[Whole-body MRT studies--the results of the "Multicenter study on the evaluation of magnetic resonance tomography"]. / MRT-Untersuchungen im Ganzkörperbereich--Ergebnisse der "Multicenterstudie zur Evaluierung der Kernspintomographie".

In a nation-wide multicentre study, the use of MRI was examined in its clinical context and in relation to other imaging methods and as a routine examination. It was found that examinations of the brain and spine were predominant. MRI is accurate for certain problems affecting the extremities, the pelvis and the liver. The findings also show that there are still technical limitations. A high degree of accuracy can only be obtained if there is adequate information concerning individual problems. Technical limitations include the frequent appearance of artifacts. In its present state, MRI can only supply additional information, concerning the trunk only, in certain situations. The study also outlines the development of MRI and the role of other conventional diagnostic procedures.

[Obstacles to the routine use of magnetic resonance imaging--results of a multicenter study for evaluating nuclear magnetic resonance tomography]. / Beeinträchtigende Faktoren in der MR-Routineanwendung--Ergebnisse der "Multicenterstudie zur Evaluierung der Kernspintomographie".

Within the framework of this study, the factors that impair the value of MR scans were also investigated and analysed. It was found that artifacts occurred in a large number of examinations that exercised a negative effect on the results. Artifacts are particularly seen in the thorax (54%) and upper abdomen (57%), but they have also been reported in 23% of the CNS examinations. The sources of such artifacts were mainly movements in the investigated layer, whereas other causes were negligible. In the 21,633 examined patients no undesirable effects caused by the examination were noted that would have impaired the patient or the course of the study. On an average, 88% of the examinations were performed with normal tolerance of the examination. In the remaining cases there was motor unrest or claustrophobia requiring special measures, thus hampering the course of the examination.

[NMR studies of the whole body--results of a multicenter study for evaluating nuclear magnetic resonance tomography]. / NMR-Untersuchungen im Ganzkörperbereich--Ergebnisse der "Multicenterstudie zur Evaluierung der Kernspintomographie".

Within the framework of a national multi-centre study the spectrum of application and the ranking of magnetic resonance tomography were investigated. Although scans of the brain and vertebral column still dominate the scene, it has become evident that whole body scanning via MR can be very accurate for certain indications. This is particularly true for examinations of gynaecological diseases (95%) and for lesions of the liver (97%). However, the number of problems in which MR can provide answers to diseases in the whole body range is still limited, and hence it would be most desirable to achieve further instrumental and methodical improvements.

[Evaluating magnetic resonance tomography studies of the extremities--results of a multicenter study for evaluating nuclear magnetic resonance tomography]. / Bewertung von MRT-Extremitätenuntersuchungen--Ergebnisse der "Multicenterstudie zur Evaluierung der Kernspintomographie".

MR examinations of the extremities occupy the first place in whole body examinations via MR. Highest accuracy quotas in this area with an average of 93.8% were obtained for tumours, trauma and necroses of the bones. These high accuracy quotas, however, were achieved only if sufficient preliminary information was available. Other statistically significant factors were found to be the field strength, the combined use of spin echo and fast imaging sequences and the use of the multislice technique.

Quantification and classification of echographic findings in the thyroid gland by computerized B-mode texture analysis.

Computerized B-mode image analysis was investigated in a pilot study to establish quantitative criteria for thyroid echography and to demonstrate the diagnostic accuracy of the procedure in thyroid disease. The ten per cent percentile of grey level distribution and entropy proved to be the most efficient statistical parameters. Computerized ultrasound detected all nodular lesions with an overall accuracy of about 85%. Standard diagnostic evaluation in nodular thyroid disease includes conventional ultrasound followed by needle biopsy to confirm the final diagnosis. However, in many cases, single needle biopsy is not sufficient to exclude malignant disease. Computer-aided ultrasound proved to be helpful for the identification of typical echographic characteristics in thyroid lesions with a high risk of malignancy. We suggest that needle biopsies be repeated in cases in which computerized analysis is positive, even when initial needle biopsy is negative.

Diagnostic accuracy of computerized B-scan texture analysis and conventional ultrasonography in diffuse parenchymal and malignant liver disease.

The use of a diagnostic system for ultrasonic liver tissue characterization based on computerized B-mode image analysis is clinically tested and compared with the results of conventional realtime and static grey scale liver ultrasound as independently assessed by three experienced observers. The diagnostic classes, normal, diffuse parenchymal and malignant disease, are clearly differentiated by computerized image analysis which is superior to subjective evaluation of liver echograms. Computerized analysis also renders a reliable and clinically useful diagnostic subclassification of diffuse parenchymal disease into echopattern changes prevalent in chronic hepatitis, cirrhosis/fibrosis, fatty infiltration and a mixed state of cirrhosis/fibrosis with fatty infiltration which cannot be achieved by conventional liver ultrasound.

[Gray-level histogram analysis in the sonographic diagnosis of diffuse parenchymal liver disease]. / Der Beitrag der Grauwerthistogramm-Analyse zur sonographischen Diagnostik des diffusen Leberparenchymschadens.

Increased echogenicity is one of the major diagnostic criteria for the subjective evaluation of liver echograms. In our approach to ultrasonic tissue characterisation by B-scan image analysis echogenicity is expressed by parameters deriving from the grey-level histogram. In this study the grey level histograms of a group of normals and patients with diffuse parenchymal liver disease are analysed. When employing the two parameters mean grey level (G) and standard deviation of grey levels (SG) the diagnostic accuracy was 90% (specificity 95%, sensitivity 85%). For subjective evaluation the liver echograms of a randomly selected sub-group of normals and patients were blindly evaluated by three observers who positively correlated the finding of increased echogenicity with the diagnosis of a pathological liver echogram. As demonstrated by our results the two-parameter analysis of the grey-level histogram leads to a quantification of the subjective diagnostic criterion of echogenicity, thus providing useful information for image interpretation.

Liver tissue characterization using computerized echography.

A digital echographic data acquisition and evaluation system was used in a clinical study including more than 100 patients with various local or diffuse liver tissue alterations. For diagnosis two decision methods are applied: first linear discriminant analysis and second a Bayesian method. Each A-scan is represented by a set of various numerical parameters from which only the most signal-distinctive ones are used for evaluation. The diagnostic results which will be presented and discussed have proved the method to be very sensitive and specific.