A US study suggests that computer-aided detection (CAD) does not help radiologists interpret mammograms more accurately and produces worse results than experienced radiologists using their eyes alone.

The study is published in the New England Journal of Medicine (NEJM).

Approved by the Food and Drug Administration (FDA) in 1998, CAD scans mammogram X-rays and highlights suspicious areas for the radiologist to look at in more detail. CAD is used as a supplementary tool to support the radiologist. On average CAD highlights about 4 areas on each mammogram.

However, Joshua Fenton, assistant professor of family and community medicine with the University of California, Davis Health System in Sacramento, and colleagues found that using CAD was more likely to lead to false positives, where benign growths are marked as potential cancer tumours.

The result is that women are called back for biopsies which are not necessary and this results in unnecessary cost to the hospital and also anxiety and stress to patients.

The researchers examined the relationship between CAD and mammography screening performance between 1998 and 2002 at 43 screening centres in 3 states.

The data they used came from 429,345 mammograms and 222,135 women, including 2,351 who were diagnosed with breast cancer within 12 months of being screened.

They looked at three aspects of mammography performance with and without CAD:

— specificity (how well a negative result is identified),
— sensitivity (how well a positive result is identified), and
— positive predictive value (the proportion of patients with positive test results who are correctly diagnosed).

And they also measured the overall accuracy and individual rates of biopsy and breast cancer detection.

The findings showed that:

— 7 of the 43 (16 per cent) of the centres started using CAD during the period of the study.
— Diagnostic specificity decreased from 90.2 per cent before, to 87.2 after, CAD was brought in.
— Positive predictive value decreased from 4.1 to 3.2 per cent before and after CAD.
— Rate of biopsy increased by 19.7 per cent before and after CAD.
— Sensitivity increased from 80.4 per cent before CAD to 84.0 per cent after CAD but this was not significant.
— Cancer-detection rate changed from 4.15 to 4.20 cases fer 1,000 screenings but this was not significant.
— Analyses of all 43 centres showed CAD linked significantly with lower overall accuracy compared with no CAD.

The researchers concluded that the use of CAD is:

“associated with reduced accuracy of interpretation of screening mammograms. The increased rate of biopsy with the use of computer-aided detection is not clearly associated with improved detection of invasive breast cancer”.

In an accompanying editorial a radiology professor at Beth Israel Deaconess Medical Center in Boston, Ferris Hall, said this does not mean radiologists are about to stop using CAD, but it does make a dent in the technology’s reputation.

Hall called for larger studies to look in more detail at how the technology is used and the best ways to use it to support radiologists.

He said that while CAD on average finds between two and four suspicious areas, only one in 2,000 is cancerous.

While the technology is there to help radiologists find suspicious marks on the mammogram, since it was approved by the FDA in 1998, “its effect on the on the accuracy of interpretation is unclear”.

Hall appears to favour moving toward a combination of methods including genetic risk profiling and MRI (magnetic resonance imaging).

Mammography has its place, but he described it as a “poor, two-dimensional projectional method being used to diagnose small, three-dimensional cancers”.

Also, MRI does not carry the risk of radiation exposure and it is more sensitive, although has lower specificity compared to mammography.

He alludes to the reluctance to adopt mammography when it first arrived on the scene, and suggests MRI is experiencing the same lack of acceptance: “the major problems with MRI of the breast and related magnetic resonance spectroscopy are cost and interpretive expertise”, he writes.

A mammogram is a low dose X-ray of the breast. It is used to find potential tumours and cysts before they get big enough to be felt by hand and in many countries it is a routine screening method for early detection of breast cancer in older women.

In the US, breast cancer is the second most common cancer in women, skin cancer is the most common.

According to American Cancer Society estimates, about 178,000 women will discover they have breast cancer this year, and over 40,000 will die from it.

The Society has recently announced that MRI (magnetic resonance imaging) should be used to screen women with a high risk of breast cancer.

“Influence of Computer-Aided Detection on Performance of Screening Mammography.”
Fenton, Joshua J., Taplin, Stephen H., Carney, Patricia A., Abraham, Linn, Sickles, Edward A., D’Orsi, Carl, Berns, Eric A., Cutter, Gary, Hendrick, R. Edward, Barlow, William E., Elmore, Joann G.
N Engl J Med 2007 356: 1399-1409.
Volume 356:1399-1409, April 5, 2007, Number 14.

Click here for Abstract.

Click here for more information about Mammograms and Other Breast Imaging Procedures (American Cancer Society).

Written by: Catharine Paddock
Writer: Medical News Today