Image-Based PSi Signal Acquisition From a Patient Monitor During a Medical Procedure

Introduction

In administering general anesthesia (GA), it is necessary to use substances that provide deep unconsciousness, analgesia, amnesia, and muscle relaxation. Proper induction of anesthetic agents is essential when performing a diagnostic procedure or surgery. GA and its associated activities in the human body are dynamically very complex processes. The processes involve various pharmacokinetic and pharmacodynamic mechanisms that have not been fully explored. During GA, the anesthesiologist must monitor the patient's vital signs and maintain the functions of vital organs.

In order to achieve adequate GA, substances are introduced into the patient's body in various ways.

The most common methods used in clinical practice today are intravenous anesthetic induction, which is the injection of the anesthetic into a vein, and inhalational anesthetic induction, in which the patient inhales the substance from the breathing mixture. The anesthetic technique in which the substances are injected intravenously is called total intravenous anesthesia (TIVA) (see Al-Rifai and Mulvey (2016)).

Maintaining the appropriate depth of anesthesia (DoA) is the primary goal of the anesthesiologist. Adequate depth is achieved by adjusting the dosage of the anesthetic. Consideration must be given to the pharmacokinetics and pharmacodynamics of the anesthetic and the nature of the procedure. Too deep anesthesia is manifested by a drop in blood pressure and heart rate, as well as slow postoperative awakening of the patient from GA. On the other hand, too shallow anesthetic depth leads to activation of sympathetic nerves or, in the least likely case, awakening of the patient, which must be avoided at all costs.

The monitoring and control of DoA seems to be a suitable problem to address with a regulatory approach (see Dumont (2012)). Easing the onerous task of constantly adjusting the inflow of anesthetic would definitely be beneficial for the anesthesiologist. The idea of using closed-loop control of DoA has been around since 1950, when Mayo et al. presented ether anesthesia during abdominal surgery that was based on electroencephalographic control (see Mayo, Bickford, and Faulconer, (1950)). Various researchers have attempted to develop and optimize an automated system for DoA control that would be suitable for clinical use (see Schwilden, Stoeckel, and Schuettler (1989); Mortier et al. (1998); Kenny and Mantzaridis (1999); Struys et al. (2001); Absalom, Sutcliffe, and Kenny (2002); Liu et al. (2006); Puri, Kumar, and Aveek (2007); Hemmerling et al. (2013)). Although the basic concept of the above approaches is somewhat similar, they differ in terms of the signals that represent the controlled variables. In most cases, the electroencephalogram is used as the feedback signal controlling the inflow of propofol. The signals are measured by one of several established methods, such as Narcotrend, Scale Entropy, Response Entropy, qCON, Patient State index (PSi) (see Drover and Ortega, 2006; Karer, 2020), and the most commonly used BIS index (see Martín-Mateos et al., 2013).

Several meta-analyzes have been performed comparing BIS-controlled closed-loop systems with manual propofol administration of TIVA as well as target controlled infusion (TCI). The results suggest that closed-loop systems may outperform traditional delivery methods in many respects (see Pasin et al. (2017) and Brogi et al. (2017)).

To implement a closed-loop control system for DoA, one must be able to acquire the relevant signals online and in real time. However, patient monitors are intentionally not able to connect to an external device, such as a laptop, during the procedure for safety reasons.

In this paper, we present a universal image-based system for signal acquisition from a patient monitor to measure the depth of anesthesia. In order for the signal acquisition system to be conveniently integrated with DoA modeling, simulation, and control, it operates in the Matlab-Simulink environment. Moreover, it is also manufacturer independent. In addition, it provides the anesthesiologist with a simple dashboard that displays the most important acquired signals and presents the relevant time-dependent graphs to show the signal trends.