# Time‐ and cost‐effective manufacturing of 3D templates of the pelvis and acetabulum in primary and revision hip arthroplasty using semiautomatic segmentation

## TL;DR

This paper introduces a fast and low-cost method to create 3D models of the pelvis and hip for surgical planning using open-source software and 3D printing.

## Contribution

The study presents a semiautomatic workflow for generating 3D models using free tools, significantly reducing time and cost compared to traditional methods.

## Key findings

- Semiautomatic segmentation reduced segmentation time to less than 10 minutes after a learning curve of five segmentations for native hips.
- Generated models were anatomically accurate and suitable for surgical planning with costs under 5 euros per model.
- The workflow is accessible for most clinical environments using open-source software and standard 3D printers.

## Abstract

To develop and evaluate a time‐ and cost‐efficient workflow for generating three‐dimensional (3D) printable models of the pelvis and acetabulum using semiautomated segmentation with free and open‐source software. The goal was to enable a streamlined production of individualized models suited for preoperative planning and education in primary and revision hip arthroplasty (THA).

A semiautomated segmentation protocol was created using 3D Slicer and the TotalSegmentator module, followed by targeted manual refinement. This was compared to traditional manual segmentation using computed tomography (CT) data sets from patients undergoing complex primary or revision THA. The workflow was validated by two participants, who independently performed the segmentation and 3D printing process after studying the step‐by‐step guide as provided in this article for 8 native hips and 12 pathologically altered hip joints. Segmentation time, material use and print duration were recorded. A learning curve analysis was conducted via a logarithmic regression analysis.

The semiautomated workflow reduced segmentation time considerably compared to manual segmentation. Both investigators were able to learn and apply the workflow with increasing speed over successive trials, indicating a steep learning curve. Segmentation times amounted to a mean of less than 10 min after a learning curve of five segmentations in native hip models and eight segmentations for pathologically altered joints. The generated models were anatomically accurate and suitable for surgical planning. Cost analysis showed substantial savings compared to commercial outsourcing with mean costs of less than 5 euros for both native and pathological hip models.

The presented workflow offers a fast, accessible, and low‐cost method to produce patient‐specific 3D models for THA planning. It can be adopted in most clinical environments using freely available software and standard 3D printers. This approach may help democratize access to 3D printing technology in orthopaedic surgery and lay the foundation for future clinical applications.

Level IV.