Faculty Bibliography

Protruding atheromas of the aortic arch identified by transesophageal echocardiography have been implicated as a cause of stroke in elderly patients. One hundred thirty patients greater than or equal to 65 years of age were studied with intraoperative transesophageal echocardiography to detect aortic arch protruding atheromas and determine if these patients were at higher risk for perioperative stroke. Protruding atheromas were identified in 23 (18%) of 130 patients. In 19 (83%) of these 23 patients, palpation of the aortic arch at operation did not identify significant abnormalities. Five patients (4%) had perioperative stroke. Logistic regression identified aortic arch atheroma as the only historical or procedural variable that was predictive of stroke (odds ratio 5.8, 95% confidence interval 1.2 to 27.9, p less than 0.03). A history of peripheral or cerebrovascular disease, presence of aortic calcification, cardiac risk factors, age and duration of cardiopulmonary bypass did not predict stroke. In contrast, patients with protruding atheromas with mobile components were at highest risk. There were 3 (25%) of 12 patients with a mobile atheroma who had a stroke versus 2 (2%) of 118 patients without a mobile atheroma (chi-square = 10.3, p = 0.001). Displacement and detachment of the frail, protruding atherosclerotic material by aortic arch cannulation or by the high pressure jet emanating from the cannula tip may play an important role in the creation of embolization and stroke

Registration of volumetric images from different medical imaging modalities is performed by matching surfaces using the principal axes technique. Translation, rotation, and scaling transformations are calculated by eigenvalue analysis of the scatter matrix. After applying the transformations, reslicing along comparable planes is carried out. The method is applied to a clinical case of X-ray computed tomography (CT) and magnetic resonance imaging (MRI) brain scans. The accuracy, measured as the distance between recognizable reference points in the registered CT and MRI slices, was 1.5 mm. Visual confirmation of the quality of the registration is provided by compositing the registered images. The method is simple to implement and computationally efficient; calculation of the transformation takes less than 1 s of computer time. This method requires full scan coverage in both scans and assumes local distortions are not present. Potential applications of this technique include radiation therapy, Surgical planning, functional/anatomical correlation, and retrospective studies

This paper presents a method of field computation in permanent magnets designed to generate a uniform field in the region of interest. The computation technique is particularly suitable to magnetic structures used in magnetic resonance imaging (MRI). The method is applied to the design of yokeless, yoked, and hybrid prismatic magnetic structures. In particular, it is applied to the correction of the field inhomogeneity of open magnets by means of a modulation of the magnetic material and by the introduction of high permeability materials

The distributions of the cerebral gray matter, the white matter, and the intracranial cerebrospinal fluid (CSF) were measured in 14 patients with Alzheimer disease (AD) and in 14 healthy control subjects. The measurements, derived from two specifically designed magnetic resonance inversion-recovery sequences, compensate for partial signal averaging. The percentage of the gray matter in the brains of AD patients (44.9% +/- 4.4) was significantly lower than in control subjects (50.2% +/- 3.2). The most significant reduction (P less than .001) occurred in the temporal lobes (13.8%) and a central region (12.8); the reduction in frontal lobe (11.2%) and occipital lobe (9.2%) was also statistically significant (P less than .01). There was an increase in the CSF volume in the temporal, occipital, and frontal regions; no region showed a significant difference in the white matter content. The findings of diffuse changes and temporal lobe involvement in AD are consistent with pathologic observations of cortical cell loss

Volume-based and surface-based algorithms for three-dimensional rendering of computed tomography (CT) scans of the human skull were compared in patients with craniofacial anomalies. Both methods were applied to a selected sample of 12 clinical CT studies. The number of sections ranged from 24 to 72 and the section thickness from 1.5 to 6.0 mm. Volume renderings were more prone to interpolation artifacts but captured the anatomy in greater detail. The sites of closed cranial sutures, visualized using the volume technique, were not demonstrated using the specific surface rendering technique used in this study. In both techniques the areas of thin bone appeared as gaps