To the Editor.— We appreciate the methods used by Heinsimer et al 1 in their article entitled "Supine Cross-Table Lateral Chest Roentgenogram for the Detection of Pericardial Effusion" in the June 19, 1987, issue of JAMA , but disagree with their conclusions. Based on improved sensitivity, specificity, and predictive values compared with a standard chest roentgenogram, they recommend that a supine cross-table lateral (SCTL) film be included in the evaluation of patients with suspected pericardial effusion or in patients in whom a large cardiac silhouette is detected on a standard chest roentgenogram. However, the test characteristics may not accurately reflect the true clinical utility of the additional roentgenogram view. Given the preroentgenogram probability of pericardial effusion of 60% in their study population, the positive and negative predictive values for a two-view chest roentgenogram are 73% and 44%, and 78% and 51% for the SCTL view. The incremental gain in information by