Measurement of the clearance rate of inhaled aerosols of ^99mTc-diethylenetriamine pentaacetic acid (DTPA) from distal airway to pulmonary capillary is a sensitive technique for the detection of lung injury. As the solute diffuses across the blood-gas barrier, the concentration in circulating blood increases, giving rise to a background signal superimposed on the signal from residual DTPA in the airway. Background subtraction is conventionally based on the thigh, but this tissue has the disadvantage in that its composition, in terms of the relative volumes of its extracellular extravascular and intravascular compartments (a ratio of ~4:1), is quite different from that of the lung (<1:6). With comparison to the thigh, we examined alternative regions for background, liver, and cranium, which have extravascular-to-intravascular compartment ratios much closer to these for the lung, to determine the most appropriate background for correction of the pulmonary signal. From 1 min after intravenous injection of ^99mTc-DTPA, the time-activity curves recorded by a gamma camera over the liver and lung in a group of otherwise normal cigarette smokers decreased up to 30 min after injection, with time courses that could essentially be superimposed on each other; the curve recorded over the thigh with a separate scintillation probe continued to increase. The curve recorded over the cranium had a time course similar to that for the liver and lung. Following aerosol inhalation, the lung clearance rates over the initial 7 min when background subtraction was used, based on the liver, cranium, and thigh were, respectively, 4.9 ± 2.9, 4.7 ± 2.6, and 5.4 ± 3.4 (SD) %/min, compared with 4.1 ± 2.2%/min without subtraction. The corresponding values based on 30 min of data were 3.3 ± 1.4, 3.4 ± 1.4, 4.2 ± 2.3, and 2.8 ± 1.0%/min. When the liver was used for background, the lung clearance curves were clearly multiexponential, whereas thigh correction tended to give curves that were monoexponential or even convex upward on semilogarithmic axes. With an appropriate region for background, the true shape of a lung curve can be identified, which permits the study of an intervention on the clearance while it is in progress. The intravenous DTPA, required for calibrating the background regions, can be given before inhalation of the tracer.