Deconvolution analysis in radionuclide quantitation of left-to-right cardiac shunts.

A poor bolus injection results in an unsatisfactory quantitative radionuclide angiocardiogram in as many as 20% of children with possible, left-to-right (L-R) cardiac shunts. Deconvolution analysis was applied to similar studies in experimental animals to determine whether dependence on the input bolus could be minimized. Repeated good-bolus, prolonged (greater than 2.5 sec), or multiple-peak injections were made in four normal dogs and seven dogs with surgically created atrial septal defects (ASD). QP/QS was determined using the gamma function. The mean QP/QS from ten good-bolus studies in each animal was used as the standard for comparison. In five trials in normal animals, where a prolonged or double-peak bolus led to a shunt calculation (QP/QS greater than 1.2 : 1), deconvolution resulted in QP/QS = 1.0. Deconvolution improved shunt quantitation in eight of ten trials in animals that received a prolonged bolus. The correlation between the reference QP/QS and the QP/QS calculated from uncorrected bad bolus studies was only 0.39 (p greater than 0.20). After deconvolution using a low pass filter, the correlation improved significantly (r = 0.77, p less than 0.01). The technique gave inconsistent results with multiple-peak bolus injections. Deconvolution analysis in these studies is useful in preventing normals from being classified as shunts, and in improving shunt quantitation after a prolonged bolus. Clinical testing of this technique in children with suspected L-R shunts seems warranted.