Testing our tests: Do clinical studies of diagnostic performance truly inform patient management?

Excerpt

In this issue of the Journal, we are presented with a report by Barone-Rochette et al., of the diagnostic accuracy of a novel approach to myocardial perfusion imaging for stable coronary artery disease (CAD).1 The protocol utilized was dual isotope, stress thallium-201/rest Tc-99m. The dose was 2 mCi of Tl-201, and around 8 mCi of Tc-99m sestamibi for patients < 80 Kg, and weight-based for heavier patients, resulting in an average total body effective dose of 12 mSv. Strengths of the study include the use of FFR, and not just visually interpreted coronary angiography as the comparator standard, and the fact that all patients were referred for coronary angiography. While the latter fact avoids the verification bias resulting from selective angiography in patients with abnormal scans, it introduces a bias of its own by selecting only high-risk patients as subjects for the study. This selection bias was reflected in the results, with reported patient-level sensitivity, specificity, and diagnostic accuracy of 92.8%, 69.2%, and 81.4%, respectively. …