The primary objective of the SEASCAPE project is to design, develop, and to apply a clinical implementation tool of a machine learning (ML) and artificial intelligence (AI)-based co-pilot system for the real-time monitoring and control of nociception during general anesthesia (GA).

The ultimate clinical purpose is to optimize individualized pain management by achieving precise titration of intravenous opioids (specifically remifentanil), thereby minimizing the incidence of over- and under-dosing. This optimization is projected to enhance patient outcomes, reduce opioid-related complications, and improve overall cost-effectiveness of anesthetic procedures.

The main scientific question guiding this work is: Can a novel algorithm be generated and validated to provide superior analytical precision for analgesic management by reliably differentiating genuine nociceptive responses from confounding physiological variables-such as inadequate neuromuscular blockade or changes in depth of anesthesia-thereby significantly improving the clinical decision-making framework for intraoperative nociception control? This project addresses the recognized challenge in anesthesiology: defining an objective measure to quantify nociception and antinociception during GA.

Study Population: Patients scheduled for elective surgical procedures requiring general anesthesia (GA).

Existing Intervention: The standard anesthetic regimen includes continuous intravenous infusion of the remifentanil for intraoperative analgesia, typically governed by a Target Controlled Infusion (TCI) system utilizing a pharmacokinetic/pharmacodynamic (PK/PD) model (Eleveld TCI model).

Project Focus: The research seeks to improve the accuracy and efficacy of this existing analgesic strategy by integrating a multivariate patient data stream with the newly developed SEASCAPE co-pilot AI. This aims to refine the remifentanil dose predictions beyond the current TCI model's capabilities, personalized system.

The assessment of pain during General Anesthesia (GA) constitutes a significant clinical challenge because the direct evaluation of pain is impossible due to the abolition of conscious responses. Consequently, clinicians are required to interpret a large volume of indirect physiological data (heart rate, blood pressure, EEG, skin conductance) to determine the level of nociception-the process of detecting noxious stimuli by the nervous system-with data captured from multiple, often independent, monitors.

This manual method is intrinsically inefficient and susceptible to errors in judgment. The result of this deficiency is suboptimal pain management, where the over- or under-dosing of opioids increases the risk of serious adverse events, such as the chronification of pain. This represents a public health problem that causes disability and a significant deterioration in the quality of life. The opportunity to improve this management lies in the integration of data through advanced technologies. Although pharmacokinetic/pharmacodynamic (PK/PD) models exist to estimate individualized opioid doses, they are insufficient on their own. Their validation has primarily relied on EEG parameters, without considering the total complexity of the physiological variables involved in the nociception. The information gap and the inherent inefficiency of the current manual process open the door for a disruptive solution: the development of a co-pilot system based on Machine Learning (ML)/AI, framed within the line of biotechnology to face global challenges.

This project proposes leveraging the opportunity of ML to develop a co-pilot system capable of intelligently integrating the remifentanil PK/PD model with real-time data from multiple monitors (hemodynamic, EEG, neuromuscular relaxation, and nociception). The solution aims for the individualized optimization of pain management, minimizing the risk of inappropriate dosing. Implementing this technology is crucial, as failure to do so would perpetuate the inefficiency and enormous expenses associated with poor clinical outcomes. Conversely, this system promises to improve postoperative outcomes, such as reduce the incidence of persistent pain, and generate a direct positive impact on healthcare costs and quality of care.

The SEASCAPE clinical co-pilot seeks to integrate real-time data from multiparameter monitors, EEG, TOF (Train-of-Four), ANI (Analgesia Nociception Index), ventilators, and infusion pumps via the Mindray m-Connect platform.

Assist in decision-making by classifying nociception and generating prioritized alerts ("increase," "maintain," or "decrease" remifentanil dose).

Personalize analgesia using PK/PD TCI models adjusted to individual variables, specifically differentiating between nociception, hypnosis, and muscular relaxation.

Among its competitive advantages are real interoperability, actionable clinical support, advanced personalization, and a positive operational and clinical impact. This impact allows for the reduction of postoperative complications, the shortening of surgical times, and the optimization of bed management.

The project contemplates the protection of the algorithm, software, and interface through patent and copyright, and a licensing agreement with Mindray for the use of m-Connect, supported by the UC Transfer and Development Directorate. In summary, SEASCAPE offers an innovative solution, combining AI, personalization, and real-time clinical support, with a clear plan for validation and scaling toward the market.

Nociception Pattern Identification: To identify distinct patterns of nociception by analyzing changes across multiple physiological and pharmacological variables, including hemodynamic parameters, electroencephalogram (EEG) data, ANI variability, mechanical ventilator parameters, and the estimated remifentanil dose derived from the Eleveld Target Controlled Infusion (TCI) PK/PD model.

Anesthetic State Differentiation: To identify patterns that reliably distinguish between states of inadequate anesthetic depth and insufficient muscular relaxation, utilizing changes in hemodynamic parameters, EEG, and ventilator data. These patterns will be validated using serial neuromuscular monitoring (e.g.,TOF) and the administered dose of neuromuscular blocking agent.

Clinical Usability Assessment: To determine the degree of usability (UX), strengths, weaknesses, and opportunities for improvement of the SEASCAPE co-pilot system interface from the perspective of the end-users (anesthesiologists).

The development of the co-pilot system is structured as a prospective observational design across three phases:

Phase 1: Initial Prospective Observational Cohort Study (N=30) Design: Prospective observational cohort study, adhering to STROBE guidelines. Population: 30 patients scheduled for surgery under general anesthesia, stratified by age: 10 patients (0-18 years), 10 patients (19-60 years), and 10 patients (over 60 years).

Data Collection: Gathering of demographic data, vital signs, ventilatory parameters, pharmacological data (Eleveld remifentanil PK/PD model), muscular relaxation monitoring, ANI data, and EEG data.

Purpose: To generate the initial data pipeline for the co-pilot system with labeled changes and stimuli verified by predetermined surgical time points for subsequent analysis.

Phase 2: Expanded Prospective Observational Cohort Study (N=100) Design: Prospective observational cohort study, adhering to STROBE guidelines. Population: 100 additional patients, stratified by two anesthetic techniques: Total Intravenous Anesthesia (TIVA) and Inhalational Anesthesia (Anesthesia with gases).

Purpose: To generate a larger dataset to train the co-pilot system. The goal is for the system to successfully predict a guidance tendency for remifentanil administration:

Recommendation to increase remifentanil administration. Recommendation to maintain current remifentanil administration. Recommendation to decrease remifentanil administration. Phase 3: Descriptive Prospective Observational Study (N=20) Design: Descriptive prospective observational study, following STROBE guidelines, focused on human factors and usability.

Population: Expected participation of 20 anesthesiologists (from a total of 35 at the developing center).

Purpose: To collect variables related to the anesthetic clinical environment from the anesthesiologists' perspective, identifying the strengths, weaknesses, and opportunities for improvement of the proposed interface.

Interface Development: The graphical interface will be built using a user-centered design (UCD) process, including the creation of wireframes, wireflows, and low-fidelity prototypes (e.g., in Figma) to ensure an accessible, functional, and visually coherent platform that facilitates intuitive interaction.

Data Registration Elective surgical patients (or their legal representatives) meeting the inclusion criteria will be invited to participate and will sign the informed consent documentation. Demographic data will be recorded. Known nociceptive stimuli will be introduced to provide a controlled reference for system validation without modifying the standard clinical conduct of the supervising anesthesiologist.

Sample Size Calculation A number of 30 cases has been established for the first phase. One hundred patients are required for the second phase. This calculation considers potential losses due to incomplete or nonexistent records, a surgery duration exceeding two hours, and the analysis of 16 variables. n=20 anesthesiologist for Phase 3.

Data Registration To obtain the research data, elective surgical patients (or their legal representatives) who meet the inclusion criteria and do not present the exclusion criteria will be invited to participate. Recruitment will be carried out either on the day of admission to the preoperative unit or the day before the scheduled surgery. Upon admission, general data will be collected, and the documentation will be signed according to the informed consent process.

Demographic data (age, weight, height, and gender) will be registered, and known nociceptive stimuli will be established. This is intended to contrast the common actions of the clinicians and establish a standard without modifying the usual conduct of the anesthesiologist in charge of the case.

Inclusion Criteria:

• Patients scheduled for elective surgery with general anesthesia.
• Surgeries scheduled to last at least two hours.

Exclusion Criteria:

• Patients undergoing emergency surgery.
• Pregnant women.
• Presence of a mental or intellectual disability before the hospitalization.
• Drug dependence.
• Surgeries scheduled for more than 4 hours.
• Intraoperative complications requiring changes in routine behavior.

Patients scheduled for elective surgery with general anesthesia and anesthesiologist.