This study examined the feasibility of rapid rCBF monitoring using 133Xe as a tracer during operative procedures. We compared the initial slope index derived from two bicompartmental and one monocompartmental physiological models. The single-compartment model requires only 3 min of monitoring, whereas the bicompartmental models, thought to be more reliable, require 11 min of clearance. Data were collected from 26 patients undergoing carotid endarterectomy. Approximately 20 mCi of 133Xe in saline was injected i.v. for up to five measurements per patient, for a total of 117 measurements. The robustness of the regression for the three parameters (r = 0.781-0.99, p less than 0.0001) suggests that the three parameters are closely related. This is supported by similarity of the slopes of the regression lines (between 0.944 and 1.25) and the mean +/- SD of the three rCBF models (24.9-27.5 +/- 12.0-14.3 ml 100 g-1 min-1). Similar results were obtained for individual detectors, despite the expected higher variability. For intraoperative use in surgical procedures in which physiological conditions may change rapidly and i.v. injections of tracer must be used, a rCBF index that quickly and accurately reflects flow conditions is useful. Our data suggest that the single-compartmental Wyper index may be used to provide information about cerebral perfusion that is as accurate and robust as bicompartmental models, but requires only one-quarter of the data collection time.