Gd. Tourassi et al., IMPROVED NONINVASIVE DIAGNOSIS OF ACUTE PULMONARY-EMBOLISM WITH OPTIMALLY SELECTED CLINICAL AND CHEST RADIOGRAPHIC FINDINGS, Academic radiology, 3(12), 1996, pp. 1012-1018
Rationale and Objectives. The authors improved the noninvasive diagnos
is of acute pulmonary embolism (PE) by studying the clinical and chest
radiographic findings of patients suspected of having PE and correlat
ing those findings with the physicians' clinical impression. Methods.
A stepwise linear discriminant algorithm was developed on the basis of
1,064 patients from the Prospective Investigation of Pulmonary Emboli
sm Diagnosis (PIOPED) study to select clinical and chest radiographic
findings with the highest diagnostic power in patients suspected of ha
ving PE. Subsequently, a linear classifier and a nonlinear artificial
neural network were developed to help diagnose PE on the basis of the
reduced number of findings. Results. Both classifiers produced a stati
stically significant improvement (A(2) = 0.77 +/- 0.02) in the clinica
l performance of the PIOPED physicians (A(2) = 0.72 +/- 0.02). Results
are also presented separately for groups of patients classified on th
e basis of the difficulty lever of their ventilation-perfusion lung sc
ans. Conclusion. Two computer-aided diagnostic tools were developed to
assist physicians in the assessment of the pretest likelihood of PE b
y using an optimally reduced number of findings.