Errors in image interpretation are a common problem in diagnostic radiology. Although many published articles provide trainees with the means to correctly interpret imaging studies, they do not provide a framework for understanding why and how errors occur. In this article, we propose a classification system that allows categorization of errors, which we hope can serve as a basis for peer review, self-education, and quality improvement programs. Our scheme incorporates elements of a classification system proposed by previous authors but also includes novel categories. In this article, we show the usefulness of our scheme by applying it to a specific, and particularly problematic, diagnosis in emergency radiology, namely that of dural sinus thrombosis.