The focus of this review is to analyze the expanding role of artificial intelligence (AI) applications within abdominal hernia surgery, encompassing inguinal, ventral, and incisional hernias. AI, including machine learning (ML) and deep learning (DL), is being used to enhance surgical planning, provide intraoperative guidance, and improve postoperative outcomes. Generative AI (GAI) and large language models (LLMs) are also contributing to surgical education and research by enabling the creation of tailored surgical strategies and facilitating the analysis of complex medical data. A review of relevant literature from PubMed and Google Scholar indicates that AI has the potential to improve surgical precision, predict complications, and aid in surgical training. However, further research is needed to address limitations in data availability and to fully explore the ethical implications of AI in surgery.

Artificial intelligence (AI) refers to the replication of human cognitive abilities within machines, a process that yields solutions that are typically safer, more precise, and more powerful than those based on conventional approaches. Practical applications of AI are diverse, encompassing autonomous vehicles, recommendation engines, speech recognition software, and search technologies. The phrase 'artificial intelligence' was initially coined in 1956 (Zhang and Lu), and the underlying concept has subsequently driven significant social advancements, offering numerous advantages to humanity [1,2]. The field of hernia surgery and its integration of AI has been the subject of extensive review, particularly concerning the specific roles played by Machine Learning (ML) and Deep Learning (DL) [3-7]. AI is a broad category that incorporates several techniques, such as ML, Neural Networks (NN), and DL. While traditional ML models operate on organized and labeled information for prediction and typically demand considerable human involvement, DL—a more specialized type of ML—can work with raw, unclassified data, like images or video, to independently identify patterns and reach advanced inferences. This capability is achieved through the use of neural networks, which employ layers between the initial data input and the resulting output to grasp complex relationships [5]. Consequently, the relationship can be understood as hierarchical: AI represents the broadest domain, ML is a specific domain within AI that focuses on learning from data, DL is a subset of ML that employs deep NNs, and NNs themselves provide the structural framework that underpins deep learning (Figure 1).

Figure 1: The connection between these concepts follows a top-down structure, progressing from the general to the particular: Artificial Intelligence (AI) represents the broadest field, followed by Machine Learning (ML, a category of AI that focuses on data-driven learning), then Deep Learning (DL, a specialized form of ML that uses deep neural networks), and finally the neural networks (NNs) themselves that form the basis of deep learning capabilities.

Another benefit of AI is its ability to ensure greater productivity by facilitating the rapid execution of simple and repetitive tasks. Unlike humans, machines do not experience fatigue when performing tasks that would be tedious for humans. In addition, AI helps reduce the number of human errors [3]. Artificial intelligence is a revolutionary technology that significantly affects multiple aspects of surgery, including patient procedures, research, education, and management [8]. A key component is Generative AI (GAI), which uses neural networks and algorithms to synthesize new content, such as images, text, and even surgical protocols. Due to its capacity for providing innovative approaches to complex healthcare issues, this technology is poised to be truly transformative in the medical field [9]. Although still a developing field, the use of GAI is growing rapidly, indicating significant potential for surgeons [4]. Large Language Models (LLMs) such as GPT-3 and GPT-4 (Generative Pre-trained Transformer, OpenAI), LaMDA (Language Model for Dialogue Applications, Google), PaLM (Pathways Language Model, Google), and BLOOM (BigScience workshop and Hugging Face) represent GAI solutions that utilize a deep learning framework called the Transformer. The transformer framework is particularly well suited for handling sequential data, making it highly suitable for applications involving natural language processing (NLP) [4]. The primary objective of this research is to evaluate the potential of AI in various types of hernia surgery, with the exception of diaphragmatic and hiatal hernias. The specific hernia types that will be examined in this manuscript are inguinal, ventral, and incisional hernias.

To identify relevant literature, we conducted searches in PubMed and Google Scholar, and then examined the bibliographies of the retrieved articles to identify additional relevant studies. The keywords and their combinations used in our search strategy were: "artificial intelligence" combined with "abdominal hernia", "procedures", "abdominal wall", "inguinal hernia", "ventral hernia", "incisional hernia", and "hernia surgery". Publications were selected based on their direct relevance to the study topic. This included focusing on the full spectrum of abdominal wall reconstruction: preoperative preparation, surgical imaging and techniques, and follow-up care. Articles were rejected if they did not cover both AI in surgery and hernia repair, or if they were not peer-reviewed. We also excluded studies that did not involve human subjects and duplicate publications. The search was limited to articles published in English. All contributing authors agreed that the articles selected for this review were relevant to the topic.

Transforming Abdominal Wall Surgery with Generative Ai: Practice, Education, And Research

Generative AI is not intended to substitute the proficiency and knowledge of experienced surgeons; instead, it aims to augment their capabilities and improve the standard of care delivered to individuals undergoing hernia surgery. Generative AI is positioned to be a significant source of support for seasoned surgeons in clinical practice, offering several distinct advantages that can lead to enhanced results for patients [4]. A key method through which generative AI supports skilled surgeons is in the area of surgical preparation. Given the inherent complexity of hernia surgery, where each patient case presents distinct challenges, GAI can examine an individual's medical background, imaging results, and other pertinent details to develop tailored surgical strategies. By integrating this information, AI can furnish surgeons with crucial insights, assisting them in making well-informed choices regarding the surgical method, technique, and possible complications. These AI-developed plans serve as a guide that can substantially improve surgical accuracy. Furthermore, the potential of GAI is most evident in facilitating immediate decision-making during the surgical procedure itself [10]. Even highly skilled surgeons can face unforeseen circumstances during operations [4]. Researchers led by Elhage et al. [11] conducted a quality improvement analysis to evaluate deep learning's effectiveness, utilizing imaging data to predict the difficulty level of abdominal wall reconstruction surgery. A central goal was to project the necessity of component separation, alongside forecasting the likelihood of lung and wound complications. This study involved comparing a surgical complexity model based on a convolutional neural network with a validation set of Computed Tomography images. To summarize, this study revealed that the convolutional neural network-based DL model demonstrated greater accuracy than the assessments of expert surgeons in predicting the surgical complexity of abdominal wall reconstruction procedures [11]. In another article by Hassan et al. [12], the authors demonstrated the efficacy of machine learning models (MLMs) in predicting hernia recurrence, surgical site events, and 30-day readmission rates. The authors concluded by advocating for the incorporation of MLMs into the preoperative assessment of patients undergoing abdominal wall reconstruction. Generative AI presents a significant prospect for progressing surgical education and research, fundamentally changing the methods by which future surgeons are trained and how medical knowledge is acquired through the examination of extensive data collections [13]. This groundbreaking technology has the capacity to connect conventional medical instruction with the requirements of contemporary healthcare. In the context of surgical training, GAI can be crucial in preparing the upcoming generation of surgeons. AI-powered simulations are especially significant in this aspect [4]. Moreover, AI technology makes a substantial contribution to surgical research by facilitating the analysis of large and intricate medical data sets [10]. Specifically, within hernia surgery, AI can glean significant information from electronic medical records, medical imaging, and patient outcome data [4]. We believe that advancing AI's progress in hernia surgery necessitates the global expansion of patient databases. Increasing both the volume and diversity of these data sets is crucial, as it will provide a broader range of information for training AI models, ultimately leading to more accurate and comprehensive predictions [5].

AI Applications in Primary Abdominal Wall Hernia Repair

The incorporation of AI into hernia surgery has progressed through its capacity to enhance medical imaging and robotics. AI can improve computer vision by utilizing image acquisition and enhancement applications that support image-guided surgical procedures and computer-assisted diagnosis [14]. Robotic technology in hernia surgery allows for minimally invasive interventions, offering the dual benefit of reduced costs and shorter hospitalizations [15]. A key observation is the continuous growth of AI integration into medical imaging and robotics, a trend that persists despite implementation challenges.

Leveraging AI to Advance Inguinal Hernia Surgery

Inguinal Hernia Surgery is the medical intervention carried out to correct an inguinal hernia. An inguinal hernia develops when tissue, such as a portion of the bowel or fatty tissue, pushes through a weakened region in the abdominal muscles within the groin area, specifically through the inguinal canal. This canal serves as a pathway for the spermatic cord to descend to the testicles in males, and contains the round ligament of the uterus in females. The goal of inguinal hernia surgery is to reposition the protruding tissue and reinforce the abdominal wall by strengthening the weakened area to prevent the hernia from returning. The selection of the surgical approach is influenced by several elements, including the hernia's size and type, the patient's general health condition, and the surgeon's expertise [16]. Being a procedure frequently conducted worldwide, this type of repair is a major topic. Among the many identified studies, one report successfully illustrated AI's role in inguinal hernia surgery by employing an artificial neural network (ANN) to successfully anticipate patient results after the operation [17]. Another study assessed the efficacy of a convolutional neural network (CNN) in identifying the vas deferens [18]. This research showed that CNN could recognize and label images of the vas deferens during laparoscopic inguinal hernia surgery [18]. Additionally, O’Brien et al. utilized a Network Neural Model to predict the occurrence of long-term skin and soft tissue infections after hernia surgery [19]. The research by Alonso-Silverio et al. showed that AI can be applied to inguinal hernia surgery to create a laparoscopic training platform for providing online education to surgeons. The system's design integrated Python programming, an artificial neural network (ANN), and Raspberry Pi to deliver the training. The study's findings indicated that the system had the potential to boost surgeons' self-assurance and is applicable in settings with limited resources [20]. In a recently published study by Takeuchi et al. [21], the main aim was to develop an automated phase recognition system based on deep learning (DL) to identify surgical stages in Transabdominal Preperitoneal (TAPP) procedures and to investigate the link between surgical proficiency and the length of each stage. An AI model was trained to automatically recognize surgical phases from videos, and the study evaluated the correlation between how long each phase lasted and surgical skill levels. A fourfold cross-validation method was used to assess the AI model's performance, achieving accuracy rates exceeding 85%. In a comparative study by Mito et al. [7], the AI-driven visualization of anatomical reference points during peritoneal dissection in TAPP surgery demonstrated high precision and offered a sufficiently dependable system for surgical assistance. Moreover, in a scoping review by Taha et al. [3], the authors presented a thorough overview of the current aims concerning the incorporation of AI within the domain of hernia surgery. They emphasized the possible uses and advantages of AI in areas like medical imaging and surgical education. However, the authors also noted the limited quantity of published research on this particular subject, suggesting a deficiency in the current body of literature. Consequently, these points suggest that the field of inguinal hernia surgery has leveraged AI to enhance procedures and ensure the provision of high-quality patient care by anticipating and managing potential complications proactively [3].

Applications of AI in Ventral Hernia Surgery

Ventral hernia surgery is the medical procedure undertaken to correct a ventral hernia. A ventral hernia arises when tissue or an organ, such as a portion of the intestine, protrudes through a weak spot or opening in the abdominal muscles. This often develops at the location of a prior surgical cut but can also occur spontaneously. The purpose of ventral hernia surgery is to return the displaced tissue to its proper place within the abdomen and then strengthen the compromised area of the abdominal wall to prevent the hernia from recurring. This strengthening process typically involves stitching the muscle layers together and frequently includes the placement of a synthetic mesh to provide additional support [22]. The incorporation of AI in ventral hernia surgery has been noteworthy, although the application of this concept remains relatively new in this specific area. In their study, López-Cano et al. [23] proposed that utilizing AI methods could enhance patient care by integrating and analyzing a vast number of ventral hernia cases and categorizing them according to severity and priority. The presence of AI in ventral hernia surgery is demonstrated through various studies, among which we highlight the research by Elhage et al., which aimed to apply three deep learning models to forecast the complexity of the procedure and the occurrence of wound infections following a ventral hernia repair [12]. The results of this study indicated that the three image-based models, using computed tomography scans, were effective in predicting surgical complexity and more precise than the assessments made by experienced surgeons [12]. A study by Ayuso et al. [24] focused on developing and comparing deep learning models to predict infrequent but severe postoperative complications after abdominal wall reconstruction, using a database to identify patients with preoperative computed tomography scans. All the aforementioned points suggest that AI holds the promise of improving various aspects of ventral hernia surgery.

Smart Incisional Hernia Surgery: The Integration of AI

Incisional hernia surgery is the medical intervention performed to correct an incisional hernia. An incisional hernia is a specific kind of hernia that develops at the location of a prior surgical cut in the abdominal area. This type of hernia arises when the integrity of the abdominal wall musculature is compromised at a prior incision site, which permits underlying tissues or organs to push through the defect [25]. The occurrence of incisional hernias is increasing, largely due to the rise in obesity. Recognizing this challenge, specialists have taken the initiative to incorporate AI to enhance patient safety and forecast complications, thus avoiding adverse outcomes. Relevant to this effort, one study established that the implementation of ML algorithms in incisional hernia surgery aided surgeons in making sound decisions during the procedure, leading to a zero-complication rate following repair [26]. Considering all the aforementioned points, it's evident that models designed for the training of surgeons hold significant importance. In this context, we refer to the study by Zipper et al., which developed a model aimed at ensuring competency-based surgeon training. Their approach demonstrated strong reliability and success in improving the technical skills of surgeons [27]. Licari and colleagues [28] performed a crucial analysis using a Support Vector Machine (SVM) to identify the elements contributing to the recurrence of incisional hernias within a cohort of 154 patients. The technique demonstrated high predictive value, achieving 86.25% sensitivity and 86.67?curacy. It can therefore be concluded that incorporating AI into surgical management for incisional hernias will lead to improved outcomes and a lower incidence of recurrence.

A primary limitation of this study is the limited availability of comprehensive information on the subject. The literature search yielded a relatively small number of relevant documents for inclusion, suggesting that this area remains less explored than its importance might warrant. Another limitation of this review is the exclusive inclusion of articles published in English. This criterion may have contributed to the limited number of sources identified, as relevant information may exist in articles written in other languages.

It is clear that there is a significant gap in the literature due to the limited number of articles found suitable for this study. Further research is essential to determine how hernia surgeons and educators can effectively utilize AI. In addition, since the advent of AI, ethical considerations have been a prominent aspect of discussions surrounding this technology. While the potential benefits are clear, there are concerns about liability in the event of complications. The substantial advantages that AI offers to surgeons in abdominal wall procedures are clearly demonstrated across three primary domains: clinical practice, educational programs, and research efforts. Future researchers investigating this topic should focus on exploring the impact and advancement of AI in hernia surgery. It is recommended that researchers conduct original studies by developing and evaluating AI algorithms to take advantage of their various benefits in inguinal, ventral, and incisional hernia surgery.

Acknowledgments

The author does not have anyone to acknowledge.

Authors’ contributions

Made substantial contributions to conception and design of the study and performed data analysis and interpretation: Golubovic I, Vukadinovic A, Vanja Pecic.

Performed data acquisition, as well as provided administrative, technical, and material support: Golubovic I.

Availability of data and materials 

Not applicable.

Financial support and sponsorship

None.

All authors declared that there are no conflicts of interest.

Ethical approval and consent to participate

Not applicable.

Consent for publication

Not applicable.