OBJECTIVETo evaluate the accuracy and direct medical cost of different screening modalities of physical examination (PE), mammography (MAM), and high-frequency ultrasonography (US) for breast cancer among Chinese women.

METHODSThis study was designed as a blindly carried outcom parative parallel screening trial for breast cancer among Chinese women aged 35 years or older. Physical examination was conducted with the subjects in the first round of breast cancer screening. The negative cases were followed up approximately 1 year later and moved into the second round of screening. Using the results of biopsies and 1-year follow-up as the gold standards, the accuracy and cost indexes of different screening modalities were calculated. Data were analyzed by McNemar test.

RESULTSA total of 2 471 eligible women were included in this trial, and 14 breast cancers were identified among them. Mammography enabled to detect 11 cancers, high-frequency ultrasonography detected 9 cancers and physical examination detected 8 cancers. Considering the three modalities separately, MAM identified the fewest suspicious cases (52 cases) and detected the most cancers (11 cases). Using US alone at the first stage, followed by MAM when indicated, offered the highest specificity (99.7%)and correct positive predictive value (11.4%), meanwhile the sensitivity was 78.6%, and the mean costs of the screening modality in urban and rural areas were 291, 210 yuan and 886,050 yuan per cancer case detected.

CONCLUSIONSThe strategy of screening with US alone at the first stage, followed by MAM when indicated, may be the most suitable modality of breast cancer detection in most regions of China, but the cost is still too high to develop the breast cancer screening in some low incidence regions with limited health resources.