Page 23 - Read Online

Page 259 Toossi et al. Art Int Surg 2024;4:258-66 https://dx.doi.org/10.20517/ais.2024.27

surgery is associated with high complication rates in both the short and long term. These observations make
ASD an ideal candidate for leveraging the significant potential offered by artificial intelligence and machine
learning (ML).

Computational techniques have been used in the past several years to process large datasets and create
complex mathematical models to determine the relationship between different variables affecting the
outcomes of surgery. The idea behind ML, a subset of artificial intelligence, is to develop a system similar to
the human brain to learn from clinical and radiographic data and apply the knowledge to new situations. In
other words, ML employs computer algorithms to learn from data and past experiences, enabling the
creation of intelligent models. These algorithms enable computers to identify patterns in datasets without
relying on predefined rules, allowing them to learn relationships from the data and make predictions or
decisions based on that knowledge. It has been shown that validated ML risk calculators can provide more
accurate and objective prognoses to adjust patient expectations during patient care than expert surgeons’
[2]
perception of risks in ASD surgery . The development of predictive models via ML algorithms for
prognosticating patient outcomes following ASD surgery represents a significant advancement over
traditional statistical models, which are more adept at identifying statistical associations between variables
rather than providing predictive value .
[3]

ML has shown promise in enhancing the accuracy and efficiency of various medical procedures, including
spinal surgery. By taking advantage of large datasets and advanced algorithms, ML can assist surgeons in
preoperative planning, intraoperative decision-making, and postoperative care, leading to improved patient
outcomes.

The aim of this narrative review is to provide an overview of the current status of ML in
enhancing spinal deformity correction surgery and its applicability in preplanning, intraoperative
guidance, predictive modeling, and postoperative risk assessment.

METHODOLOGY
The authors conducted a non-systematic review of recent literature to support their perspectives on the
applicability of ML in corrective spine surgery for adult ASD. This narrative review addresses three key
stages in surgical practice [Figure 1] where ML can be impactful, and concludes by discussing the major
challenges and future directions in the field.

Preoperative planning
Appropriate preoperative patient selection significantly impacts patient satisfaction, individualized decision-
making by surgeons, and hospital resource utilization. Identifying patients with favorable outcomes
preoperatively is a challenging task. Traditional statistical methods, such as multiple regression analyses, are
better suited for hypothesis testing rather than predicting individual patient outcomes. In contrast, ML
algorithms can readily identify patterns within large datasets without the need to test a specific
[4]
hypothesis . However, this advantage of ML algorithms comes at the cost of interpretability. Predictive
models generated by ML are more difficult to interpret than risk factors identified by traditional statistical
[3]
tests .
ASD patients exhibit significant heterogeneity in demographics, comorbidities, spinal pathologies, and
genetic factors. Traditional outcome predictive models often overlook these individual variabilities, leading
to suboptimal predictions. However, ML models excel in accounting for these differences. By analyzing
comprehensive datasets that include detailed individual patient profiles, these models can generate
personalized predictions, enhancing clinical decision-making and patient outcomes.