Corrective osteotomies are the indicated treatment for complex forearm malunions. The use of 3D computer-assisted preoperative planning offers significant advantages, such as improved comprehension of the multiplanar deformity and high accuracy. However, the clinical application of computer-assisted preoperative planning methods is limited due to the considerable time, effort, and expertise required. This emphasizes the necessity for a tool that can generate clinically feasible osteotomy plans, complying with patient-specific anatomical reconstruction goals. To address this challenge, this research developed an automatic planning tool for corrective osteotomies of radius malunions, requiring minimal user interaction. By automatically registering the pathological and contralateral bone models, the tool provides insight into the degree and nature of the deformity. An Evolutionary Algorithm is implemented to optimize patient-specific osteotomy plans by minimizing bone protrusion near the osteotomy plane. The automatic planning tool yields patient-specific osteotomy plans including the osteotomy plane location and orientation, as well as the required reduction of the distal part after the osteotomy cut. These plans ensure accurate alignment of both the proximal and distal radial joint surfaces. The developed tool was validated on 15 patient cases. The osteotomy plans generated by the automatic tool were compared to those planned manually in the past. Objective validation, based on residual alignment errors of the entire radius bone, often favored the manual planning approach. However, the automatic tool consistently provided osteotomy plans with more accurate alignment of the distal joint surface. Additionally, a blinded qualitative validation was conducted with a highly experienced orthopedic surgeon, who rated all osteotomy plans on a scale of 1 to 10. The results indicated that the automatic tool is not yet capable of generating osteotomy plans with feasibility scores equivalent to manual planning. However, the feasibility scores differed by only one point on most patient cases. The main areas that require improvement to consistently produce clinically feasible osteotomy plans include the incorporation of osteosynthesis plate fixation, consideration of the relationship with the ulna, and the option for double-cut osteotomies. In conclusion, the developed tool is capable of generating clinically feasible osteotomy plans, serving as a valuable starting point for patient-specific plans that can be further refined manually.

Introduction Fractures of the diaphysis of the radius and/or ulna are most common in children from 5 to14 years old and take up to 40% of all pediatric fractures. Patients with a diaphyseal fracture of the forearm can develop a malunion: healing of the bones in non-anatomical position. This can lead to pain, cosmetic differences and a limitation of pronation and/or supination. A malunion with an angulation of at least 15degrees leads to a limitation in rotation of the forearm in 60% of the cases. However, it is not known how amalunion leads to a rotational limitation and why certain patients have a predominant supination limitation and some have a predominant pronation limitation. It is hypothesized that the distance between the bonescan explain the limitation: a too small distance leads to bone impingement and blocking the rotation, a toolarge distance leads to contracture of the central band, a ligament between the radius and ulna.AimThe aim of this research is to explain the loss of rotational function in malunited forearms by using akinematic model in which bone impingement and contracture of the central band can be recognized.MethodFifteen (n=15) patients were included who developed a malunion after a one-sided, both-bonediaphyseal fracture of the forearm during childhood (age < 18) which led to a range of pronation and/orsupination lower than 50 degrees. Their range of motion was measured and CT-scans were made of bothforearms, from which three-dimensional bone surface models were retrieved. A kinematic model for prona-tion and supination of the forearm was developed in which the patient specific anatomy was used to detectbone impingement, measure central band length (CBL) and measure minimal interosseous distance (MID)between the radius and ulna. Bone impingement and CBL were used for prediction of the range of motion ofthe malunited forearms, MID was used to compare the distance between the bones at maximum supinationand pronation between the affected and unaffected forearms of the patients. Central band length relative tothe neutral position was calculated in unaffected forearms to define a threshold for contracture. Bone im-pingement was defined as overlapping of the bone surface models. The root mean squared error (RMSE)between in vivo measured range of pronation, supination and full range and the predicted values is calcu-lated.ResultsAll fifteen patients showed bone impingement as reason for limiting pronation, fourteen patientsshowed contracture of the central band as reason for limiting supination. By setting the threshold at 103%of the relative central band length, the pronation, supination and full range of thirteen patients could bepredicted with a RMSE between 15.5 and 17.9 degrees. Bone distance was significantly lower in malunited forearms than in unaffected forearms in maximum pronation. In supination this effect was much less clear.The kinematic model showed an error less than one millimeter and one degree for translation and rotation compared to cadaveric scans in different pronation and supination positions.Conclusion The kinematic model showed that bone impingement and central band contracture are the best explanations for limiting pronation and supination of malunited forearms. Prediction is difficult because the kinematic model uses the neutral position as starting point, which is not always clear because the kinematics of forearm rotation still has some unknowns and the central band origin and insertion is now located based on a cadaveric study