Faculty Bibliography

PURPOSE/OBJECTIVE:To investigate the reproducibility and transferability of texture features between MR centers, and to compare two feature selection methods and two classifiers. MATERIALS AND METHODS/METHODS:Coronal T1-weighted MR images of the knees of 63 patients, divided into three groups, were included in the study. MR images were obtained at three different MR centers. Regions of interest (ROIs) were drawn in the bone marrow and fat tissue. Then texture analysis (TA) of the ROIs was performed, and the most discriminant features were identified using Fisher coefficients and POE+ACC (probability of classification error and average correlation coefficients). Based on these features, artificial neural network (ANN) and k-nearest-neighbor (k-NN) classifiers were used for tissue discrimination. RESULTS:Although the texture features differed among the MR centers, features from one center could be successfully used for tissue discrimination in texture data on MR images from other centers. The best results were achieved using the ANN classifier in combination with features selected by POE+ACC. CONCLUSION/CONCLUSIONS:The differences in texture features extracted from MR images from different centers seem to have only a small impact on the results of tissue discrimination.

OBJECTIVE:Our aim was to determine the value of different MRI planes independently and in combination for assessment of acromial shape. MATERIALS AND METHODS/METHODS:Sixty-one patients with subacromial impingement syndrome who had undergone acromioplasty after failure to respond to conservative treatment were included in the study. Parasagittal T2-weighted MR images and outlet view radiographs of the affected shoulders were acquired preoperatively. Three-dimensional models of all acromions were constructed from the MR images, and the Bigliani type of acromion depicted by these models was determined. Results were compared with the acromial type assessed during acromioplasty. To provide a reliable reference for further processing and correlation, we used only those 56 acromions with agreement on acromial shape between intraoperative findings and 3D models. Then, acromial shape was determined for three MRI slice positions (S-1, lateral acromial edge; S-2, just lateral of acromioclavicular joint; and S-3, lateral portion of acromioclavicular joint), for a combination of S-1 and S-2, and for the radiographs. RESULTS:Kappa coefficients were 0.36 (36%) for S-1, 0.41 (41%) for S-2, and -0.10 (-10%) for S-3. For the outlet view radiographs, the kappa coefficient was 0.55 (55%), showing better correlation than any single slice position. Best results, however, were achieved with a combination of S-1 and S-2, with a kappa coefficient of 0.66 (66%). CONCLUSION/CONCLUSIONS:For determination of acromial shape, outlet view radiographs are superior to any single MRI slice position, but inferior to a combination of two MRI slices (S-1 and S-2). If a single MRI slice is being used, the slice position just lateral to the acromioclavicular joint is recommended.

OBJECTIVE:The purpose of this study was to describe a largely observer-independent computer-assisted method for accurate quantitative analysis of bone marrow edema. MATERIALS AND METHODS/METHODS:Ten patients with bone marrow edema of the knee were included in the study. Coronal STIR images of the affected knees were obtained using a 1.0-T MR scanner. Size and signal intensity of the bone marrow edema were assessed on the basis of gray-scale value analysis and calculation of a threshold value for differentiating normal and edematous bone marrow. All measurements were carried out three times for statistical analysis. RESULTS:The intraobserver coefficient of variation was 0.89% for the volume and 0.94% for the signal intensity of the bone marrow edema, showing the small impact of manual interference on results produced with this method. CONCLUSION/CONCLUSIONS:A computer-assisted method for quantification of bone marrow edema has been described. Intraobserver variation was very low, indicating excellent reproducibility of results. Although the method is too time-consuming for clinical use, it is recommended for research purposes.