Intraobserver Variability: Should We Worry?

摘要

Many papers have identified concerns about intraobserver variability of repeat outlining by the same clinician. These variations in individual performance in turn make it challenging to determine values for interobserver variability since these depend largely on the assumption that each observer's oudine is accurate. Aside from the concerns about inaccuracy, variability is a potential component of the planning target volume margin and thus minimization of this has the potential to reduce normal tissue dose and morbidity. One accepted measure of intraobserver agreement since I960 1 has been the Kappa (k) correlation coefficient, which varies from 0 (agreement by chance) to 1 (full agreement). The accepted subdivisions of kappa 2 are "excellent" (0.81-1.00), "good" (0.61-0.80), "moderate" (0.41-0.60), "fair" (0.21-0.40), and "poor" (0-0.20). It is clear from the evidence base that kappa is common to many aspects of medical practice. Despite the kappa assumptions concerning observer independence 3, it has been used extensively to report both intraobserver and interobserver variability in the interpretation of CT imaging data. Table 1 summarizes the results of these studies from the last 10 years.