The EuroNet paediatric hodgkin network - modern imaging data management for real time central review in multicentre trials.

Since 2007, children and adolescents with Hodgkin lymphomas are treated in the Europe-wide EuroNet-PHL trials. A real time central review process for stratification of the patients enhances quality control and efficient therapy management. This process includes reading of all cross-sectional-images. Since reference evaluation is time critical, a fast, easy to handle and safe data transfer is important. In addition, immediate and constant access to all the data has to be guaranteed in case of queries and for regulatory reasons. To meet the mentioned requirements the EuroNet Paediatric Hodgkin Data Network (funded by the European Union - Project Number: 2007108) was established between 2008 and 2011. A respective tailored data protection plan was formulated. The aim of this article is to describe the networks' mode of operation and the advantages for multi-centre trials that include centralized image review.

Whole-body MRI for primary evaluation of malignant disease in children.

OBJECTIVE: The aim of this study was to compare diagnostic accuracy of whole-body (WB) MRI to a combined reference standard of conventional cross-sectional imaging methods and FDG-PET in the detection of malignant disease spread in children. MATERIALS AND METHODS: 24 children (age between 5 and 18 years) with malignant diseases (mainly Hodgkin's lymphoma and different types of sarcoma) initially examined with conventional cross-sectional imaging methods (ultrasound, computed tomography, or magnetic resonance imaging) were examined prospectively with whole-body MRI (1.5T) and FDG-PET. Studies were read by two nuclear medicine physicians (FDG-PET) and two radiologists (WB-MRI) independently in a blinded manner and each study type was evaluated in consensus. The reference standard was defined as pathological lesions detected in the same location both in FDG-PET and another conventional cross-sectional imaging method. RESULTS: Overall 190 lesions were detected by WB-MRI and 155 lesion were found by FDG-PET. 106 lesions fulfilled the criteria of the reference standard (42 osseous and 64 extraosseous lesions) from which 102 were detected by WB-MRI (sensitivity of 96%). All bone lesions were detected and extra-skeletal lesions were identified with a sensitivity of 93.8%. Overall 88 lesions detected by WB-MRI were not part of the reference standard from which 33 were lesions of the peripheral skeleton not imaged by conventional cross-sectional imaging studies. 4 lesions of the reference standard were not identified by WB-MRI which were all lymph nodes. CONCLUSION: WB-MRI is a radiation free imaging technique with high sensitivity for the detection of malignant disease spread in particular beneficial for children. In patients with suspected bone lesions it should be considered for initial disease evaluation prior to specific and regional imaging methods to reduce the overall number of imaging examinations and radiation exposure.

[RadLex - German version: a radiological lexicon for indexing image and report information]. / RadLex - deutsche Version: ein radiologisches Lexikon zur Indexierung von Bild- und Befunddaten.

PURPOSE: Since 2003 the Radiological Society of North America (RSNA) has been developing a lexicon of standardized radiological terms (RadLex) intended to support the structured reporting of imaging observations and the indexing of teaching cases. The aim of this study was to translate the first version of the lexicon (1 - 2007) into German and to implement a language-independent online term browser. MATERIALS AND METHODS: RadLex version 1 - 2007 contains 6303 terms in nine main categories. Two radiologists independently translated the lexicon using medical dictionaries. Terms translated differently were revised and translated by consensus. For the development of an online term browser, a text processing algorithm called morphosemantic indexing was used which splits up words into small semantic units and compares those units to language-specific subword thesauri. RESULTS: In total 6240 of 6303 terms (99 %) were translated. Of those terms 3965 were German, 1893 were Latin, 359 were multilingual, and 23 were English terms that are also used in German and were therefore maintained. The online term browser supports a language-independent term search in RadLex (German/English) and other common medical terminology (e. g., ICD 10). The term browser displays term hierarchies and translations in different frames and the complexity of the result lists can be adapted by the user. CONCLUSION: RadLex version 1 - 2007 developed by the RSNA is now available in German and can be accessed online through a term browser with an efficient search function. This is an important precondition for the future comparison of national and international indexed radiological examination results and the interoperability between digital teaching resources.

Automatic detection and quantification of ground-glass opacities on high-resolution CT using multiple neural networks: comparison with a density mask.

OBJECTIVE: We compared multiple neural networks with a density mask for the automatic detection and quantification of ground-glass opacities on high-resolution CT under clinical conditions. SUBJECTS AND METHODS: Eighty-four patients (54 men and 30 women; age range, 18-82 years; mean age, 49 years) with a total of 99 consecutive high-resolution CT scans were enrolled in the study. The neural network was designed to detect ground-glass opacities with high sensitivity and to omit air-tissue interfaces to increase specificity. The results of the neural network were compared with those of a density mask (thresholds, -750/-300 H), with a radiologist serving as the gold standard. RESULTS: The neural network classified 6% of the total lung area as ground-glass opacities. The density mask failed to detect 1.3%, and this percentage represented the increase in sensitivity that was achieved by the neural network. The density mask identified another 17.3% of the total lung area to be ground-glass opacities that were not detected by the neural network. This area represented the increase in specificity achieved by the neural network. Related to the extent of the ground-glass opacities as classified by the radiologist, the neural network (density mask) reached a sensitivity of 99% (89%), specificity of 83% (55%), positive predictive value of 78% (18%), negative predictive value of 99% (98%), and accuracy of 89% (58%). CONCLUSION: Automatic segmentation and quantification of ground-glass opacities on high-resolution CT by a neural network are sufficiently accurate to be implemented for the preinterpretation of images in a clinical environment; it is superior to a double-threshold density mask.