Cancer history has progressed from indifference or even downright marginalization (of cancer patients in the past), to surgical removal of the tumors (sometimes without even knowing what to remove), to drug therapy (sometimes the therapy is worse than the cancer), to prevention, to screening and early detection. All have been well received as they became popular and evidence based and all have been suggested at the time to be the answer to curing cancer for good, but NONE have delivered that promise. Let me first say that perhaps the call to eliminate cancer once and for all has been too hasty (how can we cure things we don’t understand). The second point is that even with screening and trying to detect cancer earlier is fought with lots of problems. That is what I will spend some time addressing next.
Screening or attempting to use medical tests of one kind or another to assist in detecting cancer earlier than it is usually detected (e.g. pateint develops symptoms and goes to see his/her doctor) has been one of the most recent advances in the push to cure cancer. Unfortunately, it has turned out that the way we do screening or the way it has been done in the past is frequently not as informative/useful as it could be. Screening is designed to be used int he general population to help detect cancer earlier so that treatments can be done sooner so that patients can live longer and reduce mortality (death rates).
One important thing to consider first is that NO screening test is going to be 100% accurate. Some will pick up cancers that don’t exist and this is known as false positives. These need to be kept at a minimum as it would funnel patients who are healthy into anti-cancer treatments. On the other side of the coin, some test are not as sensitive and fail to pick up some people with true cancers and call them as non cancer. This is known as false negatives. These types of tests tend to put fewer patients into the positive pile and thus can results in cancer patients failing to get proper treatment and only get treatment when they have symptoms (too late in some cases). As you can imagine, both false negatives and false positives are bad, but in terms of mortality the latter might be more severe. So, the conclusion with this part is that no matter what kind of screening test you have, no test is perfect and the only way to determine how good it is…is to test this screening tool on a very large population of patients.
This then leads me to the next part. How are these trials done. When X-ray was discovered as a potentially effective tool for breast cancer screening, some US doctors started performing clinical trials with a large number of women patients in the 1940’s and 1950’s. Women were placed into a trial and the women who got X-ray screening were told about the trial and the women who were put into the control group (no X-ray but wait to detect the cancer naturally if they get it) were not told anything. So, you have one X-ray screening arm and a control arm. Sounds reasonable, no? Unfortunately, as caring physicians they placed women who were at higher risk of getting breast cancer (prior cancer that was removed or those with a family history) into the X-ray screening trial. As nurses and or doctors in the cancer clinics they of course felt obligated to do this as they did not feel that it would be fair to put them into the control (no X-ray) arm. Although, this is a very understandable reaction, this type of ‘caring’ is actually very bad for the statistical outcome of the trial. As you can imagine, the X-ray arm is going to be very different (the composition of women) that they control arm. The women getting screened will be those who are likely to have more serious outcomes as they have more numbers of high risk individuals. You can not perform a randomized trial if you weigh one group differently with the other group. One must conduct trials without ANY prejudice. This is very hard to do.
Huge breast cancer trials where then conducted in Canada with large numbers of women to ask if screening using x-ray helped to reduce mortality. This time they had a ledger where every other women was randomized into the two different groups. This time for many reasons women who were more prone to getting cancer were placed into the control arm…perhaps to overcompensate . This also made the data that came out hard to interpret. In both examples, it turned out that the scientific/clinical community refused to agree on the outcome of the trials as the way women were picked was not random.
Only later when the Scandinavians held massive trials with 100% complete randomization of women into the two arms of the trial did the results of the breast screening trial turn out valid and agreeable by all. In one medium-sized city in Sweden all from one clinic, thousands of women were placed into this screening clinical trials. The results showed that there was a very tiny benefit in mortality for women who had been screened versus those who did not get screening for breast cancer. This was very disappointing for many who thought that screening would help save millions of lives. However, when statisticians looked at the data in terms of age….a huge difference was noted. Women above 50 benefited from screening while those below 50 did not. This was later statistically validated in many centers around the world.
So, although this story focuses on breast cancer screening, it also holds true for prostate screening. One, tests are not perfect by any means and they must be very sensitive before they can be adopted. Two, clinical trials must be carried out before clinicians can definitively trust any new screening method. Three, the clinical trials must be done correctly, balanced, and without bias.
Thanks and again, I hope this has been useful to you. For more information please see Cancer Made Simple!