Computer-based diagnostic tools may serve as valid second opinions for pinpointing pulmonary embolism, according to a study in the August issue of the American Journal of Roentgenology.
Dr. James Scott and his co-authors from Massachusetts General Hospital in Boston used computers running artificial neural network-based algorithms to process data obtained from anterior and posterior ventilation and perfusion images, and compared the results to human observation.
The computer was pitted against three nuclear medicine specialists, using digital data from 100 patients with normal findings on chest radiographs who also had undergone radionuclide ventilation-perfusion scanning and pulmonary angiography.
According to the results, the computer and clinicians correctly identified the majority of cases of large pulmonary emboli. The authors concluded that artificial neural networks can be helpful, adjuvant diagnostic tools, particularly for training purposes.
By AuntMinnie.com staff writers
July 20, 2000
Related Reading
Can the utility of V/Q scans for embolism be improved?, February 16, 2000
New CAD scheme for lung nodules reduces false positives, July 4, 2000
Copyright © 2000 AuntMinnie.com
