Automated interpretation of ventilation-perfusion lung scintigrams for the diagnosis of pulmonary embolism using artificial neural networks.

The purpose of this study was to develop a completely automated method for the interpretation of ventilation-perfusion (V-P) lung scintigrams used in the diagnosis of pulmonary embolism. An artificial neural network was trained for the diagnosis of pulmonary embolism using 18 automatically obtained features from each set of V-P scintigrams. The techniques used to process the images included their alignment to templates, the construction of quotient images based on the ventilation and perfusion images, and the calculation of measures describing V-P mismatches in the quotient images. The templates represented lungs of normal size and shape without any pathological changes. Images that could not be properly aligned to the templates were detected and excluded automatically. After exclusion of those V-P scintigrams not properly aligned to the templates, 478 V-P scintigrams remained in a training group of consecutive patients with suspected pulmonary embolism, and a further 87 V-P scintigrams formed a separate test group comprising patients who had undergone pulmonary angiography. The performance of the neural network, measured as the area under the receiver operating characteristic curve, was 0.87 (95% confidence limits 0.82-0.92) in the training group and 0.79 (0.69-0.88) in the test group. It is concluded that a completely automated method can be used for the interpretation of V-P scintigrams. The performance of this method is similar to others previously presented, whereby features were extracted manually.

Nonlinear diffusion filtering of images obtained by planar laser-induced fluorescence spectroscopy.

The application of nonlinear anisotropic diffusion filtering to reduce noise and enhance contours in images obtained by two-dimensional planar laser-induced fluorescence (PLIF) spectroscopy is presented. In this process the diffusion coefficient is locally adapted, becoming negligible as object boundaries are approached. Noise is efficiently removed, and object contours are strongly enhanced. The technique is demonstrated with PLIF images obtained from the OH radical recorded in turbulent flames. We show that nonlinear diffusion is suitable as a preprocessing step, before image segmentation becomes possible, and we demonstrate how the technique is applied for the quantitative extraction of flame reaction boundaries from PLIF data.

A new radiographic method for evaluating the degree of sliding in devices used in hip-fracture surgery.

PURPOSE: To find a practical method for estimating the degree of sliding that occurs in screw-plate devices used in hip-fracture surgery. Greater understanding of the sliding mechanisms in different fracture types should improve surgical technique and reduce the failure rate. METHODS AND RESULTS: In dynamic screw-plate devices, the lag screw slides inside the barrel of the plate. A recent innovation allows the barrel-plate to slide inside a side-plate, thus making possible a combined fracture compression along the neck and the shaft of femur. The lengths of the different parts and the angle of a device in vivo, measured on a radiograph, depend on the position of the femur relative to the photographic film and the roentgen source. We obtained these measurements with a ruler and a protractor from sequential a.p. radiographs of the hip and implemented them in a special computerized program that used the principles of the scaled orthographic and the central projection models. These calculations provided the correct amount of sliding by the lag screw and by the barrel-plate within the side-plate. CONCLUSION: The method presented here can establish the real degree of sliding in screw-plate devices from standard a.p. radiographs independently of the position of the hip.

Ultrasound Doppler vector tomography measurements of directional blood flow.

An experimental system has been developed to verify the possibility of detecting flow activity using a technique called ultrasound Doppler vector tomography. A tomography algorithm is used to reconstruct blood flow vector fields using data from computer-controlled ultrasound continuous-wave Doppler scanning equipment. The result is a picture in which the brightness variations represent the reconstructed values of the curl of the velocity field (inverted delta x v). Continuous ultrasound is transmitted into a region with flow activity and the Doppler-shifted signals are received. To obtain measurement data suited for fan beam tomography, the scanning is performed in a plane from points encircling the region. Reconstructions have been achieved using measurement data from two different flow phantoms. A comparison between the experimental results and simulations shows good conformity.