Revolutionizing Hepatobiliary Surgical Planning: A Randomized Comparative Analysis of Deep Learning-Enhanced 3D Printing Versus Traditional Digital Simulations

Abstract Design: Our study was conducted in three phases to evaluate the efficacy of 3D-printed liver models in surgical planning, culminating in a randomized controlled trial that compared the outcomes of physical 3D models with digital simulations in hepatobiliary surgery. Methods: Phase one involved developing and validating 35 3D-printed liver models, assessing their timeliness, cost, precision, and alignment with digital models. Phase two optimized the 3D reconstruction process using deep learning algorithms to compare AI-assisted automatic segmentation with manual methods, enhancing efficiency and accuracy. Phase three, a randomized controlled trial, divided 64 patients into two groups to receive surgical planning based on either physical 3D models or digital simulations, providing an unbiased comparison of surgical outcomes such as resection extent, operation duration, blood loss, and hospitalization. The trial was registered at ClinicalTrials.gov (NCT06526754). Results: Phase One quickly produced 3D-printed liver models that matched intraoperative findings, featuring high transparency and detailed vascular structures, within 3.52 hours and at approximately $152 each. Phase Two demonstrated that the AI group significantly reduced reconstruction time to an average of 303.50 minutes versus 557 minutes for the manual group. In Phase Three, patients planned with 3D models had significantly less intraoperative blood loss, highlighting the models' utility in enhancing surgical accuracy. Conclusion: Integrating deep learning with rapid 3D printing significantly reduced the production time of precise liver models, enhancing surgical planning and execution. This innovation boosts surgical safety and efficiency and offers a scalable, cost-effective approach that could revolutionize hepatobiliary surgery within precision medicine.