Multidisciplinary management of the bone and joint infection complicating treatment of an open fracture of the lower limb.

Bone and joint infections (BJI) of the lower limb can cause functional sequelae and in some cases have an impact on patient's life prognostic. One of the main objectives of multidisciplinary consultation team meetings (MTM) in the treatment of bone and joint infections is to provide an appropriate medical-surgical care, pooling skills of different organ specialists: infectious disease physicians, microbiologists, orthopedic surgeons and plastic surgeons. Treatment is based on aggressive debridement, bone stabilization, adequate antibiotic therapy, long-term coverage of the loss of skin substance and close clinical monitoring. The authors present their multidisciplinary diagnostic and therapeutic approaches to BJI complicating an open fracture at a referent center in the management of complex bone and joint infections.

Multidisciplinary team meeting for complex bone and joint infections diagnosis: The PHICTOS study.

BACKGROUND: In France, the most severe bone and joint infections (BJI), called "complex" (CBJI), are assessed in a multidisciplinary team meeting (MTM) in a reference center. However, the definition of CBJI, drawn up by the Health Ministry, is not consensual between physicians. The objective was to estimate the agreement for CBJI classification. METHODS: Initially, five experts from one MTM classified twice, one-month apart, 24 cases as non-BJI, simple BJI or CBJI, using the complete medical record. Secondly, six MTMs classified the same cases using standardized information. Agreements were estimated using Fleiss and Cohen kappa (κ) coefficients. RESULTS: Inter-expert agreement during one MTM was moderate (κ=0.49), and fair (κ=0.23) when the four non-BJIs were excluded. Intra-expert agreement was moderate (κ=0.50, range 0.27-0.90), not improved with experience. The overall inter-MTM agreement was moderate (κ=0.58), it was better between MTMs with professor (κ=0.65) than without (κ=0.51) and with longer median time per case (κ=0.60) than shorter (κ=0.47). When the four non-BJIs were excluded, the overall agreement decreased (κ=0.40). CONCLUSION: The first step confirmed the heterogeneity of CBJI classification between experts. The seemingly better inter-MTM than inter-expert agreement could be an argument in favour of MTMs, which are moreover a privileged place to enhance expertise. Further studies are needed to assess these results as well as the quality of care and medico-economic outcomes after a MTM.

FCN-based approach for the automatic segmentation of bone surfaces in ultrasound images.

PURPOSE: A new algorithm, based on fully convolutional networks (FCN), is proposed for the automatic localization of the bone interface in ultrasound (US) images. The aim of this paper is to compare and validate this method with (1) a manual segmentation and (2) a state-of-the-art method called confidence in phase symmetry (CPS). METHODS: The dataset used for this study was composed of 1738 US images collected from three volunteers and manually delineated by three experts. The inter- and intra-observer variabilities of this manual delineation were assessed. Images having annotations with an inter-observer variability higher than a confidence threshold were rejected, resulting in 1287 images. Both FCN-based and CPS approaches were studied and compared to the average inter-observer segmentation according to six criteria: recall, precision, F1 score, accuracy, specificity and root-mean-square error (RMSE). RESULTS: The intra- and inter-observer variabilities were inferior to 1 mm for 90% of manual annotations. The RMSE was 1.32 ± 3.70  mm and 5.00 ± 7.70 mm for, respectively, the FCN-based approach and the CPS algorithm. The mean recall, precision, F1 score, accuracy and specificity were, respectively, 62%, 64%, 57%, 80% and 83% for the FCN-based approach and 66%, 34%, 41%, 52% and 43% for the CPS algorithm. CONCLUSION: The FCN-based approach outperforms the CPS algorithm, and the obtained RMSE is similar to the manual segmentation uncertainty.

Reproducibility of length measurements of the lower limb by using EOS™.

BACKGROUND: Lower-limb discrepancy following total hip arthroplasty is the third-most common reason for patient dissatisfaction in orthopaedic surgery. Therefore, accurate planning and evaluation methods are mandatory. The main aim of this study was to evaluate the reliability of the EOS™ system by establishing and comparing the reproducibility of lower-limb automatic and manual 3D measurements. We hypothesized that the reproducibility of the lower-limb measurements is similar regardless of the method used and with an agreement higher than 0.95 for the length parameters. MATERIALS AND METHODS: This study utilized an EOS radiological database of 112 patients. Two independent observers performed two rounds of lower-limb measurements twice, either in manual 3D or automatic 3D mode. The intra- and inter-observer reproducibility was evaluated by the calculation of the intra-class coefficient for each measurement method. The methods were then compared. RESULTS: The intra- and inter-observer reproducibility for length measurements found with the manual and automatic 3D methods was always > 0.98. There was no significant difference in the reproducibility between the two measurement modes, with the exception of the offset, hip-knee-shaft, and neck-shaft angles. CONCLUSION: Our results indicate a very good reproducibility of EOS™ length measurement, regardless of the method used. Automated 3D mode is preferred for the collection of angular and offset measurements. Furthermore, manual mode measurements are not affected by surgical history. Level of evidence IV.

Influence of experience on intra- and inter-observer reproducibility of the Crowe, Hartofilakidis and modified Cochin classifications.

INTRODUCTION: A previous study demonstrated that the reproducibility of the Crowe (Cr), the Hartofilakidis (Ha) and the modified Cochin (Co) classifications were comparable. However, there were differences with a trend that suggested the influence of experience. Therefore, we performed a prospective study to investigate whether experience influenced the reproducibility of the commonly used developmental dysplasia of the hip (DDH) classifications. HYPOTHESIS: The hypothesis was that the intra- and inter-observer reproducibility scores would be higher in the senior group than the junior group, and particularly for the modified Co classification. METHODS: Four seniors and four residents classified 104 A/P pelvic radiographs (200 hips) two times using the Cr, Ha and Co classification systems. RESULTS: For intra-observer reproducibility, the average weighted concordance coefficients [95% confidence intervals] were for the senior and the junior groups: 92.2 [88.6-95.7] and 92.6 [87.9-97.2] for Cr, 92.1 [88.7-94.6] and 92.0 [87.7-96.3] for Ha, 94.2 [91.8-96.6] and 94.1 [91.5-96.6] for Co. The average weighted Kappa (95% confidence intervals) were 0.8 [0.71-0.88] and 0.79 [0.68-0.89] for Cr, 0.77 [0.74-0.81] and 0.75 [0.62-0.88] for Ha, 0.82 [0.76-0.89] and 0.80 [0.74-0.87] for Co. The junior inter-observer reproducibility multi-rater Kappa (list A:list B) were 0.57:0.50 (Cr), 0.47:0.53 (Ha), 0.42:0.42 (Co). Senior multi-rater Kappa were 0.53:0.49 (Cr), 0.40:0.34 (Ha), 0.40:0.43 (Co). CONCLUSIONS: Contrary to our hypothesis, the experience of the observer did not affect the intra- and inter-observer reproducibility of the three classification systems.

Design and evaluation of an actuated knee implant for postoperative ligament imbalance correction.

In Total Knee Arthroplasty (TKA), the collateral ligament tensioning stage cannot be standardised for all patients and relies heavily on the surgeon's experience and perception. Intraoperative inaccuracies are practically unavoidable and may give rise to severe postoperative complications, leading to the need for revision surgery already a few years after primary TKA. This work proposes a novel instrumented tibial component able to detect collateral ligament laxity conditions right after primary TKA and, if needed, to compensate for them in the postoperative period. A miniaturised actuation system, designed to be embedded in the tibial baseplate, was initially evaluated by means of 3D simulations and then fabricated as a full-scale prototype. Stability and force sensors tests carried out on a knee simulator allowed to assess the effectiveness of the proposed design under normal working conditions and provided valuable insights for future work and improvements.

Can computer-assisted surgery help restore leg length and offset during THA? A continuous series of 321 cases.

INTRODUCTION: Total hip arthroplasty (THA) can bring about complications - particularly leg length differences - that are becoming increasingly litigious. Computer-assisted orthopedic surgery (CAOS) can help optimize the procedure, but its ability to effectively restore leg length is controversial. As a consequence, we carried out a study to determine: (1) its contribution to meeting leg length and offset objectives, (2) its reliability, by evaluating the correlation between radiological and navigation data, (3) its safety, by evaluating navigation-specific and non-specific complications. HYPOTHESIS: CAOS will help to restore leg length within ± 5 mm in more than 80% of cases. MATERIAL AND METHODS: A series of 321 continuous cases of cementless THA implanted through the posterolateral approach using CAOS was analyzed retrospectively. With a minimum 1 year follow-up, we evaluated whether the leg length and offset goals were achieved, how well the navigation and radiology data were correlated and whether navigation-specific and non-specific complications occurred. Based on our hypothesis that 80% of patients would have less than 5 mm leg length difference and the null hypothesis (PA = P0) with an alpha of 0.05, 200 observations were required to achieve a power of 90%. RESULTS: The leg length and offset objectives were achieved in 83.3% and 88% of cases, respectively. Twenty-two patients required a heel wedge to compensate for leg length differences. The correlation between the radiology and surgical navigation data was satisfactory - the Pearson coefficient was 0.79 for length and 0.74 for offset. Intraoperative and postoperative complications or adverse events were found in 14.6% of cases; these were specific to CAOS in 12.1% of cases and non-specific in 2.5% of cases. CONCLUSION: This study shows the relevance of CAOS for achieving preoperative leg length objectives, with good correlation between navigation and radiology data, and without major complications. LEVEL OF EVIDENCE: IV - retrospective study.

Comparison of the reliability of leg length and offset data generated by three hip replacement CAOS systems using EOS™ imaging.

INTRODUCTION: Computer-assisted orthopaedic surgery (CAOS) theoretically will help to optimise total hip arthroplasty (THA) results. Although CAOS systems have become more sophisticated, they are not widely used, partially because of their suspect reliability. To assess reliability of these systems, we carried out a study with multiple objectives in mind: (1) establish and compare the accuracy of the leg length (LL) measurement from three CAOS systems; (2) analyse the correlation of LL and offset data generated by these CAOS systems with those of the EOS™ imaging system; (3) determine if the goals of leg length restoration with ±2 and ±5 mm were achieved; (4) evaluate why certain cases fails. HYPOTHESIS: The three CAOS systems have the same accuracy for LL, and their error is less than or equal to 0.6mm. MATERIAL AND METHODS: We retrospectively studied 106 cases of primary THA where preoperative and postoperative measures of leg length had been performed with an EOS™ imaging system. The cases were placed in three groups, depending on which CAOS system had been used: group A (Amplivision™, amplitude), group B (Hip Express™, Brainlab), group P (THS™, Praxim). The accuracy of the leg length data was calculated by finding the difference between the data from each CAOS system and the gold-standard EOS measurements. RESULTS: The leg length accuracy was -0.846 [-5 to 9], -0.675 [-9 to 18] and 0.542 mm [-5 to 13], respectively for groups A, B and P. The accuracy was significantly lower in group A than B (P=0.044) and group P (P=0.038). The Pearson correlation coefficient for CAOS and EOS measurements was 0.189, 0.701 and 0.891 for leg length and 0.668, 0.202 and 0.680 for offset, for groups A, B and P, respectively. No difference between groups were observed relative to the leg length objectives being achieved within ±2 mm (P=0.61) and ±5 mm (P=0.314). There were no differences in terms of the number of CAOS failures: three in group A, one in group B and three in group P (P=0.06). CONCLUSION: The Praxim™ and Brainlab™ CAOS systems had similar accuracy for leg length measurements, and both were better than the Amplitude™ system. Only the Praxim™ had an error of less than 0.6mm. All the CAOS systems had values less than 1mm, which is considered excellent. LEVEL OF EVIDENCE: Level IV, retrospective study.

Comparison of the inter- and intra-observer reproducibility of the Crowe, Hartofilakidis and modified Cochin classification systems for the diagnosis of developmental dysplasia of the hip.

INTRODUCTION: Developmental dysplasia of the hip (DDH) leads to multiple treatment challenges during adulthood. Surgical treatment is mainly based on radiographic evaluation of the anatomical alterations. Several classification systems have been described in the published English scientific literature, but the French Cochin classification has not been used very much. Its primary advantage lies in its ability to intricately describe the DDH alterations with a large number of grades. We hypothesized that the inter- and intra-observer reproducibility of the SOFCOT-modified Cochin classification system was equal to that of the Crowe and Hartofilakidis classifications. MATERIAL AND METHODS: Five French orthopaedic surgeons who were DDH experts classified 94 A/P pelvis radiographs (179 hips) using the Crowe (Cr), Hartofilikadis (Ha) and modified Cochin (Co) systems. This evaluation was repeated a second time one month later. The intra-observer reproducibility was determined with weighted Kappa and concordance coefficients. The inter-observer reproducibility was performed by calculating the multirater Kappa coefficient on each of the two data series. RESULTS: For the intra-observer reliability, the average weighed concordance coefficients (95% CI) were 88.62-94.52 for Cr, 89.43-93.80 for Ha and 92.14-95.71 for Co. The average weighed Kappa coefficients (95% CI) were 0.70-0.85 for Cr, 0.67-0.82 for Ha and 0.75-0.83 for Co. For the inter-observer reliability, the Kappa for each assessment round was 0.57 and 0.48 for Cr, 0.43 and 0.44 for Ha, and 0.43 and 0.37 for Co. DISCUSSION: The intra- and inter-observer reliability for the modified Cochin classification system is the same as the one for the Crowe and Hartofilakidis classifications. The theoretical advantage of this classification system should be confirmed by comparing the findings with intra-operative anatomical observations. LEVEL OF PROOF, TYPE OF STUDY: IV.

A miniaturised actuation system embedded in an instrumented knee implant for postoperative ligament imbalance correction.

During Total Knee Arthroplasty surgery, the orthopaedic surgeon has to set up proper balance conditions for the two lateral ligaments of the knee. Such ligament tensioning procedure is performed manually and mainly depends on the surgeon's experience. Unfortunately, inaccuracies are unavoidable and may give rise to serious postoperative complications. In the worst-case scenario, the only solution to this problem is represented by revision surgery. In order to cope with this problem, this work proposes a novel instrumented tibial component able to detect knee imbalance conditions in the postoperative period. A miniaturised actuation system embedded in the tibial baseplate allows to restore optimal balance conditions without resorting to revision surgery.

Automatic registration of pre- and intraoperative data for long bones in minimally invasive surgery.

The Minimally Invasive Procedures (MIP) in orthopedics have grown rapidly worldwide, as clinical results indicate that patients who undergo MIP typically experience minimized blood loss, smaller incision and shorter hospital stays. For most MIP, a preoperative 3D model of the patient anatomy is usually generated in order to plan the surgery. The challenge in MIP consists in finding the correspondence between the preoperative model and the actual position of the patient in the operating room, also known as image-to-patient registration. This paper proposes a real-time solution based on ultrasound (US) images: the patient anatomy is scanned by an US probe. Then, the segmentation and the extraction of bone contours from US images result in a 3D point cloud. The Poisson surface reconstruction method provides a 3D surface from 2D US data which will be registered with the preoperative model (CT volume) using the principal axes of inertia and the Iterative Closest Point robust (ICPr) algorithm. We present quantitative and qualitative results on both phantom and clinical data and show a mean registration accuracy of 0.66 mm for clinical radius scan. The promising registration results show the possible use of the proposed registration algorithm in clinical procedures.

Recognition-based segmentation and registration method for image guided shoulder surgery.

For any image guided surgery, independently of the technique which is used (navigation, templates, robotics), it is necessary to get a 3D bone surface model from CT or MR images. Such model is used for planning, registration and visualization. We report that graphical representation of patient bony structure and the surgical tools, interconnectively with the tracking device and patient-to-image registration are crucial components in such a system. For Total Shoulder Arthroplasty (TSA), there are many challenges, The most of cases that we are working with are pathological cases such as rheumatoid arthritis, osteoarthritis disease. The CT images of these cases often show a fusion area between the glenoid cavity and the humeral head. They also show severe deformations of the humeral head surface that result in a loss of contours. This fusion area and image quality problems are also amplified by well-known CT-scan artifacts like beam-hardening or partial volume effects. The state of the art shows that several segmentation techniques, applied to CT-Scans of the shoulder, have already been disclosed. Unfortunately, their performances, when used on pathological data, are quite poor [1, 2]. The aim of this paper is to present a new image guided surgery system based on CT scan of the patient and using bony structure recognition, morphological analysis for the operated region and robust image-to-patient registration.

A 2D 3D registration with low dose radiographic system for in vivo kinematic studies.

The knowledge of the poses and the positions of the knee bones and prostheses is of a great interest in the orthopedic and biomechanical applications. In this context, we use an ultra low dose bi-planar radiographic system called EOS to acquire two radiographs of the studied bones in each position. In this paper, we develop a new method for 2D 3D registration based on the frequency domain to determine the poses and the positions during quasi static motion analysis for healthy and prosthetic knees. Data of two healthy knees and four knees with unicompartimental prosthesis performing three different poses (full extension, 30° and 60° of flexion) were used in this work. The results we obtained are in concordance with the clinical accuracy and with the accuracy reported in other previous studies.

Registration of low dose bi-planar acquisitions for motion analysis.

In this paper, we introduce a 2d-3d registration method for searching the motion of knee bones. We use a low dose bi-planar acquisition system that provides us with simultaneous frontal and profile radiographs in different positions, and the 3d volume reconstruction of the standing position. The purpose here is to reduce the user intervention during the motion tracking. The registration method is based on the central slice Fourier Transform theorem. Motion results with rotations and translations using synthetic data are shown.

Rotational femoral osteotomies using an endomedullary saw.

INTRODUCTION: The goal of this retrospective study is to analyze a series of ten patients (11 osteotomies) who underwent closed femoral rotational osteotomy using an endomedullary saw; stabilization was achieved by a locked centromedullary nail. We report the indications, technical aspects, clinical and radiological results as well as intercurrent complications with this surgical technique. HYPOTHESIS: Femoral endomedullary osteotomy is a safe procedure to correct malrotations. PATIENTS AND METHODS: Eleven femoral rotational osteotomies using an endomedullary saw were performed on ten patients, between January 1999 and July 2007. The indications were post-traumatic rotational malunions or congenital rotational deformities. The angular deformity averaged 33.5 degrees (24 degrees -52 degrees ). They were divided into internal rotation (ten cases) or external rotation (one case). One patient required a bilateral rotational osteotomy because of a congenital femoral malrotation combined to bilateral trochlear dysplasia. Rotational correction was, in two patients, simultaneously associated with a closed lengthening osteotomy. Clinical and radiological follow-up averaged 4 years and 9 months (26-104 months). The angular corrections obtained by these rotation osteotomies were calculated by CAT scans. RESULTS: Ten out of eleven osteotomies allowed a correction within a 4 degrees range in relation to the physiological femoral neck anteversion values (or to the contralateral side in the case of a healthy opposite lower extremity). There was no bone, joint, skin, or soft tissues infection, no pseudoarthrosis and no delayed outgrowth. We observed a transient neurological complication in the area of the pudendal nerve, during a combined rotational-lengthening osteotomy, as well as a bilateral femoral fracture during the bilateral rotational osteotomy. In all patients, consolidation occurred within a 3- to 5-month delay. The subjective results showed that eight out of nine patients (one was lost to follow-up) were satisfied or very satisfied with their operation, their functional recovery and the aesthetic aspects of their scars. DISCUSSION: Closed rotational osteotomies in adults represent a reliable, effective, safe and reproducible procedure for the correction of femoral torsion problems, whether they are post-traumatic or congenital. These results can be obtained only by respecting the indications and by applying a rigorously planned technique, including expertise in centromedullary nailing. LEVEL OF EVIDENCE: Level IV, retrospective therapeutic study.

Femoral offset: anatomical concept, definition, assessment, implications for preoperative templating and hip arthroplasty.

UNLABELLED: BACKGROUND OBJECTIVE: Femoral offset is supposed to influence the results of hip replacement but little is known about the accurate method of measure and the true effect of offset modifications. MATERIAL AND METHODS: This article is a collection of independent anatomic, radiological and clinical works, which purpose is to assess knowledge of the implications of femoral offset for preoperative templating and total hip arthroplasty. RESULTS: There is a strong correlation between femoral offset, abductors lever arm and hip abductor strength. Hip lateralization is independent of the femoral endomedullary characteristics. The abductors lever arm is highly correlated to the gluteus medius activation angle. There were correlations between femoral offset and endomedullary shape. The hip center was high and medial for stovepipe metaphysis while it was lower and lateralized for champagne - flute upper femur. A study was performed to compare the femoral offset measured by X-ray and CT-scan in 50 patients, demonstrated that plain radiography underestimates offset measurement. The 2D templating cannot appreciate the rotation of the lower limb. Taking into account the horizontal plane is essential to obtain proper 3D planning of the femoral offset. A randomized study was designed to compare femoral offset measurements after hip resurfacing and total hip arthroplasty. This study underlined hip resurfacing reduced the femoral offset, while hip replacement increased offset. However, the reduction of femoral offset after hip resurfacing does not affect the function. A pilot study was designed to assess the results of 120 hip arthroplasties with a modular femoral neck. This study showed that the use of a modular collar ensures an easier restoration of the femoral offset. A cohort of high offset stems (Lubinus 117 degrees) was retrospectively assessed. The survival rate was slightly lower that the standard design reported in the Swedish register. Finally, the measurement of offset and leg length was assessed with the help of computer assistance. The software changed the initial schedule (obtained by templating) in 29%. CONCLUSION: Therefore, femoral offset restoration is essential to improve function and longevity of hip arthroplasty. CT-scan is more accurate than plain radiography to assess femoral offset. Hip resurfacing decreases offset without effect on function. Modular neck and computer assistance may improve intraoperative calculation and reproduction of femoral offset. Increasing offset with a standard cemented design may decrease long-term fixation. Level IV: Retrospective or historical series.

[Septic arthritis after arthroscopic cruciate ligament reconstruction: in vitro comparison of the behavior of two types of interference screws towards the Staphylococcus aureus adherence]. / Arthrite septique après ligamentoplastie du pivot central sous arthroscopie : comparaison in vitro du comportement de deux types de vis d'interférence vis-à-vis de la fixation du Staphylococcus aureus.

PURPOSE OF THE STUDY: We report the results of an experimental study designed to investigate the behaviour of two types of interference screws (bioabsorbable versus titanium) in a context of Staphylococcus aureus infection. The main objective was to study one of the possible sources of failure for the treatment of acute arthritis of the knee after arthroscopic reconstruction of the cruciate ligament. MATERIAL AND METHODS: We used six interference screws made of titanium and six bioabsorbable screws (PLLA-PDLLA). All screws measured 9mmx25mm, were cannulized and sterilized. These screws were submitted to a protocol elaborated in cooperation with the bacteriologists. After contamination with a strain of S. aureus, the screws were washed four times in saline solution to eliminate germs adsorbed in the aqueous phase. The last step was trypsination to detach germs remaining fixed onto the screws and contained in the biofilms of glycocalix. A germ count was made after each step. For each screw, we determined the difference (Delta) corresponding to the number of germs really adherent to the screw. Mann-Whitney analysis was performed. RESULTS: On average, the germ count in the aqueous phase was 0.0855.10(5)/ml for the titanium screw versus 0.223.10(5)/ml for the bioabsorbable screw. The mean count of germs fixed in the biofilm (mean Delta) was 17.695+/-3.88.10(5) for the titanium screw and 45.86+/-3.61.10(5) for the bioabsorbable screw. The difference was statistically significant (p=0.0039). DISCUSSION: Our experimental results support the efficacy of abundant arthroscopic lavage in a context of infection, confirming the results of very recent studies. However, irrespective of the type of material used, bioabsorbable screws are more prone to persistent microbial adherence than titanium screws. It is hypothesized that the hydrophilic, more porous microstructure of bioabsorbable screws favors bacterial adherence. In practice, this implies that arthroscopic washout must be as abundant and as complete as possible, yet may not be sufficient to clean the entire surface of potentially contaminated screw positioned deep in a bony tunnel.

[Tibial valgus osteotomy]. / Ostéotomie tibiale de valgisation.

The tibial valgus osteotomy whatever its technique has a survival rate of about 85 % to 10 years, if we consider the reoperation as a criterion of failure, with a confidence index at 78%. The age, weight, sex and functional signs have no impact on the outcome. We have found no evidence in the preoperative radiographic assessment, neither the medial pinch, or varus epiphyseal neither varisant gap, which could be a failure and a reoperation before the tenth year. Good results were observed significantly when there is an over-valgus at least 3 degrees of global axis of the lower limb. This corresponds to a valgus epiphyseal by more than 2 degrees . The substantial reduction in the gap varisant that lowers the overall time varisant below 200 kg cm provides the same positive results. The outcome will depend directly on the accuracy of the calculation of the preoperative correction performed and the quality of surgical achievement. Because of the need for precision, navigation technique appears as reliable, simple which makes it also possible to monitor the front slope and tibial rotation induced. The osteosynthesis must be stable and rigid to avoid postoperative loss of correction.