CAD falls short in breast cancer screening

December 2, 2005

When it comes to spotting suspicious lesions on a screening mammogram, computer-aided detection is no match for a dedicated breast imaging specialist, according a large study by researchers at Yale, presented Thursday at the RSNA meeting.

When it comes to spotting suspicious lesions on a screening mammogram, computer-aided detection is no match for a dedicated breast imaging specialist, according a large study by researchers at Yale, presented Thursday at the RSNA meeting.

Researchers conducted a prospective study of 5875 consecutive screening mammograms performed at Yale's Breast Imaging Center between February and September 2005. Two-thirds of the studies were performed on conventional film-based systems and one-third on digital equipment.

 

 

 

 

 

 

 

 

 

 

 

 

 

Researchers at Yale's dedicated breast imaging facility have found poor correlation between CAD markings and interpretations by dedicated breast imagers.

FIGURE 1. Spiculated mass not prompted by CAD was found to be infiltrating ductal carcinoma. FIGURE 2. Calcifications not prompted by CAD were determined to be DCIS. (Provided by L. Bryan-Rest)

 

The study compared CAD markings with interpretation by the facility's radiologists, whose experience in breast imaging ranged from four to 20 years. Researchers looked at the correlation overall and compared characterization masses, calcifications, density, and architectural distortion. They also reviewed follow-up diagnostic mammograms, but these were not available in all cases.

Of the total screened, radiologists' interpretations indicated there were 735 actionable lesions in 637 patients. CAD markings correlated with these interpretations in just 48% of lesions.

"There is poor correlation between CAD and actionable findings. These findings erode confidence in CAD," said Dr. Lara Bryan-Rest, who presented the study findings. "Complete reliance on CAD by non-experts is strongly discouraged," she added.

The level of correlation varied depending on the type of lesion. CAD markings correlated with radiologist interpretation in 65% of calcification cases but only 43% of masses.

"CAD was better with calcifications," Bryan-Rest said.

The Yale study adds to a conflicting body of evidence about the value of CAD. A large prospective study of 12,000 mammograms published in 2001 showed a benefit of 20% for breast cancer detection. Conversely, a large retrospective study of 115,000 screening mammograms published in 2004 showed little benefit.