Blood Flow Estimation from Invasive Arterial Blood Pressure: a Black-Box Approach

The central arterial blood pressure (ABP) waveform is the result of the summation of a forward wave (P_f ) generated by ventricular contraction and a reflected wave (P_b) from arterial reflections. Wave separation analysis (WSA) is the gold standard for assessing P_b and P_f, and involves the simultaneous registration of ABP and arterial blood flow (ABF). However, the measurement of ABF is typically invasive or not commonly available in clinical settings. Therefore, several methods have been proposed to estimate ABF from ABP waveform, the most common being the triangular flow waveform approximation. The main drawback of these methods is that they rely on morphological features of ABP waveform, which may become difficult to detect in conditions of severe cardiovascular alterations, such as in critically ill patients. The objective of this study is to develop a black-box model for ABF estimation from ABP and to test its performance in acute critical conditions such as during sepsis. An autoregressive with exogenous input (ARX) model was implemented to estimate ABF from the measured ABP and its performance in assessing WSA indices was compared with the triangular flow approximation. The results show that the black-box modeling approach may be superior to the triangular flow method (Mann-Whitney U test p-value<0.05 when comparing WSA indices obtained with triangular flow with respect to the reference). This suggests that the black-box approach holds promise for overcoming challenges in clinical settings where ABF data may be unavailable.