Various diagnostic strategies for the assessment of pulmonary embolism were developed using the results of scintigraphic examinations in 169 patients who had both scintigraphic studies and pulmonary angiography. Likelihood ratios for several perfusion and ventilation-perfusion patterns were first obtained, and Bayes' Theorem was then applied using a range of prior probabilities for pulmonary embolism. Resulting posterior probabilities ranged from under 10% to over 90%. The former values held for patients with small perfusion defects and no ventilation study, regardless of their prior probabilities for pulmonary embolism. The latter values held for patients with average to high prior probabilities and ventilation-perfusion mismatches. (April 1980)
Journal of Nuclear Medicine
1980
http://jnm.snmjournals.org/content/21/4/319.long