The Positive Predictive Value of BI-RADS Microcalcification Descriptors and Final Assessment Categories
The Positive Predictive Value of BI-RADS Microcalcification Descriptors and Final Assessment Categories
复制标题
DOI：
10.2214/ajr.09.3423
复制
发表时间：
2010-05-01
影响因子：
5
通讯作者：
Sayre, James W.
中科院分区：
文献类型：
--
作者：
Bent, Chris K.;Bassett, Lawrence W.;Sayre, James W.
客观的。本文的目的是根据数字乳腺 X 线摄影环境中的 BI-RADS 描述符回顾性评估微钙化恶变的可能性。材料和方法。该研究纳入了 146 名患有钙化的女性，她们在 2005 年 4 月至 2006 年 7 月期间接受了影像引导活检。活检前获得的数字化乳房 X 线照片由两名对活检结果不知情的乳腺影像专家独立分析。不一致描述的病变通过协商一致解决。一位放射科医生提供了 BI-RADS 最终评估分数。结果。活检的总体阳性预测值为28.8%。各个形态学描述符预测恶性肿瘤的风险如下：细线性/分支，23 例中有 16 例（70%）；细多形性，50例中有14例（28%）；粗异质，10例中有2例（20%）；无定形，51 例中有 10 例（20%）；典型良性，12 例中为零 (0%)。 Fisher-Freeman-Halton 精确检验显示形态描述符 (p < 0.001) 和分布描述符 (p < 0.001) 之间具有统计显着性。根据 BI-RADS 评估类别，恶性肿瘤的阳性预测值如下： 2 类，0%；类别 3，0%； 4A类，13%； 4B类，36%； 4C类，79%；第 5 类，100%。结论。 BI-RADS 形态和分布描述符可以帮助评估全视野数字乳房 X 光检查检测到的微钙化的恶性风险。连续 BI-RADS 类别（4A、4B 和 4C）的阳性预测值有所增加，验证了细分可以更好地评估可疑微钙化的恶性肿瘤可能性。
OBJECTIVE. The purpose of this article is to retrospectively assess the likelihood of malignancy of microcalcifications according to the BI-RADS descriptors in a digital mammography environment.MATERIALS AND METHODS. The study included 146 women with calcifications who underwent imaging-guided biopsy between April 2005 and July 2006. Digital mammograms procured before biopsy were analyzed independently by two breast imaging subspecialists blinded to biopsy results. Lesions described discordantly were settled by consensus. One of the radiologists provided a BI-RADS final assessment score.RESULTS. The overall positive predictive value of biopsies was 28.8%. The individual morphologic descriptors predicted the risk of malignancy as follows: fine linear/branching, 16 (70%) of 23 cases; fine pleomorphic, 14 (28%) of 50 cases; coarse heterogeneous, two (20%) of 10 cases; amorphous, 10 (20%) of 51 cases; and typically benign, zero (0%) of 12 cases. Fisher-Freeman-Halton exact testing showed statistical significance among morphology descriptors (p < 0.001) and distribution descriptors (p < 0.001). The positive predictive value for malignancy according to BI-RADS assessment categories were as follows: category 2, 0%; category 3, 0%; category 4A, 13%; category 4B, 36%; category 4C, 79%; and category 5, 100%.CONCLUSION. BI-RADS morphology and distribution descriptors can aid in assessing the risk of malignancy of microcalcifications detected on full-field digital mammography. The positive predictive value increased in successive BI-RADS categories (4A, 4B, and 4C), verifying that subdivision provides an improved assessment of suspicious microcalcifications in terms of likelihood of malignancy.