| 摘要 |
The cardiac ejection fraction EF is estimated based on the parameters of a model of an indicator (such as thermal) dilution through the heart. A channel model is defined by an upstream indicator injector, such as a heater, positioned preferably in the right atrium/ventricle and a downstream indicator concentration sensor, such as a thermistor. The preferred model is a lagged normal transfer function, which has as one of its output parameters the indicator decay constant tau of the blood channel. The heart rate HR is also measured. The ejection fraction EF is then calculated continuously as EF=1-exp(-60/( tau *HR)). The cardiac output CO is also estimated in the model, preferably based on an estimate of the zero-frequency gain of the lagged normal transfer function. End diastolic volume EDV is then also calculated continuously as a function of CO, HR and EF. In applications where the step response of the indicator concentration sensor is slow relative to the indicator decay, the transfer function Hs of the sensor is determined ahead of time. Parameters characterizing Hs are then stored in a memory device and the measured transfer function data is scaled by Hs prior to modelling to provide the most accurate EF estimate. |
