequently, perhaps every 5 to 10 years.  At least two large randomized trials of flexible sigmoidoscopy screening are in progress.  In the first of these randomized trials, the PLCO trial is evaluating the efficacy of examinations every 3 years in 74,000 men and women, 55 to 74 years of age, and an equal number of controls.  In the United Kingdom, a trial of 200,000 men and women, 55 to 64 years of age, is evaluating one sigmoidoscopy delivered to approximately 65,000 adults randomized for screening.  Meanwhile, supporting evidence for efficacy comes from several case-control studies in health plans in northern California and Wisconsin and from American veterans.
3.    Colonoscopy and double-contrast barium enema.  No direct evidence supports the efficacy of either colonoscopy or double-contrast barium enema.  A strong rationale for the use of colonoscopy is based on its superior sensitivity and specificity.  Because there may never be a randomized trial mounted to evaluate the efficacy of colonoscopy directly, choices about its use will depend more on concerns about its safety, the capacity of the health care system to make it available, and its cost effectiveness.  Colonoscopy has risks:  Approximately 1 in 1000 patients experience perforation, 3 in 1000 have major hemorrhage, and 1 to 3 in 10,000 die as a result of the procedures.  Double contrast barium enema is an alternative for examining the entire colon and finds its usefulness in situations in which individuals cannot tolerate endoscopic procedures and in which the relevant expertise exists.

Multiple studies of the cost-effectiveness of colorectal cancer screening are consistent in concluding that the cost per year of life saved is within the limits accepted for most preventive procedures from a societal perspective (approximately $50,000).  However, no one modality stands out as superior.

Alternatives to current methods are being investigated and have been reviewed.  Immunohistochemical screening tests of stool may offer more sensitive and specific alternatives to guaiac beased testing.  The molecular detection of DNA mutations in cells exfoliated from neoplasms in the stool may be a highly sensitive and specific approach that is noninvasive and thus more acceptable to patients.

Newer detection techniques include “virtual” colonoscopy that uses CT of the prepared colon and avoids the invasiveness and discomfort of conventional optical colonoscopy.  Fenlon et al provided data on a study in which a crossover design was used; 100 patients were given virtual colonoscopies immediately before conventional colonoscopies.  Test performance was compared and was similar, with a threshold of detection for virtual colonoscopy at approximately 5 mm.  Pickhardt et al reported on test performance using a similar design but with a much larger population (1233 asymptomatic patients) and one that offers more potential for generalizing findings to average-risk patients.  They showed that a three-dimensional endoluminal display can achieve 93.% sensitivity and a specificity of 96.0% for polyps at least 10 mmm in diameter compared to optical colonoscopy on the same asymptomatic average-risk subjects.  These and other studies of newer technology will be commanding much attention in the near future.

Meanwhile, despite the evidence supporting the value and cost effectiveness of traditional colorectal cancer screening procedures, compliance with guidelines remains low, at approximately 30% of the population; disparities in screening rates between race and ethnic groups are increasing.  Access to screening varies greatly by region of the country and the availability of trained personnel.  However, the primary barriers to increased compliance are behavioral for physicians and for patients and associated with the complexity of and perceptions about the nature of the recommended procedures.  Efforts to improve compliance should focus on getting individuals over 50 years screened at least once by any modality, rather than on the superiority of any one test.  Choices among individuals vary consistently regarding which test is preferred.

Cervical Cancer Screening

A steady 70% decline in mortality from cervical cancer has been observed since the mid-century after the introduction of widespread Papanicolaou (Pap) cytologic screening.  This is a major success story for cancer control in the United States.  In 2003, there were an estimated 12,200 new cases of invasive cervical cance4r and 4100 deaths.  Although mortality from this cancer is no longer as common as in the past in the United States, cervical cancer is still the second most common cancer worldwide.  In developed countries, fatalities from this disease should be entirely avoidable with currently available technology.  Incidence and mortality are higher in women with no prior screening, those with concurrent human papillomavirus (HPV) infection, and women of lower socioeconomic status.

Newer methods of screening make use of increased understanding of the essential etiologic role of approximately 15 oncogenic types of HPV.  Various combinations of HPV DNA testing; newer cytologic methods, especially liquid-based cytology; and traditional Pap smears offer opportunities to improve sensitivity and specificity of cervical cancer screening. However, traditional cytologic screening with Pap smears remains the primary method of screening.

The value of Pap smear screening is of little doubt, even though there has never been an RCT to confirm its efficacy.  In the absence of an RCT, e3vidence for the effectiveness of cytologic screening has come from observed trends in countries with large national screening programs and case-control studies in varied geographic areas.  When to start, when to stop, and the interval between tests have been important questions within the context of traditional cytologic screening.

Current U.S. recommendations are that screening should start approximately 3 years after the onset of sexual activity, at prevalence of HPV infection in young, sexually active women and the frequency of low and high grade sqamous intraepithelial lesions.  However, up to 70 of high risk HPV infections are transient in young women in their 20s, and 90% of low grade squamous intraepithelial le4sions regress in this age group.  Further more, incidence is not measurable until 20 to 24 years of age, when it is still only 1.7 per 100,000 per year.  Given the low incidence of serious disease and the high likelihood of regression of early dyplastic lesions in younger women, there is concern that screening adolescents and women in their 20s could lead to over diagnosis, aggressive treatment, and unnecessary harm from ablative surgical procedures.  To address this reality, some European countries (e.g., Denmark) do not start screening until age 30 years.

More than one-half of invasive cervical cancers occur in women who have never been screened, or at least it within the previous 5 years.  Progress toward further reductions in death from this cancer could be made by concerted efforts to reach and screen such women.  However, evidence is limited on the questions of the optimal interval between screening test.  After its review of evidence, the U.S. Preventive Services Task Force left the interval variable from 1 to 3 years.  The principal study on which this judgment was based was a comprehensive analysis of large-scale screening programs and case-control studies that used various designs and methods but which showed increased efficacy with shorter intervals up to annual testing.  Sample size, even in this large analysis, was insufficient to distinguish efficacy with shorter intervals up to annual testing.  Sample size, even in this large analysis, was insufficient to distinguish efficacy betwee3n intervals of every 2 versus every 3 years. A case-control study in a large, stable health plan population confirmed that annual screening is more likely to pickup invasive cervical cancers than is using an interval of every 2 years, and likewise for 2 versus 3 year intervals.  The advantage was very small, and cost-effectiveness studies are needed. For women who have had hysterectomies for benign conditions, the likelihood of detecting vaginal dyplasia is extremely low and the false positive rate high.  The continued practice of screening at any interval in this group is inappropriate.

Data are also inadequate on which to base recommendations for an upper age limit for regular screening.  Although there is concern that Pap testing is less sensitive in order women because of a receded squamocolumnar junction of the cervix, the primary reason for invasive cervical cancer in older women is still lack of any screening.

International studies using DNA testing have demonstrated that almost all cervical cancers are associated with infection by sexually transmitted HPV.  However HPV infection is highly transient in young women, and screening protocols that use HPV testing are still being evaluated.  Protocols that add HPV DNA testing to the traditional Pap test for cytologic results of atypical squamous cells of undetermined significance (ASCUS) have demonstrated increased sensitivity over repeated Pap smear testing.  Among 3488 women with ASCUS followed for 2 years in the ASCUS/LSIL (low-grade squamous intraepithelial lesion).  Triage Study, HPV DNA testing was more sensitive for detecting cervical intraepithelial neoplasia 3 or invasive cancer and just as specific as repeat cytology.  Comparisons of the multiple strategies now available, including visual screening, conventional cytology, liquid-based cytology, and HPV DNA testing, have found that either liquid-based cytology or HPV DNA testing provides a better balance between sensitivity and specificity for cervical intraepithelial neoplasia 3+ than conventional methods. However, although these new technologies are promising, the results are highly dependent on local resources and prevalence of disease.  In most situations, clinicians are still likely to have their greatest effect by ensuring full coverage and follow-up for women who receive traditional cytologic screening.  The challenge for the future may be less of a technical nature and more dependent on local finances and screening policies.

Skin Cancer Screening

The incidence of skin caner has increased worldwide, with U.S. incidence data mirrioring this trend.  In the United States, the incidence rate for melanoma has increased approximately 4% per year since the early 1970s, with a 162% increase in male melanoma cases and 95% in women.  It is unclear whether this increase is due to actual changes in prevalence or is partly a function of increased awareness, with subsequent diagnosis, improved reporting by tumor registries, or both.  In 2003, 63,400 new cases of skin cancer and 12,000 deaths were projected, with 54,200 new cases of melanoma (skin) and 7600 deaths.  Melanoma now ranks sixth in incidence among cancers in males and seventh in incidence among cancers in females.  Approximately 800,000 nonmelanoma skin cancers are diagnosed each year.  The United States lags behind many other countries in the creative application of interventions to reduce the incidence of and mortality from melanoma and other skin cancers.  Australia, which has the highest reported incidence of melanoma anywhere, has mounted successful population-based programs that have had dramatic effects.

Experts have not agreed about screening guidelines for skin cancer.  The U.S. Preventive Services Task Force recommends routine screening for individuals at high risk (eg. Those who have a family or personal history of skin cancer, clinical evidence or precursor lesions, and increased exposure to sunlight).  The Task Force neither defines what is meant by routine screening nor reports on the specific recommendations for skin self-examination.  The ACS recommends a cancer-related checkup, including a skin examination, every 3 years, and more frequently for persons at risk.  The NCI also recognizes the benefits of skin cancer screening but offers no specific guidelines for such screening.  Only the study demonstrates evidence regarding skin self-examination.  Although it showed a decrease in mortality, there were limitations that preclude making recommendations on the basis of this study alone.

Prostate Cancer Screening

Prostate cancer is the most commonly diagnose cancer among men in the United States and is the second leading cause of male cancer deaths.  It is estimated that in 2003, 220,900 new cases of prostate cancer were identified, with 28,900 deaths.  However, consensus is lacking aobut recommendations for prostate cancer screening.  The controversy has be raised for several reasons.  First, there is no definitive evidence that prostate cancer screening results in improved clinical outcomes, especially a reduction in mortality from the disease.  Second, the rise in the incidence of prostate cancer from 1989 to 1992 (see http://www.seer.cancer.gov) was due largely to dection of latent, asymptomatic cases with uncertain clinical relevance, thus putting the value of screening in doubt.

Screening tests for Prostate cancer
The two main screening modalities for prostate cancer include digital rectal examination (DRE) and serum prostate-specific antigen (PSA; concentration more than 4ng/dl) and endorectal (transrectal) ultrasonography (TRUS).  The most widely used and oldest technique for detection of prostate cancer is the DRE ranges in estimates of sensitivity and specificity are reported, Estimates suggest a PPV of 15% to 30% and sensitivity of approximately 60%. Ultimately, only one in three patients with a positive DRE has prostate cancer.  With the development and application of intraluminal (rectal) probes with high resolution, studies have shown that small, nonpalpable malignant lesions of the prostate could be detected.  TRUS has fallen short of expectations, with large variation in reports of sensitivity and specificity, both ranging from 41% to 79%.  Despite this, TRUS is considered an excellent ancillary modality to increase accuracy of biopsy over the digital guidance alone.

PSA is a blood test that allows for earlier detection of many prostate cancers, with sensitivity of up to 80% to 85%.  Unfortunately, PSA tests have low specificity, resulting in high rates of false-positives.  Interest in the PSA grew in the late 1980s as PSA levels were shouwn to drop to undetectable levels after prostatectomy.  However, normal PSA values are found in approximately one-third of men with localized cancers, and PSA levels are often elevated in men with noncancerous conditions, such as benign prostatic hyperplasia.  Only approximately 7% of prostate cancers detected by screening are microfocal and low grade.

Some investigators have argued that integration of DRE with determination of PSA levels and the use of TRUS in selected cases should improve prostate cancer detection.  Using age-specific PSA ranges may be a promising strategy for increasing PSA sensitivity.  Prostate cancer screening is more controversial in older men; concerns about quality of life outweigh potential benefit of screening.  Several reports indicate that men over age 70 years are unlikely to benefit from PSA testing.

Nonrandomized Studies
Two nonrandomized studies are ongoing to evaluate screening tests for prostate cancer [the ACS National Prostate Cancer Detection Project (ACS-NPCDP) and a multi-center study headquartered at Washington University involving 6 university medical centers].  Results of the ACS-NPCDP trial indicate that a combined modality approach (DRE, TRUS and PSA) to prostate cancer detection yields high levels of early detection.  Data from the Washington University study also indicate that screening with PSA or PSA in combination with DRE, or both, is reasonable for detecting carcinomas at an early stage.  Continued follow-up to evaluate longer-term morbidity and mortality data from these studies will not be available until the end of the first decade of the twenty-first century.  Thus, clinicians and patients must make decisions in the absence of RCT data. 

Randomized Clinical Trials

The NCI – funded PLCO trial is a 16-year randomized control study that began in 1993.  It is accruing 74,000 men aged 60 to 74 years and has a design power of 90% to determine 20% reduction of prostate cancer mortality.  PLCO will provide important information about the efficacy of screening.  A second large randomized trial, the European Randomized Study of Screening for Prostate Cancer and PLCO trials are collaborating to share data to increase statistical power.  Results from these trials are expected in 2005 to 2008.   Finally, based on a randomized study in Sweden of 9972 men aged 50 to 65 years, it was concluded that it is safe to screen biennially in men with PSA of less than 2 ng/mL, with different screening intervals determined based on baseline PSA.  Further confirmation of appropriate screening intervals for PSA screening is needed.

As is the case for other kinds of cancer screening, there is concern about over diagnosis for prostate cancer.  In fact, it is an even greater potential problem for prostate than for some other kinds of cancer screening because of studies showing that many men die or prostate cancer without ever having known they had the disease.  Draisma et al attempted to provide some insight into the potential over diagnosis problem using models to simulate lead time.  They estimated that annual screening from age 55 to 67 years would result in an over detection rate of 50% and the lifetime prostate cancer risk was increased to 80%.  People will argue whether the model is correct, and all models have imprecision.  Nevertheless, this is further caution about the limitations of screening for prostate cancer, the need for better understanding of the biology of the disease, and the parallel need for better screening tests.

Trends and What They Mean
National data from 1990 to 1966 show that prostate cancer incidence peaked in 1992 at 190.8 per 10,,000 and declined at an average rate of 8.5% from 1992 to 1996.  A series of related reports in the Journal of the National Cancer Institute, based on data from the Surveillance, Epidemiology, and End Results Cancer Registry program show a decline in incidence of regional stage disease, as well as a decline in incidence-based mortality trends for localized and distant stage disease.  Statistical methods were applied to consider the effect of screening by limiting some analysis to the contribution from cases diagnosed since 1987, when widespread screening using the PSA test had occurred.  In a review of published data from five prospective trials, treatment of localized disease was associated with a marked decrease in prostate cancer deaths.  Thus some evidence shows improved prognosis for screen-detected cases.  However, alternative interpretations, such as the possibility that cause-of-death mis-classification could explain these findings, cannot be ruled out.

Prostate Cancer Screening Recommendations

In light of the limitations of evidence regarding prostate cancer screening, it is important to consider the natural history of this disease and subgroups of men at high risk for developing prostate cancer.  These considerations are critical for determining public health policy.  Men’s preferences regarding testing vary, and these preferences also are important in deciding on screening.  Most medical organizations now stress the importance of individualizing screening decisions based on factors such as patient’s preference, health history, and health status.

Lung Cancer Screening

Lung cancer screening is not recommended on a population basis due to lack of evidence that any available screening procedure, even for smokers, can identify tumors early enough to reduce mortality. This remains a major challenge to research and technology because of the tremendous burden caused by this cancer, including among ex-smokers. In 2003, there were estimated 171,900 new cases and 157,200 deaths, making it by far the most common killer from cancer in men and in women.

None of the four randomized trials conducted during the 1960s and 1970s reduced mortality significantly over no screening. The Mayo Lung Project, the primary trial contributing to this evidence, demonstrated that screening with either chest x-rays or chest x-rays plus sputum cytology lowered the stage at presentation and increased survival, but neither approach had any effect on lung cancer mortality. Although lack of connection between improved survival and the absence of a mortality benefit can be attributed to lead-time and length biases, these studies have been criticized for other methodologies reasons. Extended mortality follow-up of participants in the Mayo Lung Project suggested that over diagnosis, the identifications of clinically unimportant lung cancer lesions might have occurred.

Low-dose CT scanning is a new and potentially efficacious method for early detection of lung cancer. This noninvasive technique, which creates an image of the entire thorax during a single held breath with a low radiation dose, is being offered by an increasing number of imaging facilities for older smokers and former smokers. However, to date, evidence of high sensitivity comes only from observational studies that are susceptible to lead-time and length-time bias. The possibility of over diagnosis and concerns of harm from CT screening have also been raised, and there is, as yet, insufficient evidence to support mass lung cancer screening with this procedure. Nevertheless, the potential of this technology and the enormous societal burden imposed by lung cancer motivated the NCI to begin a large, randomized controlled trial to assess the effect of low dose CT screening are ongoing and may add to the evidence produced by the NCI-supported trial, which will take approximately 10 years to produce results. Meanwhile, decisions about use of CT scans for lung cancer screening remain a matter of individual judgment between physicians and patients. Several reports suggest the need for caution in adopting lung CT screening, especially in view of aggressive direct-to-consumer marketing of the procedure. In a detailed decision analysis, Mahadevia et al showed that even if efficacy of helical CT is shown, it is unlikely to be cost effective. Like many tests, helical CT may be able to identify small lesions. However, questions remain about the mortality impact and cost effectiveness. Thus, the ongoing trial is of critical import. Efforts to prevent initiation, especially by youth, and cessation of tobacco use remain the physicians best tool for combating lung cancer.

Adherence to Cancer Screening

Over the years, substantial research has been conducted to understand people’s barriers to cancer screening. Several lessons are clear. First, physician recommendations are the most important factor in motivating people t to be screened. Even then, however, many people will not get necessary tests. Several reviews show that simple reminders and letters, delivered in print or by telephone, can double or triple the odds that people will attain needed tests. This is true even among vulnerable populations. Interventions often are needed at several levels: individual, provider, and health system. The Centers of Disease Control and Prevention’s Task Force on Community Preventive Services has reviewed interventions to promote screening and publishes regular updates (http://cdc.gov).

Some people are at risk for being un- or under screened. These include recent immigrants. People without health insurance, people with no usual source of health care, and people with very low incomes. Although age and ethnicity are important determinants for some cancers, this is not the case for others.

The best source of U.S. data on cancer screening practices is the National Health Interview Survey. In the 2000 National Health Interview Survey, 70% of women aged 40 years and older reported having recent mammograms, 82% of women aged 25 and older reported having recent Pap tests, 41% of men and 37.5% of women reported having either an FOBT within the past year or colorectal endoscopy within the past 5 years, and 41% of men aged 50 years and over had a PSA test within the past 5 years. Rates of regular screening for all these tests are considerably lower than reports of recent screening.

Future of Screening

Many challenges lie ahead for cancer screening. Better detection methods are urgently needed, and molecular detection methods may surpass current techniques. At the same time, more effort is required to encourage adherence to proven cancer screening modalities. Adherence to screening is less than optimal for the entire major recommended screening techniques. With discovery of mutations in susceptibility genes that predispose to cancer, new challenges in cancer screening arise with the need for appropriate screening recommendations and programs for high-risk subgroups. Not only are those who have inherited mutations for cancer susceptibility at higher risk for developing some cancers, but, often the age at onset among such individuals is younger than age at onset in the general population. This creates challenges for those who recommend and promote screening regimens. The example of this disease. Moreover, there are no good population data on which to base guidelines. Further study is needed to establish efficacious screening protocols for those who are genetically predisposed to cancer.

Among the challenges of the future is how to evaluate new screening technologies in a world where large RCTs may be increasingly difficult to conduct. In screening, as in other areas of medicine and public health, the inclination to recommend screening tests on the basis of an intriguing and promising study must be balanced by a careful assessment of the evidence.