[18F]Fluorodeoxyglucose Uptake by Positron Emission Tomography for Diagnosis of Suspected Lung Cancer

Open Access
Publisher Website
Google Scholar
Abstract
Verification bias is an increasingly appreciated problem that occurs when investigators fail to account for incomplete measurement of a gold standard test among patients referred for a diagnostic test.1,2 For example, sensitivity and specificity of prostate-specific antigen screening for prostate cancer in men younger than 60 years are claimed to be 57% and 60%, respectively.3 However, after adjustment for verification bias, the sensitivity of prostate-specific antigen screening decreased to 18%, and specificity increased to 98%. Verification bias has also been shown to lead to substantial overestimation of sensitivity of exercise electrocardiography4 and noninvasive imaging5 for diagnosis of coronary artery disease.