Skip to main content
Fig. 5 | Insights into Imaging

Fig. 5

From: Limitations of cardiothoracic ratio derived from chest radiographs to predict real heart size: comparison with magnetic resonance imaging

Fig. 5

ROC curve of CTR and increased LVEDV. ROC curve illustrating diagnostic accuracy of multiple CTR cut-off values to detect an increased LVEDVi. The X-axis shows true positives for each cut-off (an increased LVEDV on cardiac MRI is seen by the defined cut-off). The Y-axis shows false positives (normal LVEDV on cardiac MRI, which is characterised as increased based on a given CTR value). Higher CTR values give few false positives at the expense of many false negatives (many patients with an increased LVEDV are missed). Lower CTR values diagnose most patients with an increased LVEDV at the expense of many false-positive results. The area under the ROC curve describes the overall diagnostic power of CTR. An ideal test would have no false positives or false negatives (AUC = 1 described by the 90° dashed line). Random guessing would render 50% of true positives and 50% of false positives (AUC = 0.5) and is represented by the no-discrimination line (45° dashed diagonal line). CTR cardiothoracic ratio, LVEDVi left ventricular indexed end-diastolic volume, AUC area under the curve. Other parameters are available in Additional file 1: Fig. S1

Back to article page