Text

Objectives and questions: Over- and undertrimming in open acetabular rim trimming can compromise stability or leave residual impingement, increasing revision risk. The objective of this study was to verify the accuracy of a novel personalized 3D-printed guide on patient-specific 3D-printed bone models and on cadaveric hips.

Material and methods: We conducted a prospectively planned experimental accuracy study on 24 hips: 20 patient-specific 3D-printed hip models based on computed tomography (CT) scans from our database (deformities: acetabular protrusion, acetabular retroversion, posterior overcoverage) and 4 cadaveric hips. Guides were designed in CAD software and manufactured using stereolithography (SLA), while the patient-specific pelves were produced using fused deposition modeling (FDM). A CT-based 3D plan defined the target postoperative lateral center-edge angle (LCEA), lunate surface area change, sectoral center-edge angles, and trim depth. The planned resection geometry was transferred into the guide design. After open guide-assisted rim trimming, deviations from the plan were calculated as postoperative minus planned values; negative deviations indicate over-trimming. LCEA and lunate surface area were measured once per hip; sectoral angles and trim depth were measured at each of 12 clock-face positions per hip. Accuracy was assessed using Bland-Altman analysis, reporting mean deviation (bias) and 95% limits of agreement (LoA). Group differences (3D-printed models vs cadavers) were compared using Mann-Whitney U tests for LCEA and lunate surface area and mixed-effects models for sectoral angles and trim depth. Data are reported as mean ± SD; α=0.05.

Results: LCEA deviation was −0.2° ± 2.2° on 3D-printed models (n=20 hips; LoA −4.5° to +4.0°), with 13/20 (65%) within ±2°; cadavers showed +2.0° ± 1.4° (n=4 hips; p=0.03). Lunate surface area deviation was −8 ± 81 mm² on 3D-printed models versus +89 ± 110 mm² on cadavers (p=0.13). Sectoral angle deviation was −0.3° ± 1.5° on 3D-printed models (n=100 positions; LoA −3.3° to +2.7°), with 94/100 (94%) within ±2°; cadavers showed −4.0° ± 4.3° (n=29 positions; p<0.001). Trim depth deviation was +0.1 ± 0.6 mm on 3D-printed models (LoA −1.1 to +1.3 mm), with 96% within ±1 mm; cadavers showed +1.5 ± 2.1 mm (p<0.001).

Discussion and conclusions: The personalized 3D-printed guide reproduced the planned rim resection with near-zero mean deviations for LCEA, lunate surface area, sectoral angles, and trim depth on patient-specific 3D-printed bone models. Accuracy decreased on cadaveric hips with cartilage-seated guides, suggesting guide seating on cartilage as the key translational challenge. Clinical validation and design optimization are warranted.