Jump to content

new lung cancer detection formula


Recommended Posts

New Lung Cancer Risk Assessment Formula

(March 19, 2003) A new lung cancer detection formula is available for those people 50 years and older, who have smoked at least half a pack a day for at least 25 years. The formula, created by researchers from New York's Memorial Sloan-Kettering Cancer Center, is available on www.mskcc.org and will help doctors identify high risk patients. The data can also be used to make necessary health care decisions, such as whether a patient should seek the controversial spiral CT scan.


Link to comment
Share on other sites

Thanks for the info. I used to go along with the medical profession about the spiral cat scans. But I have reversed my position.

It seems to me that they should be able to do a scan. If there are areas of suspician then wait a few weeks and do the scan again. Compare the baseline against the new scan. If they continually see the same lesions and possible growth, then obviously malignant. Since it is a low dose scan then getting them more often may not be a concern.

Also there is a device called rapidscan 2000. It is essential an x-ray that is digitized. Digital signal processing and neural nets are used to detect lung cancer by processing the digital image. The sensitivity and specificity approach a CT scan. I think it is somewhere around 80%, 80%

Link to comment
Share on other sites


I'm with you on this one. I think the medical position on using CT scans for lung cancer screening is a bunch of hogwash to obfuscate the stigma issue. I get a mammogram every year, have my teeth X-rayed, etc., etc. Your points about waiting and doing another comparison scan are certainly valid, because that is PRECISELY what they do with many of us LC survivors right now!!! What's the difference, except that many, many more people could be helped earlier??? I know, I know, I'm preaching to the choir here :roll: Always enjoy your research and points of view.

Link to comment
Share on other sites


Enjoy your posts too. I do get concerned about getting too much exposure to radiation from tests though as x-rays, etc do also cause cancer. Curie and a few of the early x-ray experimenters all got cancer. Although, the newest tests and therapy seem to pinpoint the radiation better. And hopefully doctors are keeping track of their patients exposure (I doubt it)

I still wish my research could have helped my mom. I just hope this sh** wont be here when my daughters are older


Link to comment
Share on other sites

Role of CT Screening for Lung Cancer, Part I: Pros and Cons Persist

from Medscape Radiology

Posted 03/21/2003

Claudia Henschke, MD, PhD

At the European Congress of Radiology (ECR) held in Vienna, Austria, in mid-March, Claudia Henschke, MD, PhD, provided attendees at a lung cancer management symposium with an update on the latest results from her group's world-renowned Early Lung Cancer Action Program (ELCAP) study.

Dr. Henschke noted that the ELCAP study continues to gather compelling evidence to support the role of screening CT in high-risk patients, especially in the follow-up of patients who have a small pulmonary nodule detected on an annual study. However, the consensus of the investigators discussing this topic at the ECR meeting was that additional data were needed to support the widespread use of screening.

Dr. Henschke, Professor of Radiology at Weill Medical College, Cornell University, New York, NY, is one of the preeminent researchers in the world in the area of CT scanning for the detection of lung cancer, has published extensively in the clinical literature and in other venues that address this topic, and is a featured speaker at major medical meetings, such as CHEST and RSNA, that target this area of investigation.

To place the key issues surrounding the current role of screening CT in a meaningful clinical perspective for imagers, pulmonologists, oncologists, and other caregivers at the ramparts of lung cancer detection and management, Medscape sat down to talk with Dr. Henschke at her offices in New York City.

Medscape: Dr. Henschke, you've been at the epicenter of the CT-screening-for-lung-cancer debate for a long time. What has prompted your focus and work in this arena over such an extended period?

Dr. Henschke: I decided to concentrate on chest radiology. We started a weekly oncology conference in the late 1980s. In the early 1990s, when helical CT scanning was first introduced, we started to see cases of tiny nodules in the chest and it was apparent that nobody was quite sure what they were. The introduction of this technique meant that we could go up through the chest in a single breath-hold scan lasting 20 seconds. As more CT studies were being done, it became clear that there was no information base from which to formulate a diagnosis. We decided to study the frequency with which pulmonary nodules were found and how many of them were cancers. This naturally led us to the topic of screening for lung cancer.

Medscape: The now familiar indeterminate pulmonary nodule.

Dr. Henschke: That's correct. There was a dearth of information in the literature, we had no real knowledge of what early lung cancer looked like, and we did not know how normal, usual pulmonary nodules would appear on CT. I organized several off-site retreats for participants of our thoracic conference where these issues were discussed. From that beginning, we decided that lung cancer screening needed to be reevaluated in the light of this new technology. Our group started developing our protocols and grant applications in 1992 and we were finally able to get started in 1993. The aim of our research was to explore the use of CT technology in screening for lung cancer. However, we also were hoping to offer a base of knowledge that would refine the practice of chest radiology for this indication.

Medscape: The debate around CT screening seems to reach a flashpoint each time a new study is published. The recent JAMA article that touched on lung cancer screening of older adults using CT was not supportive of the technology in older smokers, citing factors such as cost and the lack of verifiable evidence that the use of the technique has a positive impact on clinical outcomes.

Dr. Henschke: The article in JAMA highlighted from a national perspective the many considerations that should be addressed in evaluating CT screening programs. And I thought that aspect of the study was good. The part that I was less enamored with relates to the many assumptions made in the modeling of the study and the conclusions that were drawn from these theoretical data, which were flawed. Basically, the study investigators drew broad assumptions about almost every aspect of the study. And anyone can make any model produce whatever information they want, depending on what assumptions are made about the values of the variables in the model. This group did not show an extensive sensitivity analysis. So, for example, on the basis of their assumptions they thought that there would only be a modest mortality benefit, but these were highly unrealistic assumptions. In the absence of any real data, one can reach any conclusion.

In the research we are undertaking in the Early Lung Cancer Action Program (ELCAP), we think that the cost-effectiveness of CT screening is a critical issue; and we are using our actual patient data to assess the cost per life-year saved.

Our data are very different from the findings of the recent JAMA study and many other studies being reported in the literature. We come to a different conclusion. To us, even under the worst-case scenario, CT screening is going to prove to be cost-effective. Our data show that the cost is going to be about $2500 per life-year saved. Even if you took the costs up to $10,000 per life-year saved, this approach would be highly cost-effective.

All of the other screening procedures for other cancers -- breast cancer, colon cancer, cervical cancer -- are more expensive than CT screening for lung cancer. Two factors come into play when we look at the costs of screening for lung cancer with CT: First, there's the cost of the examination, which averages around $300. If the center employs a good protocol, such as the one that we have published on our Web site and encourage people to follow, then you don't intervene unnecessarily. Furthermore, when you detect a lesion that signifies early lung cancer, the cost of treating the disease at that stage is at least half the cost of treating late-stage lung cancer. Targeting a specific patient population is critical in diagnosing this form of cancer, which is why we have focused our efforts on older, high-risk people -- smokers and former smokers. When our study is completed and published, the conclusions based on our real data are going to be very different from what is currently being disseminated in the clinical literature.

Medscape: And the ELCAP study is ongoing?

Dr. Henschke: Yes. The validation of a screening approach for any disease is a complex process. There are lots of considerations that need to be addressed; for example, you need to first define the target population, so that screening provides more benefit than harm. Furthermore, it's not just measuring that one initial, low-dose imaging study in a vacuum. CT scanning needs to be properly positioned in a diagnostic continuum employed to arrive at a diagnosis. The diagnostic work-up needs to be carefully specified. In fact, one of the primary concerns we have with the current national lung cancer screening trial is that the role of CT among an array of diagnostic measures is not defined with precision, which leads to a misunderstanding of the role of the test as only one part of a broader approach linked to smoking cessation, lifestyle modification, and close follow up. If I just simply say, "Report the findings of low-dose CT in this instance," and if I don't place those results in the proper context, the results will be flawed.

Medscape: To follow up on that thought, can you outline for us what you think will distinguish the results to come from ELCAP from other national clinical trials related to this issue?

Dr. Henschke: We have very good results because we keep retooling our screening protocol on the basis of the new capacities of the technology and our advancing knowledge base regarding the findings of CT studies. When we embarked on the study we were using a single-slice helical CT scan, which acquired thirty 10-mm slices to image the entire chest. At that time, that was the extent of the study. Today, we can acquire 1-mm slices and generate 300 images; this technology will continue to develop.

The difference from yesterday to today is that I get much more information from that first scan. However, I can't proceed with the rest of the diagnostic work-up in the same way that I did before the improved technology, as it also uncovers findings that have nothing to do with cancer. It's important for the public to recognize that lung cancer screening is an evolving process that has changed dramatically in the past 3 or 4 years. We have advanced from single-slice CT to 4-slice CT to 8-slice CT to 16-slice studies. The improved technology is really changing the ways in which we measure the growth of nodules to better identify early lung cancer. In fact, when we review our data we see that our diagnoses track the latest advances in technology. The better the scanners become, the earlier I can see the lung cancer. So, to arrive at results that have a high degree of accuracy, you have to be continuously evaluating and updating the imaging protocol and reassessing your results.

Medscape: So it's a dynamic rather than a static process.

Dr. Henschke: Absolutely a dynamic process. One of the key challenges inherent in a trial that is the length and extent of the ELCAP study is that our results today are very different from what we were seeing 10 years ago; 10 years from now it's going to be very different as well, as the technology and techniques improve.

Medscape: A consistent refrain in imaging, new technology as the key market driver.

Dr. Henschke: Yes. The diagnosis and treatment of lung cancer is in the midst of a technological revolution that will change the practice patterns of everyone involved in the care of these patients. In the future, we will have the software that will allow for automated reading and assisted automatic reading of these scans. This will help us to determine whether the nodule has changed within 1-3 months after the initial scan to see whether the patient needs to be subjected to additional testing or other, more invasive procedures. We'll be able to verify that the pulmonary nodule is malignant at a very early, almost preclinical stage, which will lead to improved cure rates and cost-reduction by limiting duplicate or unnecessary testing.

A side benefit of our research is that our work has helped many patients who are enrolled in our program to stop smoking. I would show them their lung scans and the damage that occurred from smoking, and the next day they'd call me up and say, "You know, that was very helpful, because every time I lit a cigarette I thought about my lungs and what they looked like."

Because of this finding, we performed a study that was published in Preventive Medicine. In this study we demonstrated that by talking with people -- smokers and former smokers -- and showing them their CT study and then enrolling them in our lung cancer screening program in which there is an excellent chance that any lung cancers they develop will be detected early, this process literally encouraged them to quit smoking.

Medscape: Is that because instead of talking about hoping to quit smoking, which is kind of an abstraction, you're giving them a specific visual corollary of the damage being done?

Dr. Henschke: That's part of it. Some people don't want to see their lungs get worse after you show them the effects of emphysema. Some people want to be sure that they don't have cancer or won't get cancer; for this group of patients, enrolling in a screening program for early detection then compels them to stop smoking. It's a multifaceted problem. We are undertaking further studies in this area, and in fact, we see the screening program that we offer as inextricably linked to smoking cessation.

Addiction to nicotine is worse than any other addiction. What we emphasize to the patients in our program is that once you've been diagnosed with lung cancer you enter the category of highest risk, which really necessitates CT screening on an annual basis.

Medscape: Recent guidelines from CHEST seem to indicate that there is little evidence that CT screening for lung cancer has any impact on patient outcomes or survival rates. The group recommended that CT screening not be undertaken until more data surface to support a screening scenario as part of a plan to manage this disease. Could you address this?

Dr. Henschke: The efficacy of any screening program is judged in part by the cure rates that result. That critical mass of data takes time to accumulate and be analyzed in order to determine how many lives you save by employing a screening approach. We are just about there in terms of the data; we have been using CT scans to screen patients for approximately 10 years, and other investigators in Japan have also been screening patents for that length of time. Actually, I think that if we pooled our data with the Japanese investigators', we would be able to answer the question of the impact of CT screening on patients' outcomes right now, or in the very near future.

When you can detect through CT screening more than 80% of cases of stage I lung cancer -- and that's really been duplicated throughout the world now -- that's a profound paradigm shift in the detection and management of this disease. The usual finding of stage I lung cancer is only between 5% and 15% throughout the world. So, when you have a shift in early detection rates from about 15% to over 80%, that's compelling. We feel that there is enough evidence to justify CT screening of people in high-risk categories. The challenge lies in determining exactly who is a high-risk patient and then linking imaging to an overall plan to detect and manage disease. This is best undertaken in a multidisciplinary scenario in which the radiologist, oncologist, pulmonologist, and other caregivers discuss the results of the study among the care team and with the patient, where there is a specific plan for follow-up, and where a precise protocol for imaging and clinical intervention is implemented.

The emphasis in this area should not be solely on the assessment of the validity of screening CT screening; we feel that screening has to be considered as part of a continuum of evaluation and care that includes quality assurance parameters for the acquisition of the study and yearly follow-up imaging in high-risk patients, a clear diagnostic work-up in the event that disease is detected, and clinical pathways for intervening with surgery and/or other therapies.

Medscape: So, screening is only a portion of a much broader context that includes smoking cessation and the implementation of cost-effective therapeutic protocols. A lot of the literature on this topic, especially in radiology, seems to deal with screening in a vacuum, as if it were an end point instead of a beginning.

Dr. Henschke: Exactly. That's why we call it "screening." If you think of the evaluation of potential lung cancer as just a test, that's not the way you should be practicing screening in any case.

Screening CT is only the initial measure in a whole sequence of diagnostic tests that either confirm that you are cancer-free, or that you have a finding that needs to be monitored carefully over time and subjected to follow-up testing. There are 2 critical aspects that comprise an effective screening sequence: First, after a single negative screening study in a high-risk patient, you can't just stop and think that that patient is not going to have lung cancer; and second, that the successful implementation of this approach relies on a whole series of tests and clinical decisions. There are certain types of cancers that might grow very fast or that we aren't so good at detecting. Now, luckily it's only in a small proportion of cases that we don't detect the cancer early enough.

What we have observed in our study is that in instances in which we detect lung cancer early, there is a positive impact on clinical outcome. As I mentioned earlier, we've been detecting lung cancers for over 10 years and many of our patients are doing very well. We believe that our data, which will be available within a reasonable amount of time, will validate the utility and cost-effectiveness of CT screening for early lung cancer.

Coming Soon: Part II of Medscape's conversation with Dr. Henschke will showcase the objectives of the ELCAP study and highlight how that endeavor integrates a broad clinical continuum of CT screening, smoking cessation, patient risk stratification, and follow-up protocols with newly emerging treatment strategies for managing the early stages of lung cancer.

Claudia Henschke, MD, PhD, Professor of Radiology, Weill Medical College, Cornell University, Division of Chest Imaging, New York Presbyterian Hospital, New York, NY

Link to comment
Share on other sites

It seems like there are many people on this board whose cancer was picked up on a plain ole chest xray ordered for something else. Also many of the people I talked to while I was undergoing treatment had their LC picked up by a chest xray, including me. So why don't PCPs order annual chest xrays for their smoking patients? Ct scans are more expensive and less available. Is there anyone on this board whose cancer was picked up FIRST on a CT?

Just an observation. JudyB

Link to comment
Share on other sites

Judy, I also was fortunate to have my tumor picked up by a simple chest xray. The problem is however that people should not be led to believe if the chest xray is clean you are cancer free. If you have a very solid tumor it will show, but lung cancer is not always a solid tumor, so it does not show especially early which is what we need to survive. Chest xray for some is better than nothing but so many others will have lung cancer and I think will be worse off because they think that they are cancer free . We really need to push for the CT, the pet scan or the discovery of a new screening tool that is very effective in finding tumors .

Link to comment
Share on other sites

The biggest problem with x-rays is that is takes about a 3 cm lesion to show up. Far too late! There is a new thing called rapidscan 2000 that digitizes the x-ray and uses image processing to look for smaller lesions.

Helical CT scans are only $300! That is not too expensive. From what Dr Henschke said they can get 16 1mm slices. that is about 30 times smaller than what a x-ray would catch

Link to comment
Share on other sites


Thanks bunches for the Dr. Henschke article! As I've noted elsewhere, I am a real fan of hers because her work is so sound scientifically and yet her writing always conveys to me those human elements of empathy and social concern too.

I hope you'll post the second part of the article when it is available???

Smiles, :D

Link to comment
Share on other sites

I am a computer scientist working on telecommunications equipment.

So I am at the computer all day. In between writing code I search the net

I have two daughters who keep me very busy and a wife who works for the public health service.

Glad to hear I am helping. I wish I could have helped my mom more but in the end everything is in Gods hands.

I am considering going back to school part time to learn about proteomics (how genes produce protiens) and maybe try to get a job at Human Genome or another biotech doing bioinformatics. That will be a while off, but I hope and pray that by the time my wife is older there will be no more cancer

Take care,


Link to comment
Share on other sites

This explains some of the cons to screening, but I think the biggest issue (do we know if a lesion is aggressive) can be easily addressed by multiple scans. If the lesions stay the same, then watch and wait. If the growth is fast, then resection.

Tumor markers are another story and it discusses this also


Michael N. Pollak & William D. Foulkes about the authors

Michael N. Pollak and William D. Foulkes are in the Programs of Cancer Prevention and Cancer Genetics, McGill University, Montreal, Quebec, Canada H3T 1E2.

correspondence to: Michael N. Pollak michael.pollak@mcgill.ca

Population-based screening seems to be a common-sense strategy for controlling cancer, but recent reports have raised controversy concerning the benefits of common screening procedures. Intense efforts to develop and evaluate novel screening technologies are underway; however, effective use of any screening method must take into account any underlying biological considerations. What are these biological issues, and what challenges do clinicians face in screening for common cancers?

Screening populations for the presence of an asymptomatic illness will clearly be useful if the screening method is convenient and inexpensive, if false negatives and false positives are rare, and if treatment that was initiated at the early, screening-detected stage is more successful than treatment initiated later in the natural history of the illness. Although it is true that the outcome of cancer treatment is almost always more favourable when treatment is initiated early, it is simplistic to extrapolate that, in general, population screening will provide a practical method for cancer control. Indeed, recent reports have raised controversy concerning the benefits of common screening procedures, such as breast self-examination (see TIMELINE), mammography and prostate-specific antigen (PSA) measurements1-3.

Timeline | Breast self-examination – changing attitudes

Important issues include lack of benefit of early detection of disease that cannot be effectively treated and, conversely, early detection of conditions that have a low or unknown probability of causing clinical disease later in life. Screening procedures must be evaluated not only from the point of view of individuals at risk, but also from the point of view of populations at risk — conclusions obtained from rigorous assessments of screening methods might not always correspond to expectations.

As long as our ability to treat advanced cancer remains inadequate, there will be obvious pressure to implement effective screening strategies, and new technologies for detecting cancer — for example, spiral computed tomography (CT) imaging for lung cancer detection and proteomic serum assays for ovarian cancer detection — are being developed at present4, 5. However, success of a screening method in detecting cancers before they are clinically apparent is not necessarily sufficient for a screening method to be useful.

Biological issues affect screening

Often, research agendas that are designed to improve cancer screening methods emphasize technology development (for example, serum proteomics or high-definition digital mammography), and assign relatively little attention to the biological characteristics of cancers that influence the efficacy of screening efforts. Yet, the task of developing and implementing effective screening-based methods for cancer control is challenging for biological reasons that deserve consideration.

Stepwise progression to malignancy. The fact that carcinogenesis rarely involves a single transforming step from normal cells to aggressive cancer cells complicates the issue of screening. A method that identifies only fully transformed neoplastic cells with metastatic potential might indeed result in earlier detection than would be the case if patients were assessed only by clinical examination. However, a significant proportion of such lesions might have already resulted in micrometastases when detected, thereby limiting the benefits of early surgical removal and the entire screening exercise (Fig. 1). Although many screen-detected cancers have good prognostic features6, the fact that a cancer is detected by screening does not guarantee curability.

Figure 1 | When in the natural history of a cancer will screening be useful?

This varies according to the biological behaviour of the tumour. In the illustration, 'M' indicates the timepoint at which aggressive tumour cells metastasize. In the case of tumour 1, screening and resection at time A will result in cure of an otherwise lethal cancer, whereas screening and resection at time B will not prevent death from metastasis. In the case of tumour 2, screening at time A will represent a technical triumph, as the primary lesion is small and perhaps hard to detect, but will not be helpful, as this cancer has already metastasized. Detection at time B will be of even less value. In the case of tumour 3, the natural history is non-aggressive: successful screening at either time point will probably lead to interventions, possibly with associated risks, for a cancer that would never have caused a clinically important illness.

Screening methods that attempt to avoid this hazard by detecting the presence of cells earlier in the process of stepwise carcinogenesis face different challenges. There is evidence that some partially transformed cells progress very slowly (over decades) towards a clinically aggressive phenotype. Furthermore, carcinogenesis occurs in a parallel manner among at-risk cell lineages. In many organs, early neoplastic lesions are much more common than aggressive cancers. For example, polyps are more frequent than cancers in the colon7, and there is evidence that approximately 40% of all men over the age of 60 years have prostate neoplasia that is detectable at autopsy, but only 3% have lesions that will affect their lifespan8, 9.

PSA measurement is an important example of a sensitive screening method that can detect the presence of neoplastic cells relatively early in the transformation process, compared with most imaging-based screening techniques10. The rate of neoplastic progression in prostate cancer, as in other cancers, varies considerably between tumours. There is evidence that progression of a large subset of prostate neoplastic lesions towards an aggressive phenotype is slow, requiring years or even d ecades. If PSA or more sensitive screening methods were to allow identification of all men with some form of neoplasia, a situation would arise in which a large proportion of the elderly male population would require intervention. Such screening results would not represent false positives in a technical sense, as neoplasia would, in fact, be present, but the results would be problematic clinically as the natural history of many (but not all) screening-detected lesions would be favourable.

In clinical practice, PSA screening often leads to difficult choices for men who discover the presence of lesions of uncertain prognosis. Clinicians who advise men found to have elevated PSA on screening must first rule out frank false positives (related to prostatitis, for example), usually by biopsy. If neoplasia is confirmed, clinical judgment — taking into account the patient's age, co-morbid conditions and clues regarding prognosis, such as the Gleason score — is needed to decide if intervention is in the patient's best interest. Many patients, understanding that their lesion might or might not result in clinical illness if left untreated, choose to accept treatments with non-trivial morbidity 'just to be sure'. However, this might not provide the anticipated piece of mind, as many treated men experience anxiety concerning risk of recurrence11, 12. Although there is a risk that PSA screening might fail to be of value because of highly aggressive interval cancers that can arise between sequential PSA measurements, in practice this is less frequently encountered than the detection of neoplastic lesions that are unlikely to result in clinical illness if left untreated, represented by tumour 3 in Fig. 1.

In contrast to certain screening methods for other cancers in which insufficient sensitivity is a significant limitation, PSA screening represents an example of a screening method in which the main limitation is that only subsets of those identified on screening require intervention. The overall benefit of PSA screening programmes must take into account both the potentially life-saving benefits of curative early intervention in men with aggressive prostate cancer, and the cost and morbidity of identifying and treating lesions that are not destined to cause significant clinical illness. The discovery of genetic, serum or tumour markers that could be used in conjunction with PSA to allow screening for the presence of the subset of lesions that are likely to lead to clinically important illness could rapidly alter the landscape of prostate cancer screening. Research in this area is ongoing: recent data raises the possibility that serum analytes that are related to insulin growth factor (IGF) physiology might be related specifically to the risk of aggressive prostate cancer13, 14.

This research topic is challenging because there is likely to be a strong stochastic component to the fate of individual partially transformed cells — some might accumulate further mutations that will result in malignant progression, whereas others might remain stable or even accumulate mutations that result in apoptotic death. A concept that deserves study is the notion that germline genetic factors — distinct from somatic-cell genetics — might influence the speed of stepwise carcinogenesis, and so offer clues to identifying individuals in whom early lesions are particularly dangerous. Defective DNA repair represents one obvious example in this regard, but more subtle influences involving the effects of hormonal milieu or the probability of survival as compared with apoptosis of at-risk cells also deserve investigation.

These challenges are more complex than conventional concerns of 'false-positive' and 'false-negative' screening tests. Conventional evaluations focus on the ability of a screening test to detect a lesion, with less emphasis on the definition of biological characteristics of lesions that are worth detecting. All would agree that identifying a cyst as a cancer is a false positive (and one that will be relatively easy to remedy with improved technology), but in a biological sense identifying and targeting for surgery an early cancer that is not destined to ever cause illness is also a false positive.

Heterogeneity. Heterogeneity between cells of a single macroscopic cancer is well documented. Not only is there heterogeneity in the sense of different cell types (such as stromal, vascular and neoplastic) that comprise a tumour, but, more importantly, heterogeneity in gene expression is usually considerable even in analyses that are confined to the frankly neoplastic cells of a lesion15. A common example for clinicians is the heterogeneity of oestrogen-receptor expression among the individual malignant breast epithelial cells that comprise a breast cancer.

This heterogeneity complicates the development of tumour markers because the very notion of expression of a 'marker' by a 'tumour' is a simplification of the biological reality that, during the natural history of a particular neoplasm, the percentage of cells that express a given antigen will vary. Neoplastic progression towards aggressive cancer occurs simultaneously among many clonal populations, which independently undergo somatic-cell mutational events.

Sophisticated proteomic analysis of neoplastic cells obtained from a tumour specimen by laser capture microdissection might minimize concerns related to heterogeneity between neoplastic and stromal cells, but it does not address the problem of heterogeneity within the neoplastic cell population. Data obtained from a single cancer represent an integration of expression information from many individual cells. Gene-expression signatures indicating aggressive neoplastic behaviour are being developed, but, if the proportion of cells with the signature is small, the signal might be obscured by a large majority of sample cells that do not express the signature. So, negative results for a serum marker that is used for screening might indicate an absence of cancer, but also the presence of a cancer at a stage where only a small percentage of the cancer cells express the marker. A special case that might show promise is the use of marker proteins that are necessary for an aggressive phenotype, rather than those that are simply associated with malignancy.

Small lesions can be aggressive. The notion that cancer deaths in a population can be significantly reduced by public-health programmes based on the credo 'find it early and cut it out' is simplistic, because the size of the primary lesion is only imperfectly related to metastatic potential. Small-cell lung cancer (SCLC) provides an example in which the probability of metastatic dissemination of even very small neoplasms is high. Small breast cancers are less likely than SCLC to metastasize early; however, the probability of metastatic spread before detection of breast cancers by screening is high enough that large numbers of women are offered adjuvant hormonal or chemotherapies following resection of apparently localized lesions. The fact that such adjuvant therapies improve survival represents medical progress, but, conversely, indicates that current screening does not always detect breast cancer early enough. Also problematic is the fact that, at present, no screening method, even accompanied by detailed ancillary investigations of screening-detected lesions, identifies with certainty those breast cancers in which surgical cure can be guaranteed, rendering adjuvant therapy unnecessary. Consequently, many women who are, in fact, surgically cured must undergo adjuvant therapies. Conversely, chemotherapy might not be offered to women with very small breast cancers. These tumours usually have a very good prognosis, but, in some circumstances, such as in the presence of BRCA1 mutations, small size and negative nodal status might not guarantee a good prognosis16. Identifying women who are BRCA1-mutation carriers before diagnosis (or at least before treatment) might result in different therapeutic decisions. Recent progress17 in defining breast cancer gene-expression signatures that are associated with poor prognosis even in the presence of small tumour size represents an important step towards the successful selection of patients for adjuvant treatment. However, this emphasizes one of the challenges to effective screening: certain tumours might be life-threatening even if they are found before they become large.

Varied risk, customized screening? Cancer screening methods might be applied universally according to age and gender, to selected subpopulations or to individuals. There are well-known groups (such as those with genetic predisposition or accidental carcinogen exposure) in which it might make sense to consider screening strategies that are different from those applied in the general population. However, the fact that a subgroup has a higher risk of cancer does not necessarily mean that more intensive screening will result in clinical benefits. Benefits will depend on the ability of the screening method to detect cancers at a curable stage.

This problem is exemplified by women who carry germline BRCA1 or BRCA2 mutations. Such women have a particularly high risk of developing breast cancer. The average age of onset of BRCA1/2-related breast cancer is also at least 15 years younger than in the general population, so if mammography is going to influence survival in these women, it must be started no later than the fourth, or early in the fifth, decade of life. It might be expected that the positive predictive value of mammography would be higher for women at increased risk than for women at average risk. This might be offset, however, by their younger age of onset — when the breasts are expected to be denser and mammography is less effective at detecting neoplasia — and the generally unfavourable biological characteristics of BRCA1/2-related breast cancers18. The typical BRCA1 tumour — a high-grade, oestrogen-receptor-negative, p53-positive cancer occurring in a young woman — has precisely the characteristics of those tumours that are most likely to be diagnosed in the interval between screening examinations19, and recent data show that BRCA1 tumours often present as 'interval cancers'20 and are aggressive16. Interestingly, in the positive Swedish two-counties mammography trial, there was no survival benefit for women in the screened arm who developed high-grade invasive ductal carcinomas at 50 years of age or younger21. Furthermore, in a retrospective cohort of women who had been diagnosed with breast cancer and had been tested for BRCA1/2 mutations, mammography performed before surgery was significantly less likely to show a cancer if the woman was a mutation carrier and was either under 50 years of age at diagnosis, had a small tumour, or both22. So even if frequent screening were to be performed, BRCA1/2-related breast tumours might be missed by mammography. These findings, when taken together, indicate that accepted techniques, such as mammography, might not significantly reduce breast cancer mortality in BRCA1/2-mutation carriers. Other imaging modalities, such as magnetic resonance imaging (MRI), will probably outperform mammography in BRCA1/2 carriers23, 24, but the question of the appropriate screening interval for BRCA1/2-mutation carriers remains unanswered, and MRI studies in BRCA1/2-mutation carriers have yet to report medium-term follow-up data. Even if these studies show the benefits of MRI, it might not be practical to shorten screening intervals to the extent that is necessary to detect curable cancers — four breast MRIs per year, even if possible, might not be a practical method for breast cancer control in high-risk women. Surgical prevention would probably be more effective than screening25, but it has not been universally accepted by women at risk26. Increasing acceptance of surgical measures might lessen the need for effective screening for BRCA1/2-related breast cancer, but it is difficult to imagine that successful surgical intervention will completely obviate the need for better screening tests.

If it is assumed that the general population contains subpopulations that vary in the degree of benefit that screening would provide, the question of reviewing all individuals to identify the subpopulations for which screening should either be emphasized or avoided arises. 'Targeted' screening of selected subpopulations is appealing because the benefits of screening would increase if those who do not benefit — either because they are predisposed to cancers that are so aggressive that screening is unlikely to result in cures, or because they have a very low risk of cancer — were excluded. So far, few studies have addressed these issues, but interest is increasing. In breast cancer, for example, should the recommendations for women who are found to have mammographically dense breasts on initial screening be the same as for the general population? These women have an increased risk of breast cancer27, which might imply the potential for increased benefit. Conversely, mammography might be less able to accurately identify small lesions in these women24. Colorectal cancer provides another example in which we need more information to 'customize' screening recommendations on the basis of known risk factors of modest magnitude, such as polymorphisms in known tumour-suppressor genes28 or high IGF1 levels29. In the more extreme case of certain hereditary polyposis syndromes, the role for colorectal screening is limited because the risks are so high that most favour colectomy before adulthood30.

Lack of sensitive serum markers. The identification of reliable serum markers for detecting curable cancers has been a long-standing research goal31. Some markers, such as carcinoembryonic antigen (CEA), have been found to be useful aids for physicians in determining the response to treatment, but lack the sensitivity and specificity that is required for screening. PSA is more useful for screening than CEA, but its optimum use still remains controversial. Ongoing research using genomic appr-oaches to analyse gene expression in tum-ours, or using proteomic approaches in the analysis of serum, might identify more sensitive markers than those that are available at present. However, this might create dilemmas. A positive 'cancer test' in the absence of symptoms or a lesion that can be detected by imaging would present difficult decisions for patients and physicians.

Financial obstacles to screening

Colorectal cancer illustrates some of the serious financial implications of adopting a nationwide screening programme. Although it is uncertain which methods are most effective, screening for colorectal cancer is generally regarded as useful in reducing colon cancer incidence and mortality32-34. But how much will it cost?

The data indicating that screening for colorectal cancer by sigmoidoscopy can save lives are convincing, but flexible sigmoidoscopy as a first-line screening tool has not gained acceptance. This is probably because a one-time flexible sigmoidoscopy after a positive faecal occult blood test is thought to miss a quarter of all advanced cancers; these are beyond the reach of the sigmoidoscope and do not have a distal polyp in association with the proximal cancer35. Randomized studies of one-time colonoscopy are in progress, and data available at present strongly indicate that colonoscopy every 10 years or even once at age 50 years would result in a reduction in colorectal cancer mortality32. Other, non-invasive techniques, such as 'virtual colonoscopy', might replace endoscopic techniques, but they remain underevaluated at present36.

Even colonoscopy is far from perfect. Recent research has questioned the universality of a simple adenoma–carcinoma progression model of colorectal carcinoma. In particular, studies of hyperplastic polyps have indicated that pathways to colorectal cancer might be more varied than was previously thought37. These findings could limit the effectiveness of screening and polypectomy as a method of reducing colorectal cancer mortality in individuals in whom stepwise carcinogenesis proceeds quickly.

Although colonoscopy still seems to be the best alternative, the financial costs would be enormous (Box 1). In fact, the implementation of colonoscopy is not even practical at present, as there are not enough physicians performing colonoscopies. If the costs of colonoscopy could be substantially reduced, this method might become a viable option34, but, at the moment, most societies will not be able to afford universal colonoscopy screening. Once-only population-based flexible sigmoidoscopy might be a practical alternative, even though it is acknowledged that it is not an ideal test. Cost-effectiveness analyses provide support for yearly re-hydrated faecal occult blood tests, combined with a sigmoidoscopy every 5 years from age 50 years to 85 years (Ref. 33). A large multicentre randomized trail of once-only flexible sigmoidoscopy at age 60 years is now underway in the UK and Italy. The first analysis of mortality data can be expected in 2004 (Ref. 38). If this trial shows important benefits, single flexible sigmoidoscopy by nurse practitioners will deserve consideration as a practical choice for population-based colorectal cancer screening, particularly in view of the evidence that nurse practitioners are capable of performing sigmoidoscopy to the same performance level as gastroenterologists39. Of course, those who can pay for the 'best' screening test might choose more aggressive or comprehensive screening options, but this is hardly a way to determine the best approach to population-based screening, which, by its very nature, must be funded from taxation.

Elementary anatomical issues are relevant to the practicality of screening: one reason for the success of Pap screening for cancer of the uterine cervix is that screen-detected lesions can often be dealt with conveniently in a practitioner's office without the need for expensive or dangerous interventions. A finding that lung cancer mortality could be reduced by repeated spiral CT scans of at-risk populations would be problematic, particularly if the effect on mortality were small, not only because of the cost of implementing a screening programme, but also because of the need to increase bronchoscopy and/or thoracic-surgery capacity to deal with the screen-detected lesions, many of which will be false positives4.

How to evaluate screening?

Evaluation of the benefits of a screening method is complex. The first data that become available during formal or informal attempts to assess the value of a screening modality relate to the number of occult cancers detected. Detection of occult cancers is necessary but not sufficient for a screening method to be useful. An obvious hazard, particularly in efforts to screen for neoplasms such as small-cell lung cancers that tend to metastasize early, is that screening will identify occult cancers, but subsequent to metastasis. Even an impressive detection rate, in the absence of documentation of improved survival, cannot be used to justify the introduction of a screening method into clinical practice. The current controversy regarding the value of mammography provides a clear example of difficulties that are related to statistical techniques used to evaluate screening methodology2, 40.

In most cancer screening trials, disease-specific mortality is measured. In clinical trials of treatment, purists prefer to measure overall mortality, particularly when potentially dangerous treatments are involved. Recent data from Black and colleagues, summarized in Box 2, indicate that it is also preferable to consider all-cause mortality as the measured end point in trials of screening. This opinion is based on the re-analysis of 12 published, randomized trials of cancer screening (including seven of mammography) in which seven of the studies had all-cause and disease-specific mortality rates that went in opposite directions, or differed significantly in magnitude41. Interestingly, the analysis of Black et al. confirms41 that the Edinburgh breast cancer screening trial42 had highly biased randomization, as the all-cause mortality in the control group was significantly greater than the breast cancer mortality in this group. This might relate, at least in part, to social class differences in the screened and observation groups.

It is likely that all-cause mortality is a less biased end point than cause-specific mortality. Therefore, when assessing the outcome of screening trials, it is important to remember that the total number of deaths in each arm of the study is ultimately more relevant to the success or failure of screening programmes than the number of any specific cause of death. This approach has been criticized by statisticians, radiologists and epidemiologists40, 43, 44. The main argument put forward by these critics is that many of the differences between cause-specific and all-cause mortality that were noted by Black and colleagues could be due to chance, as deaths attributed to the cancer for which screening has been initiated will inevitably represent only a small fraction of the total number of deaths in a screened population. The central point that Black and colleagues re-iterate in their response is that it is inherent in screening programmes that few will be helped, some will be harmed and most will be unaffected45. Intemperate attacks46 on those prepared to question the existing dogma2, 47 indicate the entrenched nature of the views held, and indicate that these issues will not easily be resolved.

Unfortunately, the end points that provide the most rigorous answers are the most expensive and time-consuming to obtain, requiring at least years and possibly decades. The time and resources required are such that it is unrealistic to formally evaluate all proposed screening methods. Only those that show particular promise in preliminary studies and seem feasible for general application if they prove useful can be studied. These are difficult decisions, and it is a matter for debate whether, for example, trials of spiral CT imaging of smokers4 represent a wise use of resources47.

Further challenges arise in areas of rapidly evolving technology that are related to screening. If serum proteomics or novel imaging methods show promise, how should it be determined when the method is optimized to a degree that would justify a significant trial? Advocates for the introduction of new screening methods are generally correct when they predict significant technical advances over the next decade — does this argue for waiting, rather than undertaking, a trial based on soon-to-be suboptimal instrumentation?

Overall, these caveats are not meant to give the impression that progress in the development of screening methods is impossible. On the contrary, we believe that refinement of screening methods now in use, though challenging, is necessary if population-based screening is to fill its potential as a method of cancer control. If current screening methods were an unqualified success, questions as to its effectiveness would not result in such scientific and public controversy48.

Way forward

The public's view of screening is fairly straightforward: if a cancer is found early, then cure is more likely and therefore screening is a good thing. The first part of the clause is generally true, but the second part is less clear. Many recent publications do not make encouraging reading2, 3. Not only are there significant methodological problems with the interpretation of the results of trials, but also, as we have suggested, particular subgroups differ from the general population with respect to absolute cancer risk, mechanisms of carcinogenesis and the characteristics of their cancers. This means that clinical trials of screening protocols might be more informative if they stratify participants by biological and other markers of pre-existing risk. Results obtained in population-based 'all-risks' studies might not be translatable to individuals in specific risk groups and vice versa. A new look at the scientific rationale of screening studies is justified, and a renewed focus on primary prevention is warranted.

Advances in areas such as medical physics and serum proteomics are necessary but not sufficient to allow for significant improvements in the contribution of screening to cancer control. Also needed are biological and epidemiological information that will guide the application of new screening modalities in ways that will optimize reduction in cancer morbidity and mortality, as well as minimizing the number of unnecessary surgical interventions. In addition, attention must be given to the relationship between cancer screening and cancer prevention. It is imprecise to regard screening as a prevention strategy. Screening results in early detection, whereas successful prevention results in fewer invasive cancers and/or delayed carcinogenesis. Conventional screening must be combined with a treatment modality to be worthwhile.

Conventional models regard cancer as the target; in this context, the goal of screening is detection of macroscopic (but small) cancers, and subsequently undertaking curative surgery. But there might be more potential for screening than the detection of cancers that are clinically occult. Table 1 contrasts the goals of screening populations for incident cancers, screening for risk of cancer and screening to identify individuals who stand to benefit from particular risk-reduction strategies.

Table 1 | Prevention and screening: distinct but related

As pointed out by Michael Sxxx, it might be useful to consider the process of carcinogenesis as an alternative target49. In this context, screening would be directed towards identifying those individuals who have a particular propensity towards rapid rates of carcinogenesis on the basis of unusual exposure to carcinogens, a high-risk hormonal profile, a specific genetic predisposition or lifestyle factors. Having identified, through screening, those individuals with a particular mechanism that predisposes them to carcinogenesis, appropriate prevention interventions tailored to the at-risk individual might be offered.

We are far from achieving this kind of intervention in oncology, but a useful precedent is provided by everyday practice in cardiology. Screening to reduce mortality from myocardial infarction is not based on assessing the coronary arteries in asymptomatic individuals. Rather, it is commonplace to screen populations for particular modifiable risk factors that predispose to coronary artery disease (such as hypertension) and then offer appropriate prevention strategies. So, the screening is primarily designed to identify those with risk factors for heart disease, and to guide preventative interventions, rather than to identify those who already have occlusion of coronary arteries.

We speculate that the notion of screening populations to identify those who have the most to gain from specific risk-reduction strategies (as distinct from screening to detect early cancers) might become more common. A recent report50 presents clinical criteria that identify individuals who benefit most from the use of tamoxifen as a breast cancer prevention strategy. It is possible that genetic or hormonal measurements could improve these predictions. In a sense, detection of BRCA1 mutations represents a prototype of this kind of screening — a positive result indicates risk, rather than the presence of cancer, and has implications for prevention (in this case, prophylactic surgery) rather than for the treatment of cancers that are already present.

Notwithstanding the enormous challenges, efforts to use population screening to control cancer morbidity and mortality will continue as long as our success in treating advanced cancer remains limited. Some screening methods, such as Pap tests for cancer of the cervix, are clearly useful in reducing the number of individuals who come to medical attention with late-stage disease, and in reducing death rates51. Other screening methods, even those regarded in the older literature as obviously worthwhile — breast self-examination, for example — are being re-examined in the light of the notion that 'number of cancers detected' might not be the best end point. At present, the field is in flux: just as the usefulness of older screening techniques and the methods used to evaluate their benefits are being questioned, basic research is providing new imaging methods and serum markers, some of which deserve clinical evaluation.


Box1 | Colorectal cancer screening: financial implications

Annual cost estimate for universal colonoscopy at age 50

Number of individuals in the United States who will turn 50 during a 12-month period: 3,000,000 (Ref. 52)

Cost of screening colonoscopy per person: US$ 1,000 (Ref. 58)

Annual cost of universal population-based screening by colonoscopy for 50-year olds, excluding costs related to follow-up of abnormalities detected: US$ 3,000,000,000

Cost estimate in terms of dollars needed to save 1 year of life: US$ 10,000–25,000 per year of life saved53

Box 2 | Evaluation of screening techniques

The argument for 'all-cause' mortality versus 'disease-specific' mortality41:

Screening trials often measure 'disease-specific' mortality as the key outcome measure. This approach might be inadequate to fully reflect the impact of a screening test.

Deaths caused, both directly and indirectly, by screening tests themselves might be incorrectly attributed to other causes. These deaths caused are often not accounted for as disease-related mortality.

Deaths from cancer in the control group might be misattributed to other causes, or deaths in the screened group due to other causes are attributed to the screened cancer and are not accounted for as 'all cause'. This results in a 'sticky diagnosis' because the cancer diagnosis 'attracts' other diseases that are, in fact, the true cause of death.

This indicates that disease-specific mortality might be a suboptimal end point for screening trials and that a reduction in disease-specific mortality cannot be cited as strong evidence of efficacy when the all-cause mortality is the same or higher in the screened group.

When all-cause mortality is considered to be an end point, it might be increasingly difficult to identify the benefits of screening for late-onset cancers such as ovarian, prostate and colorectal cancer, especially in an ageing population.

Recent data on the use of hormone-replacement therapy as a way of reducing mortality rates in post-menopausal women, and the use of tamoxifen to prevent breast cancer, have illustrated the problems with using measures other than all-cause mortality in evaluating an intervention that is designed to lower cancer mortality.



Cancer.gov: breast cancer | colon cancer | ovarian cancer | prostate cancer | small-cell lung cancer

LocusLink: BRCA1 | BRCA2 | p53 | PSA


McGill Program in Cancer Genetics | McGill Program in Cancer Prevention



Baxter, N. Preventive health care, 2001 update: should women be routinely taught breast self-examination to screen for breast cancer? CMAJ 164, 1837-1846 (2001). | PubMed |


Olsen, O. & Gotzsche, P. C. Cochrane review on screening for breast cancer with mammography. Lancet 358, 1340-1342 (2001). | Article | PubMed |


Woolf, S. H. The accuracy and effectiveness of routine population screening with mammography, prostate-specific antigen, and prenatal ultrasound: a review of published scientific evidence. Int. J. Technol. Assess. Health Care 17, 275-304 (2001). | Article | PubMed |


Mulshine, J. L. Opinion: Screening for lung cancer: in pursuit of pre-metastatic disease. Nature Rev. Cancer 3, 65-73 (2003). | Article | PubMed |


Petricoin, E. F. et al. Use of proteomics patterns in serum to identify ovarian cancer. Lancet 359, 572-577 (2002). | Article | PubMed |


Nicholson, P. W. & Harland, S. J. Survival prospects after screen-detection of prostate cancer. BJU Int. 90, 686-693 (2002). | Article | PubMed |


Neugut, A. I., Jacobson, J. S. & Rella, V. A. Prevalence and incidence of colorectal adenomas and cancer in asymptomatic patients. Gastrointest. Endosc. Clin. N. Am. 7, 387-399 (1997). | PubMed |


Neal, D. E. & Donovan, J. L. Prostate cancer: to screen or not to screen. Lancet Oncol. 1, 17-24 (2000). | Article | PubMed |


Whitmore, W. F. Localized prostate cancer: management and detection issues. Lancet 343, 1263-1267 (1994). | PubMed |


Bok, R. A. & Small, E. J. Bloodborne biomolecular markers in prostate cancer development and progression. Nature Rev. Cancer 2, 918-926 (2002). | Article | PubMed |


Lofters, A., Juffs, H. G., Pond, G. R. & Tannock, I. F. 'PSA-itis': knowledge of serum prostate specific antigen and other causes of anxiety in men with metastatic prostate cancer. J. Urol. 168, 2516-2520 (2002). | PubMed |


Tannock, I. F. Eradication of a disease: how we cured symptomless prostate cancer. Lancet 359, 1341-1342 (2002). | Article | PubMed |


Pollak, M. Insulin-like growth factors and prostate cancer. Epidemiol. Rev. 23, 59-66 (2001). | PubMed |


Chan, J. M. et al. Insulin-like growth factor-I (IGF-I) and IGF binding protein-3 as predictors of advanced-stage prostate cancer. J. Natl Cancer Inst. 94, 1099-1106 (2002). | Article | PubMed |


Greller, L. D., Tobin, F. L. & Poste, G. Tumor heterogeneity and progression: conceptual foundations for modeling. Invasion Metastasis 16, 177-208 (1996). | PubMed |


Foulkes, W. D. et al. Primary node negative breast cancer in BRCA1 mutation carriers has a poor outcome. Ann. Oncol. 11, 307-313 (2000). | Article | PubMed |


Van de Vijver, M. J. et al. A gene-expresssion signature as a predictor of survival in breast cancer. N. Engl. J. Med. 347, 1999-2009 (2002). | Article | PubMed |


Chappuis, P. O., Nethercot, V. & Foulkes, W. D. Clinico-pathological characteristics of BRCA1- and BRCA2-related breast cancer. Sem. Surg. Oncol. 18, 287-295 (2000). | Article |


Porter, P. L. et al. Breast tumor characteristics as predictors of mammographic detection: comparison of interval- and screen-detected cancers. J Natl Cancer Inst 91, 2020-2028 (1999). | Article | PubMed |


Brekelmans, C. T. et al. Effectiveness of breast cancer surveillance in BRCA1/2 gene mutation carriers and women with high familial risk. J. Clin. Oncol. 19, 924-930 (2001). | PubMed |


Tabar, L., Duffy, S. W., Vitak, B., Chen, H.-H. & Prevost, T. C. The natural history of breast carcinoma: what have we learned from screening? Cancer 86, 449-462 (1999). | Article | PubMed |


Goffin, J., Chappuis, P. O., Wong, N. & Foulkes, W. D. Re: Magnetic resonance imaging and mammography in women with a hereditary risk of breast cancer. J. Natl Cancer Inst. 93, 1754-1755 (2001). | Article | PubMed |


Stoutjesdijk, M. J. et al. Magnetic resonance imaging and mammography in women with a hereditary risk of breast cancer. J. Natl Cancer Inst. 93, 1095-1102 (2001). | Article | PubMed |


Warner, E. et al. Comparison of breast magnetic resonance imaging, mammography, and ultrasound for surveillance of women at high risk for hereditary breast cancer. J. Clin. Oncol. 19, 3524-3531 (2001). | PubMed |


Meijers-Heijboer, H. et al. Breast cancer after prophylactic bilateral mastectomy in women with a BRCA1 or BRCA2 mutation. N. Engl. J. Med. 345, 159-164 (2001). | Article | PubMed |


Julian-Reynier, C. M. et al. Women's attitudes toward preventive strategies for hereditary breast or ovarian carcinoma differ from one country to another: differences among English, French, and Canadian women. Cancer 92, 959-968 (2001). | Article | PubMed |


Boyd, N., Lockwood, G., Byng, J., Tritchler, D. & Yaffe, M. Mammographic densities and breast cancer risk. Cancer Epidemiol Biomarkers Prev 7, 1133-1144 (1998). | PubMed |


Laken, S. J. et al. Familial colorectal cancer in Ashkenazim due to a hypermutable tract in APC. Nature Genet. 17, 79-83 (1997). | PubMed |


Ma, J. et al. Prospective study of colorectal cancer risk in men and plasma levels of insulin-like growth factor (IGF)-I and IGF-binding protein-3. J. Natl Cancer Inst. 91, 620-625 (1999). | Article | PubMed |


Giardiello, F. M., Brensinger, J. D. & Petersen, G. M. AGA technical review on hereditary colorectal cancer and genetic testing. Gastroenterology 121, 198-213 (2001). | PubMed |


Sidransky, D. Emerging molecular markers of cancer. Nature Rev. Cancer 2, 210-219 (2002). | Article | PubMed |


Anderson, W. F. et al. Colorectal cancer screening for persons at average risk. J. Natl Cancer Inst. 94, 1126-1133 (2002). | Article | PubMed |


Frazier, A. L., Colditz, G. A., Fuchs, C. A. & Kuntz, K. M. Cost-effectiveness of screening for colorectal cancer in the general population. JAMA 284, 1954-1961 (2000). | Article | PubMed |


Budenholzer, B. Cost-effectiveness of colorectal cancer screening. JAMA 285, 407 (2001). | Article | PubMed |


Lieberman, D. A. & Weiss, D. G. One-time screening for colorectal cancer with combined fecal occult-blood testing and examination of the distal colon. N. Engl. J. Med. 345, 555-560 (2001). | Article | PubMed |


Pineau, B. C. et al. Validation of virtual colonoscopy in the detection of colorectal polyps and masses: rationale for proper study design. Int. J. Gastrointest. Cancer 30, 133-140 (2001). | PubMed |


Jass, J. R., Whitehall, V. L., Young, J. & Leggett, B. A. Emerging concepts in colorectal neoplasia. Gastroenterology 123, 862-876 (2002). | Article | PubMed |


Atkin, W. S. et al. Design of a multicentre randomised trial to evaluate flexible sigmoidoscopy in colorectal cancer screening. J. Med. Screen. 8, 137-144 (2001). | Article | PubMed |


Schoenfeld, P. et al. Accuracy of polyp detection by gastroenterologists and nurse endoscopists during flexible sigmoidoscopy: a randomized trial. Gastroenterology 117, 312-318 (1999). | PubMed |


Kopans, D. B. The most recent breast cancer screening controversy about whether mammographic screening benefits women at any age: nonsense and nonscience. AJR Am J Roentgenol 180, 21-26 (2003). | PubMed |


Black, W. C., Haggstrom, D. A. & Welch, G. All-cause mortality in randomized trials of cancer screening. J. Natl Cancer Inst. 94, 167-173 (2002). | Article | PubMed |


Alexander, F. E. et al. 14 years of follow-up from the Edinburgh randomized trial of breast-cancer screening. Lancet 353, 1903-1908 (1999). | Article | PubMed |


Church, T. R., Ederer, F. & Mandel, J. S. Re: All-cause mortality in randomized trials of cancer screening. J. Natl Cancer Inst. 94, 861 (2002). | Article | PubMed |


Gail, M. H. & Katki, H. A. Re: All-cause mortality in randomized trials of cancer screening. J. Natl Cancer Inst. 94, 862 (2002). | Article | PubMed |


Black, W. C., Haggstrom, D. A. & Welch, H. G. Re: All-cause mortality in randomized trials of cancer screening. J. Natl Cancer Inst. 94, 865-866 (2002). | Article |


Rogers, L. F. Screening mammography: target of opportunity for the media. AJR Am J Roentgenol 180, 1 (2003). | PubMed |


Mahadevia, P. J. et al. Lung cancer screening with helical computed tomography in older adult smokers: a decision and cost-effectiveness analysis. JAMA 289, 313-322 (2003). | Article | PubMed |


Kolata, G. Breast cancer: mammography finds more tumors. Then the debate begins. New York Times (9 Apr 2002).


Sxxx, M. D. & Suh, N. Chemoprevention: an essential approach to controlling cancer. Nature Rev. Cancer 2, 537-543 (2002). | Article | PubMed |


Veronesi, U. et al. Italian randomized trial among women with hysterectomy: tamoxifen and hormone-dependent breast cancer in high-risk women. J. Natl Cancer Inst. 95, 160-165 (2003). | Article | PubMed |


Franco, E. L., Duarte-Franco, E. & Ferenczy, A. Cervical cancer: epidemiology, prevention and the role of HPV infection. CMAJ 164, 1017-1025 (2001). | PubMed |


US Bureau of Census. An Aging World: 2001 [online], (cited 24 Feb 2003), (2001).


Pignone, M., Saha, S., Hoerger, T. & Mandelblatt, J. Cost-effectiveness analyses of colorectal cancer screening: a systematic review for the US Preventive Services Task Force. Ann. Intern. Med. 137, 96-104 (2002). | PubMed |


Greenwald, P. et al. Estimated effect of breast self-examination and routine physician examinations on breast cancer mortality. N. Engl. J. Med. 299, 271-273 (1978). | PubMed |


Huguley, C. M. & Brown, R. L. The value of breast self-examination. Cancer 47, 989-995 (1981). | PubMed |


Newcomb, P. A. et al. Breast self-examination in relation to the occurrence of advanced breast cancer. J. Natl Cancer Inst. 83, 260-265 (1991). | PubMed |


Thomas, D. B. et al. Randomized trial of breast examination in Shanghai: final results. J. Natl Cancer Inst. 94, 1445-1457 (2002). | Article | PubMed |


Smith, R. A., Cokkinides, V. & Eyre, H. J. American Cancer Society guidelines for the early detection of cancer, 2003. CA Cancer J. Clin. 53, 27-43 (2003). | PubMed |


The authors would like to thank R. Narasimhadevara for her contribution to this review. W.D.F. is a Principal Investigator of the Canadian Genetic Diseases Network. M.N.P. holds the Alexander Goldfarb Research Chair in Medical Oncology at McGill University.

To benefit from the full content of Nature Reviews Cancer every month, simply take out a subscription - click here for details.

back to top


© 2003 Nature Publishing Group

Link to comment
Share on other sites


WHEW!!! Very convoluted issue with no easy black and white answers!! :?

Sooooo, yes, yes, I would encourage you to go back to school and pursue a new career and help solve these confounding issues (8)).

As Katie says, we are lucky to have you with us. :D

Best to you and your precious family.

Link to comment
Share on other sites

Hey John,

Well, I third what Katie and Tiny have said about being "LUCKY TO HAVE YOU ON THIS BOARD"! I think I may have thanked you before, but It doesn't matter how many times I may have thanked you, More is BETTER!!!

Your a Bless and such a wonderful asset to all of us that need some education on this subject. THANK YOU!!

I'm with you, I sure would like to see this cancer (crap) long gone in a few years!!! You must have been a WONDERFUL help to your mom!! I can tell you were a wonderful son!! I'm sure she was and still is very proud of you.

Warm and Gentle Hugs,

Link to comment
Share on other sites

I thank everyone here also. I have learned a lot from everyone here. Not only about cancer, but I learn that people do care and are compassionate.

I am lucky to have a mom who passed on her smarts. There were probably a couple hundred people at her funeral. She was a social worker who did real estate on the side and then full-time. One of her real estate friends said she was always honest and caring. One of her clients closed right around Thanksgiving. On Thanksgiving day, not only did she cook a meal for them. She cooked a completly separate meal and cook ours also.

Link to comment
Share on other sites

Hi John,

I'm so sorry to hear about your Mom. :( She sounds like an amazing, special person, like yourself. I too have been reading your posts and am thankful for your information. I'm glad that you are with us, your Mom, wife and daughters are lucky to have such a dedicated family man on thier side. Thanks for sharing with us and I think you would be an excellent additon to the medical field.

Best Wishes,


Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

  • Create New...

Important Information

By using this site, you agree to our Terms of Use.