Several preoperative planning tools in computer-assisted surgery in acetabular fractures have been proposed. Moreover, all these preoperative planning tools are based on geometrical repositioning with their own limitations. The aim of this study was to evaluate the value of our prototype virtual planning tool using a rigid biomechanical model to predict failure in fracture reduction.

Between November of 2015 and June of 2016, 10 patients were operated by the main author for acetabular fracture in our institution. To validate our biomechanical model planning tool, biomechanical simulation was performed for each patient immediately after the surgery. Reduction quality was assessed on post-operative CT scans. A 3D model of the acetabular fracture was build out of the CT images using the non-commercial software Itksnap. Then a biomechanical model implemented within the non-commercial Artisynth framework was used to perform virtual reduction. Surgical approach and surgical strategy according to the operative report were simulated. The simulated reductions and the surgical reductions were compared.

The same reductions were obtained during surgery and biomechanical simulation in the 10 cases. For 7 cases, reduction was achieved by anterior surgical approach and so was the simulation. For 3 cases, reduction was achieved by posterior surgical approach and so was the simulation. The biomechanical simulation found similar results using the same surgical strategy with 9 anatomical reductions (90%) and one imperfect reduction (10%). The mean duration to perform acetabular planning surgery was 24 +/− 9 min [16–38].

Our virtual planning tool using a rigid biomechanical model can predict success or failure in fracture reduction according to the surgical approach and the surgical strategy.

Introduction

We report a single-centre, prospective, randomised study for pedicle screw insertion, by using a Computer Assisted Surgery (CAS) technique with three dimension (3D) intra-operative images intensifier versus conventional surgical procedure.

In recent years internal fixation of the spine by using posterior approach with minimally invasive and percutaneous technique were increasingly used in trauma. The percutaneous surgery lose information and navigation is supposed to provide better data because the lost information is found again. We hypothesise that a percutaneous minimal invasive dorsal procedure by using 3D intra-operative imaging for vertebral fractures allows short operating times with correct screw positioning and does not increase radiation exposure.

59 patients were included in this prospective, monocentric and randomised study. 29 patients (108 implants) were operated on by using conventional surgical procedure (CP) and 30 patients (72 implants) were operated on by using a 3D fluoroscopy-based navigation system (3D fluo). In the two groups, a percutaneous approach was performed for transpedicular vertebroplasty or percutaneous pedicle screws insertion. In the two groups surgery was done from T4 level to L5 levels. Patients (54 years old on average) suffered trauma fractures, fragility fractures or degenerative instabilities. Evaluation of screw placement was done by using post-operative CT with two independent radiologists that used Youkilis criteria. Operative and radiation running time were also evaluated.

With percutaneous surgery, the 3D fluo technique was less accurate with 13.88% of misplaced pedicle screws (10/72) compared with 11.11% (12/108) observed with CP. The radiation running time for each vertebra level (two screws) reached on average 0.56 mSv with 3D fluo group compared to 1.57 mSv with the CP group. The time required for instrumentation (one vertebra, two screws) with 3D fluo was 19.75 minutes compared with CP group 9.19 minutes. The results were statistically significant in terms of radiation dose and operative running time (p < 0.05), but not in terms of accuracy (p= 0.24).

With percutaneous procedures, 3D fluoroscopy-based navigation (3D fluo) system has no superiority in terms of operative running time and to a lesser degree in terms of accuracy, as compared to 2D conventional procedure (CP), but the benefit in terms of radiation dose is important. Other advantages of the 3D fluo system are twofold: up-to-date image data of patient anatomy and immediate availability to assess the anatomical position of the implanted screws.

Purpose of the study: We present the clinical and radiologic results of a prospective series of 20 patients who had percutaneous osteosynthesis of the thoracolumbar spine using a longitudinal system with monoaxial screws.

Material and methods: Twenty patients, 12 women, were treated from February 2008 to February 2009. Mean age was 44 years (15–61 years). Fractures were from T4 to L5, five patients had two levels fractured. There were 18 Magerl A and two type B. None of the patients had neurological problems. The reduction was achieved with a postural method and instrumentation with monoaxial screws using the percutaneous longitudinal method. Two levels were instrumented for 12 patients, three levels for three and five levels for four. Vertebroplasty was associated in nine patients. Pain (VAS), vertebral kyphosis (VK), regional angle (RA) were assessed pre-, postoperatively and at last follow-up. Operative time, blood loss, and duration of hospital stay were noted. Pedicle implantations were controlled on the scanner.

Results: Mean follow-up was ten months (3–19 months). In the preoperative phase, the VAS was 7/10, VK 17 (8–26), RA 16.6 (4–30). Postoperatively VAS was 4/10), VK 8 (4–15), RA 10.4 (−3 to 27). AT last follow-up VK was 8.5 and RA 12. Mean operative time was 90 minutes (40–180). On the postoperative scan, 23 of the 106 screws implanted had an extraosseous position (21%) including seven cranial screws. None of the patients developed a neurological deficit postoperatively. The first rise from bed was on day 1 in 14 patients. Mean duration of hospital stay was eight days (6–35).

Discussion: This reduction method provides a gain in VK (−9), a value between orthopaedic treatment (−8 gain) and surgical treatment as reported at the Sofcot in 1995 (−11.1 gain). The reduction was maintained well without loss at last follow-up.

Conclusion: This technique allows assembly of more or less long rigid configurations depending on the injury levels, with satisfactory reduction which is sustained over time. Blood loss is zero with little pain in the postoperative period.

Metal-on-metal hip arthroplasty is nowadays a well spread technique for hip replacement. It is a technically demanding procedure with sine qua non steps. Most of the large Australian, British and north-American clinical outcomes found about 20% early failure, within 3 months, during the steep part of the learning curve. In a biomechanical study on cadaver we showed that valgus and version placements should be appropriate in order to get the most effective strength in compression. But valgus placement is drastically limited by lateral neck notching which leads to early fracture. Anteversion also should be manage to address the cam-type shape of the ventro-lateral femoral neck, since most of the young patient with hip arthritis present retroverted misalignment of the head along the neck axis. The size of the femoral head is another issue because big head component needs a big acetabular cup. Sparing large amount of bone on the femoral side might lead to bone lost on the pelvic side. In order to enhance the placement of the femoral component – smallest one, in valgus, without retroversion and without neck notching – manufacturers propose mechanical device based on neck intraoperative palpation. Digitalized versions of this principle are also available. These systems still demand experienced surgeon to make slight adjustment on entry point and trajectory of the guide-wire. Imageless computer navigation based on proximal femur palpation and atlas is interesting on almost normal bone but could be inaccurate on altered bone, especially cam-type shape. Navigation based on CT scan gives exact 3D information and accurate planning but is still time consuming. Navigation on 2D fluoroscopic view shows good clinical results with only about 20 minutes more than a standard procedure.

We proposed a bone morphing^® procedure with emphasis of surface palpation on head-neck junction to get accurately the personal shape of each femur. Preliminary results on pathological bones showed safe reaming of the head without notching.

Following orthopaedic reconstruction and cranial neurosurgery, spine surgery is now entering its low invasive period. When, in 90’s, computer went routinely available in the surgical field, the main goal was to help surgeons operate on with more accuracy some difficult but standard procedures. The surgery was “computer aided”. The displayed tools on 2D or 3D images allowed surgeons to avoid permanent intra operative landmarks. Once patient personal anatomy was capture into the machine and the tools calibrated, the surgeon was able to plan and optimised ideal trajectories without direct vision to check tools position. “Navigation” starts to be more obvious to describe this intra operative control. Anyway, we still needed large exposure to get the full bone surface in order to build a 3D surface based model. This model optically localised was matched using rigid or elastic algorithm with preoperative CT scan model or bone morphing^®. Ultrasound recognition of the soft tissue/bone interface let think about trans cutaneous palpation. However, automatic segmentation of the bone surface never lead to commercially available soft. Only X-ray is commonly use during surgery to help surgeon to see tools and bone without surgical exposure. Fluoroscopy allows percutaneous trajectory as iliosacral screwing, vertebroplasty, fracture nailing et caetera. Radiation exposition could therefore be an issue for patient but also for surgeon. Fluoronavigation is a good response to percutaneous surgery. In spine no transversal view could be available. Surgeons should make mental reconstruction of the volume to perform the right trajectory. Industrial proposed intra operative tomography on C-arm with 3D reconstruction. It works well for limbs, but in thoracic and lumbar spine the large amount of surrounding soft tissues leads to low quality images. Flat panel X-ray receptor are a path to get more accurate images. Other perspectives are circular intra CT scan. The cost and the volume of machines stops the spread of such device.

Robots are used by knee surgeons but abandoned by hip surgeons. In spine tool holder robot are available in order to place a pedicular drill guide. Matching with bone is based on fluoroscopy.

Spine navigation could be useful to e-leaning and simulators too. The training of percutaneous surgery is long, because of mental matching between fluoroscopic 2D projections and the vertebra volume. We need a simulator allowing 3D virtual trajectory checked on AP and lateral view to short the learning curve.

Purpose of the study: Appropriate treatment of displaced supracondylar fractures of the distal humerus in children remains a controversial topic. Blount reduction followed by percutaneous or open pin fixation have been widely used. The purpose of this study was to analyze outcome after open surgical treatment of these fractures in pediatric trauma victims.

Material and methods: The study included all pediatric patients who underwent surgical treatment for displaced supracondylar fractures of the distal humerus over a ten year period. Fractures were classified III or IV according to Lagrange and Rigault. Cross pinning was used in all cases, via a posterior approach or a double lateral and medial approach. The mechanism of the fracture and pre- and postoperative vascular and neurological complications were noted. The long-term assessment included standard x-rays of the elbow joint (ap and lateral views) and a physical examination to search for misalignment and residual neurological disorders.

Results: We identified 110 patients, 61 boys and 49 girls, mean age 7.61 years (range 2–15 years). There were 96 grade IV fractures and 24 grade III. Mechanisms were: sports accident (n=44), fall from height (n=36), fall from own height (n=30). A neurological complication was observed in 29 children, skin opening in three and regressive vascular damage in six. A posterior approach was used for 95 patients and a double approach for 15. There was one revision for secondary displacement. Five patients developed transient paresthesia of the ulnar nerve which resolved without sequela. Three patients presented a moderately hypertrophic or deformed callus which had little functional impact. One patient with an open fracture required surgerical arthrolysis for stiffness six months after fracture.

Discussion and conclusion: Open surgery is a very reliable treatment for supracondylar elbow fractures with a low rate of short- and long-term complications. Ulnar nerve palsy, the classical complication of percutaneous cross pinning, can be attributed to the medial pin (7–16% of cases in the literature). The Blount method and Judet or Métaizeau fixations can sometimes be complicated by secondary displacement or a deformed callus, complications which were almost never observed in our series. The results obtained in this series favor our approach for open surgery for the treatment of displaced supracondylar fractures of the distal humerus in children.

Purpose of the study: Dislocation of total hip arthroplasty remains a frequent complication, occurring for 0.5% to 10% of implants depending on the series. In about 30% of the cases, the orientation of the acetabular cup is involved. It is sometimes difficult to visualize the acetabular landmarks during surgical procedures performed for revision or with a minimally invasive technique. The surgeon uses the position of the pelvis on the operative table as a guide. It can be noted however that the patient’s weight bearing on the table is not constant during the entire operation, potentially changing the position of the pelvis during the procedure. We evaluated the use of a visual referential visible within the operative field for implanting the prosthetic cup.

Material and methods: The method materialized the anterior plane of the pelvis then transferred geometrically this plane for display on the ipsilateral iliac crest. The pelvis was masked under a drape, in lateral decubitus. The cup was implanted 12 times using the plane of the floor as the reference, 8 times using the acetabular rim as the reference, and 10 times using the iliac reference. The goal was 20° anteversion in the sagittal plane and 45° inclination in the frontal plane. The position of the pelvis was randomized. The final positions of the cup, of the iliac reference, of the anterior plane of the pelvis and of the floor were recorded with an optical system. Spearman’s test was used to search for correlations.

Results: Using the floor referential, mean anteversion was 21.8° (15–30.9°) and mean inclination 43.2° (37–47.6°). Using the acetabular referential, mean anteversion was 21.7° (18.1–26.6°) and mean inclination 45.8° (40.9–48.6°). With the iliac referential, mean anteversion was 20.3° (17.3–25.5°) and mean inclination 43.3° (41.1–44.8°). Mean error between the pelvic plane and the iliac referential was, for anteversion −0,32° (−1.07 to 0.8°) and for inclination, −0.1° (−0.95 to 1.43°). Implantation with the iliac referential was not correlated to the position of the pelvis. When the plane of the floor was used, the position of the implant was correlated with pelvic anteversion (p< 0.01) and inclination (p< 0.01).

Discussion: Insertion of the cup was independent of the position of the pelvis within a 3D referential in the operative field. In addition to computer-assisted navigation, simple tools can be developed to improve the surgeon’s perception in difficult indications, especially when they can provide satisfactory accuracy. A clinical feasibility study is currently under way.

Purpose of the study: Hip resurfacing with a metal-on-metal bearing gives good mid-term clinical results. The design of the femoral piece has an effect on implant longevity, as does the vitality of the underlying cephalic bone. Computer-assisted surgery has been helpful in position the implant but the choice of the best position is still empirical. Prosthesis designers recommend valgus, but with too much there is a risk of a superolaterl notch which would weaken the neck. This leads the surgeon to use a larger femoral implant, and consequently to resect more acetabular bone. Anteversion is not evaluated. The purpose of this study was test mechanically different valgus/varus, anteversion/retroversion positions of the femoral implant.

Material and methods: We implanted 15 femurs made of resin which were geometrically and mechanically identical. The following angles were tested: varus/valgus (−10°, neutral, +10°, +20°) and ante/retroversion (−10°, neutram, +10). A valgus notch (+20° and +10°) and a varus notch (−10°) were simulated. The femurs thus prepared were tested with Instron 8874. Load at failure was noted as well as the type of fracture: distance from the fracture line to the greater trochanter (FGT). Student’s t test was applied.

Results: All of the femurs fractured at the neck. The fracture was closer to the implant (FGT: 11.0 mm) for the 20° valgus implantation (p< 0.05). The displacement was lesser with a valgus notch (mean 2.2 mm) then without a notch (mean 3.3 mm (p< 0.05). The varus notch had no effect. The failure load was lower for 20° valgus (1236 N, range 1117–1356N) then for the other angles (1664N, range 1142–2113 N) with near statistical significance (p=0.08). Retroversion had no effect. Anteversion allowed greater displacement (4.1 mm) and supported greater loading (1879 N) before failure.

Discussion: This study, unlike clinical studies, did not demonstrate any static mechanical superiority of the valgus position for the femoral piece. Another study on cadaver bones is planned for confirmation. Clinical studies reflect the vitality of the trabecular bone supporting the implant, a vitality which could be stimulated by the valgus position.

Purpose: Hip resurfacing techniques utilize mechanical jigs to align a femoral guide-pin that directs implant placement. Errors in alignment may lead to premature failure. The purpose of this study was to compare femoral guide-pin placement using a computer-assisted surgical (CAS) navigation system to a currently available manual alignment device.

Methods: A computer-assisted navigation system was developed by our group. Target guide-pin position was determined for each cadaveric hip from radiographs. A guide-pin was driven into each hip using either the CAS method or a manual alignment jig (Durom; Zimmer) by a novice or expert surgeon respectively. Radiographic pin position was compared to the target and accuracy was compared between the two techniques.

Results: Guide-pin insertion trajectory using the CAS system was significantly less variable in varus/valgus alignment than the mechanical jig (2.0° SD vs. 5.5° SD; p < 0.05). Ante/retroversion variability was also lower using CAS (4.4° SD vs. 7.7° SD) as was alignment error (CAS: 2.0° ± 2.2° SD valgus vs. Durom: 3.3° ± 5.5° SD varus, p=0.17; CAS: 0.1° ± 4.6° SD anteverted; Durom 3.2° ± 7.7° SD retroverted, p=0.48) but not significantly. Both methods proved accurate in placing the pin within 2 mm from the centre of the neck axis. Procedure time was similar between the two methods

Conclusions: Computer-assisted surgical navigation significantly improves reliability of guide-pin placement. This technique may help achieve better femoral implant alignment regardless of experience and contribute to improving hip resurfacing outcomes.

Funding : Other Education Grant

Funding Parties : Zimmer/UBC Research Fund

Purpose: To examine whether neutral or valgus placement results in greater fracture strength ex vivo, when the femoral neck is notched superolaterally as sometimes occurs during hip resurfacing arthroplasty.

Methods: We loaded 10 paired fresh-frozen notched proximal cadaveric femora (8F/2M, 66 to 80 years) to failure. In each case, the right femur was implanted, using bone cement, with a machined resurfacing component aligned neutrally with respect to the femoral neck whereas the left femur was implanted at 10° valgus. The superolateral notch was 3 mm wide by 2 mm deep directly beside the component. Potted femurs were loaded to failure using an Instron materials testing machine. All 20 femora were scanned using Dual-Energy X-Ray Absorptiometry.

Results: The effect of neutral versus valgus placement was complex. (1) Two pairs slowly crushed; the remaining femurs exhibited a clear fracture. When only the fracture-type failures were analyzed, valgus placement resulted in fracture loads on average 22% higher than for neutral placement (paired t-test, p< 0.05). All femurs failed within the notch, as occurs clinically. (2) Femurs with high bone density (BMD> 0.65 g/cm2) showed a significant increase in fracture load (p< 0.05) whereas femurs with low BMD (< 0.65 g/cm2) were unaffected by component placement. BMD was highly correlated with fracture load (Pearson r=0.87, p=0.0003). (3) The greatest improvements occurred when the neck-shaft angle was relatively low, 128°–132°. (4) Two of ten femurs required larger head sizes at 10° valgus.

Conclusions: (1) Fracture load was primarily controlled by bone quality (BMD); (2) nevertheless, varus/valgus placement did affect the fracture load; (3) the magnitude and direction of this effect depended on fracture type, bone mineral density and the original neck-shaft angle; (4) for the level of bone quality typical of patients undergoing hip resurfacing, and for low-to-average neck-shaft angles (up to 132°), the fracture load for 10° valgus placement was significantly higher than for neutral placement. Funding: Other Education Grant

Purpose: The purpose of this work was to study the reliability and the precision of a lumber vertebra reconstruction method using images obtained from a 3D statistical model and two calibrated radiograms. The technique is designed for surgical approach to the lumbar spine and implantation of osteosynthesis material using enhanced-reality technology.

Material and methods: A lumbar vertebra was reconstructed on several specimens using images issuing from a 3D statistical model and two calibrated radiograms. The images obtained from the model of this lumbar vertebra to be reconstructed constituted the preoperative images. Intra-operative images corresponded to two calibrated radiograms acquired with a fluoroscope using advanced technology (silicium receptor). The model was equipped with reflecting patches which can be detected in space using a 3D optical system. Correspondence between the 3D statistical model and the two calibrated radiograms was achieved with appropriate software. Navigation views were displayed on the screen to guide surgical tools at the vertebral level. Pedicular screws were implanted into several anatomic specimens to evaluate the reliability and precision of the system. The exact position of the implanted screws was established with computed tomography.

Results: This system demonstrated its reliability and precision for the reconstruction of a lumbar vertebra from a 3D statistical model and two calibrated radiograms. All the implanted screws were perfectly positioned in the pedicles. Precision was to the order of 1 mm.

Discussion: This method is a passive system not requiring intraoperative intervention. Reconstruction of a lumbar vertebra from a preoperative 3D statistical model and two intra-operative calibrated radiograms avoids the need to identify anatomic landmarks and/or surface points on the vertebra to be reconstructed. The level of precision is very similar to that obtained with CT-based systems. Preoperative CT is not needed for navigation.

Conclusion: With this system, new generation fluoroscopic equipment should appear in the operating room, allowing acquisition of successive calibrated images. The digital data could then be matched with statistical anatomic data, avoiding the need for preoperative imaging (CT or MRI). Progressive introduction of intra-operative ultrasound to replace the calibrated radiograms should open a new approach for percutaneous surgery of the lumbar spine.

Purpose: This study analysed nerve trunk injury associated with posterior fractures of the pelvic girdle, distinguishing initial post-trauma damage from morbidity correlated to treatment by reduction and iliosacral screw fixation.

Material: Fifty bone or ligament injuries to the posterior pelvic girdle were identified in 44 patients. Management included initial external reduction and differed fixation using iliosacral screws inserted under fluorescence guidance.

Methods: The metameric examination of the lumbar and sacral trunks (L2, L3, L4, L5, S1, S2, S3) was performed at admission when the patient was conscious. The postoperative work-up included a complete neurological exam and computed tomography assessment of the screw trajectory. The quality of the reduction was quantified on the anterioposterior view of the pelvis. At last follow-up, evolution of symptoms ± EMG, Trendenburg gait, Mageed score, QMS score and pain (assessed on a visual analogue scale) were recorded.

Results: Preoperatively, 14 deficits of the nerve trunks were identified. The neurological status was unknown for eleven bone and ligament injuries because the patients were sedated at admission. Postoperatively, 28 deficits were identified. Fourteen (50%) involved the lumbosacral trunks L4 and L5, five the S1 root, six L4, L5, S1 territories, and three L5 to S4.

Computed tomography demonstrated 15 extraosseous screws lying anteriorly to the sacral ala or in the sacral canal. These extraosseous screws were associated with neurological deficits in nine cases without a preoperative diagnosis. In six cases, the extra-ossesous screw was not associated with any postoperative deficit. In five cases, neurological lesions diagnosed after the operation were not associated with an extra-osseous screw. Twenty-six neurological lesions were reviewed at a mean follow-up of 25 months: improvement was observed in 19, no change in five and aggravation in two.

Conclusion: Initial diagnosis of neurological injury with precision of the localisation can be established for only half of pelvic girdle fractures. The main mechanism involves stretching of the lumbosacral trunk by displacement of the sacral ala. Injury to the superior gluteal nerve is often associated. Closed reduction or compression of a nerve trapped in the fracture gap during screw fixation could be a second mechanism. Finally, rigorous screw insertion is necessary to avoid extra-osseous trajectories lying anteriorly to the sacral ala.

Purpose: A new software based on computed tomography data has been developed for individual unicompartmental reconstruction. The puropse is to achieve custom-made anatomic reconstruction of the medial compartment of the knee and to restore the physiological kinetics of each specific knee. This system can also be used to provide individualised tools for implantation

Material and methods: Custom-made unicompartmental prostheses and their corresponding implantation patterns were designed by this software. The data were fed to a rapid prototyping machine to produce a two-piece (cobalt-chromium and polyethylene) unicompartmental prosthesis and the guides for anatomic cuts specific for each knee. This method was tested on ten cadaver specimens in order to verify the precision of the reconstruction and alignment. A specially designed software was used to establish 3D reconstructions of the knee specimen taking precisely into account the cartilage surfaces. Virtual bone cuts were established to obtain the future seat for the implant. The data of the virtual bone cut were recovered to obtain a brute form of the future implant. The joint surfaces which still presented the degenerative lesions were repaired and finalised by interpolation and by use of the mirror data obtained from the contralateral knee. An individual support system was then installed to perform the bone cuts established virtually. The final data were then fed to a rapid prototyping machine and the aleasing machine to produce a two-piece unicompartment prosthesis and the guides for the anatomic cuts specific for each knee. The operation was then performed via a medial parapatellar 8cm incision. The implants were inserted without other fixation onto the knees and tested immediately to determine function. Anatomic reconstruction of the knee was checked with standard x-rays and CT images compared with the initial data.

Results: The implants could be inserted rapidly and easily. The impact of exhibited an excellent press fit and reproduced the initial morphology of the healthy articulation.

Discussion: Movement of the knee prosthesis reproduced all the characteristic physiological movements. Furthermore, the knee prosthesis obtained with this method allowed individualised and minimalised bone cuts, optimising congurency and contact zones between the bone and the implant and thus reducing the difficulties of surgical alignment. This system appears to offer high-performance restoration of individual physiological joint kinetics.

Purpose: The purpose of this work was to compare the precision and reliability of screw fixation using two different guiding systems. The first system was based on computed tomography (CT) imaging and the second on digitalized fluoroscopic imaging.

Material and methods: Between 1998 and 2000, 88 patients underwent spinal fixation for diverse disease states (idiopathic scoliosis in 43, and fracture, spondylolisthesis or instability in 45). Pedicular screws (n = 223) were inserted in levels T4 to S1. The passive CT navigation system was used for 73 patients (177 pedicular screws) and the fluoroscopic navigation system for 15 (46 pedicular screws). An independent observer identified the position of the pedicular screws on the postoperative CT.

Results: Among the 73 patients who underwent a CT-guided procedure (177 pedicular screws) the rate of incorrect screw position was 6.2% (11/117) with = 2 mm penetration of the cortical. Among the 15 patients who underwent a fluoroscopy-guided procedure (46 pedicular screws), the rate of incorrect screw position was 17% (8/46) again with = 2 mm penetration of the cortical. For scoliosis patients, the rate of erroneous screw insertion was 6% for CT navigation and 28% for fluoroscopic navigation. For fractures and degenerative instability, the rates were 6% and 11% respectively.

Discussion: The passive nature of the two navigation systems used do not induce any peroperative constraint on the surgeon. With the CT system, landmarks have to be collected peroperatively on the posterior arch of the operated vertebra, a step that is not needed for the fluoroscopic system. The two techniques appear to be reliable for insertion of pedicular screws. We did not have any neurological disorders in this series. It can be recalled that the conventional method produces a 15 to 40% rate of erroneous insertion. The CT system provides better results for all types of diseases; the improvement is about 6%.

Conclusion: With CT-navigation, a large portion of the per-operative radiographs are no longer necessary. Operative time is slightly longer than for the classical procedure due to the collection of the 3D information, particularly important for scoliosis. With the fluoroscopy system, no special preoperative imaging is required. Two or three peroperative radiographs are sufficient, limiting irradiation during insertion of the pedicular screws. The fluoroscopic system does not however provide 3D images.

Purpose: We present a prospective review of 30 unstable pelvic ring fractures treated with iliosacral screw fixation under fluoroscopic guidance with or without anterior osteo-synthesis.

Material and methods: Thirty patients were included between January 1997 and June 2000. They were all treated in an emergency setting by traction with or without a pelvic clamp. Osteosynthesis was deferred for a mean eight days. Percutaneous iliosacral screw fixation was used in all patients associated with symphsis osteosynthesis in eight. The radioscopic technique used three views: inlet, outlet and lateral. There were three Tile B fractures, 26 Tile C fractures and one sacral fracture. The ISS was 30.8/75. Eleven patients experienced haemorrhagic shock, 16 had injuries to the lower limbs, four had surgical urological injuries, and seven had emergency arterography. A neurology examination and CT scan were obtained pre- and postoperatively. Clinical assessment used a visual analogue scale, the Majeed score, and the WHO score. Peroperative radiation was quantified in minutes, Kvots and mA.

Results: Fifty-one screws were implanted, 24 patients with two screws. There were 12 outside the bone and seven potentially iatrogenic lesions among the 18 neurological lesions observed. Mean radiatio was 1.03 min per patient and 0.6 min per screw. Mean follow-up was 24 months (9–50). Clinically the mean scores were: Majeed 8.5/100, WHO 0.7/3, visual analogue scale 3.2/10). Twenty patients suffered pain related to associated injuries. There were three dismantelings including one requiring revision surgery. Deformed callus was: anterior vertical translation 2.9 mm, posterior vertical translation 4.5 mm and horizontal anterior translation 3 mm. There was one complication for the symphysis fixation.

Discussion: These results are comparable to those reported in the literature. Complications are less frequent with this method. External reduction is good when achieved early. This series represents a learning curve where the 12 screws in extraosseous positions occurred during our first 16 cases. Implanting two screws per articulation increases the risk of extraosseous screws. Screw insertion is safer with the lateral view.

Conclusion: This series demonstrates that iliosacral percutaneous screw fixation is a valid method for the treatment of severe fractures of the pelvic ring. The rigorous method required is emphasised.