Quantifying Local Lung Perfusion and Ventilation Using Correlated SPECT and CT Data

• Published in: The Journal of nuclear medicine (1978) Vol. 35; no. 5; pp. 784 - 792
• Main Authors: Damen, Eugene M.F, Muller, Sara H, Boersma, Liesbeth J, de Boer, Roel W, Lebesque, Joos V
• Format: Journal Article
• Language: English
• Published: United States Soc Nuclear Med 01-05-1994
• Subjects: Algorithms; Humans; Krypton Radioisotopes; Lung - diagnostic imaging; Lung - physiopathology; Lung Diseases - diagnostic imaging; Lung Diseases - physiopathology; Stochastic Processes; Technetium; Tomography, Emission-Computed, Single-Photon; Tomography, X-Ray Computed; Ventilation-Perfusion Ratio
• ISSN: 0161-5505; 1535-5667

A clinically applicable method for quantifying lung perfusion and ventilation on a subregional (local) scale from SPECT scans in order to estimate local lung function in patients with pre-existing pulmonary disease and to monitor local treatment effects was developed and evaluated.
SPECT 99mTc perfusion and 81mKr ventilation images were corrected for photon attenuation and scatter effect with a postreconstruction correction method incorporating a variable-effective linear-attenuation coefficient calculated from spatially-correlated CT data. A new algorithm was developed to quantify local ventilation from the SPECT data, which, in contrast with other algorithms, makes no assumptions on ventilation homogeneity over the lung. The quantification procedure was applied to clinical data from patients with a normal lung function and from patients suffering from radiation-induced pulmonary dysfunction.
The calculated attenuation correction factors on the observed number of counts in the lung range from 2.0 to 3.0 and 2.3 to 3.5 for 81mKr and 99mTc, respectively, showing a systematic increase from the diaphragm to the lung apex. As a result of this correction, the values of local perfusion and ventilation differ 10%-15% from values calculated without attenuation correction. The calculated values of the local ventilation are 10%-50% lower than those found by quantification algorithms which assume homogeneous ventilation.
The methods presented here are robust with respect to uncertainties in the input parameters and yield realistic values for perfusion and ventilation distribution in the lung with an intrinsic accuracy (largely determined by count statistics) of about 10%.