Artificial Intelligence Methods for Surgical Site Infection: Impacts on Detection, Monitoring, and Decision Making

A novel high accuracy model for automatic surgical workflow recognition using artificial intelligence in laparoscopic totally extraperitoneal inguinal hernia repair (TEP)

INTRODUCTION

Artificial intelligence and computer vision are revolutionizing the way we perceive video analysis in minimally invasive surgery. This emerging technology has increasingly been leveraged successfully for video segmentation, documentation, education, and formative assessment. New, sophisticated platforms allow pre-determined segments chosen by surgeons to be automatically presented without the need to review entire videos. This study aimed to validate and demonstrate the accuracy of the first reported AI-based computer vision algorithm that automatically recognizes surgical steps in videos of totally extraperitoneal (TEP) inguinal hernia repair.

METHODS

Videos of TEP procedures were manually labeled by a team of annotators trained to identify and label surgical workflow according to six major steps. For bilateral hernias, an additional change of focus step was also included. The videos were then used to train a computer vision AI algorithm. Performance accuracy was assessed in comparison to the manual annotations.

RESULTS

A total of 619 full-length TEP videos were analyzed: 371 were used to train the model, 93 for internal validation, and the remaining 155 as a test set to evaluate algorithm accuracy. The overall accuracy for the complete procedure was 88.8%. Per-step accuracy reached the highest value for the hernia sac reduction step (94.3%) and the lowest for the preperitoneal dissection step (72.2%).

CONCLUSIONS

These results indicate that the novel AI model was able to provide fully automated video analysis with a high accuracy level. High-accuracy models leveraging AI to enable automation of surgical video analysis allow us to identify and monitor surgical performance, providing mathematical metrics that can be stored, evaluated, and compared. As such, the proposed model is capable of enabling data-driven insights to improve surgical quality and demonstrate best practices in TEP procedures.

The detection and prediction of surgical site infections using multi-modal sensors and machine learning: Results in an animal model

Introduction

Surgical Site Infection (SSI) is a common healthcare-associated infection that imposes a considerable clinical and economic burden on healthcare systems. Advances in wearable sensors and digital technologies have unlocked the potential for the early detection and diagnosis of SSI, which can help reduce this healthcare burden and lower SSI-associated mortality rates.

Methods

In this study, we evaluated the ability of a multi-modal bio-signal system to predict current and developing superficial incisional infection in a porcine model infected with Methicillin Susceptible Staphylococcus Aureus (MSSA) using a bagged, stacked, and balanced ensemble logistic regression machine learning model.

Results

Results demonstrated that the expression levels of individual biomarkers (i.e., peri-wound tissue oxygen saturation, temperature, and bioimpedance) differed between non-infected and infected wounds across the study period, with cross-correlation analysis indicating that a change in bio-signal expression occurred 24 to 31 hours before this change was reflected by clinical wound scoring methods employed by trained veterinarians. Moreover, the multi-modal ensemble model indicated acceptable discriminability to detect the presence of a current superficial incisional SSI (AUC = 0.77), to predict an SSI 24 hours in advance of veterinarian-based SSI diagnosis (AUC = 0.80), and to predict an SSI 48 hours in advance of veterinarian-based SSI diagnosis (AUC = 0.74).

Discussion

In sum, the results of the current study indicate that non-invasive multi-modal sensor and signal analysis systems have the potential to detect and predict superficial incisional SSIs in porcine subjects under experimental conditions.

Clinically relevant features for predicting the severity of surgical site infections

Surgical site infections are hospital-acquired infections resulting in severe risk for patients and significantly increased costs for healthcare providers. In this work, we show how to leverage irregularly sampled preoperative blood tests to predict, on the day of surgery, a future surgical site infection and its severity. Our dataset is extracted from the electronic health records of patients who underwent gastrointestinal surgery and developed either deep, shallow or no infection. We represent the patients using the concentrations of fourteen common blood components collected over the four weeks preceding the surgery partitioned into six time windows. A gradient boosting based classifier trained on our new set of features reports, respectively, an AUROC of 0:991 and 0:937 at predicting a postoperative infection and the severity thereof. Further analyses support the clinical relevance of our approach as the most important features describe the nutritional status and the liver function over the two weeks prior to surgery.

Situating Artificial Intelligence in Surgery: A Focus on Disease Severity

OBJECTIVES

Artificial intelligence (AI) has numerous applications in surgical quality assurance. We assessed AI accuracy in evaluating the critical view of safety (CVS) and intraoperative events during laparoscopic cholecystectomy. We hypothesized that AI accuracy and intraoperative events are associated with disease severity.

METHODS

One thousand fifty-one laparoscopic cholecystectomy videos were annotated by AI for disease severity (Parkland Scale), CVS achievement (Strasberg Criteria), and intraoperative events. Surgeons performed focused video review on procedures with ≥1 intraoperative events (n = 335). AI versus surgeon annotation of CVS components and intraoperative events were compared. For all cases (n = 1051), intraoperative-event association with CVS achievement and severity was examined using ordinal logistic regression.

RESULTS

Using AI annotation, surgeons reviewed 50 videos/hr. CVS was achieved in ≤10% of cases. Hepatocystic triangle and cystic plate visualization was achieved more often in low-severity cases (P < 0.03). AI-surgeon agreement for all CVS components exceeded 75%, with higher agreement in high-severity cases (P < 0.03). Surgeons agreed with 99% of AI-annotated intraoperative events. AI-annotated intraoperative events were associated with both disease severity and number of CVS components not achieved. Intraoperative events occurred more frequently in high-severity versus low-severity cases (0.98 vs 0.40 events/case, P < 0.001).

CONCLUSIONS

AI annotation allows for efficient video review and is a promising quality assurance tool. Disease severity may limit its use and surgeon oversight is still required, especially in complex cases. Continued refinement may improve AI applicability and allow for automated assessment.

Application of Artificial Intelligence in Medicine: An Overview

Artificial intelligence (AI) is a new technical discipline that uses computer technology to research and develop the theory, method, technique, and application system for the simulation, extension, and expansion of human intelligence. With the assistance of new AI technology, the traditional medical environment has changed a lot. For example, a patient's diagnosis based on radiological, pathological, endoscopic, ultrasonographic, and biochemical examinations has been effectively promoted with a higher accuracy and a lower human workload. The medical treatments during the perioperative period, including the preoperative preparation, surgical period, and postoperative recovery period, have been significantly enhanced with better surgical effects. In addition, AI technology has also played a crucial role in medical drug production, medical management, and medical education, taking them into a new direction. The purpose of this review is to introduce the application of AI in medicine and to provide an outlook of future trends.