[Self-organizing neural networks for automatic detection and classification of contrast (media) enhancement of lesions in dynamic MR-mammography]. / Selbstorganisierende neuronale Netze zur automatischen Detektion und Klassifikation von Kontrast(mittel)-verstärkten Läsionen in der dynamischen MR-Mammographie.

PURPOSE: Investigation and statistical evaluation of "Self-Organizing Maps," a special type of neural networks in the field of artificial intelligence, classifying contrast enhancing lesions in dynamic MR-mammography. MATERIAL AND METHODS: 176 investigations with proven histology after core biopsy or operation were randomly divided into two groups. Several Self-Organizing Maps were trained by investigations of the first group to detect and classify contrast enhancing lesions in dynamic MR-mammography. Each single pixel's signal/time curve of all patients within the second group was analyzed by the Self-Organizing Maps. The likelihood of malignancy was visualized by color overlays on the MR-images. At last assessment of contrast-enhancing lesions by each different network was rated visually and evaluated statistically. RESULTS: A well balanced neural network achieved a sensitivity of 90.5 % and a specificity of 72.2 % in predicting malignancy of 88 enhancing lesions. Detailed analysis of false-positive results revealed that every second fibroadenoma showed a "typical malignant" signal/time curve without any chance to differentiate between fibroadenomas and malignant tissue regarding contrast enhancement alone; but this special group of lesions was represented by a well-defined area of the Self-Organizing Map. DISCUSSION: Self-Organizing Maps are capable of classifying a dynamic signal/time curve as "typical benign" or "typical malignant." Therefore, they can be used as second opinion. In view of the now known localization of fibroadenomas enhancing like malignant tumors at the Self-Organizing Map, these lesions could be passed to further analysis by additional post-processing elements (e.g., based on T2-weighted series or morphology analysis) in the future.

[Automated detection of spleen volume by spiral CT scans using neural networks and "fuzzy logic"]. / Automatische Berechnung des Milzvolumens aus Spiral-CT-Daten mit Hilfe neuronaler Netze und "Fuzzy Logik".

PURPOSE: To assess spleen segmentation and volumentry in spiral CT scans with and without pathological changes of splenic tissue. METHODS: The image analysis software HYBRIKON is based on region growing, self-organized neural nets, and fuzzy-anatomic rules. The neural nets were trained with spiral CT data from 10 patients, not used in the following evaluation on spiral CT scans from 19 patients. An experienced radiologist verified the results. The true positive and false positive areas were compared in terms to the areas marked by the radiologist. The results were compared with a standard thresholding method. RESULTS: The neural nets achieved a higher accuracy than the thresholding method. Correlation coefficient of the fuzzy-neural nets: 0.99 (thresholding: 0.63). Mean true positive rate: 90% (thresholding: 75%), mean false positive rate: 5% (thresholding > 100%). Pitfalls were caused by accessory spleens, extreme changes in the morphology (tumors, metastases, cysts), and parasplenic masses. CONCLUSIONS: Self-organizing neural nets combined with fuzzy rules are ready for use in the automatic detection and volumetry of the spleen in spiral CT scans.

Automatic detection of ground glass opacities on lung HRCT using multiple neural networks.

The purpose of this study was to implement neural networks and expert rules for the automatic detection of ground glass opacities (GG) on high-resolution computed tomography (HRCT). Different approaches using self-organizing neural nets as well as classifications of lung HRCT with and without the use of explicit textural parameters have been applied in preliminary studies. In the present study a hybrid network of three single nets and an expert rule was applied for the detection of GG on 120 HRCT scans from 20 patients suffering from different lung diseases. Single nets alone were not capable to reliably detect or exclude GG since the false-positive rate was greater than 100 % with regard to the area truly involved, more than 50 pixels throughout, and the true-positive rate was greater than 95 %. The hybrid network correctly classified 91 of 120 scans. Mild GG was false positive in 15 cases with less than 50 pixels, which was judged not clinically relevant. The pitfalls were: partial volume effects of bronchovascular bundles and the chest wall. Motion artefacts and diaphragm were responsible for 11 misclassifications. Hybrid networks represent a promising tool for an automatic pathology-detecting system. They are ready to use as a diagnostic assistant for detection, quantification and follow-up of ground glass opacities, and further applications are underway.