Automatic cardiac MR image segmentation using edge detection by tissue classification in pixel neighborhoods.

A highly sensitive edge detector has been developed that uses tissue classification of pixels based on analysis of data in their local neighborhoods. In conjunction with recursive region growing, it has been used successfully to define regions of interest (ROI) when applied specifically to gradient echo MR images of the heart. The detector adapts to nonuniformity by carrying out an independent analysis at each location. If two tissues are present in a neighborhood and the pixel at that location cannot be classified with the seed pixel, a region edge has been crossed and recursion is stopped. No geometric assumptions relating to object shape such as definition of a region center and radial search are required. The detector was applied to multi-slice, multi-phase images of the heart from 26 subjects. A segmentation strategy specified slice processing order, graded ROIs, and used successfully detected ROIs to guide subsequent detection. Segmentation of all images resulted in a 90.3% median edge pixel detection efficiency.

Right and left ventricular volumes and function after acute pulmonary hypertension in intact dogs.

A canine model was developed to record right (RV) and left ventricular (LV) volumes and high-fidelity pressures during acute pulmonary hypertension without the need for major surgery. In this study, new methodology was applied to record high-fidelity RV and LV pressures during cinemagnetic resonance imaging of the heart before and after acute pulmonary hypertension in six anesthetized intact dogs in which the pericardium and thorax were never disturbed by any surgical procedure. After pulmonary embolus, RV systolic pressure increased from 27 + 2 (SD) to 43 +/- 8 mmHg (P < 0.01) as LV systolic pressure decreased (97 +/- 17 to 76 +/- 3 mmHg; P < 0.05). Stroke volume (26 +/- 7 to 21 +/- 5 ml; P < 0.05) and RV ejection fraction (45 +/- 9 to 28 +/- 3%; P < 0.01) decreased as LV ejection fraction was unchanged (50 +/- 5 to 52 +/- 5%; P = NS). LV end-diastolic pressure decreased from 11 +/- 4 to 7 +/- 3 mmHg (P < 0.05), and RV end-diastolic pressure increased from 6 +/- 3 to 11 +/- 3 mmHg (P < 0.01). RV end-diastolic volume increased from 57 +/- 14 to 75 +/- 20 ml (P < 0.01) as LV end-diastolic volumes decreased from 53 +/- 11 to 42 +/- 10 ml (P < 0.01), resulting in no change in total ventricular volume at end diastole (111 +/- 24 to 116 +/- 28 ml). The observed mean decrease of 4.0 mmHg and 11 ml in LV end-diastolic pressure and volume, respectively, was associated with no change in total ventricular volume.(ABSTRACT TRUNCATED AT 250 WORDS)

Ventricular coupling via the pericardium: normal versus tamponade.

OBJECTIVE: The aim was to examine how regional variations in pericardial pressure affect the mechanical coupling between the ventricles. METHODS: Canine hearts from 14 dogs (14.5-18 kg) were removed and placed in cold cardioplegia solution. Balloons were inserted into the left and right ventricles and the atria. Pericardial pressure over the left ventricle (Pclv) and the right ventricle (Pcrv) was measured with thin balloon catheters. Ventricular and pericardial pressures were measured, and ventricular and pericardial coupling was calculated, under control conditions and with increases in pericardial tension and fluid. RESULTS: At baseline, regional differences in pericardial pressure occurred [Pclv > Pcrv, 4.0(SD 0.9) v 2.9(0.6) mm Hg, p < 0.05]. Ventricular coupling via the pericardium was defined as delta Pclv/delta Pcrv for right ventricular volume increases and delta Pcrv/delta Pclv for left ventricular volume increases. This ratio increased more after increasing right ventricular volume than after increasing left ventricular volume [delta Pclv/delta Pcrv > delta Pcrv/delta Pclv, 1.14(0.33) v 0.51(0.15), p < 0.05]. Increasing the pericardial tension by clamping the pericardium increased pericardial pressures, yet did not alter the regional variations in pressure [Pclv > Pcrv, 8.4(2.2) v 6.4(2.5) mm Hg, p < 0.05] or pericardial coupling [delta Pclv/delta Pcrv > delta Pclv/delta Pcrv, 1.18(0.46) v 0.54(0.16), p < 0.05]. In contrast, creating a mild tamponade increased pericardial pressures, eliminated regional differences in pressure, and altered the coupling between ventricles [delta Pclv/delta Pcrv approximately delta Pclv/delta Pcrv, 0.95(0.11) v 1.05(0.08), p = NS]. These regional differences in pericardial pressure might have a geometrical basis. In four in vivo canine experiments using cine magnetic resonance, the short axis radius of curvature for the right ventricle was greater than for the left ventricle [38.3(4.4) mm v 29.2(3.8) mm, p < 0.05]. CONCLUSIONS: The pericardium partially protects right ventricular filling: regional differences in pericardial pressure normally occurred with lower pericardial pressure over the right ventricle, and left to right ventricular coupling was less. This protection of right ventricular filling was lost with even a small pericardial effusion.