PhD Scientific Days 2026

Surgical Medicine

Integration of AI and Parametric Scripting in the Workflow of 3D Surgical Guide Design

Előadó neve

Dr. Hajnal, Benjamin

Neptune code

MM14PU

Előadó munkahelye

3D Lab, National Center for Spinal Disorders

Előadó telefonszáma

+36705989752

Előadó e-mail címe

hajnal.beni@gmail.com

Az előadás címe

Integration of AI and Parametric Scripting in the Workflow of 3D Surgical Guide Design

Szerző(k) neve és munkahelye

Dr. Benjamin Hajnal¹, Agoston Jakab Pokorni¹, Arnold Tomcic², Prof. Dr. med. Dezsö J. Jeszenszky, PhD³, Dr. Peter Endre Eltes, PhD¹

1: 3D Lab, National Center for Spinal Disorders
2: University of Medicine and Pharmacy of Târgu Mureș
3: National Center for Spinal Disorders

Bemutatás módja

Szóbeli

Szekció

Surgical Medicine

Language of the presentation

English

Preferred session

Surgical Medicine

Összefoglaló szövege

Introduction
Patient-specific surgical guides (PSSGs) manufactured via 3D printing offer a precise, cost-effective alternative for pedicle screw placement. However, labor-intensive computer-aided design (CAD) workflows remain a critical bottleneck limiting Point-of-Care adoption.
Aims
Our aim was to develop and validate a semi-automated, Python-based parametric scripting workflow for PSSG design, evaluating its efficiency and in vitro geometric accuracy against an optimized manual baseline.
Methods
Ten retrospective patient cases (104 screw trajectories) were included, spanning complex spinal deformities, cervical pathologies, and a normative lumbar control. Commercial AI-based segmentation served as a prerequisite step. A Python script developed within the 3-matic environment automated socket generation, surface wrapping, Boolean operations, and documentation. Manual designs were produced by a single expert biomedical engineer as a best-case control. Surgical guides were printed via stereolithography and tested in vitro on FDM-printed anatomical models by a spine surgeon. Post-interventional CT scans were analyzed for entry point distance and angular deviation. Non-inferiority was assessed using mixed-effects models with predefined margins of 0.5 mm and 1.0°.
Results
AI-assisted segmentation reduced total segmentation time from 90.0 ± 49.2 to 37.6 ± 34.6 minutes. Average design time decreased from 37.7 ± 15.7 minutes (manual) to 3.0 ± 1.0 minutes (automated) per template. Mean entry point distance and angular deviation were 0.55 ± 0.35 mm and 1.73 ± 0.81° (manual) versus 0.46 ± 0.33 mm and 1.79 ± 1.19° (automated). The automated method passed both non-inferiority tests. However, 0.9% and 2.8% of automated trajectories exceeded clinical safety thresholds of 2.0 mm and 5.0°, respectively, attributed to reduced fitting surface area, and smooth or heavily undercut surface anatomies.
Conclusion
Parametric CAD scripting reduces surgical guide design time by over 92% while maintaining non-inferior average accuracy. Identified failure modes on specific anatomical surfaces require targeted algorithmic refinement. The developed framework provides a scalable, auditable foundation for Point-of-Care manufacturing and prospective clinical validation.
Funding
No external funding declared.

University

Semmelweis University

Supervisor

Dr. Peter Endre Eltes, PhD

Publication of my abstract

I do not give consent to the publication of my abstract on the website of the congress.

phd.section.field

after finishing doctoral studies with absolutorium (PhD)

Kind

Szabad

Status

elfogadva

Accepted presentation method

szóbeli

Előadás fájl jóváhagyás

nem rendelkezett róla

Előadó

7389

Start

16:45

End

16:55