A logistic regression model based on the national mammography database format to aid breast cancer diagnosis.
A logistic regression model based on the national mammography database format to aid breast cancer diagnosis.
复制标题
DOI：
10.2214/ajr.07.3345
复制
发表时间：
2009-04
期刊：
影响因子：
--
通讯作者：
Burnside ES
中科院分区：
文献类型：
--
作者：
Chhatwal J;Alagoz O;Lindstrom MJ;Kahn CE Jr;Shaffer KA;Burnside ES
基于国家乳房X线摄影数据库（NMD）格式的描述符，使用逻辑回归创建乳腺癌风险估计模型，该模型可以帮助制定乳腺癌早期检测的决策。机构审查委员会免除了这项符合 HIPAA 的回顾性研究的知情同意要求。我们根据 1999 年 4 月 5 日至 2004 年 2 月 9 日期间使用乳房成像报告和数据系统 (BI-RADS) 词典和 NMD 格式报告的 18,270 名患者的 48,744 项研究中的 62,219 例连续乳房 X 光检查记录的乳房 X 光检查特征和人口统计数据创建了两个逻辑回归模型。州癌症登记结果与我们的数据相匹配，作为参考标准。癌症的概率是两个模型的结果。 Model-2 是使用 Model-1 中的所有变量以及放射科医生的 BI-RADS 评估代码构建的。我们使用 10 倍交叉验证来训练和测试模型，并计算受试者工作特征 (ROC) 曲线下面积 (Az) 来衡量性能。将两种模型与放射科医生的 BI-RADS 评估进行比较。放射科医生的 Az 值为 0.939 ± 0.011。 Model-1 的 Az 为 0.927 ± 0.015，Model-2 的 Az 为 0.963 ± 0.009。在 90% 特异性下，Model-2 (90%) 的敏感性显着优于放射科医生 (82%) 和 Model-1 (83%)（P<0.001）。在 85% 的敏感性下，模型 2 的特异性 (96%) 显着优于放射科医生 (88%) 和模型 1 (87%) (P<0.001)。我们的逻辑回归模型可以有效区分良性和恶性乳腺疾病，并确定与乳腺癌相关的最重要特征。
To create a breast cancer risk estimation model based on the descriptors of National Mammography Database (NMD) format using logistic regression that can aid in decision-making for early detection of breast cancer. Institutional Review Board waived this HIPAA-compliant retrospective study from requiring informed consent. We created two logistic regression models based on the mammography features and demographic data for 62,219 consecutive cases of mammography records from 48,744 studies in 18,270 patients reported using the Breast Imaging-Reporting and Data System (BI-RADS) lexicon and NMD format between 4/5/1999 and 2/9/2004. State cancer registry outcomes matched with our data served as the reference standard. The probability of cancer was the outcome in both models. Model-2 was built using all variables in Model-1 plus radiologists’ BI-RADS assessment codes. We used 10-fold cross-validation to train and test the model and calculate the area under the receiver operating characteristic (ROC) curves (Az) to measure the performance. Both models were compared to the radiologists’ BI-RADS assessments. Radiologists achieved an Az value of 0.939 ± 0.011. The Az was 0.927 ± 0.015 for Model-1 and 0.963 ± 0.009 for Model-2. At 90% specificity, the sensitivity of Model-2 (90%) was significantly better (P<0.001) than that of radiologists (82%) and Model-1 (83%). At 85% sensitivity, the specificity of Model 2 (96%) was significantly better (P<0.001) than that of radiologists (88%) and Model-1 (87%). Our logistic regression model can effectively discriminate between benign and malignant breast disease and identify the most important features associated with breast cancer.