Screening asymptomatic patients for cancer is based on a simple premise: cancers that are detected early before they produce symptoms should be easier to cure than those that progress to become clinically detectable. Yet, this seemingly simple idea is fraught with problems. There are several factors that contribute to the difficulty of proving the value of cancer screening. First of all, any test will have false negatives, in which case a patient’s cancer goes undetected. Secondly, tests for cancer are often falsely positive because cancer is relatively rare in an asymptomatic population, meaning that many patients with positive test results will not have cancer. It is unwise, however, to ignore a positive test, so additional testing is warranted. This additional testing may include radiation exposure, surgery, and other testing that may place the asymptomatic patient at risk for harm, not to mention the psychological harm of telling someone, “you might have cancer.”
The most unexpected finding that has emerged from screening studies for breast, lung, and prostate cancer is that an imprecisely defined fraction of the cancers that are detected by screening do not progress to cause symptomatic disease or to kill the patient. For some as yet unknown reason, some tumors simply stop growing. The dilemma is how do we tell the difference between cancers that can kill you and those that cannot. To meaningfully assess the value of cancer screening, the most relevant and clearest measure is an improvement in overall survival. Yet measuring that value may take many decades, so surrogate end points are presumed but not established. Consequently, screening tests that have adverse side effects are widely recommended, they may be expensive, and have not been shown to improve overall survival. Moreover, some, if not many, people may be harmed.
Take, for example, screening for breast cancer. Some women have extremely dense breast tissue. This makes mammography more difficult and more likely to miss a cancer. Because of this, women with very dense breasts have an increased risk of breast cancer. Ultrasound has not proved to be helpful in these situations, so some practitioners have resorted to using MRI scanning to enhance screening. MRI screening is more sensitive for breast cancer screening than mammography, but far less specific. Last November, The New England Journal of Medicine published a study comparing breast cancer screening using mammography alone vs. mammography plus MRI in women with very dense breasts. Of the women who underwent MRI screening, 10% were recalled for abnormal scans, and of those, in the women who had biopsies based on their MRI findings, only 26% had breast cancer, making the false positive rate for MRI screening 74%! However, almost all of the cancers found by MRI screening were stage 0 or 1, lymph node negative, and ductal carcinoma in situ, which would have been undetected by mammography alone.
The ultimate test of value of MRI screening in women with very dense breasts is overall survival, and we won’t have that answer for a very long time, so the dilemma remains. Women with dense breasts and negative mammography who undergo MRI screening have a very low risk of having breast cancer (16.5 per 1000 in this study) and a very high rate of false positive MRI results (74%), but they will have very treatable cancers discovered and cured, which would not have been found with mammography alone.
By Peter Galvin, MDBLOG COMMENTS POWERED BY DISQUS