Patient-specific 3D-Printed jigs for high tibial osteotomy: a preclinical simulation study using affordable 3D printing

Study design

This preclinical investigation was structured to evaluate the feasibility of affordable 3D printing for HTO (Figure 1). The primary outcomes of this proof-of-concept study were corrections achieved in MPTA and PPTA with respect to the virtual plans, making them paired continuous measurements. Therefore, sample size estimation was based on the precision of the 95 % confidence interval (CI) of the mean paired difference [18]. For paired data, the 95 % CI sample size can be calculated as:

where, SD is the expected standard deviation of the paired difference, and t is the critical value of the Student’s t-distribution. Using an SD of 1.5° reported by Hiranaka et al. [11] for MPTA measurements in PSI-guided HTO, and specifying a desired precision (half-width) of ±1.25°, the sample size was estimated using the large-sample critical value t=1.96:

Figure 1:

Schematic chart of the HTO planning and simulation process, illustrating the workflow from CT scan acquisition to surgical simulation using 3D-printed jigs and tibia prototypes.

Accordingly, a minimum of six paired tibial measurements was required to achieve the desired CI precision. To account for practical variability and to strengthen the estimation, eight tibiae were included in the study. Following the institutional ethical clearance (Ref. No: IEC-INT/2025/Study-3149), high-resolution CT data of eight full-length adult tibiae were obtained from the department repository. The DICOM format scans were imported into Materialise Mimics software, a widely recognised platform for medical image processing. Within this environment, the DICOM files were segmented and converted into 3D CAD models (STL format), enabling detailed virtual planning of HTO. A board-certified orthopaedic surgeon methodically defined surgical parameters, including cutting trajectories, K-wire placements, spacer arrangements, and target values for the MPTA and PPTA. This virtual planning phase was critical, establishing the reference against which physical outcomes would be measured, simulating the preoperative process used in clinical settings.

3D printing

Patient-specific jigs were designed to fit each tibia model precisely, incorporating guides for osteotomy cuts and drill placements (Figure 2). These jigs were fabricated using an Anycubic Photon Mono X™ (Hongkong Anycubic Technology Co., Limited), a resin-based 3D printer costing approximately $550, selected for its affordability and high surface resolution. Printing was performed with PLA-based photopolymer resin (∼$40 per kg), with parameters optimised for accuracy: a layer height of 0.05 mm, an exposure time of 8 s per layer, a lifting speed of 60 mm/min, and a 2-s light-off delay. Concurrently, tibia prototypes were printed on a Creality Ender 5 Plus™ (Shenzhen Creality 3D Technology Co., Ltd.), a fused deposition modelling (FDM) printer priced around $600, using PLA filament (∼$15 per kg). The printer is one of the most economical, globally available solutions for FDM printing of models of the size of a human tibia. Printing parameters included a 0.2 mm layer height, 20 % infill density, and a 45° raster angle, balancing strength and material efficiency [19]. For post-processing, resin jigs were washed in isopropyl alcohol and cured under UV light to eliminate residual monomer. Both printers adhered to manufacturers’ calibration standards, ensuring reproducibility. This dual-printer approach allowed the creation of precise jigs and robust bone models at a fraction of the cost of industrial systems. A brief description of the associated costs, space and power specifications [20] of the two printers is presented in Table 1.

Figure 2:

Virtual cutting (a), spacer placement (b), designed jig (c), and its attachment view (d).

Surgical simulation

Medial opening wedge high tibial osteotomy (MOWHTO) was simulated using the 3D-printed tibial models and corresponding patient-specific instrumentation (Figure 3). The instrumentation comprised the patient-specific cutting jigs and a spacer designed to reproduce a predetermined medial opening wedge. All procedures were performed by a board-certified orthopaedic surgeon experienced in deformity correction. The surgeon was blinded to the virtual preoperative plan and relied exclusively on the patient-specific instrumentation to achieve the intended correction.

Figure 3:

3D printed bone and cutting jig (a), K-wire-aided attachment (b), wedge opening (c), spacer placement (d), and plate fixation (medial and anterior views) (e–f).

Each 3D-printed tibial model was secured on a stable work surface to permit controlled execution of the osteotomy. The patient-specific cutting jig was positioned on the medial aspect of the proximal tibia, conforming to predefined bony landmarks. Proper seating of the jig was confirmed visually. The jig was then fixed to the bone model using two 1.5 mm Kirschner wires inserted through pre-designed guide apertures. Prior to proceeding, the surgeon confirmed the absence of perceptible relative motion between the jig and the bone model, ensuring stable jig fixation.

An oscillating saw was introduced through the cutting slot of the jig, and the medial opening wedge osteotomy was performed along the predefined osteotomy plane. The cut was advanced in a controlled manner up to the planned lateral cortical hinge, with care taken to preserve hinge integrity. Following completion of the osteotomy, the cutting jig was removed, and a conventional high tibial osteotomy spreader was inserted into the osteotomy gap. The wedge was gradually opened under direct visual supervision, with incremental distraction performed until the patient-specific spacer could be introduced. No navigation systems or external alignment aids were utilized to evaluate the intrinsic accuracy of the jig-guided correction.

Once the desired opening was achieved using the spacer, the osteotomy was stabilised with a standard medial locking plate and screw construct applied to the proximal tibia. Fixation was intended solely to maintain the achieved alignment during subsequent measurements. There were no failed fittings and repeat prints. As the procedure was performed on a standalone model, no physiological loading, soft-tissue constraints, or biological healing considerations were incorporated. The simulation was therefore limited to assessing the geometric accuracy and reproducibility of the planned correction.

Accuracy assessment

Postoperative models were scanned using the Atos Core 300™ (GOM GmbH), a high-resolution 3D scanner, which generated detailed STL files of the realigned tibiae. This step captured the physical outcomes of the simulated surgery, allowing for a direct comparison with the virtual plan. Accordingly, the accuracy was evaluated by superimposing postoperative STL files onto preoperative virtual plans within Materialise 3Matic [21] (Figure 4(a)).

Figure 4:

Accuracy assessment of postoperative tibial alignment using 3D model superimposition. (a) Superimposition of the postoperative tibial STL model obtained from high-resolution 3D scanning onto the preoperative virtual plan, with a color-coded surface deviation map illustrating differences between planned and achieved corrections. (b) Measurement of the medial proximal tibial angle on the overlapped models. (c) Measurement of the posterior proximal tibial angle (PPTA) on the overlapped models.

For MPTA measurement, the tibia was oriented in a true anteroposterior position, as confirmed by the symmetric appearance of the medial and lateral tibial plateaus and cortices, with the tibial shaft centrally aligned and no visual evidence of axial rotation [22] (Figure 4(b)). The tibial mechanical axis was defined by a line connecting the centre of the proximal tibial metaphysis to the centre of the distal tibial articular surface. The proximal tibial joint line was drawn as a line tangent to the tibial plateau, connecting the highest points of the medial and lateral tibial plateaus. The medial angle between these two lines was recorded as the MPTA.

For PPTA measurement, the tibia was oriented in a true lateral position with superimposition of the tibial plateaus (Figure 4(c)). The proximal tibial joint line was defined by connecting the anterior and posterior margins of the tibial plateau. The sagittal plane–modified tibial mechanical axis was constructed from a point located at approximately one-fifth of the proximal tibial joint line from the anterior cortex to the centre of the distal tibial articular surface. The posterior angle between these two lines was recorded as the PPTA [23]. The deviations in planned and achieved angles were recorded, and mean errors were calculated (Table 2). The following statistical tests were performed in MATLAB (R2024, Natick, Massachusetts: The MathWorks Inc.): (i) Paired t-tests for systematic bias, (ii) Bland-Altman plots for agreement, (iii) Linear regression (achieved ∼ planned) for proportional bias (H₀: slope=1), and (iv) Two one-sided t-tests (TOST) for ±1° equivalence.

MPTA outcomes

The mean difference between achieved and planned MPTA was −0.05° ± 1.32° (SD), with an absolute mean error of 1.13° (range: −1.45° to +2.23°). A paired t-test confirmed no significant systematic bias (p=0.912). Bland-Altman analysis revealed 95 % limits of agreement from −2.66° to +2.55° (Figure 5(a)), which lie within the clinically accepted ±3° tolerance for high tibial osteotomy [16], 17]. Linear regression demonstrated a strong positive correlation between planned and achieved values (R²=0.77, slope=0.59 [95 % CI: 0.264–0.921], intercept=35.452°, RMSE=0.901°) (Figure 5(b)). The slope was significantly less than 1.0 (p<0.05), indicating proportional under-correction at higher planned angles [11]. Two one-sided tests (TOST) confirmed statistical equivalence to zero within a ±1° margin (p=0.042; 90 % CI of mean difference: [−0.944°, +0.836°]), demonstrating no clinically meaningful systematic error despite the reduced slope.

Figure 5:

Planned vs. achieved MPTA. (a) Bland-Altman plot; mean bias: −0.05°; 95 % LoA: −2.66° to +2.55° (all within ±3° tolerance). (b) Regression analysis: R²=0.77, slope=0.59 (p<0.05 vs. 1.0), RMSE=0.901°.

PPTA outcomes

Achieved-minus-planned mean difference was +0.57° ± 0.38° (SD), with an absolute mean error of 0.57° (range: +0.02° to +1.10°). The paired t-test showed a statistically significant over-correction (p=0.004). Bland-Altman analysis demonstrated tight 95 % limits of agreement from −0.17° to +1.31° (Figure 6(a)), well within clinical tolerances. Linear regression revealed an excellent correlation (R²=0.99, slope=1.04 [95 % CI: 0.970–1.112], intercept=−2.947°, RMSE=0.352°) (Figure 6(b)). The slope was not significantly different from 1.0 (p>0.05), and the intercept was not significantly different from zero, confirming proportional and additive accuracy. TOST equivalence testing confirmed statistical equivalence within a ±1° margin (p=0.007; 90 % CI of mean difference: [+0.318°, +0.824°]), assuring no clinically relevant bias despite the slight over-correction.

Figure 6:

Planned vs. achieved PPTA. (a) Bland-Altman plot; mean bias: +0.57°; 95 % LoA: −0.17° to +1.31°. (b) Regression analysis: R²=0.99, slope=1.04 (p>0.05 vs. 1.0), RMSE=0.352°.