The authors report their experience with the use of Tc-99m DTPA aerosol following a perfusion lung scan. The study includes 422 consecutive patients with suspected pulmonary embolism. The final diagnosis was determined by (a) clinical follow-up for 2 months or more, (b) pulmonary angiography, or (c) autopsy. There were 79 patients (19%) who had a normal or near-normal perfusion study and in whom no aerosol study was required. Interpretation groups were classified and divided as follows: Normal or low probability 281 (66.5%) High probability 75 (18.0%) Intermediate 60 (14.0%) Technically inadequate 6 (1.5%) There was autopsy or angiographic confirmation of 72 patient studies with confirmation of the scan diagnosis in 29 of 31 classified as normal or low probability and 24 of 25 classified as high probability. The background perfusion albumin activity was not computer-subtracted from the combined aerosol-perfusion image. Technical improvements included the use of eight standard views for both the perfusion and the subsequent aerosol scan, and the use of 75 mCi (2,775 MBq) of Tc-99m DTPA in the radioaerosol nebulizer. This allowed for easy accumulation of more than three times the count rate in a posterior aerosol image when compared with the previous posterior perfusion image. The authors' experience shows that the perfusion lung scan followed by this radioaerosol technique is a reliable means to evaluate suspected pulmonary embolism.