Surgeon-oriented three-dimensional planning for supracondylar humeral malunion

Malunion of supracondylar humeral fractures is often associated with complex three-dimensional (3D) rotational deformities involving the coronal, sagittal, and transverse planes. Pre-operative planning for corrective osteotomy typically requires computer-aided design engineers, limiting surgeons' ability to independently understand deformity patterns and adjust surgical strategies intraoperatively. A practical, surgeon-oriented method for quantifying 3D rotational deformities is, therefore, needed. Herein, we established a surgeon-oriented computational framework that enables quantitative assessment of 3D deformities based on computed tomography-derived 3D bone models and allows surgeons to independently perform pre-operative planning using 3D-computer-aided design software.

We developed a 3D rotation analysis method based on mirrored bone models of the unaffected humerus. After superimposition of proximal segments, object-specific coordinate systems were defined using a centroid and 2 anatomical landmarks. Rotational deformities were calculated using 3 approaches: wedge osteotomy-based rotation (Method A), 3D rotational osteotomy (Method B), and coronal-plane-only rotation (Method C). Validation was performed using computed tomography data from 5 patients with extra-articular supracondylar humeral malunion. Robustness was assessed by calculating rotational deformities for all 21 landmark-pair combinations derived from 7 predefined landmarks and by evaluating sensitivity to coordinate rounding.

Across all patients, rotational components calculated using Methods A and B showed minimal variability across landmark-pair selections, with median within-patient ranges below 0.02° for all components. Method C demonstrated a larger absolute range but remained consistent relative to the magnitude of rotation. Rounding landmark coordinates to 1 decimal place resulted in only minor changes in calculated rotations, whereas integer-level rounding produced substantial deviations, particularly in transverse-plane components. In addition, the coronal-plane rotational components derived from Methods A and B closely matched the single-axis rotation obtained by Method C, with absolute differences within 3° across all patients.

This study demonstrates a robust and surgeon-oriented method for quantifying 3D rotational deformities in extra-articular supracondylar humeral malunion. The method is insensitive to landmark selection and clinically realistic coordinate precision, which may reduce dependence on dedicated engineering support or outsourced services. Furthermore, consistency between 3D and simplified coronal-plane correction estimates supports the practical applicability of this approach in routine surgical decision-making.