[Second reading of breast imaging at the hospital department of radiology: reasonable or waste of money?]. / Mammadiagnostische Zweitbefundung in der Radiologischen Klinik: sinnvoller Mehraufwand oder Ressourcenverschwendung?

PURPOSE: To assess the effect of a second diagnostic reading of breast imaging at a university department of radiology. MATERIAL AND METHODS: The diagnostic reports of first readers from different private radiology practices and the reports of second readers from the university department of radiology were compared with the histological results (n = 214) and outcome of follow-ups for 4 years (n = 74) in 236 patients (mean age 55 years). BI-RADS categories were used for this purpose. The initial examinations had been performed because of symptoms (n = 117), early detection outside an organized screening program (n = 102), evaluations following breast cancer therapy (n = 13) and unknown primary tumors (n = 4). In addition, the number of complementary examinations and the influence of a second reading on patient management were evaluated. RESULTS: A total of 140 lesions were malignant and 148 were benign. Of the 288 lesions, 49 % were classified identically in the second reading; 36 % (79/217) of the lesions initially classified as BI-RADS 4 and 5 were downgraded to benign; and 41 % (29/71) of the lesions classified as BI-RADS 1 to 3 were upgraded as suspected of being malignant. The kappa value between the first and second readers was 0.34 with respect to each BI-RADS category separately and 0.18 with respect to categories 1 to 3 (benign) versus 4 and 5 (malignant). A second reading increased the sensitivity from 81 % (114/140) to 96 % (135/140) and the specificity from 30 % (45/148) to 78 % (116/148). Second readers detected 23 additional malignant lesions, changed two lesions correctly classified as malignant to benign categories and caused 6 additional false-positive findings. In all, 38 surgical biopsies could be prevented, one biopsy was erroneously delayed and three unnecessary excisional biopsies were initiated. In the study group, 49 MRI examinations revealed 5 additional malignant lesions and positively influenced surgical planning in 28 patients. In addition to these MRI examinations, 221 patients were examined using ultrasonography, 62 patients using complementary mammography, and 7 using galactography. CONCLUSION: The benefit of a second reading outweighs the expenditure.

[Contribution of early systole to total antegrade flow volume in breath-hold phase-contrast flow measurements]. / Anteil des frühen systolischen Flussanstiegs am antegrad fliessenden Gesamtvolumen bei Phasenkontrast-Flussmessungen in Atemanhaltetechnik.

PURPOSE: To evaluate the contribution of early systole for the assessment of antegrade aortic flow volume by breath-hold velocity-encoded magnetic resonance (MR) flow measurements. MATERIALS AND METHODS: Expiratory breath-hold fast low-angle shot (FLASH) phase-contrast flow measurements (temporal resolution tRes 61 msec, shared phases) perpendicular to the proximal ascending aorta and short axis true fast imaging with steady-state precession (TrueFISP) cine MR ventriculometry (tRes 34.5 msec) were performed in ten healthy male volunteers on a 1.5 T MR system (Sonata, Siemens Medical Solutions). Antegrade aortic flow volume (AFV) and left ventricular stroke volume (LV-SV) were evaluated using Argus Ventricular Function and Argus Flow Software, version MR 2002B (Siemens Medical Solutions). A beta release of Argus Flow MR 2004A allowed interpolation of the flow up-slope during early systole to the preceding R-wave trigger. The respective intraindividual median differences between the AFV of each flow evaluation and LV-SV as well as between both AFV measurements were calculated and compared using the sign test for paired samples. RESULTS: Non-interpolated AFV significantly deviated from LV-SV (p = 0.006), underestimating the latter by 13.1 mL (13 %). Interpolating aortic flow during early systole significantly increased AFV by 10.8 mL (13 %) compared to the flow evaluation which did not take early systole into account (p = 0.006). AFV with interpolation of early systolic flow agreed well with LV-SV (median difference - 3.0 mL or - 3 %, respectively), and no significant difference between these measurements was found (p = 1.0). CONCLUSION: Flow during early systole contributes substantially to total forward flow volume in the ascending aorta. Interpolation of the early systolic up-slope is therefore recommended for the evaluation of breath-hold phase-contrast flow measurements.

[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.

Improved artificial neural networks in prediction of malignancy of lesions in contrast-enhanced MR-mammography.

The aim of this study was to evaluate the capability of improved artificial neural networks (ANN) and additional novel training methods in distinguishing between benign and malignant breast lesions in contrast-enhanced magnetic resonance-mammography (MRM). A total of 604 histologically proven cases of contrast-enhanced lesions of the female breast at MRI were analyzed. Morphological, dynamic and clinical parameters were collected and stored in a database. The data set was divided into several groups using random or experimental methods [Training & Testing (T&T) algorithm] to train and test different ANNs. An additional novel computer program for input variable selection was applied. Sensitivity and specificity were calculated and compared with a statistical method and an expert radiologist. After optimization of the distribution of cases among the training and testing sets by the T & T algorithm and the reduction of input variables by the Input Selection procedure a highly sophisticated ANN achieved a sensitivity of 93.6% and a specificity of 91.9% in predicting malignancy of lesions within an independent prediction sample set. The best statistical method reached a sensitivity of 90.5% and a specificity of 68.9%. An expert radiologist performed better than the statistical method but worse than the ANN (sensitivity 92.1%, specificity 85.6%). Features extracted out of dynamic contrast-enhanced MRM and additional clinical data can be successfully analyzed by advanced ANNs. The quality of the resulting network strongly depends on the training methods, which are improved by the use of novel training tools. The best results of an improved ANN outperform expert radiologists.

[Automatic classification of liver segments according to Couinaud: development of a new algorithm and evaluation spiral CT data]. / Automatische Klassifikation der Lebersegmente nach Couinaud: Entwicklung eines neuen Algorithmus und Evaluierung an Spiral-CT-Datensätzen.

PURPOSE: To develop a software tool that analyzes the anatomy of the portal vein branches and assigns segmental and subsegmental branches according to Couinaud's classification system and to evaluate its accuracy. MATERIALS AND METHODS: The algorithm was developed in C++ on a PC. The algorithm recognizes the three major branching patterns of the portal vein. Segmental and subsegmental branches are assigned to 8 segments following Couinaud and encoded by 8 colors. The software was evaluated using CT data sets of 39 patients. After the individual segmental anatomy of each patient was determined by an experienced radiologist, automatic classification was performed and the results were compared on a branch by branch basis. RESULTS: The numbering was accurate according to Couinaud's system in 358 of 409 segmental and subsegmental branches (88 %). The assignment failed in 51 of 409 branches due to unexpected anatomy or software problems. CONCLUSION: Automatic classification of portal vein branches and their appendant parenchyma is possible. The automatic designation of liver segments enables the three-dimensional visualization of the segmental anatomy. In the future, automatic analysis might facilitate the reporting and communication of CT findings.

[Combination of low and high resolution T1-weighted sequences for improved evaluation of morphologic criteria in dynamic contrast enhanced MRI of the breast]. / Eine Kombination niedrig und hochauflösender dynamischer T1-gewichteter Sequenzen zur besseren Beurteilung der Morphologie Kontrastmittel aufnehmender Läsionen in der MRT der weiblichen Brust.

PURPOSE: Presentation of a new protocol for simultaneous acquisition of both low and high resolution T 1 -weighted images of breast lesions for dynamic contrast-enhanced MR mammography. Demonstration of possible diagnostic improvement with representative measurements in patients with suspected breast cancer by adding morphologic parameters from high resolution sequences to the analysis of the signal-time curve. MATERIALS AND METHODS: Dynamic MR imaging was performed with a 1.5 T system (Magnetom SONATA, Siemens Medical Systems, Germany) and the manufacturer's double-breast coil. Coronal T 1 -weighted 3D FLASH sequences (spatial resolution 1.25 x 1.25 mm 2; slice thickness 1.7 mm) were acquired once before and five times after administration of contrast medium (Gd-DTPA, 0.15 mmol/kg) injection. In addition, a high resolution T 1 -weighted 3D-FLASH sequence (spatial resolution, 0.63 x 0.63 mm 2) was obtained before administration of contrast medium and after the third post-contrast low-resolution sequence. Except for the acquisition matrix, all imaging parameters were identical for both 3D pulse sequences. To assure comparison of the measured signal intensities for both T 1 -weighted sequences, calibrating phantom measurements were performed using a dilution series of Gd-DTPA. RESULTS: Phantom measurements demonstrated similar signal intensities and enhancement pattern for both sequences. A combined protocol consisting of both pulse sequences can be employed and does not interfere with the signal-time curve analysis. By measuring one high resolution sequence 3:18 minutes after administration of contrast medium, morphologic features can be evaluated without interference from barely enhancing surrounding tissue. The overall study time is not increased. The improved spatial resolution slightly increases the severity of motion artifacts. CONCLUSION: The new protocol is a clever way to improve the measurement of morphologic features without relevant loss of dynamic information. It is superior to converting the entire investigation to high resolution sequences and does not add any costs by not extending or duplicating the investigation. How much the new protocol can improve the specificity or sensitivity of MR-mammography is currently investigated on a larger patient group.