PURPOSE: To determine the diagnostic accuracy of bedside chest radiography in patients who develop severe pulmonary complications while undergoing therapy for leukemia. MATERIALS AND METHODS: The authors studied 45 patients, aged 21 years and younger, who died of leukemia or of treatment complications and for whom autopsy findings were available. Pulmonary findings at autopsy comprised pneumonia (n = 25), adult respiratory distress syndrome (ARDS) (n = 16), hemorrhage (n = 38), infarction (n = 18), and leukemic cellular infiltration (n = 11). Four radiologists who were unaware of the autopsy diagnoses independently rated antemortem bedside chest radiographs. RESULTS: Diagnostic accuracy for each disease was as follows: ARDS, 0.81 +/- .03 (standard error); all pneumonias, 0.56 +/- .04; hemorrhage, 0.47 +/- .07; infarction, 0.50 +/- .12; and pulmonary leukemic cellular infiltration, 0.38 +/- .12. CONCLUSION: The specific radiographic appearance of ARDS permits excellent diagnostic accuracy.
PURPOSE: To determine the diagnostic accuracy of bedside chest radiography in patients who develop severe pulmonary complications while undergoing therapy for leukemia. MATERIALS AND METHODS: The authors studied 45 patients, aged 21 years and younger, who died of leukemia or of treatment complications and for whom autopsy findings were available. Pulmonary findings at autopsy comprised pneumonia (n = 25), adult respiratory distress syndrome (ARDS) (n = 16), hemorrhage (n = 38), infarction (n = 18), and leukemic cellular infiltration (n = 11). Four radiologists who were unaware of the autopsy diagnoses independently rated antemortem bedside chest radiographs. RESULTS: Diagnostic accuracy for each disease was as follows: ARDS, 0.81 +/- .03 (standard error); all pneumonias, 0.56 +/- .04; hemorrhage, 0.47 +/- .07; infarction, 0.50 +/- .12; and pulmonary leukemic cellular infiltration, 0.38 +/- .12. CONCLUSION: The specific radiographic appearance of ARDS permits excellent diagnostic accuracy.