An in-depth review of the application of artificial intelligence in the perioperative management of gastrointestinal surgery

Introduction Artificial Intelligence (AI) refers to the intelligent science of machine demonstration based on reinforcement learning and around the use of algorithms.

AI uses available databases called "big data" to formulate algorithms.
Analysis based on these algorithms or other data facilitates early diagnosis, accurate risk assessment, intraoperative management, automatic drug delivery, prediction of anesthesia and surgical complications, and postoperative As a result, perioperative management is effectively carried out and treatment costs are reduced.

The use of AI is the way forward for perioperative management of major surgery in the future.

Application of AI in preoperative management 1.
Pathological diagnosis Digital pathology can help remote centers obtain more opinions on certain pathological problems through the transmission of remote report slide images.

Due to the large number of specimens and the ability of the microscope to focus on areas, even pathologists will inevitably miss certain areas.

AI helps analyze the entire field and detect subtle changes.

It can identify celiac disease, Helicobacter pylori, liver fibrosis classification, colon polyp classification, dysplasia classification, predict the 5-year overall survival rate of colorectal cancer from pathology, and predict microsatellite instability.

But its disadvantage is the large amount of data storage and the need for efficient networks, computers and equipment.

2.
An important achievement achieved by AI in endoscopy is the evaluation of a large number of endoscopic images obtained through wireless capsule endoscopy, and the integrated ultrasound system in the capsule for evaluation.
As described by Sonopill, due to the problem of abdominal organ deformation, The integration speed of AI in gastrointestinal pathology imaging is not as fast as in breast or lung diseases.

3.
Surgical decision-making Surgical decision-making includes patient participation in management, improving patient satisfaction and compliance, and allowing patients to cope with any complications.

IBM Watson Oncology aims to help oncologists understand the latest evidence and guide decision-making.

Despite the relevant publicity, it is not well applied to gastrointestinal tumors (GI).

4.
Risk assessment Existing risk assessment models, such as major adverse cardiac events, revised cardiac risk index, and the risk of Gupta myocardial infarction or cardiac arrest, underestimate the risks involved.

In addition, the widely used American Society of Anesthesiologists (ASA) classification also has subjective assessment errors.

AI and machine learning (ML) platforms can obtain data from a huge database of results, and accurately identify and predict risks based on existing electronic health record (EHR) variables.

The MySurgeryRisk platform uses EHR data of 285 variables and has been proven to predict perioperative risks more accurately than clinical judgments.

ML has also proven that it can accurately predict mortality, sepsis, and acute kidney injury using intraoperative data.

Emergency surgery risk prediction optimization tree (POTTER) calculator is an ML platform based on the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database.
By integrating with EHR data, it can identify and prevent surgical site infections (SSI).
, Sepsis, pneumonia, urinary tract infection, heart complications and other risks.

Similarly, the ML algorithm can recommend available anesthetics based on data from previous operations (including patient risk factors and postoperative results) to further prevent complications of anesthesia.

Application of AI in intraoperative management 1.
Strengthen anesthesia induction and maintenance intubation robots, such as the Kepler intubation system, use video laryngoscopes and robotic arms to place endotracheal intubation, and the success rate is as high as 91%.

The McSleepy automatic intravenous infusion machine uses the bispectral index (BIS) to maintain the depth of anesthesia (DOA) and vital signs by administering propofol, anesthetics, and muscle relaxants.

At present, automatic anesthesia is still in its infancy and is not yet ready to be used in general practice.

The closed-loop anesthesia delivery system is another BIS-based automatic delivery system that has been proven to be effective and efficient.

2.
Surgery stage recognition ML and data annotation system can be used to recognize the operation stage and identify deviations or delays in the operation steps.

3.
Avoid close mistakes in surgical navigation technology or computer-assisted abdominal surgery.
Use preoperative or intraoperative imaging to track surgical instruments and help explore hidden surgical anatomy to improve safety and improve clinical outcomes.

Application of AI in postoperative management 1.
Predicting risk Natural language processing (NLP) allows ML algorithms to use EHR data and use it to predict results.
ML algorithms have been proven to accurately predict superficial SSI, inter-organ space SSI, sepsis, liver and pancreas Or bleeding after colorectal surgery.

2.
Help reduce complications AI has proven to use prediction platform algorithms to predict the risk of pancreatic fistula (POPF) after pancreaticoduodenectomy.

The predicted POPF risk can guide clinical management to prevent or mitigate adverse outcomes.

At the same time, ML can be used to predict urine volume and fluid status after infusion in patients with sepsis or in the ICU, helping to avoid fluid overload and oliguria.

3.
Reducing postoperative pain AI also focuses on minimizing postoperative pain and improving patient comfort.

In a randomized controlled trial (RCT) involving 50 patients, the Nociception level (NOL) index was used to monitor the pain sensation of patients during surgery, and there was no difference in the use of opioids (fentanyl and morphine) after surgery.
, But the postoperative pain score of the NOL group increased by 1.
6 points.

Conclusion The application of AI in clinical systems is increasing, and it is becoming a powerful tool in the field of healthcare.
Its ultimate goal is to provide patients with the highest quality treatment.

AI is the direction of future development, and its application fields will continue to expand.
I hope that with the assistance of AI, doctors can better contribute to China's medical career and escort the health of the people! Yimaitong compiled and compiled from: Solanki SL, Pandrowala S, Nayak A, et al.
Artificial intelligence in perioperative management of major gastrointestinal surgeries[J].
World J Gastroenterol 2021; 27(21): 2758-2770