Abstract 18655: Lack of Accuracy of Chest Radiography in Predicting Left Ventricular Lead Position when Compared to Chest Computed Tomography: Potential Implications for Cardiac Resynchronization Therapy
Background: While cardiac resynchronization therapy (CRT) improves outcomes in patients with systolic heart failure and prolonged QRS duration, up to 33% of patients do not respond clinically. Inappropriate left ventricular (LV) lead position is a major contributor to CRT nonresponse, with anterior and apical LV lead positions associated with worse outcomes. Chest x-ray (CXR), a 2-dimensional technique, has traditionally been utilized to ascertain LV lead position. Computed tomography (CT) is considered the gold standard to localize objects in 3-dimensional space. We sought to ascertain the accuracy of CXR-based LV lead position analysis compared to contrast-enhanced chest CT in patients with biventricular pacing.
Methods: We selected 47 consecutive CRT patients that had an available CXR and a chest CT (performed for unrelated indications after CRT). CXR LV lead position was classified as anterior, lateral or posterior and as basal, mid or apical using lateral and posterior-anterior (PA) projections. Multi-planar reformatting of CT images was used to generate four chamber, axial and sagittal projections for categorization of LV lead position (similar to CXR). CXR and CT analysis was done in a blinded fashion.
Results: The median age was 65+15 years with 60% men and 52% with an ischemic etiology. The baseline median ejection fraction was 28%+6 with a median QRS duration of 166 msec+38. On Chest CT, 12.8% of LV leads were anterior, 19.1% lateral and 68.1% posterior. LV leads were basal in 27.7% of patients with 59.5% categorized as mid and 12.8% as apical. CXR correctly identified the LV lead position on lateral projection in only 51% of cases and on PA projection in 53% of cases. CXR correctly identified the LV lead position on both orthogonal projections in only 38% cases. CXR overestimated the number of apical LV leads (29.8% vs. 12.8%, p < 0.05) and underestimated the number of anterior LV leads (2.1% vs. 12.8%, p < 0.05).
Conclusion: In CRT, LV lead position predicted by CXR has limited correlation with LV lead position determined by chest CT. This has potential implications for LV lead position relocation and the ultimate success of CRT.
- © 2010 by American Heart Association, Inc.