Scans are becoming more and more appealing as a way to view the future. For those who find no disease during various scans, the window to the future seems rosy. For those with an abnormality, the view is starker. Readily available scanning now includes electron-beam CT, total-body CT scans, chest-only CT scans, virtual-colonoscopy scans, MRI coronary artery scans, PET scans of the brain or total body, and combinations of these. Sometimes scans promise more than can be delivered and reassurance may be false. The opposite also occurs where abnormal findings on scans lead to opportune intervention and to good outcomes.

Certain things about scans will only become known as independent studies accumulate enough data to answer some key questions. Do scans make a difference in the quality of length of life? Do scans carry risks (e.g. radiation exposure) that mitigate the number of times a scan can be done? If so, then what are the optimum number and the optimum timing for different scans? Who should be scanned and by what methods? When should the first scan take place? If a first scan is valuable, when should subsequent scan(s) be done?

Who should be scanned? The answer is that profiling of patients is advisable because some patients have higher risk for certain conditions than others. For some scans the answer is clear. For example, it is clear that young patients with no history of heart disease and no risks of heart disease will not benefit from ultra-fast electron-beam scanning. Nearly all such patients will have a negative scan and thus no new knowledge will have been gained. On the other hand, a middle-aged male with a strong family history of coronary disease, high cholesterol, and unhealthy life-style that may include smoking, obesity, hypertension, or diabetes may learn about a high calcium score in his coronary arteries and can be told reliably that the possibility of a future heart problem is very high unless intervention occurs to change the future. If such a high-risk patient has a negative calcium score, then his risk of developing a coronary obstruction is minimal. However, the risk is not totally negligible because some obstructions are purely cholesterol, and it is now known that heart attacks often occur in cholesterol plaques without calcification or significant narrowing. Indeed, myocardial infarctions (heart attacks) are thought to occur when a susceptible plaque has ruptured, inducing an acute clotting reaction leading to an arterial obstruction. The resolution on an angiogram of the coronary arteries is the diagnostic standard for viewing obstructions. The resolution of non-invasive MRI angiography has improved to include the secondary branches of the coronaries but is not yet widely available. Moreover, the heart is a moving target and acquiring high-resolution images is still difficult. A comparison of MRI angiography with ultra-fast electron beam calcium scanning has not been done to see if decisions are more opportune with one screening method than another. Whichever scan is better will prevail but they may provide equally useful information.

When should the first scan be obtained? For most scans there is no good answer to this question. The problem in our current state of knowledge is that we have not identified when to trigger a scan except in certain specific patients. For example, even a scan as old as mammography still has controversy about whether to scan women in their 40s.

We know that doing routine CT or MRI of the brain on asymptomatic people has a very low yield. The same is true of scanning all patients with headaches, who should have a higher prevalence of disease. Even in patients known to have an early breast cancer who sometimes get brain metastases, routine brain scans are still very low yield and are usually not ordered. The same is true of scans of the liver in these patients. However, patients with known lung cancers have higher risks and are usually scanned because the chance of a small brain metastasis is large enough to warrant obtaining the scan. However, in patients who do not have known lung cancer but who have a known risk of lung cancer, e.g. in very heavy smokers, a random brain scan is still not advised because the yield is low. In fact, it is only recently that such patients were shown to benefit from CT scanning of the lungs to find small nodules, which might be cancers or potential cancers. The issue is not yet settled how often to rescan such patients in order to watch all the nodules. It makes sense that the more nodules present the higher the risk of one being a cancer. But even the number or the sizes of the nodules only raises the question of how soon and how frequently to rescan i.e. should these be rescanned in 3 months, 6 months, or longer. Usually, 6 months is the recommendation in patients with one or more candidate nodules of a small size. Since some heavy smokers have several small nodules, it may not be possible to remove them all so another way of monitoring must be utilized. In patients with a solitary nodule it is common practice to use PET scanning to identify the nodules most likely to be cancer; i.e. the hot nodules. PET scanning has a much lower false positive rate; nevertheless, it is known that some nodules showing no activity on PET scans are still cancer. This setting is one where the PET scan is chosen to elucidate another finding but not as a first scan. Since CT and MRI routine scanning of the brain has been a very low yield, it is correct to surmise that routine PET scanning will also be of low yield.

PET scanning has a role in screening for Alzheimer's disease. In patients suspected of having Alzheimer's disease, the PET scan can confirm the diagnosis of Alzheimer's without requiring invasive procedures or waiting a long time for the disease to be fully manifest. However, in patients with no symptoms and no family history of Alzheimer's disease, is there a role for routine screening? Alzheimer's disease has no current effective therapy to prevent its onset. A positive test does not fully correlate with disease severity or provide a prognosis. If no family history exists, then there may not be any reason to perform such a screen. If a patient has cognitive or memory deficits, then PET scans may provide a definitive explanation.

Additionally, PET scans are very good at finding hidden cancers, especially in the lungs, when a nodule has already been found. However, definite data does not yet exist to support utilizing PET scanning of the chest or other areas as a routine screen when no other screening studies have been done. Even if a scan is negative, we know that most lung cancers have a fast rate of growth, and a follow-up scan would be needed to detect a tumor smaller than the current limits of resolution. Therefore, how long would a negative screening PET scan give reassurance to the patient without providing a false sense of security?

The same question can be asked about screening CT scans of the chest, which are advocated by some for people profiled with high risk for lung cancer: e.g., strong smoking history, industrial exposures, or family history.

For abdominal scanning there are some findings that can be life-saving. But the finding of an unexpected cancer is serendipity itself. It is just as likely to find such a cancer too late as it is to find it opportunely. It is also as likely to have performed the scanning too early before it is visible at all. Some cancers are especially notorious in this regard. For example, some radiologists say that even a small 1-2 cm pancreatic cancer is already a futile problem. For renal or kidney cancer, it is clear that getting a head-start by finding red blood cells in routine urinalysis alerts us about patients at risk and that scanning this select population may lead to more opportune discovery. In ovarian cancer, however, blood testing of CA125 has been disappointing as a screening maneuver, even when coupled with routine pelvic ultrasounds. This is another form of cancer which may not announce its presence until quite late. Unless there is a strong family history or another risk factor such as breast cancer, it is not yet clear that scanning is productive. For prostate cancer there is a sensitive way of profiling patients: the PSA test. If the PSA is elevated then scanning of several types is targeted directly on the prostate. Even in cases of proven prostate cancers scans such as MRI, CT, and even PET do not detect the known cancer, let alone possible metastases. Such scans help in staging but recent information suggests multiple types of negative scans often do not find metastases in the pelvic nodes when surgery is later done. We hope for a future battery of blood tests that detect risks so that targeted scans can be selected to find the source of the risk. In the meantime the role of the proliferating scans has not been satisfactorily determined.

Another popular scan is CT colonography, also called virtual colonoscopy. It is popular because it is less invasive that actual colonoscopy in the sense that sedation is not needed and insertion of an instrument, which may rarely cause perforation, is not required. Many advertising web sites show perfect pictures of the lumen of the colon and show findings of tumors and small polyps. What is not emphasized is that the lower limit of resolution is in the 5-7 mm range. Of course, the larger the polyp, the higher the associated risk of colon cancer. If a polyp is 20 mm the risk is especially high. Once again the issue of timing is crucial. If a polyp is 45 mm and is not identifiable because it is below the resolution of the CT scan, when should the next scan be performed to find it? If a small polyp is found at 6 mm, would it be necessary to remove it now or would it be acceptable to wait, rescanning in 3, 6, or 12 months to see if it changes. If it did not change would that exonerate the polyp as a potential cancer? Probably not, but the point is made that it is not self-evident what future actions should be taken. If no polyps are present on CT scan, when should the next scan be done to be sure that an invisible 4-5 mm polyp has not begun to grow wildly, i.e. should it be yearly, biannually, etc.? The same issues have been discussed regarding colonoscopy itself. If there have been no polyps on colonoscopy, can one safely wait 5 years or even 10 years to look again? Most specialists subscribe to 5-10 years, but discussion is ongoing. For the time being, an unequivocal recommendation for CT colonography cannot be given. Clearly, if polyps are identified then colonoscopy itself is required to remove the polyps. However, I have many cases where polyps were identified on old-fashioned barium enemas which could not be found on colonoscopy done to remove the so-called "known" polyps. In such cases of false-positive barium enema or CT, the colonoscopy would have been normal if it had been the method of screening. Does a negative scan change the predictive value? Cost, invasive instruments, and anesthesia also need to be factored into recommending one procedure over another.