Braindex Cerebral Tissue Oxygen Saturation (SctO2) Measurement Algorithm Assessment and Improvement

SPONSOR BRAINDEX sas Parc Eurasanté 12 avenue Pierre Mauroy 59120 Loos

INVESTIGATOR COORDINATOR Professor Julien AMOUR Institut de Perfusion, de Réanimation et d'Anesthésie de Chirurgie Cardiaque Paris Sud (IPRA) Hôpital Privé Jacques Cartier 6 avenue du Noyer Lambert 91300 Massy, France

TITLE In vivo optimization study of the algorithm for measuring cerebral tissue oxygen saturation (SctO2) by Braindex compared with a mixed reference value of jugular venous and systemic arterial saturation during proven loss of cerebral auto-regulation.

ID-RCB NUMBER 2022-A02770-43

JUSTIFICATION / BACKGROUND The incidence of postoperative cognitive impairment (POCD) in cardiac surgery is of the order of 30-50%. Loss of cerebral autoregulation is a factor in POCD and cognitive decline. Cerebral near-infrared oximetry (NIRS), a commercially available non-invasive monitoring system, aims to detect and limit the loss of cerebral autoregulation during intraoperative hemodynamic instability.

NIRS currently measures cerebral tissue oxygen saturation (SctO2) over a patchy tissue volume, and corresponds to blood sampling comprising arterial, venous and capillary components.

The performance of cerebral oximetry devices currently on the market has been clinically validated to a limited extent, often in healthy volunteers with no comorbidity, or in cardio-vascular situations (hypoxemia tests) outside any surgical procedure. For a number of devices, test results have been extrapolated and marketed without any real per-procedural clinical validation, and without being able to fully assert the correlation between a real loss of cerebral autoregulation, which is difficult to affirm, and the fall in SctO2. Beyond clinical validation, the various devices present a number of limitations and technical imperfections, notably a restricted sampling zone, a low number of transmitters and receivers, and variable signal processing algorithm quality. All these factors may explain the mixed results of recent clinical meta-analyses concerning the benefits of this technology.

The BRAINDEX sensor offers the advantage of an optical imager equipped with 60 LEDs emitting 6 wavelengths (660, 730, 780, 810, 850 and 900nm) and 28 photodiodes (optical sensors). This set of optodes translates into a measurement of 420 optical vectors per cerebral hemisphere, enhancing the resolution of cerebral oxygenation measurement and the number of analysis zones compared with current devices on the market.

Initial efficiency tests have already been carried out in silico using simulation models and on a test bench. We now wish to optimize and calibrate the signal analysis algorithm for measuring cerebral tissue oxygen saturation (SctO2) in vivo, in a context of proven loss of cerebral autoregulation, and compare it with a mixed value of jugular venous and systemic arterial saturation, a commonly accepted reference.

In this sense, percutaneous aortic valve replacement or TAVI (Sapiens valve, Edwards) offers the decisive advantage of generating a standardized and perfectly reproducible circulatory arrest, leading to a planned loss of cerebral autoregulation, throughout the duration of aortic valve occlusion by the valve deployment balloon.

The aim of the research is to optimize the Braindex algorithm in vivo during the TAVI procedure, and in particular during the loss of cerebral autoregulation linked to circulatory arrest systematically induced by valve deployment. This study will also rely on the measurement of jugular venous oxygen saturation (SvjO2) and arterial oxygen saturation (SaO2), which are the reference measurements for the previous devices.

OBJECTIVES The primary objective of this research is to optimize an algorithm for estimating SctO2 (SctO2Exp.), in vivo, in patients scheduled for a femoral TAVI procedure and systematically exposed to a period of loss of cerebral autoregulation during deployment of the percutaneous aortic valve.

Secondary objectives are:

• To evaluate the quality of EEG (electroencephalogram) signals from electrodes in the BRAINDEX sensor,
• Evaluate the possibility of calculating the ANI (Analgesia Nociception Index) from PPG (photoplethysmography) signals obtained with the BRAINDEX sensor
• Study the technical performance of the BRAINDEX sensor.

RESEARCH SCHEME This is an exploratory, prospective, interventional, open clinical investigation aimed at developing and optimizing the algorithm for estimating SctO2Exp.

INCLUSION CRITERIA

• Patient, male or female, over 18 years of age
• Patient undergoing percutaneous aortic valve replacement with aortic valve balloon deployment (Edwards Sapien valve) under simple sedation with remifentanil in AIVOC, with spontaneous ventilation
• Patient affiliated to or entitled under a social security scheme
• Patient having received written and informed information about the study and having signed a free and informed consent to participate in the study.

NON-INCLUSION CRITERIA

• Emergency surgery
• Patient weighing less than 40kg
• Patient with history of ischemic stroke
• Patient with a contraindication to catheterization arterial catheters
• Patient with a contraindication to the Braindex device
• Protected patient: adult under guardianship, curatorship or other legal protection deprived of liberty by judicial or administrative decision
• Pregnant, parturient or breast-feeding women
• Patient hospitalized without consent

RESEARCH PROCEDURE The research procedure consists in optimizing the Braindex algorithm for measuring in vivo SctO2Exp before, during and after a proven loss of cerebral autoregulation linked to circulatory arrest programmed by percutaneous aortic valve deployment via transvalvular balloon, on a vigorous patient sedated with remifentanil in IVCA.

Continuous measurement of SvjO2, the usual reference measurement for commercially available devices, will be performed by implanting an optical fiber (CEVOX,PULSIONMedicalSystemsSEHans-Riedl-Str.17 85622 Feldkirchen Germany) through an internal jugular catheter, in addition to the conventional external electro systolic drive catheter. The following parameters will be recorded every 2 seconds: invasive blood pressure (systolic, mean and diastolic), heart rate, SaO2 on the pulse wave. SvjO2 will be recorded continuously throughout the implantation procedure. The times at which the surgical incision begins, the onset of sinus tachycardia (overdriving) induced by external electro-systolic stimulation, the start of inflation of the intra-avalvular aortic balloon, maximum balloon occlusion, the end of complete balloon deflation, the end of stimulation and the end of skin closure will be recorded.

Jugular venous and arterial gasometry (1 ml per sample) will be taken to perform initial calibration of the CEVOX fiber at the start of the procedure (skin incision), then before overdriving, then when overdriving is stopped and finally at the end of skin closure, to ensure that there is no derivation of the estimated continuous value of the CEVOX fiber.

ADDED ACT Placement of the CEVOX probe for continuous SvjO2 monitoring will be performed under local anaesthetic, in addition to remifentanil sedation, under ultrasound guidance.

Installation of the BRAINDEX non-invasive forehead sensor.

EVALUATION CRITERIA

Primary endpoints :

• NIRS (Near-infrared spectroscopy) signals recorded via the BRAINDEX sensor to optimize the algorithm
• SctO2Exp value obtained from jugular measurement to optimize algorithm.

Secondary endpoints :

• EEG signals recorded via BRAINDEX sensor and three-axis acceleration information via accelerometer
• PPG (photoplethysmocardiography) signals via BRAINDEX sensor and ANI signals via ANIv2 monitor
• Sensor operation, optode detachment, electrode detachment

STUDY SIZE As this is an exploratory study, no statistical hypothesis could be defined. As the main objective of the study is to optimize an in vivo algorithm for estimating SctO2Exp, it is important to obtain a sample representative of the diversity of hemodynamic situations observed during TAVI procedures, which offer the advantage of being highly standardized and reproducible. The inclusion of 40 patients should enable us to achieve this diversity. The inclusion of 40 patients should enable us to achieve this diversity.

INTENDED NUMBER OF CENTERS One center: Institut de Perfusion, de Réanimation et d'Anesthésie de Chirurgie Cardiaque Paris Sud (IPRA) at Hôpital Privé Jacques Cartier (Massy).

DURATION OF RESEARCH

• Length of inclusion period: 10 months (projected start date of IC: 01/03/2023)
• Duration of each participant's participation: duration of the surgical procedure with follow-up of adverse events during the 24 postoperative hours.
• Total research duration: 10 months and 24 hours (projected end of CI: 05/01/2024).

STATISTICAL ANALYSIS OF DATA Analyses will be performed on all included patients. Main analysis NIRS signals collected over the entire operative period of the TAVI procedure will be used to develop the algorithm for calculating SctO2Exp.

The raw data will be denoised using a butter Worth 0.05Hz order 6 low-pass filter and subtracting potential artifacts, external light interference and high-frequency interference (physiological and caused by the electromagnetic environment). The regional oxygen concentration SctO2Exp will then be calculated using the modified Beer-Lambert law, expressed by the equation below, on all optical vectors, In order to optimize the algorithm obtained, the concentrations thus estimated, SctO2Exp, will be compared with the reference measurement, SctO2Ref (estimated as being equal to 25% SaO2 and 75% SvjO2), and will be used to group together the distances and vectors required for a more accurate estimate of oxygen saturation.

Secondary analysis

Evaluation of EEG signal quality:

• EEG signal quality in the face of physiological and environmental artefacts environmental artifacts associated with the TAVI procedure (patient handling, electromagnetic environment, etc.), will be assessed by calculating a signal-to-noise ratio defined by the power density in the spectral band of interest (0.1 - 4hz) divided by the remaining spectral content and signal saturation indicators (considered saturated if greater than ±200mV). The proportion of saturated ratio will be described, - Signal variations in relation to voluntary or involuntary movements of the patient's head measured along 3 axes will be calculated. A variation of more than 0.5g in any axis will be considered as a head movement (voluntary or involuntary). The proportion of variation over 0.5g will be described.

Assessment of the possibility of calculating ANI :

• The PPG signals recorded by the BRAINDEX sensor will be used to calculate ANI. The agreement between the ANI obtained from the PPG signals via the BRAINDEX sensor and the ANI obtained from the ANIv2 monitor will be assessed using the intraclass correlation coefficient (ICC) and the graphical method of Bland and Altman. The ICC and its 95% confidence interval will be determined using a 2-factor random analysis of variance, patient and method.

Study the technical performance of the BRAINDEX sensor:

• The operational reliability of the BRAINDEX sensor will be assessed using a log file indicating potential abnormal operation of the system
• The identification of optode detachment during measurement will be evaluated by estimating the variance of ambient noise with diodes off
• The indication of electrode detachment during measurement will be evaluated by estimating the impedance of the electrodes by the acquisition component

EXPECTED BENEFITS The expected spin-offs of this research are to optimize the BRAINDEX sensor's SctO2Exp calculation algorithm in vivo, under actual patient management conditions, in proven situations of loss of cerebral auto regulation.