【摘要】 目的:探讨全数字化乳腺摄影(DM)计算机辅助诊断(CAD)在不同乳腺结构内检出乳腺癌的价值。方法:185例经病理证实的单乳单灶性乳腺癌及179例正常乳腺均行DM检查,根据BI-RADS将所有乳腺分为非致密组和致密组。将所有乳腺摄影图像采用CAD法进行诊断,计算CAD的诊断敏感性,记录CAD在病例组及正常组的平均每例假阳性标记数,并进行不同结构乳腺组间及病例组与正常组的比较。结果:CAD检出乳腺癌的敏感度为88.6%;在非致密组与致密组中检出乳腺癌的敏感度分别为97.4%和82.4%,两组比较,差异具有统计学意义(P<0.05)。在病例组和正常组中CAD的假阳性标记数的中位数(最小值,最大值)分别为1(0,12)个和2(0,8)个,两组间的差异有统计学意义(P<0.05)。结论:CAD检出乳腺癌的敏感性较高,乳腺密度可能影响DM对单纯肿块型乳腺癌的检出。更多还原

【Abstract】 Objective:To assess the value of full-field digital mammography(DM)with computer-aided detection system(CAD)for the diagnosis of breast carcinoma in breasts with different parenchyma structures.Methods:Mammograms of 185cases with pathology proved solitary breast carcinoma in unilateral breast and these of 179breasts of normal women were collected.All breasts were classified into dense and non-dense parenchyma groups according to BI-RADS classification.CAD were applied in all mammograms.The sensitivity and false-positive marks of CAD were evaluated and the comparison of cancer with normal cases,as well as comparison of non-dense group with dense parenchyma group were performed.Results:The overall sensitivity of CAD in detecting breast carcinoma was 88.6%.The sensitivity of CAD in nondense and dense parenchyma group was 97.4%and 82.4%respectively,with significant statistic difference(P<0.05).The median(maximum,minimum)of false positive marks was 1(0,12)in carcinoma cases and 2(0,8)in normal cases,with significant statistic difference.Conclusion:CAD shows a high sensitivity in detecting breast carcinoma.Breast parenchyma density can impact the detection of mass lesion in breast cancer.更多还原