Bringing Artificial Intelligence to the operating room: edge computing for real-time surgical phase recognition

Optimizing intraoperative AI: evaluation of YOLOv8 for real-time recognition of robotic and laparoscopic instruments

The accurate recognition of surgical instruments is essential for the advancement of intraoperative artificial intelligence (AI) systems. In this study, we assessed the YOLOv8 model's efficacy in identifying robotic and laparoscopic instruments in robot-assisted abdominal surgeries. Specifically, we evaluated its ability to detect, classify, and segment seven different types of surgical instruments. A diverse dataset was compiled from four public and private sources, encompassing over 7,400 frames and 17,175 annotations that represent a variety of surgical contexts and instruments. YOLOv8 was trained and tested on these datasets, achieving a mean average precision of 0.77 for binary detection and 0.72 for multi-instrument classification. Optimal performance was observed when the training set of a specific instrument reached 1300 instances. The model also demonstrated excellent segmentation accuracy, achieving a mean Dice score of 0.91 and a mean intersection over union of 0.86, with Monopolar Curved Scissors yielding the highest accuracy. Notably, YOLOv8 exhibited superior recognition performance for robotic instruments compared to laparoscopic tools, a difference likely attributed to the greater representation of robotic instruments in the training set. Furthermore, the model's rapid inference speed of 1.12 milliseconds per frame highlights its suitability for real-time clinical applications. These findings confirm YOLOv8's potential for precise and efficient recognition of surgical instruments using a comprehensive multi-source dataset.

[Artificial Intelligence in Head and Neck Surgery: Potentials, Challenges, and Ethical Considerations]

BACKGROUND

The growing prominence of Artificial Intelligence (AI) in medicine introduces both transformative possibilities and potential challenges. Our study focuses on the current status and perceptions of AI in Head and Neck Surgery (HNS), examining its utilization, benefits, ethical concerns, and protective measures.

OBJECTIVES

The study aims to illuminate the existing landscape of AI in HNS in Germany.

MATERIALS AND METHODS

Conducted through a questionnaire, key aspects include its current usage, potential applications (e.g., diagnosis, surgical planning), anticipated benefits (e.g., improved patient care ), and significant ethical concerns (e.g., miscalculations by AI, data privacy).

RESULTS

The survey reveals limited AI adoption in HNS, with substantial potential for improvement. Ethical considerations, especially miscalculations by AI and data privacy, emerge as central issues. The survey emphasizes the crucial role of physician control and the need for legal oversight to address concerns related to AI integration. While AI's presence in HNS is modest, the study identifies opportunities for enhancement. Ethical guidelines and practitioner-centric control are vital for discussions surrounding AI integration.

CONCLUSIONS

This research underscores the significance of ethical considerations and practitioner control in the context of AI in surgical practices. It highlights the demand for targeted training to empower practitioners in navigating the complexities of AI technologies in healthcare, ensuring responsible and patient-centric implementation.

Enhancing Surgical Safety: Microbiological Air Control in Operating Theatres at University Medical Centre Maribor

BACKGROUND

the aim of the study was to assess microbiological air quality in operating theatres by determining the level of microbiological contamination of the air and critical surfaces using the passive air sampling method and compliance of the operating theatre staff with infection control measures.

MATERIALS AND METHODS

The prospective study was conducted in the surgical block of the University Medical Centre Maribor. For two months continuously, ten operating theatres were assessed for microbial contamination of air and surfaces during quiet and active times of the day. A passive air sampling method with Petri dishes on an agar specially adapted for this purpose (plate count agar) was used. In addition, ten surgical procedures were observed to assess staff compliance with recommended practises.

RESULTS

Air samples met microbiological standards in all operating theatres. In both sampling sessions of the day (quiet and active periods), microbial contamination of the air was always within the limit of 10 CFU/m3. The average number of bacterial colonies was zero to two during quiet phases and one to four during active phases. Approximately 60% of the isolates from the operating theatres belonged mainly to the genus Staphylococcus: S. epidermidis (36% of the isolates), S. hominis (17.5%) and S. haemolyticus (5.5%). The rest were identified as Streptococcus anginosus (23%) and Bacillus sp. (18%). Pathogenic bacteria and moulds were not present. In regard to staff compliance with good surgical practise, the former varied by behaviour and function, with non-compliance in pre-operative skin preparation and operating theatre congestion being notable. The cleanliness of the environment was satisfactory.

CONCLUSIONS

Microbiological air control is extremely important for the safety and success of both surgical and postoperative practises. In spite of good results obtained in the study, further improvements in surgical staff compliance with good surgical practise are essential to reduce surgical site infections.

Robot-assisted surgery in thoracic and visceral indications: an updated systematic review

BACKGROUND

In surgical advancements, robot-assisted surgery (RAS) holds several promises like shorter hospital stays, reduced complications, and improved technical capabilities over standard care. Despite extensive evidence, the actual patient benefits of RAS remain unclear. Thus, our systematic review aimed to assess the effectiveness and safety of RAS in visceral and thoracic surgery compared to laparoscopic or open surgery.

METHODS

We performed a systematic literature search in two databases (Medline via Ovid and The Cochrane Library) in April 2023. The search was restricted to 14 predefined thoracic and visceral procedures and randomized controlled trials (RCTs). Synthesis of data on critical outcomes followed the Grading of Recommendations, Assessment, Development, and Evaluation methodology, and the risk of bias was evaluated using the Cochrane Collaboration's Tool Version 1.

RESULTS

For five out of 14 procedures, no evidence could be identified. A total of 20 RCTs and five follow-up publications met the inclusion criteria. Overall, most studies had either not reported or measured patient-relevant endpoints. The majority of outcomes showed comparable results between study groups. However, RAS demonstrated potential advantages in specific endpoints (e.g., blood loss), yet these findings relied on a limited number of low-quality studies. Statistically significant RAS benefits were also noted in some outcomes for certain indications-recurrence, quality of life, transfusions, and hospitalisation. Safety outcomes were improved for patients undergoing robot-assisted gastrectomy, as well as rectal and liver resection. Regarding operation time, results were contradicting.

CONCLUSION

In summary, conclusive assertions on RAS superiority are impeded by inconsistent and insufficient low-quality evidence across various outcomes and procedures. While RAS may offer potential advantages in some surgical areas, healthcare decisions should also take into account the limited quality of evidence, financial implications, and environmental factors. Furthermore, considerations should extend to the ergonomic aspects for maintaining a healthy surgical environment.