Attenuation correction in pulmonary and myocardial single photon emission computed tomography

Sammanfattning: The objective was to develop and validate methods for single photon emission computed tomography, SPECT, allowing quantitative physiological and diagnostic studies of lung and heart. A method for correction of variable attenuation in SPECT, based on transmission measurements before administration of an isotope to the subject, was developed and evaluated. A protocol based upon geometrically well defined phantoms was developed. In a mosaic pattern phantom count rates were corrected from 39-43% to 101-110% of reference. In healthy subjects non-gravitational pulmonary perfusion gradients, observed without attenuation correction were artefacts caused by attenuation. Pulmonary density in centre of right lung, obtained from the transmission measurement, was 0.28 ± 0.03 g ml-1 in normal subjects. Mean density was lower in large lungs compared to smaller ones. We also showed that regional ventilation/perfusion ratios could be measured with SPECT, using the readily available tracer 133Xe. Because of the low energy of 133Xe this relies heavily upon attenuation correction. A commercially available system for attenuation correction with simultaneous emission and transmission, considered to improve myocardial SPECT, performed erroneously. This could lead to clinical misjudgement. We considered that manufacturer-independent pre-clinical tests are required. In a test of two other commercial systems, based on different principles, an adapted variant of our initial protocol was proven useful. Only one the systems provided corrected emission count rates independently on phantom configuration. Errors in the other system were related to inadequate compensation of the influence of emission activity on the transmission study.