|Home | About | Journals | Submit | Contact Us | Français|
Redirective, periacetabular osteotomies (PAO) represent a group of surgical procedures for treatment of developmental dysplasia of the hip (DDH) in skeletally mature and immature patients. The ultimate goal of all procedures is to reduce symptoms, improve function and delay or prevent progression of osteoarthritis. During the last two decades, the understanding of the underlying pathomechanisms has continuously evolved. This is mainly attributable to the development of the femoroacetabular impingement concept that has increased the awareness of the underlying three-dimensional complexity associated with DDH. With increasing knowledge about the pathobiomechanics of dysplastic hips, diagnostic tools have improved allowing for sophisticated preoperative analyses of the morphological and pathobiomechanical features, and early recognition of degenerative changes, which may alter the long-term outcome. As redirective, PAO are technically demanding procedures, preoperative planning is crucial to avoid intraoperative obstacles and to sufficiently address the patient-specific deformity. Although conventional radiography has been used for decades, it has not lost its primary role in the diagnostic work-up of patients with DDH. Furthermore, an increasing number of modern imaging techniques exists allowing for assessment of early cartilage degeneration (biochemical magnetic resonance imaging) as well as 3D planning and computer-based virtual treatment simulation of PAO. This article reviews the literature with regard to the current concepts of imaging of DDH, preoperative planning and treatment recommendations for redirective, PAO.
The natural course of developmental dysplasia of the hip (DDH) has been reported to be poor, with early development of osteoarthritis (OA) requiring joint replacement surgery [1–4]. Depending on the patient’s age and skeletal maturity, different pelvic osteotomies have been proposed to surgically treat this condition. All techniques aim at improving hip pain and function by addressing the abnormal biomechanics of the hip. The goal is to relieve pain and to delay or prevent progression of OA to preserve the natural hip joint over time. Generally, pelvic osteotomies can be categorized as augmentation procedures and reorientation procedures. Augmentation procedures include shelf osteotomies , the Salter osteotomy  and the Chiari osteotomy . Redirective, periacetabular osteotomies (PAO) include rotational/spherical osteotomies [8, 9], triple osteotomies [10, 11] and the Bernese PAO . The general principles for preoperative planning do not differ among the different types of PAOs. However, most of recently published studies related to preoperative radiographic planning for redirective PAOs refer to the Bernese PAO. Therefore, direct implementation of the here listed planning principles should be done with caution when performing spherical/rotational or triple osteotomies. Skeletal maturity is a prerequisite when performing the Bernese PAO or spherical/rotational procedures as the osteotomies cross the triradiate cartilage complex in skeletally immature patients and may lead to early acetabular growth arrest and joint incongruency. This does not apply to triple osteotomies as the growth plates of the innominate bone remain uncompromised making triple osteotomies the procedure of choice in skeletally immature patients. PAOs render the potential of a free 3D re-orientation of the acetabulum. Due to the ability of performing an almost unlimited reorientation in the spatial room, severe dysplastic deformities [13, 14] and other complex hip deformities such as sequelae of Legg Calve Perthes disease [15, 16] or proximal femoral deficiency  can be treated with redirective, PAO. Additionally, acetabular retroversion causing pincer-type impingement can be corrected when performing an anteverting PAO [18, 19]. However, PAO are technically demanding procedures that are prone to treatment errors and complications if performed without adequate expertise. Patient selection, correct indication and thorough preoperative planning are crucial for successful outcome. The current article reviews the literature highlighting preoperative 2D and 3D imaging tools for preoperative planning of redirective, PAO.
DDH has been classically defined as deficient anterolateral acetabular coverage with a decreased surface area of the acetabular, articular surface and a steep orientation of the acetabular roof . The radius of the acetabular curvature frequently exceeds that of the femoral head resulting in joint incongruency (sloping roof dysplasia) . Short- or flat-roof dysplasia, in contrast, is represented by normal inclination of the acetabulum, whereas overall femoral head coverage is reduced [21, 22]. Short- or flat-roof dysplasia has been aetiologically attributed to recurrent impingement from an asphericity of the femoral head leading to growth and remodelling abnormalities of the iliac acetabular epiphysis after closure of the triradiate cartilage complex . The forces around the hip concentrate on the edge of the superior acetabular socket resulting in static overload [20, 23]. This condition is aggravated by frequently occurring instability with subluxation of the joint leading to a further increase of the peak pressures onto the articular cartilage particularly in hips with steep roof dysplasia . The resulting intra-articular damage pattern comprises compensatory hypertrophy of the labrum with subsequent partial or complete tears , cartilage damage [25–27] and ultimately OA . Short- or flat-roof dysplasia can be associated with osseous fatigue fractures of the superior acetabular margin . Since the widespread recognition of femoroacetabular impingement (FAI) , the understanding of the pathological hip biomechanics has evolved. This has increased the awareness of the underlying 3D problem associated with DDH. In this context, DDH was found to be associated with acetabular retroversion [29–31] resulting in a mixed picture of DDH and FAI . Both undercorrecting and overcorrecting femoral head coverages were found to predict failure after Bernese PAO [32, 33]. Additionally, it was shown that the articulating surface area of the acetabulum is decreased in dysplastic hips compared with normal hips [34, 35]. The resulting decreased contact area between the acetabulum and the femoral head remains unchanged after PAO. A dysplastic acetabulum can therefore only be corrected as close as possible to a normal acetabulum. Furthermore, surgical treatment of sloping roof dysplasia and short/flat roof dysplasia may vary considerably due to the pathomorphologic differences of these two entities (pelvic osteotomies versus treatment of impingement) . To date, there are no universal guidelines regarding target ranges of specific radiographic parameters to achieve an optimal correction. Alongside preditive factors associated with surgical accuracy (over-/under-correction), further negative predictors of outcome have been identified in previous studies including increased age at surgery, preoperative pain und reduced hip function, pre-existing OA, and hip subluxation [17, 32, 33, 36]. However, this is in contrast to other studies that reported favourable outcomes in older patients and patients with radiographic signs of preoperative OA after rotational/spherical and Bernese PAO [37–39].
Conventional radiography is the primary imaging modality for preoperative planning. The routine preoperative work-up includes an anterior-posterior (AP) radiograph of the pelvis and a lateral projection of the hip [40, 41] (i.e. cross-table view, frog-leg view, Lauenstein view , Dunn Rippstein Müller view [43, 44]). A false-profile view , a true lateral pelvic view and functional views serve as adjuncts. In the initial study by Ganz et al. , the authors suggested an AP view of the pelvis, a false-profile view, and abduction views as the minimally required preoperative projections for the Bernese PAO. However, they noted that computed tomography might add to a more precise 3D planning [12, 46]. It must be emphasized that despite the increasing availability of modern imaging technologies, conventional radiography has not lost its role in the preoperative work-up of dysplastic hips and remains a mandatory imaging tool for the assessment of DDH, treatment planning and postoperative evaluation of the achieved correction. Furthermore, the majority of radiographic factors (pre- and postoperatively) predicting outcome after PAO have been identified and validated on conventional radiographs [17, 32, 33, 36].
A variety of radiographic parameters associated with DDH exist to assess femoral head coverage, orientation of the acetabulum, joint congruity, femoral head sphericity, femoral torsion and the neck shaft angle (Table I). On the AP pelvis radiograph, lateral acetabular coverage is assessed by the lateral center edge angle (LCE ), the acetabular index (AI ) and extrusion index (EI ; Figs. 1 and and2).2). Anterior femoral head coverage can be assessed with the anterior CE  on the false-profile view and the anterior acetabular wall index  on the AP pelvis radiograph. Posterior coverage and acetabular version are assessed by the presence or absence of posterior wall sign  (Fig. 1), a crossover sign , the ischial spine sign , the retroversion index (RI)  and the posterior acetabular wall index  on the AP pelvic radiograph. A previous study suggested reference values for acetabular over- and undercoverage on the AP pelvis radiograph obtained in a standardized fashion with standardized pelvic tilt and rotation (Table I) . Eleven common radiographic parameters in patients with hip dysplasia undergoing PAO (n=26), isolated cam type FAI and isolated pincer-type FAI undergoing surgical hip dislocation were assessed. The intersection of the distribution curves determined the threshold values for each parameter (Table I).
Standardized acquisition techniques of plain radiographs are of upmost importance as factors such as film-tube distance, patient-film distance, centring and direction of the X-ray beam and pelvic orientation [50–55]. Most radiographic parameters of the hip have been defined on AP pelvic radiographs with the central beam directed to the radiographic centre of the pelvis defined as the midpoint between upper border of symphysis and midpoint between anterior superior iliac spines . Interpreting radiographic parameters on different projections such as low centred AP pelvis projections, hip centred projections or fluoroscopic images should be performed with caution because the projected angles can vary significantly from the standardized AP pelvis radiograph . Pelvic orientation during image acquisition (i.e. pelvic tilt, rotation and obliquity) has considerable effects on many radiographic parameters associated with DDH [51, 53, 56]. A previous study investigated the effects of tilt and rotation on a selection of standard radiographic parameters  and found anterior and posterior acetabular coverage [52, 57], the crossover and posterior wall sign  and the RI [40, 51] to vary considerably, whereas the LCE , AI , EI , ACM angle , Sharp angle and craniocaudal coverage [52, 57] were independent of pelvic tilt and orientation . Furthermore, it was reported that pelvic tilt is increased in supine AP pelvic radiographs compared with standing AP pelvic radiographs resulting in differences of the projected acetabular version, whereas the LCE angle remained unchanged .
Notably, due to the decreased area of the acetabular articular surface in dysplastic hips, meeting these reference values for normal acetabula after PAO is rarely possible. For instance, normalizing the AI to neutral does not always completely correct the LCE to a normal range. While further abduction would increase the LCE, this may result in a negative AI leading to pincer-type FAI . A previous study revealed that only 6% of hips undergoing PAO for DDH met the previously determined normal ranges of six predefined radiographic parameters (LCE, AI, EI, computerized anterior, posterior and total coverage) . Optimal acetabular reorientation was therefore defined if at least four of these six parameters were in the normal range resulting in improved survivorship after 10-year follow-up .
Joint congruity is assessed on the AP pelvis radiograph. The distance of the most medial contour of the femoral head to the ilioischial line has been shown to be 5–15mm . This distance increases with increasing degree of subluxation. Another indicator for joint incongruity and subluxation is an interrupted Shenton line (Fig. 2) . Abduction views can be used to assess joint incongruity. Due to subluxation in DDH, lateral joint space narrowing is frequently observed on the AP pelvis radiograph. In abduction, normal joint space restores due to a realignment or enlocation of the subluxed hip whereas persistent joint space narrowing is a relative contraindication for surgery because it indicates cartilage damage (Fig. 3). Apart from the assessment of subluxation, abduction views simulate the anticipated degree of acetabular reorientation.
Femoral head asphericity is a frequent feature in dysplastic hips, which can lead to secondary cam type FAI after PAO [64, 65]. It was shown previously that concomitant correction of an asphericity of the femoral head resulted in a better outcome 10years after PAO . Radiographic parameters for femoral head sphericity include the alpha angle , the head-neck offset ratio [40, 66, 67] on lateral projections of the hip and the pistol grip deformity on the AP pelvis radiograph . Further associated pathomorphologies of the femur include concomitant valgus, varus and torsional deformities. Although CT represents the gold standard to assess femoral torsion , Dunn Rippstein Müller allow for assessment of femoral torsion [43, 44]. Approximately 10% of patients undergoing PAOs require concomitant proximal femoral osteotomies (PFO) . Ganz et al  proposed an algorithm that allows to assess the necessity for concomitant PFO. For hips with a normal joint space and sufficient coverage and congruency on abduction views, isolated PAO is indicated. In case of persistent subluxation on abduction views, the abduction view will be repeated with additional internal rotation. Congruency of the joint indicates PAO with concomitant varus and derotation PFO .
Magnetic resonance imaging (MRI) (with or without arthrogram) is routinely performed in the preoperative assessment of DDH. MRI allows 3D assessment of the pathomorphology both on the acetabular and femoral side. The use of intraarticular gadolinium contrast agent injection (MR-arthrogram) has been associated with adverse events such as post injection pain, increased risk of infection and Gadolinium-related local and systemic toxicity . However, despite improving MRI protocols, previous studies revealed higher sensitivity and specificity for the detection of intraarticular cartilaginous and labral lesions on MR-arthrogram as compared with conventional MRI [72, 73]. Radial sequences oriented circumferentially around the femoral neck axis have gained increasing popularity as they provide perpendicular projections of the acetabular wall and labrum at all positions and a circumferential impression of asphericity at the femoral head-neck junction . MRI is useful to detect alterations of the articular cartilage and the acetabular labrum (Fig. 4). A recent multicentre study assessed the status of the labrum in 553 hips undergoing PAO for DDH with concomitant arthrotomy or arthroscopy . Fifty-five percent had a hypertrophic, hypoplastic, or ossified labrum. The prevalence of labral hypertrophy was inversely related to the ACE and LCE. Additionally, 64% of all hips presented with labral tears with the majority being degenerative tears (52%) or detachment (39%) according to the grading system by Beck et al. , and only 7.2% complete labral avulsions . Leunig et al.  compared MRI of 14 dysplastic hips and 14 hips with FAI. Of the dysplastic hips, 86% revealed labral hypertrophy while none of the hips with FAI had signs of labral hypertrophy. Thus, the authors concluded that labral hypertrophy is a good predictor for the presence of dysplasia. Another recent study by Garabekyan et al.  confirmed these findings. Stratified by the degree of the LCE, the authors found increased labral length measured on MRI in hips with frank dysplasia (LCE<20) and borderline dysplasia (LCE <25° and>20°) compared with normal hips and hips with pincer morphology . The authors concluded that labral length may represent an adaptive change in response to instability and serves as a guide for clinical decision making in hips with borderline DDH.
Advanced OA has been reported as an independent risk factor for a poor outcome after PAO [17, 32, 33, 77]. MRI is useful to assess the chondral status. In particular, nonfat-suppressed fast spin echo sequences are often preferred as they provide excellent contrast between bone, cartilage, and joint fluid . Cartilage damage includes full and partial thickness defects. Serial MRI assists in detecting progressive degeneration with chondral hyperintensity and loss of normal grey scale stratification as early signs of degeneration. Often, surface fibrillation is noted prior to more substantial lesions .
Biochemical MRI allows to detect joint degeneration at an early stage. As such, delayed gadolinium-enhanced MRI of cartilage (dGEMRIC) is increasingly implemented in the clinical routine (Fig. 5). Loss of glucosaminoglycan (GAG) content in the cartilage is associated with an early change of degenerative joint disease that appears prior to structural damages detected on MRI or plain radiographs . dGEMRIC depicts reduced GAG represented by reduced T1 relaxation times. A previous study showed that a low dGEMRIC index (defined as the average T1 time measured on four femoral and acetabular coronal slices in the weight bearing zone) predicts premature failure after PAO . Regional variations in GAG content in dysplastic hips  revealed a higher failure rate after PAO with a low dGEMRIC index in the anterior portion of the joint . Additionally, it was shown that GAG content changes before and after PAO, an observation that was attributed to changed mechanical loading properties of the cartilage . Alongside dGEMRIC, T2-mapping has been introduced as further biochemical method for identifying cartilage damage based on its composition. Decreased water content and loss of collagen fibre orientation in the cartilage, both of which is found as a sign of early cartilage degeneration, results in increased T2 relaxation times [82, 83]. Hamada etal.  compared the 3D distribution of acetabular cartilage T2 mapping in nine symptomatic patients with DDH (mean LCE=5°) without radiographic signs of OA to those from asymptomatic volunteers. The authors found increased T2 values at the lateral zones of the acetabulum, with even further increases in the presence of an adjacent labral tear. Nishii et al.  investigated the change of cartilage T2 values with loading in patients with hip dysplasia. The decrease in cartilage T2 values was significantly greater in patients with DDH compared with asymptomatic volunteers.
MRI also allows assessment of the periarticular muscle mantle. It was previously shown that the iliocapsularis muscle, a small muscle overlying the anterior hip capsule, is hypertrophied in dysplastic hips. It was therefore suggested as an important dynamic stabilizer in the dysplastic hip . Furthermore, an increased iliocapsularis-to-rectus-femoris ratio has been described as a secondary feature in DDH that can lead clinical decision making in hips with borderline hip dysplasia and a concomitant cam-type deformity to identify the predominant pathology . It was shown that iliocapsularis-to-rectus-femoris ratio of 1:1 had a positive predictive value to distinguish dysplasia from protrusion of 89% for the diameter, 77% for the circumference, and 82% for the thickness of each muscle .
In the early years of PAO, conventional radiographs were based on analogue films. Radiographic parameters around the hip had to be obtained using rulers and markers. The introduction of digital radiography has simplified this evaluation. Furthermore, modern software applications were developed allowing to correct for pelvic tilt and rotation [52, 57, 88]. A validated computer analysis model (Hip2Norm) exists that allows assessing acetabular parameters associated with DDH on AP pelvis radiographs [52, 57]. Selected landmarks on 2D radiographic images such as the femoral head and acetabular centre, and the anterior and posterior wall contours are manually defined. The software reconstructs these landmarks in a 3D coordinate system. Pelvic malorientation can be calibrated back to neutral. Pelvic tilt is adjusted by the pelvic inclination angle on a true lateral pelvic radiograph [89–91]. If no true lateral pelvic radiograph is available, the individual pelvic tilt can be estimated with the vertical distance between the sacrococcygeal joint and the upper border of the symphysis using standard values of this vertical distance from a normal population. This distance was shown to be 47mm in females and 31.5mm in males [56, 57].
Advances of multiplanar imaging modalities have opened a new field of 3D planning methods for PAO. As far back as 1988, Klaue et al.  introduced a computed tomography(CT)-based method for evaluating femoral coverage and joint congruency in hips undergoing PAO. This method was based on outlining the contours of the femoral head, the articulating surface area of the acetabulum and the acetabular cartilage on axial CT slices. The resulting lines at each assessed level were superimposed, creating a two dimensional topographical map defining the coverage of the femoral head. The authors also introduced rudimentary virtual reorientation manoeuvres to estimate the improvement of femoral head coverage after PAO . Millis and Murphy reconstructed 3D CT scans and quantified femoral head containment by measuring classic geometric parameters in three dimensions . Janzen et al.  measured centre edge angles on 3D CT reconstructed rotational sequences as reference values for preoperative planning for adult acetabular dysplasia. Dandachli et al.  introduced a CT based method to assess femoral head coverage, acetabular inclination and anteversion normalized to the anterior pelvic plane. With increasing technologic advances, an increasing number of modern image segmentation applications have become available for 3D preoperative planning. One of these applications utilizes 3D models derived from CT data offering a diagnosis module for evaluating hip joint morphology including classic parameters (Table I), femoral head coverage and simulation of the reorientation [95–99]. The software algorithm is based on a feature for automated acetabular rim detection , definition of centre of rotation of the joint assuming equidistant cartilage thickness  and calculating the acetabular opening plane in relation to the anterior pelvic plane . Virtual reorientation is achieved by excising a sphere over the acetabulum as a simplification of the actual osteotomies [100, 101]. The operator virtually performs stepwise increments of acetabular reorientation (flexion/extension, rotation and abduction/adduction; Fig. 6). The numeric change of the conventional radiographic parameters (Table I) and femoral head coverage is displayed in real time. After the desired degree of reorientation is achieved, range of motion analyses can be performed to recognize potential impingement conflicts [32, 64, 102, 103]. An alternative 3D planning tool not depending on proprietary third-party software is the Move Forward 3D Motion Simulation (Clinical Graphics; Zimmer Biomet) . The software is based on CT or MR data and allows for 3D motion simulation and planning of the osteotomies. The software was initially developed for femoracetabular impingement motion analysis and collision detection of the osseous structures, however, a planning tool for PAO has recently been added as well. The segmentation process, editing of the 3D bone models, and motion analysis is exclusively executed by the company. An interactive PDF report is thereafter provided to the clinician .
Another approach to assess the effect of acetabular reorientation is based on methods assessing joint biomechanics. The commonality of these methods is that they rely on the estimation of peak contact pressures and contact areas. Different approaches are nowadays applied, such as discrete element analysis  or finite element analysis [106–109]. Potential benefits include the possibility of preoperative planning, visual feedback and intraoperative navigation . One of the first studies was published by Hipp et al.  who reported a significant decrease of joint contact pressures after virtual reorientation of the acetabulum. Mechlenburg et al.  measured the area of load bearing before and after PAO based on a stereologic method. An improvement of the load bearing area was observed after PAO comparable to normal hips . The research group around Armiger and Lepistö released a series of papers about a biomechanical guiding system (BGS) based on CT data for PAO [105, 112–115]. BGS combines preoperative planning with intraoperative fragment tracking, assessment of geometric parameters and real time biomechanics. In their most recent publication, the authors compared anatomic and biomechanical measurements obtained from BGS to manual, intraoperative measurements and measurements derived from postoperative CT in 12 PAO. BGS measurements of acetabular positioning were highly concordant with postoperative CT measurements while manual measurements during surgery revealed less accuracy .
Redirective, PAO render the potential for successful treatment of DDH in skeletally mature and immature patients. The common goal of these procedures is to relieve pain and to delay or prevent progression of OA to preserve the natural hip joint over time. Many factors associated with a good long-term result have been attributed to the accuracy of 3D re-orientation of the acetabulum . As dysplastic acetabula exhibit an altered shape with a shallow socket and a decreased articular surface area, femoral coverage can only be restored to the best possible degree. Recognizing the patient-specific pathoanatomy requires thorough preoperative planning to achieve the best possible reorientation depending on the individual anatomic restraints. Next to clinical examination, routine preoperative planning consists of conventional radiography to identify specific acetabular and femoral radiographic parameters, MRI/MR-arthrogram to assess chondral and labral disease, and the status of the periarticular soft-tissues. CT may be performed to better assess the 3D osseous anatomy. Furthermore, modern computer-assisted methodologies allow for a dynamic assessment of the biomechanics and treatment simulation. However, implementation of these modern techniques in the daily routine requires further development prior to their widespread clinical application.
No funding was received for this article.