Biostatistics and Biomathematics, Fred Hutchinson Cancer Research Center, Seattle, Washington 98109, USA.
Med Decis Making. 2010 Jul-Aug;30(4):484-98. doi: 10.1177/0272989X09357477. Epub 2010 Feb 10.
Statistical evaluation of medical imaging tests used for diagnostic and prognostic purposes often employs receiver operating characteristic (ROC) curves. Two methods for ROC analysis are popular. The ordinal regression method is the standard approach used when evaluating tests with ordinal values. The direct ROC modeling method is a more recently developed approach, motivated by applications to tests with continuous values.
The authors compare the methods in terms of model formulations, interpretations of estimated parameters, the ranges of scientific questions that can be addressed with them, their computational algorithms, and the efficiencies with which they use data.
The authors show that a strong relationship exists between the methods by demonstrating that they fit the same models when only a single test is evaluated. The ordinal regression models are typically alternative parameterizations of the direct ROC models and vice versa. The direct method has two major advantages over the ordinal regression method: 1) estimated parameters relate directly to ROC curves, facilitating interpretations of covariate effects on ROC performance, and 2) comparisons between tests can be done directly in this framework. Comparisons can be made while accommodating covariate effects and even between tests that have values on different scales, such as between a continuous biomarker test and an ordinal valued imaging test. The ordinal regression method provides slightly more precise parameter estimates from data in our simulated data models.
Although the ordinal regression method is slightly more efficient, the direct ROC modeling method has important advantages in regard to interpretation, and it offers a framework to address a broader range of scientific questions, including the facility to compare tests.
用于诊断和预后目的的医学成像测试的统计评估通常采用接收器操作特征 (ROC) 曲线。ROC 分析有两种流行方法。当评估具有有序值的测试时，有序回归方法是标准方法。直接 ROC 建模方法是一种最近开发的方法，其动机是应用于具有连续值的测试。
作者从模型公式、估计参数的解释、可以用它们解决的科学问题范围、计算算法以及数据使用效率等方面比较了这两种方法。
作者通过证明当仅评估单个测试时，它们拟合相同的模型，证明了这两种方法之间存在很强的关系。有序回归模型通常是直接 ROC 模型的替代参数化，反之亦然。与有序回归方法相比，直接方法有两个主要优势：1）估计参数直接与 ROC 曲线相关，便于解释协变量对 ROC 性能的影响，以及 2）可以在此框架内直接进行测试之间的比较。可以在容纳协变量效应的情况下进行比较，甚至可以在具有不同尺度的测试之间进行比较，例如在连续生物标志物测试和有序值成像测试之间进行比较。在我们的模拟数据模型中，有序回归方法从数据中提供了略微更精确的参数估计。
尽管有序回归方法的效率略高，但直接 ROC 建模方法在解释方面具有重要优势，并且它提供了一个框架来解决更广泛的科学问题，包括比较测试的能力。
