1 Introduction

A high tibial osteotomy (HTO) is an established treatment option for tibial deformities, with good long-term survival and patient-reported outcomes.¹ Its success relies heavily on achieving the planned correction angle, as under-correction can cause recurrence of varus deformity and continued compartment overload, while over-correction may overload the contralateral compartment.^2–4 Open-wedge HTOs often show limited accuracy, with studies reporting suboptimal coronal correction,^5–7 with a general tendency to under-correct.⁶ Patient-specific three-dimensional (3D) CT-based operative planning and 3D-printed cutting guides (PSCGs) may improve HTO accuracy by enabling individualized, preoperative planning on a virtual bone model. PSCGs guide the osteotomy and screw placement, allowing surgeons to replicate the planned correction. This can be especially valuable in complex, multiplanar osteotomies requiring precise alignment in all three planes.

Although increasingly used, evidence showing superiority or non-inferiority over conventional techniques is limited. PSCGs have shown to safely achieve HTO corrections within two degrees of the planned alignment.^7–11 Two studies compared patient-specific to free-hand HTOs, with conflicting results on accuracy and outcomes.^12,13 No studies have directly compared these specific PSCGs to conventional HTO. In addition, there is no good estimate of the additional costs that 3D-planning and -printing entails.

This study compared the accuracy of CT-planned HTO usings PSCGs (PSCG group) versus conventionally executed HTO (CONV group). Thereby, we aimed to gain insight into the clinical implementation of this new innovation into the clinic, by measuring care-related costs and reporting logistics of implementation of CT-planning and PSCGs. As no prior studies have made this comparison, this study focused on uniplanar osteotomies as a first step towards future application in complex multi-planar cases.

2 Methods

This exploratory study took place from January 2021 to November 2022 at the Sint Maartenskliniek, Nijmegen, the Netherlands, in accordance with the Declaration of Helsinki.¹⁴ Ethical approval was obtained from CMO Regio Arnhem-Nijmegen (NL72556.091.20). The original study was set up as a randomized controlled trial, with n = 56 as initial sample size. The trial was stopped due to severe delay in patient inclusions and logistic challenges. Because valuable knowledge and experience was gained, it is important to share the results of the first batch of patients in this study.

All patients scheduled for an HTO at the Sint Maartenskliniek were screened and checked for eligibility to participate in this study based on the following inclusion criteria: a medial overload as a result of a varus tibial deformity (medial proximal tibial angle (MPTA) < 88°) or medial osteoarthritis (MPTA < 90°) indicated for an HTO. Further inclusion criteria: age between 18 and 65 (male) or 60 (female); maximal Kellgren-Lawrence (KL) grade 3 osteoarthrosis (OA) of the medial compartment; maximal KL grade 1 OA of the lateral compartment. Exclusion criteria: flexion contracture >10°, significant anterior-posterior (AP) or mediolateral instability as a result of ligament deficiency, smoking, a body mass index >35 and rotational or severe slope deformities in need for correction in multiple planes. As this was the first use of CT-planning and PSCGs for regular HTO surgery within our clinic, it was decided to keep the groups as simple and homogeneous as possible.

All patients provided written informed consent and were divided over the CONV and PSCG group with an allocation ratio of 1:1 using block randomization by electronic data capture system Castor EDC (https://www.castoredc.com/) with a variable block size of either 2 or 4. Patients were blinded for group allocation until after surgery. The outcome data was blinded until data analysis was completed.

A pre-operative long-leg radiograph and low-dose CT scan for PSCG modelling (including hip and ankle) were made for all patients. A patient file was created in Newclip Technics’ secure portal (Haute Goulaine, France). Anonymized patient data, including surgical side and gender, along with imaging, were uploaded to initiate the CT-guided planning and PSCG process. The surgeon specified the desired valgus over-correction (HKA) or target MPTA, and, when applicable, details such as distal or proximal tuberosity osteotomy. Based on the CT and these inputs, Newclip Technics proposed an osteotomy and plate position plan, which was reviewed and approved by the surgeon before PSCG production. In parallel, all patients underwent conventional planning, on a weightbearing long-leg radiograph, using the Miniaci method (i.e. conventional method).¹⁵

All patients underwent an open-wedge HTO, carried out by 2 experienced surgeons. The surgeons had 7 and 11 years experience in knee practice with both approximately 15 osteotomies per year, they had limited experience with the specific Newclip Technics PSCGs and the ACTIVMOTION plate. Each used the Newclip PSCGs and plate twice before starting the study to get used to the specific technique and plate. The HTO fixation method was with the ACTIVMOTION plate (Newclip Technics, Haute Goulaine, France),¹⁶ a widely used and effective fixation plate for osteotomies that is compliant with MDR and FDA approved. The CT-planned correction angle from the finalized plan was the planned correction angle in all patients. In the PSCG group the PSCG was designed to act as an intra-operative guide, but the surgeon was still expected to use the standard checks prior to bone cutting or drilling. Patients randomized to the CONV group were operated using the osteotomy technique described in the next section.

2.1

2.1 Surgical technique

2.1.1

2.1.1 Conventional osteotomy technique

Conventional osteotomies were performed using the technique by Lobenhoffer et al.¹⁷ Under fluoroscopic guidance, two k-wires were placed to outline the osteotomy angle, position and depth. The osteotomy length was measured from the K-wire, to determine the appropriate saw depth. The osteotomy cut was completed in a biplanar manner at the tibial tubercle, distal or proximal depending on patella height and amount of wedge opening. The wedge was gradually opened to the correction length specified in the CT-based plan. The ACTIVMOTION plate was fixed with six screws. Intraoperative imaging was used to verify K-wire(s) placement, if the hinged point is intact after opening of the wedge, leg alignment after opening of the wedge, plate and screw positioning in two planes before and after fixation.

2.1.2

2.1.2 PSCG osteotomy technique

The PSCG HTO was performed as described by the company (Newclip Technics, Haute Goulaine, France). After clearing soft tissue, the PSCG was positioned and secured with a K-wire, ensuring its unique, contour-specific fit (Fig. 1A). K-wire placement was verified with fluoroscopy and compared with the planning file. A second protective K-wire, crossing the hinge point, was inserted through the guide and served as a mechanical stop. The six screw holes were drilled through the guide, and temporary pins provided additional stability. The osteotomy cut was sawn through a specific portion of the PSCG, the proximal guide portion could then be removed for better posterior access (Fig. 1B). After completing the cut and removing the distal guide, a tuberosity cut (proximal or distal) was made to allow opening. The osteotomy was gradually opened using a stacked osteotome technique until the 6 pre-drilled holes were aligned with the Activmotion plate, indicating the planned correction was achieved (Fig. 1C). The aimed correction was confirmed using intra-operative imaging.

Close

Fig. 1 Images from the planning documentation of Newclip Technics (Haute Goulaine, France). a: The guide is positioned on the bone, and the K-wires are placed in the correct locations. The saw passes through the guide at the designated cutting plane, and the screw holes are drilled into the six screw holes of the guide. b: The tibia after the cut, with the proximal part of the guide removed, showing the K-wire clearly, which prevents the cut from being too deep. c: The tibia after the wedge has been opened and the Activmotion plate has been screwed in place. The yellow band is the patella tendon.

Export to PPT

3 Results

A total of 14 patients were randomly assigned to the CONV and PSCG groups (Fig. 2). Two were excluded early: one due to a failure in initiating the planning process via the online portal, and one due to planning being requested for the contralateral knee. This resulted in 12 patients included in the final analysis, see Table 1 for their characteristics.

Close

Fig. 2 CONSORT flow diagram for screening and selection of participants.

Export to PPT

The implementation of CT-guided HTO planning and surgery posed significant challenges (Fig. 3). It required extensive communication with an external party abroad via an online portal, which was time-consuming and prone to errors (e.g., planning for the wrong knee or unprocessed requests). Consequently, two patients could not undergo surgery with PSCGs and were excluded from the study. The data exchange also raised privacy concerns: patient information needed to be pseudonymized while remaining traceable to prevent medical errors. This was addressed by assigning each patient a study-specific number and manually removing identifying data from the imaging files. Finally, several hospital departments had to adapt protocols and workflows, including implementing a new CT protocol, coordinating pseudonymized data transfer with IT, and adjusting OR logistics to accommodate variable surgical times and instrument requirements.

Close

Fig. 3 All hospital steps involved from the initial decision to do a HTO until surgery. The dark blue blocks represent the departments involved with every (conventional) HTO surgery, the light blue blocks are extra departments that got involved when using CT-based planning and PSCGs. The white blocks present the extra tasks for each department when using CT-based planning and PSCGs. PSCG: Patient-Specific Cutting Guides; CSD: Central Sterilisation Department; HTO: High-Tibial Osteotomy; OR: Operating Room; IT: Information Technology department.

Export to PPT

Radiologically, the planned to achieved correction for the MPTA or PPTA was comparable between the groups. One patient in the CONV group was considered an outlier (defined as ΔMPTA/PPTAachieved-planned>3°) with a ΔMPTAachieved-planned of −3.9°. This patient experienced an intraoperative hinge fracture. For the PPTA we found 2 outliers in the CONV group and 4 in the PSCG group. Additionally, a possible learning curve effect was observed, with the ΔPPTAachieved-planned becoming progressively smaller as the cases advanced, although the third case did not follow this trend (Supplementary Material Table S2).

Interestingly, the median post-operative MPTA in the PSCG group was under-corrected and the upper limit of the IQR was just barely neutral (89.3° (87.4°, 90.4°)). See Supplementary Material Table S1 for a full overview of all radiological outcome measures.

The total surgery time for the PSCG group (73 (66–80) minutes) was significantly higher compared to the CONV group (51 (46–60) minutes), p = 0.044. Factors contributing to this longer surgery time in the PSCG group were: the specific ‘fit’ of the cutting guide on the proximal tibia, the surgeons' lower familiarity with using the PSCG, and a reduced level of confidence in the process. The latter two factors were evident in a noticeable learning curve for surgery time within the PSCG group (Supplementary Material Table S2). In none of the patients bone grafts were used.

The additional costs for the use of CT-guided planning and PSCGs entailed one extra CT-scan (€163), use of the ACTIVMOTION plate and screws instead of the regular Tomofix plate and screws (€1000 over €575) and the production and shipment of the PSCG (€425), plus an estimated €150 of additional sterilisation costs. This amounts to €1163 ($1269) additional costs.

In the PSCG group, one patient developed a deep infection with delayed wound healing at 6 weeks postoperative; this was successfully treated by a surgical debridement, antibiotics and implant retention (DAIR). Furthermore, one planning error occurred in the PSCG group where the pre-operative planning file suggested an opening wedge length in millimetres that did not match the (correctly planned) angle in degrees. The PSCG for position of the osteotomy was still used in this patient, but the opening height and correction angle were intraoperatively checked more extensively based on the conventional planning.

4 Discussion

This study was initiated because, while PSCGs were promoted to improve HTO accuracy, high-level evidence, such as a randomized controlled trial, was lacking. From a clinical perspective, such innovations would ideally be applied in complex cases requiring multi-planar corrections. Beforehand, initial evaluation of this innovation in simpler cases was deemed appropriate, which is why this study was conducted. Implementation posed logistical challenges, including time-consuming communication with an external partner, planning errors, and data protection issues. Two patients were excluded due to these complications. The process also required protocol adjustments and interdepartmental coordination across radiology, IT, and OR logistics.

Previous studies have also reported logistical challenges in implementing 3D cutting guides. One study on trochanteric osteotomies noted the need for iterative design refinement and close coordination with developers and surgeons to achieve a stable fit to complex anatomy and ensuring the guide's usability without disrupting standard surgical procedures.¹⁸ Materialise similarly reported difficulties in workflow integration, citing a steep learning curve for surgeons, the reliance on high-quality imaging, and the challenge of aligning production with surgical schedules.¹⁹ These issues, along with the initial costs and resource demands, hinder widespread adoption.

To improve clinical integration of this innovation, it is important to streamline the process. Embedding pseudonymization into the Newclip Technics software would eliminate the need for IT involvement, saving time and reducing errors. Additionally, supporting multiple plate brands would allow hospitals to maintain existing workflows. Lastly, improving the user-friendliness of the online request portal could further minimize errors and increase efficiency.

Radiological outcomes showed no overall improvement in correction accuracy with PSCGs compared to the conventional technique, with no significant difference in aimed-to-achieved correction angles in the coronal (MPTA) and sagittal planes (PPTA) or in number of outliers >3° of MPTA. One under-corrected outlier in the CONV group may have been due to an intraoperative hinge fracture, which is less likely with PSCGs due to the protective K-wire placement.^20,21 Both groups showed an unintended increase in PPTA, despite aiming to maintain the slope, consistent with previous findings in open-wedge HTO.⁵ Using PSCG could potentially solve this problem.¹⁰ However, the PSCG group also showed slope increases, likely due to limited access to the posterior cortex during early use, which made them unintentionally increase the tibial slope when trying to cut the tibia. This was addressed by removing the proximal PSCG portion before completing the cut, as advised by the manufacturer.

In terms of cost-efficiency, both the surgery time and costs were higher for the PSCG group, making it less cost-effective in simple cases. However, surgery time decreased with experience and previous research suggests a learning curve of about 10 cases,²² indicating PSCG-surgery may become more time-efficient.

This study has a few limitations. First, early termination led to a small sample size, limiting definitive conclusions. However, it offers valuable insight into PSCGs implementation in clinical practice and its challenges. Second, both surgeons had limited experience with the Newclip PSCGs and Activmotion plate, likely contributing to the observed learning curve, particularly in slope correction. Third, CT-guided planning relied heavily on surgeon input based on long-leg radiographs, with the CT used only to apply this correction, limiting the potential of truly patient-specific planning. Software-generated suggestions could enhance this, though CT scans are not weightbearing and the question still remains: what exactly is the optimal correction?²³ Lastly, patients needing complex multi-planar corrections were excluded, which is where PSCGs might offer greater benefit.

In summary: the use of PSCGs entailed a challenging implementation process, while it did not result in a more precise correction, and the costs were higher compared to conventional HTO surgery. Although the small sample size precludes definitive conclusions, the findings suggest limited added value of PSCGs in low-volume clinical settings for simple corrections. From this study, we have learned that at this early stage of evaluation (in simple cases), we encountered problems that prevent us from achieving our actual goal: complex cases. Thereby, the costs for production of PSCGs are higher for complex cases, making it more interesting from a business point of view to opt for higher volumes and lower complexity. This leads to an innovation paradox: a potentially useful innovation stumbling prematurely, hindering its routine use in clinical practice.

CRediT authorship contribution statement

Kelly Mills: Validation, Formal analysis, Data curation, Writing – original draft, Visualization. Petra J.C. Heesterbeek: Supervision, Writing – review & editing. José M.H. Smolders: Conceptualization, Methodology, Supervision, Writing – review & editing.

Patients consent

All patient provided written informed consent.

Ethical statement

This study was conducted in accordance with the Declaration of Helsinki. Ethical approval was obtained from the Medical Research Ethics Committee (CMO) Region Arnhem-Nijmegen, under protocol number NL72556.091.20.

Funding

The authors report no funding.