/ 27 November 2013

Comment: Mammograms still the gold standard of breast screening

Breast cancer expert, Justus Apffelstaedt, explains the complexities around the issue of screening for this disease.

Screening mammography was the subject of a few articles over breast cancer month and has raised considerable interest. Many of the issues, however, require some clarification. Here is an introduction to the complexities of mammographic screening issues. 

Screening for any disease can only be considered if the disease is serious, relatively common, a reliable screening test is available and can be administered to change the course of the disease for the better. 

Let's look at these prerequisites. 
Nobody will dispute that breast cancer is a serious disease. 

Is it, however, common enough to warrant screening? 

At the commonly quoted figure of a lifetime risk of breast cancer for South Africa of 1 in 28, screening does not make sense. These figures, however, are outdated and cancer diagnosis rates from two screening series in the Western Cape in both the public and the private sector reported cancer diagnosis rates similar to those in countries where the life-time risk is 1 in 8 – 10. At this level screening does make good sense. It remains to be seen, however, whether these numbers also apply to other provinces. 

There is no argument that mammography can discover breast cancer early and change the course of the disease to an extent that the vast majority of women with screen-detected breast cancers will live out their normal life-span. 

Screening programs are broadly described as being either "community" or "organised". In "community" screening, women go for screening freely, there is no central quality control, procedures are poorly defined and few if any results are reported. Conversely, in "organised" screening, which is a public health initiative, women are invited at regular intervals to come for screening on a population basis; there are stringent, central quality controls and results are often publicly reported. Much of what is known about the impact of mammographic screening stems from "organised" screening. 

In South Africa, no "organised" screening exists; the public sector does not support mammographic screening. In the private sector where "community" screening is practiced, figures from private health funders show that about 30% of eligible women had a screening mammogram within the last year.  

A screening program can only be effective if as many eligible women as possible participate. In the best organised screening programs, the participation rate is between 60% and 80%. 

One "of the best researched medical interventions"
Mammographic screening is one of the best researched medical interventions and quality benchmarks are provided by "organised" programs. The recall rate and the biopsy rate relate to harms generated by screening: Psychological stress is caused by calling the woman back for further investigations and the wait for results; physical harm to the patient and finally financial costs to the system are caused by unnecessary biopsies. As no screening test is 100% accurate, a balance must be struck between too many "false positives" and missing cancers. This sweet spot is characterised by a recall rate of about 5 – 10% in the first round of screening and 4 – 8% in the subsequent rounds of screening, a biopsy rate of about 1.5 -2% and a malignancy rate of biopsy between 20 and 50%. To reduce harms it is important that more than 90% of biopsies are needle biopsies done in the doctor's rooms in contrast to surgical biopsies in theatre, which are much more invasive and of course also more expensive. 

In order to deliver on the promise of screening of a reduced mortality rate it is paramount that cancers are discovered early. This is indicated by the following performance indicators: The proportion of non-invasive cancers, i.e. cancers that are curable with surgery alone should be 20% or higher; of invasive cancers, which need more intensive treatment, more than half should be less than 1.5 cm in diameter and more than 80% of these should not yet have spread to lymph nodes. Finally, in a well-run program, between about 4 and 10 cancers will be diagnosed for every 1000 mammograms depending on the age of the screening population and the underlying cancer load. 

There is no reason why screening centres should not publish these performance parameters to allow women to gauge how their screening centre measures up. 

Over diagnosis
A couple of further considerations deserve mention.  Much is being made of over diagnosis: These are cancers that will not lead to any serious consequences if left undiagnosed. To illustrate this point: If a woman is diagnosed at screening with a breast cancer and dies of a heart attack a day later, the breast cancer was inconsequential for her survival and did not need to be treated. Unfortunately, we cannot predict with accuracy which breast cancers are so slow growing that they will never pose a danger in the remaining survival of the patient; neither can we predict when life will end for the individual. Obviously, in the age groups targeted for mammographic screening, that is 40 – 69 years, death from other causes becomes more and more common with advancing age and the benefit from mammographic screening diminishes accordingly; this is why most public screening programs stop at age 69. 

Overdiagnosis therefore is unavoidable and is currently estimated at about 5 – 10% of all breast cancers in screening programs. In the same vein, as women diagnosed with breast cancer outside of screening programs do not die immediately, it takes time for the benefit of the earlier diagnosis through screening to manifest itself. Only in the second decade after the first screening episode does the screening effect becomes fully visible: The breast cancer mortality in women participating in screening is up to 60% lower than in those that did not participate. In our own experience in South Africa, a similar reduction in mortality was projected in women screened in our centre. 

Can we afford to screen?
Setting up an "organised" screening program requires considerable resources. In a resource-restricted environment, it is imperative that organizational and financial efficiencies are sought ruthlessly. For example, running the program, reading the images and performing the biopsies should be performed by specially trained radiographers as done in some first world countries and not be restricted to expensive specialists in radiology and surgery as is currently the case in South Africa. While the costs of installing a screening program can be fairly accurately estimated, these must then be offset against the benefits. Surprisingly, very little if any research has been done on the financial benefits of mammographic screening. These benefits would initially arise out of the less expensive primary treatment of earlier diagnosed breast cancer. Research by the author and others in 2010, and presented at the European Breast Cancer Conference during the same year, indicates that significant savings can be made in radiation, chemotherapy and biologicals. Surgery costs and hormonal therapy costs, however, were the same for screen-detected versus symptomatic breast cancer. 

Unfortunately, there is no research on the cost of treatment of metastatic breast cancer (breast cancer that has spread to other areas in the body) and the financial benefits that arise out of the savings of treatment cost of advanced breast cancer can only be speculated on: Experience teaches, that patients with advanced disease usually require multiple courses of chemo, radiation and endocrine therapy and are admitted to hospital for complications of the disease or its treatment. This probably adds up to considerable costs much exceeding those for primary therapy of the cancer. As these occur only years after the initial diagnosis, little effort has been expended to take these into consideration when calculating the systemic cost/benefit analysis of mammographic screening. It is the contention of this author that this is vital when looking at the systemic impact of screening for breast cancer and currently, incomplete data make this assessment impossible.     

Professor Justus Apffelstaedt is the head of the Surgical Oncology Unit at Tygerberg Hospital and Stellenbosch University's Faculty of Medicine. He is also the president of Breast Interest Group of Southern Africa.