OPTIMAL REFERENCE EXAM FOR MAMMOGRAPHIC INTERPRETATIONS

由于许多妇女每年进行乳房X光检查作为乳腺癌筛查的一部分，其中许多人有多个旧的乳房X光检查可与目前的乳房X光检查进行比较。在高容量快速筛查程序中，放射科医生将所有先前的乳房X线照片与当前检查进行比较是不切实际的。我们不知道有任何综合性研究将诊断性能作为参考检查的函数进行调查，也没有年龄依赖性最佳参考（例如，对于55岁及以上的女性，2年前或对于55岁以下的女性，1年前）。除了年龄依赖性最佳参考的可能性之外，在接受激素替代疗法的女性和不接受激素替代疗法的女性中，年龄依赖性最佳参考可能存在差异。在这项研究中，我们将要求七名放射科医生在四种不同的阅读模式下诊断180个病例，类似于筛选环境。除第一种模式不包括与既往检查的比较外，所有其他三种模式都将包括既往乳房X线照片，以与当前检查进行比较。两次考试之间的时间间隔为一年、两年或三年。因此，使用ROC型方法，我们将调查这些环境如何影响（或不）读者的表现，我们希望找到一个有效的和高效的阅读模式，为大多数患者。或者，我们可以发现不同的模式对于不同的患者是最佳的（例如，不同年龄组）。

As many women have annual mammograms as part of screening for breast cancer, many of them have multiple older mammograms available to compare with the current mammogram. In a high volume, rapid screening program, it is not practical for a radiologist to compare all of the prior mammograms to the current examination. We are not aware of any comprehensive study that investigates the diagnostic performance as a function of the reference examination nor is a there an age dependent optimal reference (e.g. 2 year prior for women age 55 and over or l year prior for women under 55 years of age). In addition to the possibility of an age dependent optimal reference, there may be a difference in an age dependent optimal reference in women who take hormone replacement therapy and those who do not. In this study, we will ask seven radiologists to diagnose 180 cases under four different reading modes similar to a screening environment. Except the first mode that does not include a comparison to the prior examinations, all other three modes will include a prior mammogram for comparison with the current examination. The time interval between two examinations will be either one, two, or three years. Hence, using an R0C-type methodology, we will investigate how these environments affect (or not) reader-performance, and we hope to find an effective and efficient reading mode for the majority of patients. Alternatively, we may find that different modes are optimal for different patients (e.g., for different age groups).