Recirculation subtraction for analysis of left-to-right-cardiac shunts: concise communication.

The object of this study is to improve the techniques for describing the lung dilution curve for shunt quantification by separating the effects of systemic recirculation on the curve form those of direct shunt return. The time of the systemic recirculation peak was estimated by determination of transit times from the right and left ventricles and lung. A gamma variate fit based on the distribution of points at that segment was applied to the recirculation curve and subtracted from the original lung dilution curve. Similar gamma variate fitting was performed for both primary and shunt curves. Rather than fitting the gamma variate of the shunt curve by the leading edge only, a larger portion could now be used since the trailing edge of the curve is clearer following recirculation subtraction. The algorithm is completely automatic, requiring no operator intervention or selection of curve-fitting regions. The correlation coefficient for comparison of the dilution-curve analysis with oximetry determinations was 0.92 in a series of 29 patients.