Abstract. Objectives. Three-dimensional visualisation of sonographic callus has the potential to improve the accuracy and accessibility of ultrasound evaluation of fracture healing. The aim of this study was to establish a reliable method for producing three-dimensional reconstruction of sonographic callus. Methods. A prospective cohort of ten patients with a closed tibial shaft fracture managed with intramedullary nailing were recruited and underwent ultrasound scanning at 2-, 6- and 12-weeks post-surgery. Ultrasound B-mode capture was performed using infrared tracking technology to map each image to a three-dimensional lattice. Using echo intensity, semi-automated mapping was performed by two independent reviewers to produce an anatomic three-dimensional representation of the fracture. Agreement on the presence of sonographic bridging callus on three-dimensional reconstructions was assessed using the kappa coefficient. Results. Nine of the ten patients achieved union at six months. At six weeks, seven patients had bridging callus at ≥1 cortex on the three-dimensional reconstruction; when present all united. Compared to radiographs, no bridging callus was present in any patient. Of the three patients lacking sonographic bridging callus, one went onto a nonunion (77.8%-sensitive and 100%-specific to predict union). At twelve weeks, nine patients had bridging callus at ≥1 cortex on three-dimensional reconstruction and all united (100%-sensitive and 100%-specific to predict union). Compared to radiographs, seven of the nine patients that united had bridging callus. Three-dimensional reconstruction of the anteromedial and anterolateral tibial surface was achieved in all patients, and detection of sonographic bridging callus on the three-dimensional reconstruction demonstrated substantial inter-observer agreement (kappa=0.78, 95% confidence interval 0.29–1.0, p=0.011). Conclusions. Three-dimensional fracture reconstruction can be created using multiple ultrasound images in order to evaluate the presence of bridging callus. This imaging modality has the potential to identify impaired healing at an early stage in fracture management. Declaration of Interest. (b) declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported:I declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research project

Background. There is no consensus on which glenoid plane should be used in total shoulder arthroplasty. Nevertheless, anatomical reconstruction of this plane is imperative for the success of a total shoulder arthroplasty. Methods. Three-dimensional reconstruction CT-scans were performed on 152 healthy shoulders. Four different glenoid planes, each determined by three surgical accessible bony reference points, are determined. The first two are triangular planes, defined by the most anterior and posterior point of the glenoid and respectively the most inferior point for the Saller's Inferior plane and the most superior point for the Saller's Superior plane. The third plane is formed by the best fitting circle of the superior tubercle and the most anterior and posterior point at the distal third of the glenoid (Circular Max). The fourth plane is formed by the best fitting circle of three points at the rim of the inferior quadrants of the glenoid (Circular Inferior). We hypothesized that the plane with normally distributed parameters, narrowest variability and best reproducibility would be the most suitable surgical glenoid plane. Results. No difference in position of the mean humeral center of rotation is found between the Circular Max and Circular Inferior plane (X=91.71degrees/X=91.66degrees p=0.907 and Y=90.83degrees /Y=91.7degrees p=0.054 respectively), while clear deviations are found for the Saller's Inferior and Saller's Superior plane (p < 0.001). The Circular Inferior plane has the lowest variability to the coronal scapular plane (p<0.001). Conclusion. This study provides arguments to use the Circular Inferior glenoid plane as preferred surgical plane of the glenoid. Level of evidence: Level II, Basic Science Study, Anatomical Survey

Three-dimensional (3D) printing has become more frequently used in surgical specialties in recent years. Orthopaedic surgery is particularly well-suited to 3D printing applications, and thus has seen a variety of uses for this technology. These uses include pre-operative planning, patient-specific instrumentation (PSI), and patient-specific implant production. As with any new technology, it is important to assess the clinical impact, if any, of three-dimensional printing. The purpose of this review was to answer the following questions: . What are the current clinical uses of 3D printing in orthopaedic surgery?. Does the use of 3D printing have an effect on peri-operative outcomes?. Four electronic databases (Embase, MEDLINE, PubMed, Web of Science) were searched for Articles discussing clinical applications of 3D printing in orthopaedics up to November 13, 2018. Titles, abstracts, and full texts were screened in duplicate and data was abstracted. Descriptive analysis was performed for all studies. A meta-analysis was performed among eligible studies to compare estimated blood loss (EBL), operative time, and fluoroscopy use between 3D printing cases and controls. Study quality was assessed using the Methodological Index for Non-Randomized Studies (MINORS) criteria for non-randomized studies and the Cochrane Risk of Bias Tool for randomized controlled trials (RCTs). This review was prospectively registered on PROSPERO (Registration ID: CRD42018099144). One-hundred and eight studies were included, published between 2012 and 2018. A total of 2328 patients were included in these studies, and 1558 patients were treated using 3D printing technology. The mean age of patients, where reported, was 47 years old (range 3 to 90). Three-dimensional printing was most commonly reported in trauma (N = 41) and oncology (N = 22). Pre-operative planning was the most common use of 3D printing (N = 63), followed by final implants (N = 32) and PSI (N = 22). Titanium was the most commonly used 3D printing material (16 studies, 27.1%). A wide range of costs were reported for 3D printing applications, ranging from “less than $10” to $20,000. The mean MINORS score for non-randomized studies was 8.3/16 for non-comparative studies (N = 78), and 17.7/24 for non-randomized comparative studies (N = 19). Among RCTs, the most commonly identified sources of bias were for performance and detection biases. Three-dimensional printing resulted in a statistically significant decrease in mean operative time (−15.6 mins, p < .00001), mean EBL (−35.9 mL, p<.00001), and mean fluoroscopy shots (−3.5 shots, p < .00001) in 3D printing patients compared to controls. The uses of 3D printing in orthopaedic surgery are growing rapidly, with its use being most common in trauma and oncology. Pre-operative planning is the most common use of 3D printing in orthopaedics. The use of 3D printing significantly reduces EBL, operative time, and fluoroscopy use compared to controls. Future research is needed to confirm and clarify the magnitude of these effects

Sonographic callus may enable assessment of fracture healing. The aim of this study was to establish a reliable method for three-dimensional reconstruction of sonographic callus. Patients that underwent non-operative management of displaced midshaft clavicle fractures and intramedullary nailing of tibia fractures were prospectively recruited and followed to union. Ultrasound scanning was performed at periodical time points following injury. Infra-red tracking technology was used to map each image to a three-dimensional lattice. Criteria was fist established for two-dimensional bridging callus detection in a pilot study. Using echo intensity of the ultrasound image, semi-automated mapping was used to create an anatomic three-dimensional representation of fracture healing. Agreement on the presence of sonographic bridging callus was assessed using the kappa coefficient and intra-class-correlation (ICC) between observers. 112 clavicle fractures and 10 tibia fractures completed follow-up at six months. Sonographic bridging callus was detected in 62.5% (n=70/112) of the clavicles at six weeks post-injury. If present, union occurred in 98.6% of the fractures (n=69/70). If absent, nonunion developed in 40.5% of cases (n=17/42)(73.4%-sensitive and 100%-specific to predict union). Out of 10 tibia fractures, 7 had bridging callus of at least one cortex at 6 weeks and when present all united. Of the three patients lacking sonographic bridging callus, one went onto a nonunion (77.8%-sensitive and 100%-specific to predict union). The ICC for sonographic callus between four reviewers was 0.82 (95% CI 0.68–0.91). Three-dimensional ultrasound reconstruction of bridging callus has the potential to identify impaired fracture healing at an early stage in fracture management

Total temporomandibular joint (TMJ) replacements reduce pain and improve quality of life in patients suffering from end-stage TMJ disorders, such as osteoarthritis and trauma. Jaw kinematics measurements following TMJ arthroplasty provide a basis for evaluating implant performance and jaw function. The aim of this study is to provide the first measurements of three-dimensional kinematics of the jaw in patients following unilateral and bilateral prosthetic TMJ surgeries. Jaw motion tracking experiments were performed on 7 healthy control participants, 3 unilateral and 1 bilateral TMJ replacement patients. Custom-made mouthpieces were manufactured for each participant's mandibular and maxillary teeth, with each supporting three retroreflective markers anterior to the participant's lip line. Participants performed 15 trials each of maximum jaw opening, lateral and protrusive movements. Marker trajectories were simultaneously measured using an optoelectronic tracking system. Laser scans taken of each dental plate, together with CT scans of each patient, were used to register the plate position to each participant's jaw geometry, allowing 3D condylar motion to be quantified from the marker trajectories. The maximum mouth opening capacity of joint replacement patients was comparable to healthy controls with average incisal inferior translations of 37.5mm, 38.4mm and 33.6mm for the controls, unilateral and bilateral joint replacement patients respectively. During mouth opening the maximum anterior translation of prosthetic condyles was 2.4mm, compared to 10.6mm for controls. Prosthetic condyles had limited anterior motion compared to natural condyles, in unilateral patients this resulted in asymmetric opening and protrusive movements and the capacity to laterally move their jaw towards their pathological side only. For the bilateral patient, protrusive and lateral jaw movement capacity was minimal. Total TMJ replacement surgery facilitates normal mouth opening capacity and lateral and inferior condylar movements but limits anterior condylar motion. This study provides future direction for TMJ implant design

Background. The femoral head center shift on reduction time in total hip arthroplasty (THA) causes alteration of the muscle tension around the hip joint. Many studies about the shift of the femoral head in the cranio-caudal direction or medio-lateral direction on coronal plane have been reported. It has been known widely that the shift on these directions influence tension of the abductor muscle around the hip joint. Nevertheless few studies about the three-dimensional shift including the antero-posterior direction have been reported. Purpose. The purpose of this study is to evaluate the three-dimensional shift of the femoral head center in THA using three-dimensional THA templating software. Subjects & Methods. The subjects of this study were 156 primary THA cases of 143 patients. Using CT-based three-dimensional THA templating software ZedHip® (LEXI, Tokyo Japan), simulation of optimal implantation was performed on each THA case. On case which has over anteverted or less anteverted femoral neck, a stem which has modular neck system was selected to adjust anteversion of the femoral neck. The three-dimensional shift of the femoral head center on reduction time was calculated with ZedHip®. The three-dimensional shift was resolve into cranio-caudal, medio-lateral and antero-posterior direction (Fig. 1). Furthermore the correlation between the amount of the shift and hip joint deformity was investigated. Results. The average amount of the shift on cranio-caudal direction was 9.9mm to caudal side, on medio-lateral direction was 3.1mm to medial side and on antero-posterior direction was 2.6mm to posterior side. The average total amount of three-dimensional shift was 12.9mm (Fig. 2). On Crowe type 1 hips in 88 cases, the average shift to posterior side was 3.2mm, on Crowe type 2 in 20 cases was 3.7mm and on Crowe type 3 in 13 cases was 4.0mm. Among them there was no significant difference (Fig. 3). Conclusion. At THA surgery, the femoral head center shifted three-dimensionally and the maximum amount of shift on antero-posterior direction was 16.6mm to posterior side. There was no correlation between these amounts of the shift on antero-posterior direction and anatomical deformity of the hip joint. It is important to understand the shift of the femoral head center for predicting the alteration of muscle tension around the hip joint. The shift on antero-posterior direction influences the tension of iliopsoas muscle and there is a possibility that the shift to posterior side causes anterior iliopsoas impingement after THA surgery

Aims

This systematic review aims to identify 3D predictors derived from biplanar reconstruction, and to describe current methods for improving curve prediction in patients with mild adolescent idiopathic scoliosis.

Abstract. OBJECTIVES. An unresolved challenge in osteoarthritis research is characterising the localised intra-tissue mechanical response of articular cartilage. The aim of this study was to explore whether laboratory micro-computed tomography (micro-CT) and digital volume correlation (DVC) permit non-destructive visualisation of three-dimensional (3D) strain fields in human articular cartilage. METHODS. Human articular cartilage specimens were harvested from the knee (n=4 specimens from 2 doners), mounted into a loading device and imaged in the loaded and unloaded state using a micro-CT scanner. Strain was calculated throughout the volume of the cartilage using the CT image data. RESULTS. Strain was calculated in the 3D volume with a spatial resolution of 75 µm, and the volumetric DVC calculated strain was within 5% of the known applied stain. Variation in strain distribution between the superficial, middle and deep zones was observed, consistent with the different architecture of the material in these locations. CONCLUSIONS. The DVC method is suitable for calculating strain in human articular cartilage. This method will be useful to generate chondral repair scaffolds that that seek to replicate the strain gradient in cartilage. Declaration of Interest. (b) declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported:I declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research project

Objective. In total knee arthroplasty, three-dimensional “criss-cross” line locate femoral osteotomy and conventional osteotomy were used. By comparing the two methods osteotomy in patients before and after surgery and imaging-related information data, to evaluate the recent post operative efficacy, at the same time to find out if there has clinical evidence that three-dimensional “criss-cross” line locate femoral osteotomy can be used in total knee arthroplasty. Methods. From July 2012 to July 2014, 64 patients who undertook the artificial total knee arthroplasty were divided into 2 groups: conventional osteotomy group(group A)and three-dimensional “criss-cross” line locate femoral osteotomy group(group B). In the X-ray of the two groups, it was measured that the hip-knee-ankle angle and the joint gap symmetry of 90°flexion degree. It was also measured that the two group joints range of motion. Those data were statistically analyzed. The KSS score of the two groups were compared. Results. In Group B the excellent and good rate was 93.8%, and Group A was 81.3%. The postoperative results of Group A were relatively better than Group B in limb alignment and joint mobility. There were significant differences between the prosthesis placement of the two group patients. Conclusion. The Short-term results of the three-dimensional “criss-cross” line locate femoral osteotomy group was better than the conventional osteotomy group. The reference osteotomy method of three-dimensional “criss-cross” line is very helpful to have a good result in TKA

We have measured the three-dimensional patterns of carpal deformity in 20 wrists in 20 rheumatoid patients in which the carpal bones were shifted ulnarwards on plain radiography. Three-dimensional bone models of the carpus and radius were created by computerised tomography with the wrist in the neutral position. The location of the centroids and rotational angle of each carpal bone relative to the radius were calculated and compared with those of ten normal wrists. In the radiocarpal joint, the proximal row was flexed and the centroids of all carpal bones translocated in an ulnar, proximal and volar direction with loss of congruity. In the midcarpal joint, the distal row was extended and congruity generally well preserved. These findings may facilitate more positive use of radiocarpal fusion alone for the deformed rheumatoid wrist

We used three-dimensional movement analysis by computer modelling of knee flexion from 0° to 50° in 14 knees in 12 patients with recurrent patellar dislocation and in 15 knees in ten normal control subjects to compare the in vivo three-dimensional movement of the patella. Flexion, tilt and spin of the patella were described in terms of rotation angles from 0°. The location of the patella and the tibial tubercle were evaluated using parameters expressed as percentage patellar shift and percentage tubercle shift. Patellar inclination to the femur was also measured and patellofemoral contact was qualitatively and quantitatively analysed. The patients had greater values of spin from 20° to 50°, while there were no statistically significant differences in flexion and tilt. The patients also had greater percentage patellar shift from 0° to 50°, percentage tubercle shift at 0° and 10° and patellar inclination from 0° to 50° with a smaller oval-shaped contact area from 20° to 50° moving downwards on the lateral facet. Patellar movement analysis using a three-dimensional computer model is useful to clearly demonstrate differences between patients with recurrent dislocation of the patella and normal control subjects

Using roentgen stereophotogrammetry we have recorded the three-dimensional movements of the knee during an anteroposterior laxity test in 36 patients with torn anterior cruciate ligaments and in three cadaver knees. At 30 degrees of knee flexion and before loads were applied the tibia occupied a more laterally rotated position if the anterior cruciate ligament had been injured. When the tibia was pulled anteriorly knees with cruciate deficiency rotated more laterally and were more abducted than normal knees. Posterior traction induced lateral rotation in the injured knee and medial rotation in the intact one. Precise knowledge of the three-dimensional instability of the anterior cruciate deficient knee may be important when the laxity is evaluated only in relation to one of the three cardinal axes

We report the case of a 12-year-old boy with flexion loss in the left elbow caused by deficient of the concavity corresponding to the coronoid fossa in the distal humerus. The range of motion (ROM) was 15°/100°, and pain was induced by passive terminal flexion. Plain radiographs revealed complete epiphyseal closure, and computed tomography (CT) revealed a flat anterior surface of the distal humerus; the coronoid fossa was absent. Then, the bony morphometric contour was surgically recreated using a navigation system and a three-dimensional elbow joint model. A three-dimensional model of the elbow joint was made preoperatively and the model comprising the distal humerus was milled so that elbow flexion flexion of more than 140° could be achieved against the proximal ulna and radius. Navigation-assisted surgery (contouring arthroplasty) was performed using CT data from this milled three-dimensional model. Subsequently, an intraoperative passive elbow flexion of 135° was obtained. However, active elbow flexion was still inadequate one year after operation, and a triceps lengthening procedure was performed. At the final follow-up one year after triceps lengthening, a considerable improvement in flexion was observed with a ROM of −12°/125°. Plain radiographs revealed no signs of degenerative change, and CT revealed the formation of the radial and coronoid fossae on the anterior surface of the distal humerus. Navigation-assisted surgery for deformity of the distal humerus based on a contoured three-dimensional model is extremely effective as it facilitates evaluation of the bony morphometry of the distal humerus. It is particularly useful as an indicator for milling the actual bone when a model of the mirror image of the unaffected side cannot be applied to the affected side as observed in our case

Purpose: Scoliosis is a three-dimensional deformation of the spinal column. Modern surgical techniques have attempted to address this 3D component of the problem but pre- and postoperative measurements lack precision. A solution is stereoradiographic 3D reconstruction providing 1.1 mm precision for vertebral shape and 1.4° precision for axial rotation. Material and methods: Ten patients (seven adolescents and three adults) with idiopathic scoliosis (mean 56°, range 36°–78°) were treated with an in situ arching method. A calibrated teleradiogram (AP and lateral view) was obtained before and after surgery. The spinal columns were reconstructed by stereoradiography. Six rotation angles were measured in the three planes for each vertebra and each intervertebral space, taking into account the curvatures and their apical and junctional zones. Results: Preoperatively, for thoracic scoliosis, measurements were: mean vertebral axial rotation (VAR) measured at the apex = 20°; mean lateral axial rotation (LAR) of the junctional zones = 30°; mean intervertebral rotation (IVR) = 10°. Depending on the curvatures, in situ arching yielded a 52–60% correction of the VAR at the apex, and 78–79% correction of the junctional zones. VLR of the junctional zoenes was improved 58–74%. Intervertebral sagittal rotation (ISR) at the summit (kyphosis) was improved 5.5° on the average. Discussion: Unlike computed tomoraphy where scans are obtained in the supine position, three-dimensional reconstruction of the spinal column enables a precise analysis of the loaded spine. Improvement was significant in the frontal plane with 18.3° and 21.4° improvement of the VLR for the thoracic and thoracolumbar junctional zones respectively, compared with the rod rotation where the peroperative stereophotogram showed a 9.6° and 8.6° gain respectively. There was a real improvement in VAR, differing from the literature where the rotation of the rod appears to be less pronounced. Conclusion: Three-dimensional reconstruction of the spinal column enables a segmentary analysis of scoliosis deformations. In addition, by enabling a view of the spinal column in all directions, angle measurements can be made with precision allowing repeated measurements and comparisons. This technique demonstrated the efficacy of in situ arching in improving vertebral rotation

This paper describes how advances in three-dimensional printing may benefit the military trauma patient, both deployed on operations and in the firm base. Use of rapid prototype manufacturing to produce a 3D representation of complex fractures that can be held and rotated will aid surgical planning within multidisciplinary teams. Patient-clinician interaction can also be aided using these graspable models. The education of military surgeons could improve with the subsequent accurate, inexpensive models for anatomy and surgical technique instruction. The developing sphere of additive manufacturing (3D printing functional end-use components) lends itself to further advantages for the military orthopaedic surgeon. Military trauma patients could benefit from advances in direct metal laser sintering which enable the manufacture of complex surfaces and porous structures on bio-metallic implants not possible using conventional manufacturing. “Bio-printing” of tissues mimicking anatomical structures has potential for military trauma patients with bone defects. Deployed surgeons operating on less familiar fracture sites could benefit from three-dimensionally printing patient-specific medical devices. These can make operating technically easier, reducing radiation exposure and operating time. Further ahead, it may be possible to contemporaneously 3D print medical devices unavailable from the logistics chain whilst operating in the deployed environment

Summary Statement. We used three-dimensional software to assess different anatomic variables in the femur. The canal of Femur twisted slightly below the lesser trochanter in cases with a larger angle of anteversion. Introduction. Accurate positioning of the joint prosthesis is essential for successful total hip arthroplasty (THA). To aid in tailoring of the prosthesis, we used three-dimensional software to assess different anatomic variables in the femur. Patients & Methods. We used CT imaging data of the unaffected normal side of the 25 patients (22 females, age range 30 to 81 years) who underwent THA in 2012 in our hospital. The femur was reconstructed from CT data and measured using three-dimensional modeling software (Mimics 16.0 Materialise, Leuven, Belgium). We measured ellipse fitting to the medullary canal in the axial plane of the femur at 20-mm intervals. The angle between the major axis of those ellipses and the axis of the femoral neck was measured and expressed as the canal rotation. The distance between the lesser trochanter and the center of the femoral head was measured along the Z axis. Results. The major axes of the ellipses direct to medial, front and medial side in the level of epiphysis, above isthmus and distal portion respectively in all cases. The maximum rotated level was above isthmus. The rotation angle in the proximal portion ranged from 36 to 84 degrees (mean, 60.6 degrees, SD ± 12.1). The rotation angle of the distal portion ranged from 71 to 95 degrees (mean, 86.1 degrees, SD ± 6.1). Discussion/Conclusion. The torsion of the canal varied more widely between individuals in the proximal portion than did the distal portion. In addition, the torsion of the proximal aspect, although more variable, was on average smaller when the angle of anteversion was large. Because the canal twisted slightly below the lesser trochanter in cases with a larger angle of anteversion, it is suggested that attention to the degree of anteversion of a flat prosthesis stem is warranted

We have studied the three-dimensional geometry of the proximal humerus on human cadaver specimens using a digitised measuring device linked to a computer. Our findings demonstrated the variable shape of the proximal humerus as well as its variable dimensions. The articular surface, which is part of a sphere varies individually in its orientation as regards inclination and retroversion, and it has variable medial and posterior offsets. These variations cannot be accommodated by the designs of most contemporary humeral components. Although good clinical results can be achieved with current modular and non-modular components their relatively fixed geometry prevents truly anatomical restoration in many cases. To try to restore the original three-dimensional geometry of the proximal humerus, we have developed a new type of humeral component which is modular and adaptable to the individual anatomy. Such adaptability allows correct positioning of the prosthetic head in relation to an individual anatomical neck, after removal of the marginal osteophytes. The design of this third-generation prosthesis respects the four geometrical variations which have been demonstrated in the present study. These are inclination, retroversion, medial offset and posterior offset

We evaluated the impact of stereo-visualisation of three-dimensional volume-rendering CT datasets on the inter- and intraobserver reliability assessed by kappa values on the AO/OTA and Neer classifications in the assessment of proximal humeral fractures. Four independent observers classified 40 fractures according to the AO/OTA and Neer classifications using plain radiographs, two-dimensional CT scans and with stereo-visualised three-dimensional volume-rendering reconstructions. Both classification systems showed moderate interobserver reliability with plain radiographs and two-dimensional CT scans. Three-dimensional volume-rendered CT scans improved the interobserver reliability of both systems to good. Intraobserver reliability was moderate for both classifications when assessed by plain radiographs. Stereo visualisation of three-dimensional volume rendering improved intraobserver reliability to good for the AO/OTA method and to excellent for the Neer classification. These data support our opinion that stereo visualisation of three-dimensional volume-rendering datasets is of value when analysing and classifying complex fractures of the proximal humerus

We created virtual three-dimensional reconstruction models from computed tomography scans obtained from patients with acetabular fractures. Virtual cylindrical implants were placed intraosseously in the anterior column, the posterior column and across the dome of the acetabulum. The maximum diameter which was entirely contained within the bone was determined for each position of the screw. In the same model, the cross-sectional diameters of the columns were measured and compared to the maximum diameter of the corresponding virtual implant. We found that the mean maximum diameter of virtual implant accommodated by the anterior columns was 6.4 mm and that the smallest diameter of the columns was larger than the maximum diameter of the equivalent virtual implant. This study suggests that the size of the screw used for percutaneous fixation of acetabular fractures should not be based solely on the measurement of cross-sectional diameter and that virtual three-dimensional reconstructions might be useful in pre-operative planning

We investigated the three-dimensional morphological differences of the articular surface of the femoral trochlea in patients with recurrent dislocation of the patella and a normal control group using three-dimensional computer models. There were 12 patients (12 knees) and ten control subjects (ten knees). Three-dimensional computer models of the femur, including the articular cartilage, were created. Evaluation was performed on the shape of the articular surface, focused on its convexity, and the proximal and mediolateral distribution of the articular cartilage of the femoral trochlea. The extent of any convexity, and the proximal distribution of the articular cartilage, expressed as the height, were shown by the angles about the transepicondylar axis. The mediolateral distribution of the articular cartilage was assessed by the location of the medial and lateral borders of the articular cartilage. The mean extent of convexity was 24.9° . sd. 6.7° for patients and 11.9° . sd. 3.6° for the control group (p < 0.001). The mean height of the articular cartilage was 91.3° . sd. 8.3° for the patients and 83.3° . sd. 7.7° for the control group (p = 0.03), suggesting a wider convex trochlea in the patients with recurrent dislocation of the patella caused by the proximally-extended convex area. The lateral border of the articular cartilage of the trochlea in the patients was more laterally located than in the control group. Our findings therefore quantitatively demonstrated differences in the shape and distribution of the articular cartilage on the femoral trochlea between patients with dislocation of the patella and normal subjects