Saturday, June 16, 2012

Hold that mammogram - it may not benefit all

Jun 16, 2012
 
By Andy Ho

THE commonest cancer in women here is breast cancer, which also has the highest kill rate.

There are, on average, 1,556 new cases of breast cancer and 370 deaths from breast cancer annually.

The 2010 Health Ministry guidelines urge all women between the ages of 50 and 69 to have a mammogram once every two years. The 2009 United States guidelines likewise recommend the same for women aged between 50 and 74.

Ardent worldwide pink ribbon advocacy has made it sacrilegious to say this, but these guidelines are overkill. Here's why.

In April, the National Cancer Centre Singapore (NCCS) published a study begun in 1994 that had followed 28,000 women aged 50 to 64 for a decade. It reported that women of all the three major races here with a close family history of breast cancer were twice more likely to get it than a Western female. But those with no such family history were at only half the risk of a Western female.

That is because mammography leads to 'overdiagnosis' in the lower-risk group who thus 'would benefit less from it', the NCCS said. Those in the high-risk group, however, should consult their doctors on its need.

The radiologist community immediately slammed the NCCS study as 'based solely on a single round of screening mammograms as opposed to the more robust Gail (US) model, which derived data from five annual rounds of screening mammograms across 29 centres in the US'.

But if the NCCS study has a reasonable basis, it is possible that the US model could lead to serious overdiagnosis rates.

In fact, both the NCCS and radiologists may have omitted to stress an even more important consideration in the debate.

When deciding on a screening method's utility, one must distinguish between death rates caused by breast cancer per se and death rates in breast cancer patients from all causes.

This is because women who have regular mammograms are likely to have the early, non-invasive form of breast cancer called ductal carcinoma in situ (DCIS) picked up. Some experts call it a very early breast cancer but others call it a precancer.

Left untreated, DCIS generally doesn't kill. Doctors cannot predict if it will get invasive. Hence, once diagnosed, it is invariably treated, for no woman who knows there is a cancer in her breast will do nothing about it. Because virtually all women are treated aggressively, no one knows what would happen if women opted to wait and see.

Doctors know so little about DCIS because it was not diagnosed before widespread mammography. Now its incidence has grown sevenfold compared to the 1970s, when doctors began using mammograms. Hence the 'overdiagnosis' that NCCS noted which leads 'to increased risks of unnecessary treatment or over-treatment' - that is, regular mammograms tend to lead to more surgery, chemotherapy and/or radiation. But these may lead to side effects and unanticipated complications that sometimes kill the patient.

With widespread mammography, the number of cases caught at more advanced stages hasn't dropped much at all, which a good screening test picking out early cases before they progressed should lead to. Instead, screening is picking up cases that need not be treated, namely DCIS.

One can infer that DCIS is very likely being over-treated by looking at the death rates in breast cancer patients from all causes, not just from breast cancer. This way, deaths caused by these side effects of treatment are included. And only this can tell if mammography is worthwhile. In short, count all deaths in breast cancer patients, for what matters is whether the patient lives, not what she dies from.

It was only over the past decade or so that researchers have learnt to distinguish the two. But in 1963, when the first trial of 60,000 women began in the US, this crucial difference was not recognised yet, which led scientists to conclude that, overall, mammograms saved lives.

Four other large trials at the time, also unaware about this point, also found support for mammography. This oversight has been blissfully perpetuated to this day by advocacy groups and their experts (who do know better).

In a massive 2001 meta-analysis involving 470,000 women, 'breast cancer mortality rate' was found to be 15 per cent lower in patients who had regular mammography. But their 'all-cause mortality rate' was no different from patients who had not.

This suggested that mammography could have led to enough fatalities from treatment (of DCIS) that overwhelmed the benefits of early detection. The biostatistician who led this study has produced other studies to buttress this finding.

The latest study (done by other scientists) showing this effect was published in October last year in the Archives of Internal Medicine. However, this crucial fact seldom trickles down into the media or the advocacy universe.

Earlier and regular screening disproportionately picks up DCIS that likely does not need to be treated but is anyway. One post-mortem study of women dying between the ages of 40 and 50 from all causes other than known breast cancer found that 40 per cent of them had DCIS that was detected microscopically only during the autopsy.

So these women died without ever knowing they had DCIS. If they had had a mammogram in life, it would have been picked up and they would have had unnecessary treatment that might very well have killed some of them.

So mammography is useful only if used hand-in-glove with an accurate predictor of risk. The NCCS study showed that family history best predicts a woman's risk: A female here with a mother, sister or daughter who has had breast cancer is at four times the risk compared to one with no such family history. For this group, regular mammography may be justified. For the latter - which would include most women in Singapore - it would seem unnecessary.

The sensible woman would of course do regular breast self-examination and see a doctor if any symptoms appear, such as discovering a breast lump or finding discharge from the nipple. For women with no family history and no symptoms, regular mammograms may not be particularly useful.

andyho@sph.com.sg

No comments: