The differentiation of pulmonary embolism (PE) from regional ventilatory abnormalities accompanied by reduced perfusion requires contemporary perfusion and ventilation studies. Distinguishing these conditions in ventilator-dependent patients is aided by administering a Tc-99m aerosol to characterize regional ventilation, and by performing a conventional Tc-99m MAA perfusion study. The technique uses a simple "in-house" constructed apparatus. Simple photographic techniques suffice, but computer subtraction of perfusion from the combined perfusion-ventilation image renders interpretation easier if aerosol administration follows perfusion imaging. Multiple defects can be examined in a single study. Excluding normal or near-normal perfusion studies, PE was thought to be present in eight of 16 patients after perfusion imaging alone, but in only one of eight after added aerosol imaging. Angiography confirmed the diagnosis in that patient. Of the eight patients who had abnormal perfusion but were thought unlikely to have PE from the perfusion study alone, two had normal ventilation, and subsequently were shown to have PE by angiography. Because angiography was only performed on patients who were thought to have a high probability of PE on sequential perfusion-ventilation imaging, the true incidence of PE may have been higher. Aerosol ventilation imaging is a useful adjunct to perfusion imaging in patients on ventilators. It requires an efficient delivery system, particularly if aerosol administration follows perfusion imaging, as it does in this study. The major disadvantage of aerosol imaging compared to a gas in intubated patients is the significant bronchial deposition due to retained mucus secretions.(ABSTRACT TRUNCATED AT 250 WORDS)