Reliability of computed tomography: correlation with neuropathologic findings

Correlations between psychopathology, psychological test results and computerized tomography changes in senile dementia

This study is an investigation of the relationship between morphological, neuropsychological, and clinical measurements, all presumably contributing to an expert's diagnosis of senile dementia. Morphological assessment was carried out by linear measurements taken from the photographs of the CT scanners display. The battery of tests comprised instruments able to evaluate performance (WAIS, Benton test, d2 test, Trail Making test) as well as personality (FPI) from a quantitative point of view. For clinical assessment the Assessment Scale for Gerontopsychiatry (AGP) was used. A group of 39 patients suffering from mild to severe dementia was studied. The results show strong correlations between some sub-tests of the WAIS and certain CT measurements but fail to show comparable features in using the full scale IQ. The other cognitive tests did not work as well in detecting brain atrophy. Surprisingly, the personality inventory (FPI) showed some correlations with brain atrophy. A number of correlations between morphological and clinical variables were found as well, consistent with well-known clinical experience. Topographical features will be discussed supporting the view that cerebral lesions in the progress of senile dementia take the form of distinct patterns.

A comparison between the computed tomogram and the neuropathological findings in cerebrovascular disease

The findings on computed tomography (CT) and neuropathological examinations were correlated in 87 autopsied cases of cerebrovascular disease with regard to the seize and location of lesions, experience of reviewers, and improvement in CT quality. Small infarctions less than 5 mm were very difficult to detect accurately on CT. This was largely because of the limitations in the efficiency of the CT scanner. Accurate diagnosis of medium sized infarctions was also often difficult. This was mainly due to the anatomical location of lesions, the confluence of deep and widened sulci, periventricular white matter, and structures in the posterior fossa. Large infections could be visualized easily on CT, except in their early periods or in cases with hemorrhagic infarctions. The improved accuracy of CT diagnosis for small and medium sized infarctions could not be attained by the experience of reviewers, but was only possible by instrumental improvement of CT quality.

Detection of cryptic vascular malformations by computerized tomography. Report of two cases

Two cases of cryptic vascular malformation that were not demonstrated by cerebral angiography were detected by computerized tomography. One of these patients had a cavernous angioma in the fourth ventricle with recurrent subarachnoid hemorrhages, and the other harbored a small arteriovenous malformation and intracerebral hematoma. The usefulness and limitations of computerized tomography in the identification of cryptic vascular malformations are discussed.

[Interobserver variability in CT reporting. A comparative evaluation (author's transl)]

Within a test of interobserver variability 88 CCT images were described by four physicians. In spite of a standardized documentation method, first results demonstrate that continuous quality control is necessary if the data is gathered into CCT data bases. The evaluation of uncontrolled data seems to be limited.