Position on Lung Cancer Screening
Lung Cancer Alliance (LCA) has consistently maintained that those at high risk for lung cancer should speak with their doctors about the risks and benefits of screening. Those at risk include smokers and former smokers, first degree relatives of people diagnosed with lung cancer and those with prolonged exposure to radon, asbestos, Agent Orange, radioactive materials and other lung carcinogens. If the decision is made to undergo screening by computed tomography (CT), the scan should only be done at a site which has experience in screening for lung cancer and which follows a lung cancer screening protocol based on best published practices, such as the International Early Lung Cancer Action Program (I-ELCAP) protocol.
Lung cancer causes more deaths than breast, prostate and colon cancers combined. Nearly one in every three cancer deaths is due to lung cancer. The 5-year survival rate is still only 15%. Only 16% of lung cancers are being diagnosed at an early curable stage. According to a prevalence survey by the Centers for Disease Control (CDC), 60% of lung cancer patients in 2006 had already quit smoking and an additional 17.8% had never smoked. (i) Yet lung cancer continues to be underfunded for research and early detection, primarily because it is viewed as a “self-inflicted” disease that could be virtually eliminated if no one smoked. While tobacco cessation is critical to reducing lung cancer incidence and this strategy is strongly supported by LCA through a number of initiatives, cessation alone cannot effect either the prevalence of lung cancer among former and never smokers or the very high mortality rate of newly diagnosed lung cancer patients whether they smoked or not. The limited number of survivors, coupled with the stigmatization of smokers and the vast size of the at-risk population undermines efforts to build a grass roots political advocacy army of support to change public health policy leading to substantive improvements in lung cancer outcomes.
Screening is a method of detecting a disease cancer before clinical signs or symptoms become evident. Early stage cancers can be more easily treated with a curative outcome than later stage cancers. Once a cancer has spread to other organs (metastasized) that cancer is rarely cured.
Examples of screening tests are mammograms for breast cancer, prostate specific antigen (PSA) blood tests for prostate cancer and colonoscopies for colon cancer. Despite continuing controversy over efficacy and costs, they have been widely promoted for the entire adult population, not just those at high risk. At this time, 61% of breast cancers, 91% of prostate cancers and 40% of colon cancers are being diagnosed at an early localized stage. These tests have been credited with increasing the 5-year survival rates for those cancers. The higher the rate, the better the prognosis. The 5-year survival rates for these cancers have increased to for 89% breast cancer, 99% for prostate cancer and 65% for colon cancer.
No cancer screening method is 100% accurate. Some cancers may be missed (false negatives) and some benign cancers (false positives) may be found. Every method has generated controversy and triggered decades of debate about accuracy, reliability, the impact on mortality (how many deaths are actually prevented) and cost effectiveness. The continuing debate over PSA testing and the renewed debate over mammograms for women under 50 are examples.
EVALUATIONS OF SCREENING METHODS
Screening methods for a particular disease are tested by clinical trials. Randomized controlled trials (RCT) compare the results of a screening test between two groups of people randomly assigned to one of two arms. Those in one arm will receive the test, and an equivalent number in the other arm will not. After several years, the deaths caused by the particular disease in question are counted to see if the screened arm had fewer deaths than the unscreened arm. The percentage by which deaths are lower in the screened arm is called the mortality impact. The higher the percentage the greater the lung cancer mortality benefit.
While RCT’s are considered the gold standard for testing drugs, there are conditions in which this approach is relevant and there are limitations to this methodology in evaluating a screening test. Mortality depends on many other variables beyond the screening test, such as: the treatment, how quickly the patient was treated and the quality of the surgery or treatment. The data collected can also be impacted by how long the patients were followed and how accurately the causes of death were reported.
Results from RCT’s on the same screening method can vary widely. Eight large RCT’s have been carried out on mammography, with results ranging from no benefit to a 15%-32% (depending on age) reduction in mortality. (ii)
Two RCT’s have been completed on PSA screening for prostate cancer - one in the U.S. and one in Europe - with conflicting results. The European Randomized Study of Screening for Prostate Cancer (ERSPC) showed a 20% mortality benefit (iii) while the United States Prostate, Lung, Colon and Ovarian (PLCO) trial showed none. (iv) In fact, the number of deaths from prostate cancer were slightly higher in the screened arm in the U.S. trial. (v)
Screening tests are also evaluated for their sensitivity (the probability of accurately diagnosing cancer) and specificity (the probability of accurately ruling out cancer). These too can vary widely as a function as to how these parameters are defined.
Mammograms trials have resulted in sensitivity ranges of 77-95% and specificity ranges of 94-97%.
The numbers on PSA testing vary depending on age and the cutoff value of the PSA level with sensitivity ranging from 70% to 86% and specificity from 20% to 49%. (5). But no PSA level resulted in both high sensitivity and high specificity.
The U.S. Preventive Services Task Force (USPSTF), a federally appointed and funded panel, evaluates screening data and makes recommendations. While these recommendations are not binding on standards for clinical management, they are often used by public and private insurers to determine coverage and may be the basis for minimum coverage requirements under healthcare reform. (vi)
USPSTF Rating System (vii)
A: The USPSTF recommends the service. There is high certainty that the net benefit is substantial.
B: The USPSTF recommends the service. There is high certainty that the net benefit is moderate or there is moderate certainty that the net benefit is moderate to substantial
C: The USPSTF recommends against routinely providing the service. There may be considerations that support providing the service in an individual patient. There is at least moderate certainty that the net benefit is small.
D: The USPSTF recommends against the service. There is moderate or high certainty that the service has no net benefit or that the harms outweigh the benefits
I: The USPSTF concludes that the current evidence is insufficient to assess the balance of benefits and harms of the service. Evidence is lacking, of poor quality, or conflicting, and the balance of benefits and harms cannot be determined
Currently the USPSTF gives a B rating to mammograms for women 50-74 and a C rating to mammograms for women under 50. The newer digital and magnetic resonance imaging (MRI) screening methods for breast cancer are rated I. Clinical breast exams beyond screening are rated I. The USPSTF recommends against teaching women to do breast self-exams - a D rating. (viii)
For colon cancer screening, USPSTF gives colonoscopies for people between 50 and 74 an A rating, but a C rating to the tests for people over 75 and a D rating for those over 85. Colon cancer screening by CT scanning (such as the President had in his 2009 physical check-up) and fecal DNA testing are both rated I. (ix)
PSA prostate cancer screening for men under 75 has an I rating and is not recommended for men over 75 - a D rating. (x)
Screening for lung cancer by CT scanning, chest x-ray or sputum cytology has an I rating, the same as prostate cancer screening. (xi)
IMAGING AND LUNG CANCER SCREENING
Chest x-rays produce flat two dimensional (2-D) images that are often foggy, of limited sensitivity and difficult to measure volumetrically. Tumors of less than half an inch are frequently missed, especially in out areas of the lung.
Computed tomography (CT) scanners take x-rays from various angles in a rapidly descending spiral around the chest. The digital data is reconstructed into 3-D images on a computer screen and can be measured volumetrically. Nodules as small as a grain of rice which regular x-rays would miss can be picked out.
However, the greater clarity also requires a higher radiation dose (see below). Scanners can also pick up a lot of “noise” that has to be sorted out, another reason for only going to a site with lung screening experience that adheres to the best published protocols - a step by step process - for analyzing nodules and diagnosing tumors.
Most people, especially smokers, former smokers and those living in air-polluted areas, will have abnormalities or suspicious nodules in their lungs. Some can be readily identified by their size and shape as cancers and others as non-malignant areas of calcification or scarring. But smaller nodules of indeterminate shape and density usually require a second scan a few months later. Images of the suspicious nodules from the first and second scans can be measured three dimensionally and compared side by side to check for any growth or change in volume, which usually indicates malignancy. A follow-up PET scan or an imaging-guided fine needle biopsy may be required to definitively confirm the malignancy and stage.
CT LUNG CANCER SCREENING PROTOCOL
Screening, then, is not a single snapshot. Screening is a step by step process - a protocol. involving all the component actions required to find an asymptomatic cancer in a high risk population. The CT lung cancer screening protocol was pioneered and developed over the past 17 years by the International Lung Cancer Action Program (I-ELCAP).
Researchers at fifty I-ELCAP sites around the world are continually refining the process as imaging advances come on line and as the results from ongoing trials and other research become available. The ongoing I-ELCAP observational trial (one arm trial), the largest and longest ever carried out, has demonstrated that CT screening can diagnose lung cancer at early stage 85% of the time, and those treated promptly have a ten year survival rate of 92%. Its protocol is the best published practice for lung cancer screening. The protocol, published papers on its findings and additional information on I-ELCAP research can be accessed at the I-ELCAP website. (xii)
Computer aided detection (CAD) tools can enhance the accuracy and efficiency of the CT screening process and help determine if a suspicious nodule is malignant. For example, researchers in Europe incorporated a new CAD program in their screening trial and achieved 95% sensitivity and 99% specificity - extremely high levels of accuracy and better than the levels achieved in trials of mammography or PSA screening.
CT LUNG CANCER SCREENING TRIALS
The Prostate, Lung, Colon and Ovarian (PLCO) trial was initiated by the National Cancer Institute (NCI) in 1993. Chest x-rays were used as the screening method for the lung cancer arm, but the results have not yet been made public.
The U.S. National Lung Screening Trial (NLST) launched by NCI in 2002 was ended in November 2010 because it met it's endpoint of proving that screen with CT scans can reduce lung cancer deaths by 20%. For complete results visit the NCI Website: http://www.cancer.gov/newscenter/qa/2002/nlstqaQA.
In Europe, the Netherlands and Belgium launched a collaborative RCT in 2004, referred to as the NELSON trial, an acronym for its Dutch title. The study of 16,000 people is compares outcomes in two equally divided arms - one screened with low dose CT scans and the other not screened at all. In addition to looking for the mortality impact of screening, these investigators are methodically tracking and evaluating the lesions picked up by the scans for type, shape, location and growth rate, which will yield valuable data for imaging and molecular science researchers.
Although the trial will not be completed until 2015, two papers on preliminary NELSON findings were published in 2009. (xiii) They indicate that lung cancer is being detected at Stage 1 in 70% of those diagnosed- a finding that validates the 15 year old, ongoing I-ELCAP observational study of 50,000 people at high risk for lung cancer which was started in 1992. The NELSON study is also utilizing very low dose scans with new CAD software and achieving unprecedented levels of 95% sensitivity and 99% specificity in diagnosing cancer accurately.
The Danish Lung Cancer Screening Trial, initiated in 2005, randomized smokers and former smokers two arms. Half will receive annual CT scans for five years and the other half will not. The baseline report published in 2009 indicated a high rate of lung cancers being diagnosed at early stage. Most of the early stage lung cancer patients are being treated with minimally invasive video assisted thoracic surgery (VATS) resections. (xiv) The trial is still ongoing.
Imaging procedures are useful clinical tools that can help identify and manage disease, but they also carry health risks. It is important for a patient to understand the trade-off between the potential benefits of an imaging procedure and the potential risks and to discuss the balance between potential benefits and risks with their physician. As a general rule, no imaging procedure should be done without good cause or at any higher level of risk than absolutely necessary.
A Computed Tomography (CT) scan provides a wealth of important information on internal anatomy that is often unattainable with other types of imaging procedures. However, CT scans expose a patient to x-ray radiation, which will cause damage to the areas of the body that are exposed. Fortunately, the human body has a natural mechanism for repairing radiation damaged. This mechanism is present to repair the body as a result of our continuous and unavoidable exposure to low levels of background radiation. One of the most common ways to express radiation dose is the sievert (Sv), which attempts to quantify the biological effects of radiation. In the United States, the average background radiation dose is approximately 3.0 milliSieverts (mSv) each year, with a large amount of variation depending on several factors, including higher doses for individuals living at higher elevation levels.
When considering a CT scan, the benefit of obtaining high resolution information of internal anatomy must outweigh the long-term health risks associated with additional tissue damage that is not fully repaired. When the level of radiation exposure is high, such as was the case for individuals living near the detonation sites of the atomic bombs dropped on Japan during the Second World War, the body is unable to fully repair itself leading to increased rates of cancer and other diseases. However, epidemiological studies have not demonstrated that exposure to low doses over a number of years would have the same cumulative effect as the high doses sustained by people near the bomb sites. (xv)
When the level of radiation exposure is low, such as is the case of performing a mammogram, a chest x-ray, or a low dose CT scan to detect cancer early, the benefits to the individual must once again outweigh the risks.
Men and women in the United States have an average lifetime risk of 40% of developing some kind of cancer.
In the case of a typical mammogram, the radiation dose is approximately 0.7 mSv for a single examination. A single 1.0 mSv dose of radiation is estimated to cause 1 cancer in 10,000 individuals. Thus a woman receiving 10 years of annual mammograms would then increase her lifetime risk of developing cancer from approximately 40.0% to 40.07%. Given that nearly 1 in 8 women will develop breast cancer, the benefits of regular mammography have generally been determined to outweigh the increased risk of causing cancer and thus women above the age of 40 are typically encouraged to regularly perform the procedure.
Annual low dose CT lung cancer screening should also be considered in the context of individual benefits and risks. The I-ELCAP and NELSON lung cancer screening trials utilize particularly low dose CT settings and achieve high resolution CT scans of the lungs with a dose of approximately 1.0 mSv. Similar to mammography, a person receiving 10 years of annual low dose CT lung cancer screening would increase their lifetime risk of developing lung cancer by 0.1%.
On average in the United States one in 15 people will be diagnosed with lung cancer in their lifetime, with much higher odds for those with a higher number of risk factors. Although cancer screening is under constant scientific debate, over ten years of study on screening for early lung cancer has shown the potential to detect a significant percentage of lung cancers at an early, curable stage. For individuals with elevated lung cancer risk, such as those that have previously smoked or individuals with a family history of lung cancer, the potential benefits and risks of lung cancer detection with low dose CT should be discussed and reviewed with a physician. If low dose lung cancer screening is performed, adherence to high quality protocols and procedures, as has been continuously refined and used in the I-ELCAP study, is important.
Some nodules identified as suspicious in the first scan may prove to be negative in subsequent testing. These temporary “false positives” may cause anxiety while additional testing is carried out. Similar anxiety is caused by false positive mammograms. But those who actually gone through the breast or lung cancer screening process contend it is a minor discomfort they are willing to tolerate to avoid a late stage diagnosis. A spin off study from the NELSON trial indicated that while anxiety increased for those with inconclusive initial findings, the levels drop greatly for those confirmed negative.
Controversies persist, even over long-accepted cancer screening methods such as mammography and PSA testing, and newer ones such as CT lung screening and CT colonoscopies. Issues include:
- * False positives and false negatives: how accurate is the screening method
- * Lad time bias: does screening merely find cancers earlier without reducing mortality
- * Length bias: does screening find primarily less aggressive cancers
- * Over diagnosis: does screening find too many cancers that would never be lethal
The last one is also called “pseudo-disease.” Calculations based on computer models have been used by some critics of screening to claim that over diagnosis is rampant. But these claims - like computer modeled weather reports - are mathematical projections based on several variables and certain assumptions which do not always prove to be correct in reality. Testing the hypothesis that some screen detected cancers are “pseudo-disease” with a randomized controlled trial would be nearly impossible since practically all patients diagnosed with a cancer would want to be treated. There are also ethical questions involved.
To help address the issue, LCA commissioned at actuarial study by Milliman Inc, a world-renowned actuarial firm, which analyzed the outcomes of 300,000 actual lung cancer cases in the NCI SEER registry data bases. The mortality outcomes of patients diagnosed at varying stages of lung cancer were compared with mortality records of similar sets of the general population stratified by age, sex and ethnicity.
Using the same type of analysis that insurance companies would use to set rates, the study determined that 70,000 lives a year could be saved if lung cancer were diagnosed at early stage. The study also demonstrated that most of those diagnosed and treated at early stage were indeed cured and that pseudo-disease, lead time bias and length bias were not factors. The entire study, entitled “An Actuarial Approach to Comparing Early Stage and Late Stage Lung Cancer Mortality and Survival,” can be reviewed on LCA’s website. (xvi)
These issues will take on greater urgency as healthcare costs continue to soar. Attempts to cut costs also raise difficult questions about what screening tests should be given, for which cancers and to whom. The most recent mammography controversy in the United States is a preview of how political those debates will be.
THE FUTURE OF SCREENING
The ideal screening test would find cancer early, even at pre-cancerous stages, and would also be able to differentiate between aggressive cancers and more benign tumors. For certain cancers, such as breast cancer, tests for genetic markers and errors or mutations in one of the body’s genes are already being used to identify those who may be at higher risk for cancer than the rest of the population and help people decide whether to be screened or not.
Researchers are gradually unraveling the molecular make up of cancer cells and their immediate environment and they are looking for the biomarkers of lung cancer that may be found in the blood, urine, sputum or breath. Testing the tumors for genetic mutations may lead to more individualized drug therapies.
Validation and translation of these and other promising leads into reliable diagnostic and treatment tools for lung cancer and all cancers will require a significant commitment of research funding over many years.
The first applications of imaging as a biomarker are moving towards validation. With the increased sensitivity of imaging technology, minute changes in tumor cells can be detected. So the efficacy of new drugs could be tested on smaller populations and responses gauged quickly and more accurately, thus decreasing the time and cost of drug development.
While initial medical costs associated with managing the vast at-risk population for lung cancer is substantial, the long term economic benefit of a significant mortality reduction have not been fully calculated. Nevertheless, in our society budget pressures should not be allowed to create a barrier to objectively determining the full benefit of this early detection approach. Researchers, doctors and patient advocates must maintain a united front.
Every screening test has potential risks and benefits. Imaging is the only screening tool available for lung cancer now and for the foreseeable future. Those at high risk for lung cancer and their primary care doctors need to be fully informed in order to reach a decision on screening appropriate to each individual case.