Results 1 - 10 of 26473
Results 1 - 10 of 26473. Search took: 0.038 seconds
|Sort by: date | relevance|
[en] The assessment of simulated scintigraphic pictures by means of curves based on binary decisions of individual observers called pooled ROC-curves of pooled L-ROC-curves is presented. The technique of creating the curve is described. The examples of various processing techniques and the findings of various groups of observers are compared and discussed. The images of two IAEA simulated phantoms were used to perform the evaluation. The results indicate that the pooled L-ROC-curves may serve as a measure of image quality. (orig.) 891 MG
[de]Es wird die Auswertung simulierter Szintigraphiebilder durch Kurven beschrieben, die auf binaeren Entscheidungen einzelner Beobachter basieren und als 'pooled ROC'-Kurven oder 'pooled L-ROC'-Kurven bezeichnet werden. Die Methode der Herstellung solcher Kurven wird beschrieben. Beispiele verschiedener Verarbeitungsmethoden und Befunde verschiedener Beobachtergruppen werden verglichen und diskutiert. Zur Auswertung wurden Bilder zweier simulierter Phantome der IAEO verwendet. Die Ergebnisse zeigen, dass die 'pooled L-ROC'-Kurven als Mass fuer die Qualitaet des Bildes dienen koennen. (orig.)
[en] Aim: To investigate the advantage of digital tomosynthesis (DTS) over chest radiography (CXR) and dual-energy subtraction radiography (DES) for pulmonary nodule detection according to the location and size of solid simulated pulmonary nodules (SPNs). Materials and methods: Ninety-six SPNs of variable sizes were inserted into eight different regions of a lung phantom. These regions were further classified into two groups of danger and non-danger zones based on anatomical location influencing the detection of pulmonary nodules. The 96 cases with inserted SPNs and an additional nodule-free 96 control cases all underwent CXR, DES, and DTS examinations. Three observers independently reviewed all the images. The jackknife alternative free-response receiver operating characteristic was used to analyse diagnostic performance for each technique. Results: DTS was superior to CXR and DES for detection of smaller SPNs, except in the retrodiaphragmatic and apical regions. DTS outperformed CXR and DES for detection of larger SPNs in the paramediastinal region. For 5- and 8-mm SPNs, DTS was superior to CXR and DES in the apical, paramediastinal and lateral pulmonary regions. In the retrodiaphragmatic region, the three techniques showed similar diagnostic performance regardless of the SPN size. DES was similar to DTS for detection of 8-mm SPN in the apical region. For 10- and 12-mm SPNs, CXR and DES showed similar diagnostic performance to DTS in the apical and lateral pulmonary regions; however, DTS was superior to CXR and DES in the paramediastinal region. Conclusions: DTS significantly improved the capability to detect synthetic pulmonary nodules compared with CXR and DES, for detection of smaller nodules in the apical, paramediastinal, and lateral pulmonary regions, and larger nodules located in the paramediastinal region in a thoracic phantom. - Highlights: • The location and size of simulated pulmonary nodule affects observer's detection rate of SPN. • DTS improves detectability of smaller simulated pulmonary nodule in non-danger and some danger zones. • DTS improves detectability of larger simulated pulmonary nodule in some danger zones. • CXR and DES show similar performance to DTS for detecting retrodiaphragmatic simulated pulmonary nodule.
[en] A phantom is described in which the authors have tried to have all parameters continuously adjustable and to have a phantom applicable for both steady-state ventriculography and first-pass bolus angiocardiography. In a pulsating balloon arrangement, the conformity of all parameters determined by radionuclide ventriculography was verified, and the calibration of the geometrical method for end-diastolic volume was confirmed. In the pumping mode, the accuracy in determination of the cardiac output and various aspects of combined angiocardiography and ventriculography were verified, in particular the possible ways of assessing the absorption coefficients for the sample (count basis) method of calculation of the end-diastolic volume of the ventricle. (U.K.)
[en] Thirty two head CT scans were acquired employing an anthropomorphic phantom which contains small lesions in posterior fosse, using 2 scanners, one with 64 slices 'Siemens Sensation' and other with 6 slices Philip Brilliance. Parameters as Tube current (I [mA]), Collimation (C [mm]), spectrum energy (kVp) and dynamic range were changed during studies, looking for the optimal acquisition/processing conditions which permit both, good lesion detectability and the lowest dose. Air kerma (mGy) was measured with a ionization chamber and the air kerma index (Ca,100 [mGy]) was calculated as dose index. Image quality was analyzed by 5 expert criteria using a 5 points-scale (1=poor, 2=fair, 3=good, 4=very good, 5=excellent) and also using 5 Figure of merit in the spatial and frequency domains: Contrast C(%), Contrast to Noise Ratio CNR, Signal to Noise Ratio SNR, Normalized Mean Square Error (NMSE) and the Spectral Distance (SD). Objective and subjective results were correlated. We observed that doses can be diminished until a 25% respect to the usual practice with both technologies, diminishing mainly the mAs, without affecting lesion detection. As a result, we propose an optimized protocol for each scanner as follow: The use of 250 mAs, 120 kVp and the collimation of 6 slices x 1.50 mm per rotation to detect the lesions in posterior fosse with good image quality for the Philips Brilliance 6 tested, while 150 mAs, 100 kVp and slice thickness of 3 mm were needed with the Siemens Sensation 64. (Author)
[en] A series of experiments was undertaken to evaluate the response of a positron emission transverse tomograph (PETT) to measured radionuclide concentrations similar to those encountered in human studies. The correlation between the response of the imaging system (mean PETT) number/min), and the concentration of the radioactivity producing the output data, was linear with a computed sensitivity of 2720 PETT number/min, per μCi/ml, per picture element, for a radionuclide (100% β+) contained in either of two phantoms and imaged with a resolution of 1.5 cm. It was concluded that the output data are essentially independent of the imaged object's physical dimensions for the range of 18 to 28-cm diam and faithfully reflect the regional radioactivity concentration within the object, provided valid attenuation correction is achieved and the sampled area is not compromised by the imaging system's limitations of spatial resolution
[en] A phantom is presented which permits the evaluation of streaking artifacts produced in CT reconstructions by abrupt edges. Its application is demonstrated by results obtained from nine CT scanners. It is observed that even in regions where streaking is not readily apparent, edge-induced artifacts can significantly increase the variance in the reconstruction
[en] Ultrasound strain imaging is used to measure local tissue deformations. Usually, only strains along the ultrasound beam are estimated, because those estimates are most precise, due to the availability of phase information. For estimating strain in other directions we propose to steer the ultrasound beam at an angle, which allows estimating different projections of the 2D strain tensor, while phase information remains available. This study investigates beam steering at maximally three different angles to determine the full 2D strain tensor. The method was tested on simulated and experimental data of an inclusion phantom and a vessel phantom. The combination of data from a non-steered acquisition and acquisitions at a large positive and an equally large but negative steering angle enabled the most precise estimation of the strain components. The method outperforms conventional methods that do not use beam steering.
[en] In the tomography of the lumbar part of the spine circular blurring of the interfering shadows allows the structure of the layer to be examined more distinctly than linear blurring. The spiral movement of the tomographic system, despite its more complicated tra ectory, allows producing an image similar to a zonographic one