Page 27 - Read Online

Walker et al. Art Int Surg. 2025;5:345-9 https://dx.doi.org/10.20517/ais.2025.02 Page 347

[8]
several years after their surgery, as mesh-related complications may take years to present . Similarly,
patients who are deemed at lower risk for complications may be appropriately operated on closer to home
to promote appropriate resource utilization for AWR patients.

The use of AI in hernia surgery extends beyond outcome prediction. For instance, a group led by Takeuchi
from the IRCAD institute helped to develop an AI model that identified the phases of a laparoscopic
[9]
inguinal hernia repair through video review . The phases of these operations were linked with experience
level and were postulated to help with monitoring surgical training and even increasing operating room
efficiency. Using AI in the operating room could also aid in assessing the completeness of an operation (e.g.,
the critical view of the myopectineal orifice) and in recognizing key anatomical structures, such as
identifying the semilunar line during a minimally invasive transversus abdominis release. The robotic
platform may not only adjust its instruments and camera based on visual cues, but eventually, it may be able
to assist or complete parts of the procedure. Robotics can also provide data on surgeon ergonomics and
kinematics, which can aid in surgical coaching and improve surgical performance.

While the use of AI algorithms in the preoperative and intraoperative phases of hernia surgery is already
being defined, the role of AI in the postoperative phase is less clear. As mentioned, if patients are identified
to be at risk for a postoperative complication, they may be followed by the surgical team for a longer time
period postoperatively or have their care altered at the time of surgery. For instance, a patient at high risk of
pulmonary failure may be monitored in the intensive care unit. Similarly, a patient at high risk of hernia
recurrence may receive a more robust mesh overlap or even a different type of mesh. An area that is ripe for
investigation is the use of AI to detect complications or deviations from the standard postoperative
pathway. Data from automated patient monitoring systems may be leveraged to accurately determine
patients exhibiting abnormal vital signs or physiological disturbances that may indicate impending clinical
deterioration .
[10]
While promising in its application in hernia surgery, the widespread adoption of AI has also been critiqued.
Some argue that using AI models for preoperative planning lacks reliability and promotes overreliance on a
[5]
technology that we do not fully understand . For example, in the DLMs that were developed using CT
images, we are unable to decode the black box and know what features of the images are important when
predicting outcomes. AI models are only as good as the data they are trained on. If the training data are
incomplete, biased, or not representative of diverse patient populations, the model’s predictions may be
inaccurate or misleading . This could lead to suboptimal surgical plans or overconfidence in AI-generated
[11]
recommendations. Prematurely marketing certain AI models or applications may actually have an adverse
effect on the integration of this technology in the long term and negatively impact public perception. It is
important that AI augment, rather than replace, clinical knowledge and gestalt, ensuring that patient care is
improved rather than compromised. AI is not only limited by the integrity of the data for model
development, but to this point, still relies on surgeons to have the technical skill to carry out its
recommendations. AI should enable healthcare providers to become more efficient and focus on what is
most important.

There should be a reframing of how AI is used in clinical medicine and hernia surgery. AI models should be
viewed as decision-support tools, providing additional insights to complement a surgeon’s clinical
judgment . Moreover, surgeons remain the ultimate decision-makers, critically evaluating and adapting
[12]
the model’s recommendations based on their expertise and real-time observations. AI has the capacity to
supplement rather than supplant the role of the surgeon. Having more information at the disposal of
surgeons can lead to more informed clinical decision making and improve patient outcomes. Hernia