Thursday, November 29, 2012

Screening, Cancer and the Right Questions

In a recent piece in the NCI Cancer Bulletin they discuss what appears to be the ineffectiveness of cancer screening. Now upon closer examination there are clearly many epistemological issues. The author commences her presentation by stating:


Much of the confusion surrounding the benefits of screening comes from interpreting the statistics that are often used to describe the results of screening studies. An improvement in survival—how long a person lives after a cancer diagnosis—among people who have undergone a cancer screening test is often taken to imply that the test saves lives. 

Indeed the interpretation is critical but even more importantly is the question for which the statistics reflect an answer. Let me address this in two ways; first by examining her diagrams and second by examining some cancer specifics.

First, her examples. She provides two cases.

The first case is a demonstration that by diagnosing the cancer early one obtains an improved 5 year survival but no change in time of death. Consider the case shown below where she assumes the following. A patient has lung cancer and it commences at some unspecified time and the patient dies at 70. Now assume that by some unspecified but reliable test we can identify it at say 60 years of age and act in some manner but that despite the actions and early diagnosis the patient dies at 70. She contends we see a 100% 5 year survival. True, but. Then assume we never did anything until some massive event occurred at 67 and we then diagnosis the cancer and take some unspecified actions. Then we would have a 0% 5 year survival.

Now what is wrong here? First, lung cancer is generally very aggressive so that any chance of even a three year survival is low. Indeed if there were a 10 year survival then that itself would have been a success. In fact if diagnosed at 60, by whatever means, and treated by whatever means, more than likely if untreated the patient would have died of lung cancer much earlier.

The fault here is that there is an assumption that the two cases are pari passu, equal in all ways but the early diagnosis. In fact I would argue we are comparing apples and oranges.

Now for the second case. Here she assumes that we have a progressive and a non-progressive form of some cancer. The usual victim chose especially by women is prostate cancer. So let us assume that. We then look at two cases. First with no screening. Somehow there is no treatment and the result is 400 dead and 600 alive. The protocol of diagnosis and treatment is left unstated but one can assume that the patient presents say with an inability to urinate or sever bone pain. Namely with no screening the patient waits until the symptoms are obvious and compelling.

Now the second case she uses to demonstrate bias is that if we screen, and that for every 3000 people screened who have the disease as a result of screening, 2000 have an indolent form, which we cannot ascertain at time of diagnosis, and 1000 still have the progressive form. Now she argues that using this erroneous method we have a false survival rate of 80% rather than the “true” rate of 40%.


Let me quote from the Bulletin:

"I had a brilliant oncologist say to me, 'Don, you have to understand: 20 years ago, before mammography, I'd see a patient with breast cancer, and 5 years later she was dead. Now, I see breast cancer patients, and 15 years later they're still coming back, they haven't recurred; it's obvious that screening has done wonders,'" he recounted. "And I had to say no—that biases could completely explain the difference between the two [groups of patients]."

There is a problem here. If the oncologist still sees patients after 15 years then somehow we have managed to accomplish something. After all in prior days the patient was seen at the point of massive tumor load and significant metastasis. We again are comparing apples to oranges.

Now let us step aside and consider three cancers and the screening issues.

1. Melanoma: Anyone familiar with melanoma knows that screening does work. If we waited for the patients to come with a bleeding lesion then most likely we would have a greater mortality rate. There is the question of increased incidence and its relationship to increased screening, but melanoma is one of those cancers that just have no indolent form. It is plain and simple a killer and one can deduce that the increased incidence is due to increased exogenous factors such as sun exposure, just look at Australia. Now if we were to screen, especially those of increased risk (family history, number nevi, sun exposure etc) using dermoscopy, and remove promptly and effectively all suspect lesions, then it can be argued that we can reduce the mortality substantially. Not just 5 year survival, but actual mortality due to that specific melanoma.

2. Prostate: This is often the hot potato of screening. The reason is, the opposition argues, that if biopsy is required, that it is uncomfortable and there may be some limited morbidity.

The author states:

The extreme example of length bias is overdiagnosis, where a slow-growing cancer found by screening never would have caused harm or required treatment during a patient's lifetime. Because of overdiagnosis, the number of cancers found at an earlier stage is also an inaccurate measure of whether a screening test can save lives. (See the graphic on the left for further explanation.)

The effects of overdiagnosis are usually not as extreme in real life as in the worst-case scenario shown in the graphic; many cancers detected by screening tests do need to be treated. But some do not. For example, recent studies have estimated that 15 to 25 percent of screen-detected breast cancers and 20 to 70 percent of screen-detected prostate cancers are overdiagnosed.

The problem, however, is that we do not know the slow growing from the virulent forms. We do not have genetic markers that tell is that a Gleason 6 will become metastatic in 6 weeks while the Gleason 8 is going nowhere. This is the fatal flaw in her second analysis. It assumes we know indolent from aggressive, we just do not, until after the fact. Thus we treat all as aggressive, which albeit costly may be in the end efficacious to saving lives. Yet again we let the decision be an informed patient decision.

3. Colon: The question here is; what of colonoscopies? Clearly here we have often slowly growing tumors, starting with adenomas and slowly developing into invasive cancers. Removal of adenomas is an almost certain guarantee of non-progression to a cancer. The data seem to demonstrate a saving of lives. The procedure is invasive, and at times costly, with small but existing morbidity factors. Should we ban this also?

Now the issue is; what should be the proper questions? Namely, should we continue to do 5 year survival? More importantly should we even start down the road of denying efficacy if we do not have enough information? Again consider prostate cancer. We do not know how to tell aggressive from indolent, so should we just stop PSA testing? PSA testing helps for the aggressive group, albeit a small but unidentifiable sample.

I suggest stepping back and looking at the questions. The author in her presentation I believe makes a strong argument for it, not by what she understood and said, but what was clearly not understood and left unsaid.