Page 48 - Read Online

Goyal et al. Art Int Surg. 2025;5:450-64 https://dx.doi.org/10.20517/ais.2024.108 Page 452

and accuracy, significantly contributing to better surgical outcomes. Furthermore, the concurrent
advancement of both laparoscopic and robotic techniques has been instrumental in advancing AI
integration in hernia repair.

Due to the nascent nature of AI use in IHR, there is limited literature that compiles and aggregates the
benefits and disadvantages of this practice. To the best of our knowledge, this review is the first systematic
evaluation to compile and analyze the real-time intraoperative use of AI models in IHR. While recent
discoveries show promising results, ongoing data collection and findings indicate that this area of practice
and research is continuously evolving.

The purpose of this systematic review was to evaluate the latest use of AI models in the real-time
intraoperative phase of IHR for their accuracy, applicability, and potential influence on clinical outcomes
and surgical training. Additionally, the number of surgical phases recognized by each AI model is
considered as a measure of phase-specific accuracy. This review provides a comprehensive and novel
analysis of the recent performance and transformative potential of video-trained surgical phase-recognition
AI systems, assessing their ability to enhance surgical outcomes and redefine traditional practices.

MATERIALS AND METHODS
Study design
This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic
[11]
Review and Meta-Analysis Protocols (PRISMA-P 2020) guidelines to assess the application of AI in real-
time surgical phase recognition during minimally invasive IHR. Both robotic and laparoscopic approaches,
including TAPP and TEP procedures, were included. The study protocol was registered with the
International Prospective Register of Systematic Reviews (PROSPERO) under the number
CRD42024621178. Approval from an Institutional Review Board (IRB) committee was not necessary or
indicated, as we used data from published primary studies available in the literature.

Information sources and search strategy
A comprehensive search was performed in six electronic databases: PubMed, Scopus, Web of Science,
Embase, Cochrane Library, and ScienceDirect. The final search was conducted on November 12, 2024.
Language, geographical location, and date of publication were not imposed restrictions. The search terms
included were derived from the keywords “artificial intelligence”, “machine learning”, “computer vision
systems”, “inguinal hernia”, and “herniorrhaphies”. The search strategy, covering Boolean operators and
keywords, is detailed in Supplementary Table 1. Manual searches of reference lists and relevant studies were
also performed to ensure comprehensive coverage.

Study selection
The titles and abstracts of the identified studies were screened by three independent reviewers to ensure
relevance based on the predefined inclusion and exclusion criteria. The final stage of eligibility was assessed
by reviewing the full text of potentially relevant studies. Reviewers resolved any discrepancies through
discussion and by consensus. Eligibility criteria included prospective and retrospective studies that used AI-
based video recordings for surgical phase recognition in minimally invasive inguinal hernia repair, either
laparoscopic or robotic. Both unilateral and bilateral hernia repairs were incorporated. Exclusion criteria
involved case reports, case series, reviews, commentaries, conference abstracts, and editorials. Studies
focusing solely on image-based AI models or evaluating skills without surgical phase recognition were
excluded.