China study shows 44% improvement in ultrasound CAD detection rates

Twitter icon
Facebook icon
LinkedIn icon
e-mail icon
Google icon

Results from a study conducted at the PLA301 Hospital in Beijing, China, show a significant increase in diagnostic accuracy when using Medipattern’s B-CAD v2 on lesions less than 1 cm in size. B-CAD v2 is only available outside the United States.

The study, "The Value of Computer-Aided Detection for Breast Ultra-Sonography (B-CAD) in the Diagnosis of Breast Cancer," based upon 100 confirmed biopsy cases, compared the diagnostic accuracy of three reading scenarios: a single physician, double reading with two physicians and a single physician reading with B-CAD. The accuracy of the single physician increased with B-CAD in all categories.

The results show a statistically significant 44 percent improvement in diagnostic accuracy for the single reader with B-CAD v2 on lesions less than 1 cm in size (n equals 9), according to the researchers.

"The study also finds that CAD is a very useful tool for China's current needs; a single doctor with B-CAD can perform as well as two doctors reading together. A single doctor with B-CAD is far more efficient and economical than using our limited skilled resources for double reading,” Junlai Li, MD, co-author of the study, noted in the paper.

Li, co-author, DanFai Song and colleagues said they were not aware of the pathology outcomes when reading. The cases were separated into three size categories: malignant cancers greater than 2 cm; 1 to 2 cm; and less than 1 cm. The cases were evaluated using the American College of Radiology Breast Imaging Reporting and Data System Lexicon. Doctors reading without B-CAD used the lexicon to characterize the lesions in the study and assess the diagnosis.

B-CAD found the edge of the lesion and used this information to characterize the lesion for the physician. B-CAD further consolidated the information into a detailed report documenting whether the lesion is probably benign or probably malignant, according to Medipattern.

The researchers concluded that B-CAD has been shown to correctly identify potentially malignant lesions with a high degree of sensitivity, particularly for less than 1 cm breast cancers, and improved the accuracy using the ACR BI-RADS assessment.

Therefore, according to the results, the researchers recommended that “doctors should analyze ultrasonic images of breast lesions assisted by B-CAD, and then judge the suspicious sites identified with B-CAD and make the diagnostic decision using their own clinical experience. This will increase the detection rate of breast cancer and reduce the false negative rate."

The study will be presented at the 18th annual conference of the National Consortium of Breast Centers, held March 1-5 in Las Vegas.