MEASURING CARDIAC OUTPUT THROUGH THERMODILUTION BASED ON MACHINE LEARNING

Cardiac output (CO) refers to the amount of blood ejected from a unilateral ventricle per minute and is an important measure of cardiac function. Thermodilution is the gold standard for CO measurement because of its accuracy. However, the traditional thermodilution method requires calibration of the correction factor before measurement, which makes its practical application difficult. Therefore, conducting CO measurement by using a machine-learning-based thermodilution method is proposed in this paper, and CO is regressed and predicted through the thermodilution curve by a machine learning model. In this paper, we constructed five cardiac vascular models, and three of them were randomly selected to simulate the thermodilution process. Nine features of the thermodilution curve from the time–frequency domains were extracted and fed into the multilayer perceptron model for training. On the basis of a cross-validation method, the accuracy of the final prediction model was 97.99% ([Formula: see text]%). Simultaneously, a trained neural network was used to predict the CO of the remaining two cardiac vascular models, and the resulting error was within 5%. In this paper, an experimental system consisting of a water pump, a three-way valve and a temperature sensor is also designed, and the thermodilution curves at different quantities of flow are tested and regressed and predicted with the above model, with the error being within 10%, which met the requirement for real-world use, and thus, a method was established for measuring CO by using machine-learning-based thermodilution.

Abstract 19059: Thermodilution-derived Coronary Blood Flow Pattern Immediately After Coronary Intervention as a Predictor of Long-term Clinical Outcomes in Patients With ST-segment Elevation Myocardial Infarction

Background: We reported that coronary blood flow (CBF) can be evaluated by analyzing thermodilution curve that is measured with a single pressure sensor/thermistor-tipped guidewire in the cardiac catheterization laboratory during percutaneous coronary intervention (PCI). Bimodal shape of thermodilution curve was associated with microvascular damage and predictors of left ventricular functional recovery after ST-segment elevation myocardial infarction (STEMI). However it is unknown whether the bimodal shape of thermodilution curve predicts mortality and re-hospitalization for heart failure in long term period for patients experiencing STEMI. Methods: Between September 2009 and August 2012, 97 consecutive patients with a first STEMI were prospectively enrolled in this study. Using a pressure sensor/thermistor-tipped guidewire, CBF pattern was assessed from the thermodilution-curves after successful PCI at maximum hyperemia. CBF pattern was classified into 3 groups according to the shape of thermodilution curve: a narrow unimodal (a rapid fall and rise of temperature-time curves) (n=47), a wide unimodal (a gradual fall and rise of temperature-time curves) (n=33), or bimodal (two populations with valley deeper than 20% of peak temperature drop) (n=17). Major adverse cardiac events (MACE) were defined as cardiac death and/or heart failure re-hospitalization within this study period. Results: Median follow-up period was 2.4 years. Although patients in the narrow-unimodal group and the wide unimodal group had a significantly lower incidence of MACE, patients in bimodal group had a higher risk of MACE during this study period (71, 15, 21%, p<0.001). Multivariate analysis revealed that bimodal shape of the thermodilution-curve was the only independent predictor of MACE after STEMI (hazard ratio, 8.38; 95% confidence interval, 2.13-33.00; P=0.0023). Conclusions: A bimodal shape of the thermodilution curve is associated with the poor long-term clinical outcomes. This easily assessable coronary flow pattern is useful in clinical risk stratification for STEMI patients in the cardiac catheterization laboratory immediately after PCI.