Anesthesia robots yesterday, today and tomorrow

Authors: Fu Hailong, Yuan Hongbin, Department of Anesthesiology, Second Affiliated Hospital of Naval Military Medical University

With the rapid development of science and technology, robotics has greatly changed people's lives
.
From aviation, aerospace to housekeeping services, to game entertainment, you can see the figure of
robots.
It can be said that robot technology has been deeply applied to all walks of life in society, which has had a profound impact on
people's production and lifestyle.
Since 2000, medical robot technology has entered a period of rapid development, and the famous da Vinci robotic surgical system has been widely used in
general surgery, thoracic surgery, urology and obstetrics and gynecology.

The robotic-assisted diagnostic system "Dr.
Watson" is also gradually being applied to the clinic
.
The rational use of robotics technology can make medical services more standard and efficient, safer and more convenient to operate, and it is subtly changing the diagnosis and treatment habits and learning methods
of clinicians.
In the field of anesthesiology, how to achieve accurate administration of drugs in the perioperative period, reduce the workload of anesthesiologists, and reduce the error rate of anesthesia operations have always been concerned about, and anesthesia robots have come into being
.
At this stage, anesthesia robots are mainly divided into three types according to functional positioning, namely drug delivery robots, anesthesia technology operation robots and cognitive robots
.

In a sense, the earliest prototypes of drug-delivery robots can be traced back to the target controlled infusion (TCI) system
.
The TCI system is a set of high-precision infusion pumps
that calculate the blood drug concentration by relying on the pharmacokinetic model and control the drug administration rate through software to make the blood concentration of anesthetic drugs reach the expected value.
However, the TCI pump is an open-loop system that requires manual determination and adjustment of the concentration of anesthetic drugs according to the surgical stimulus, and the constant anesthesia depth
is often not maintained during surgery.

At present, a series of intravenous drug delivery robots (also known as intravenous anesthesia robots) are mainly based on closed-loop TCI systems, which judge the depth of anesthesia and circulation by obtaining relevant target parameters in real time, such as bispectral index (BIS), auditory evoked potential, blood pressure, oxygen saturation, etc.
, and the control system monitors the change of target parameters.
Automatically adjust the dosage and rate of anesthetic drugs and vasoactive drugs in the infusion pump to achieve closed-loop feedback infusion of drugs, which can solve drug differences in
real time, precision and individualization.

In recent years, with the development of perioperative monitoring technology, intravenous anesthesia robots have gradually shifted from single input singleoutput (SISO) systems to multiple input multiple output (MIMO) systems
.
At this stage, in addition to the three major elements of intravenous anesthesia such as sedation, analgesia and muscle relaxation (referred to as muscle laxative), the target parameters of the intravenous anesthesia robot also cover basic vital signs
such as blood pressure, heart rate, and phalangeal oxygen saturation.
MIMO systems that set sedation, analgesia and muscle relaxation are gradually replacing the traditional BIS guided SISO system
.

The "McSleepy" robot, which was launched in 2013, can simultaneously monitor sedation, analgesia and muscle relaxation during anesthesia induction and anesthesia maintenance, and guide drug administration; In the same year, Johnson & Johnson's painless colonoscopy anesthesia system SEDASYS○R was approved by the US FDA for clinical use; In 2017, Beijing Yifei Huatong Technology Development Company successfully developed the anesthesia robot "Anesthesia Concerto", which predicts the degree of anesthesia sedation and analgesia based on the wavelet index, and its second-generation system can continuously monitor the sedation, analgesia index and vital sign parameters, and independently feedback by the computer system to guide the precise TCI
of perioperative intravenous anesthesia.
At present, the system is in clinical testing in China, setting a precedent
for the application of MIMO systems.
The emergence of the above products is of landmark significance
in the history of the development of intravenous anesthesia robots.

In the field of anesthesia technology operation, robots are currently mainly used in tracheal intubation and ultrasound-guided regional nerve blocks
.
In 2012, the Kepler intubation system was introduced, using its robotic arm to complete the first visible laryngoscopy tracheal intubation
on a simulator.
In 2018, China successfully developed a remote robot-assisted intubation system (RRAIS) that makes remote emergency assisted endotracheal intubation possible
.
In 2020, through laryngeal imaging and computer image automatic recognition technology, the automatic recognition of the glottis
by the tip of the fiber bronchoscope during endotracheal intubation was successfully realized.

In 2021, Chinese scientists developed a set of magnetically navigated tracheal intubation devices and completed magnetic navigation endotracheal intubation experiments
on simulated people.
With the continuous deepening of research, tracheal intubation robots are gradually moving towards automation and remoteization
.
In terms of area blocks, the Magellan robotic nerve block system, introduced in 2013, is the first robotic assistance system
for nerve block manipulation.
The system consists of a control interface, a joystick, and a robotic arm with a nerve block needle at the end of the robotic arm that assists anesthesiologists in their operations
.

Cognitive robots in the field of anesthesiology can provide clinical decision support for anesthesiologists by analyzing various data of patients, identifying clinical special cases that require human intervention during the perioperative period, and providing relevant recommendations and specific treatment plans in real time
, so it is also called clinical decision support systems (CDS).

In the field of preoperative assessment and intraoperative adverse event monitoring, the "prophet" system, introduced in 2018, predicts the risk of hypoxemia during anesthesia and analyzes risk factors in real time; By building models, researchers have been able to predict the development of perioperative hypotension with the help of arterial waveforms and electronic medical records, and use facial images to predict whether intubation is difficult
.
However, the cognitive robot currently developed is still in the stage of single model prediction, and it cannot comprehensively assess the condition and predict perioperative adverse events, but can only make recommendations to physicians, and ultimately it is up to clinicians to decide whether to adopt them
.

It is foreseeable that in the near future, the functions of anesthesia robots will be further integrated and strengthened, as anesthesiologists, how to grasp this development trend, with the help of robot technology to better serve patients is worth seriously thinking about
.
At present, the intravenous anesthesia robot has gradually adopted the MIMO system, but each closed loop is operated independently, and only the corresponding monitoring indicators guide the infusion
of the corresponding drugs.
In fact, there is cross-communication between the various circuits, for example, some intravenous anesthetics exert a synergistic effect on analgesics and muscle relaxants while performing a sedative effect
.

How to fully consider the interaction between the rings and effectively coordinate the operation of each closed loop is the focus
of the next development of the intravenous drug anesthesia robot.
For cognitive robots, how to upgrade from simple "predictive assessment" to "predictive therapy" is a problem
worthy of attention.
The current CDS is not yet able to comprehensively predict the occurrence of various adverse events, let alone independent feedback
.

How to accurately predict the timing and cause of perioperative adverse events and to intervene effectively in advance to better improve prognosis is more important
.
For anesthesia technology operation robots, how to integrate into the other two types of robots and become an artificial intelligence anesthesia robot in the full sense is the direction of
future development.
The control software of the drug delivery robot and the anesthesia technology operation robot are integrated with the cognitive robot to form the "brain" of the robot; The infusion pump of the drug delivery robot and the joystick and manipulator of the technical operation robot constitute the "hand" of the robot, and the "brain" and "hand" cooperate with each other to complete various operations more safely and efficiently to achieve the ideal anesthesia state
.

A robotic learning system introduced in 2018 can automatically locate the tip of the epidural puncture needle in the dural ultrasound image, and automatically identify the anatomical boundary markers of the epidural cavity in the ultrasound image, which greatly improves the accuracy of the needle tip positioning and the puncture success rate
.
It is conceivable that the integration of this cognitive robotics technology into the anesthesia technology operation robot could lay the foundation
for the development of an automated epidural puncture robot.
Similarly, a neural tracking method based on fully automated ultrasound images reported in 2018 can accurately track nerve positions and is of great significance for achieving the integration of fully automated technology for ultrasound-guided neural block robots
.

In fact, in clinical work, anesthesiologists face a large number of hemodynamic monitoring devices and electronic medical records, and by developing anesthesia robots with deep learning capabilities, it will greatly reduce the workload of anesthesiologists and provide patients with truly personalized treatment
.
Especially in the face of major disasters and epidemics, anesthesia robots will hopefully alleviate the clinical
pressure caused by the shortage of front-line anesthesiologists.
Although robotics can bring great convenience to clinical anesthesia work, it also brings a series of ethical issues
.
The first to bear the brunt of this is the accountability of those responsible for medical care
.
Given the current lack of laws and norms around the world that define the responsibilities of AI in healthcare services, the introduction of robots in the medical process may prompt questions about
the traditional doctor-patient relationship.
The core contract of clinical medicine, the relationship of trust between the patient and the physician, will be confused
as robotic systems are added.

When a medical malpractice occurs, which party is the responsible subject, and how to carry out accountability, it is still necessary to further formulate norms and take corresponding measures
.
With the advent of the era of big data and the development of cognitive robots, electronic medical record data needs to be continuously analyzed and integrated to support clinical decision-making, and how to protect the privacy of patients at this time is an important issue
.
Finally, the problem of occupational substitution brought about by robotics in the field of anesthesiology may also become a concern for anesthesia practitioners
.
Therefore, while developing anesthesia robots in the artificial intelligence environment, the national level should also accelerate the formulation of relevant laws and regulations and strengthen the management and control of ethical and
moral hazards.

It is foreseeable that the application prospect of robotics in the field of anesthesia is immeasurable, and it is of great significance to ensure patient safety, improve the quality of
anesthesia, achieve personalized services, reduce the work intensity of anesthesiologists and relieve their stress.
As anesthesiologists in the new era, they should be good at accepting new things and new technologies, and give full play to the advantages of artificial intelligence technology to make them assist anesthesiologists to better serve patients
.

Source: Fu Hailong,Yuan Hongbin.
Yesterday, Today and Tomorrow of Anesthesia Robots[J].
Shanghai Medical Journal,2022,45(03):146-148.