Page 70 - Ebook-Book JCMS 2025

AI-Assisted Surgical Planning in Spine Surgery: Accuracy
and Clinical Outcomes
Chawakrit Pansritoom1*, Suppawut Pruetiworanan 2
1 Faculty of Medicine, Chulalongkorn University, King Chulalongkorn Memorial Hospital
2 Department of Orthopedic Surgery, Nopparat Rajathanee Hospital
*Corresponding Author E-mail: newsc27032014@gmail.com
Background: Abstract
Spine surgery requires precise alignment targets and optimal implant selection to achieve
favorable biomechanical and functional outcomes. In recent years, artificial intelligence
(AI)-assisted surgical planning systems—utilizing machine learning and image processing
—have been developed to aid preoperative planning in spine surgery. These systems
promise to improve sagittal alignment accuracy, reduce variability in surgical decisions,
and potentially enhance patient-reported outcomes. However, their clinical effectiveness
remains uncertain. This meta-analysis aims to evaluate the accuracy and clinical outcomes
of AI-assisted surgical planning compared to conventional planning in spine surgery, with
specific focus on alignment achievement, operative parameters, and functional scores.
Methods: Following PRISMA guidelines, a systematic literature search was performed across
PubMed, MEDLINE, and Scopus until March 2025. Eligible studies included studies that
compared AI-assisted preoperative planning tools with conventional planning in spinal
fusion or deformity correction surgeries. Primary outcomes were alignment parameters
(e.g. PI-LL mismatch, SVA) and operative time. Secondary outcomes included Oswestry
Disability Index (ODI), Visual Analog Scale (VAS) scores, and revision surgery rate.
Results: Eight studies with a total of 1,276 patients were included. AI-assisted planning significantly
improved alignment accuracy, with a greater proportion of patients achieving target PI-LL
mismatch <10° (OR: 2.42 [1.38, 4.25]; p = 0.002). Operative time was reduced by a mean of
18.6 minutes (MD: -18.58; 95% CI: -25.74 to -11.42; p < 0.001). Functional outcomes also
favored AI, with greater ODI improvement at final follow-up (MD: -5.23 [−8.13, −2.33];
p = 0.0005). Revision rates showed no significant difference.
Conclusions: AI-assisted surgical planning in spine surgery improves alignment accuracy, reduces
operative time, and enhances early functional recovery. These findings support the
adoption of AI as an adjunctive tool in modern spinal surgical practice, though further
high-quality trials are warranted.
68 Joint Conference in Medical Sciences 2025