➢ A broad spectrum of hip pathological conditions, including ischemic, traumatic, inflammatory, arthritic, and neoplastic etiologies, can be detected by magnetic resonance imaging (MRI).
➢ Alteration of fluid content in tissue can be visualized by fluid-sensitive sequences such as T2, proton density, and short tau inversion recovery (STIR).
➢ Anatomy, joint alignment, marrow abnormalities, and fractures are best analyzed on a T1 sequence.
➢ Contrast agent improves the visual discrimination between the joint fluid and the capsular, labral, and osteochondral structures and shows a distension effect when applied in an intra-articular fashion.
One of the key reasons for the tremendous growth and interest in the operative management of the young adult with hip pain1,2 is advancements in high-resolution magnetic resonance imaging (MRI)3,4. Although radiographs provide critical information with regard to the osseous architecture and should always remain the first line of investigation, they are limited in their capacity to provide a detailed analysis of the other key anatomical components. MRI is the modality that provides the most comprehensive imaging of the hip joint, allowing for visualization of anatomy and detection of pathological conditions of the various structures, including the labrum, articular cartilage, synovium, and bone. The investigations of labral pathological conditions, femoroacetabular impingement, and early hip arthritis areas are studies in which MRI excels over all other imaging modalities. In this article, we review the basic principles of MRI in the evaluation of the native hip joint and the most common pathological conditions of the hip with their characteristic MRI findings.
In clinical practice, 1.5 or 3-Tesla (T) field strength MRI scanners are the most commonly used. Higher field strengths can be favorable as they provide higher signal-to-noise and contrast-to-noise ratios5,6.
MRI has multiplanar image acquisition capability. It is standard for protocols to be multiplanar with at least one sequence in each plane5. Images in the standard axial, sagittal, coronal, or any oblique planes can be directly acquired or can be secondarily constructed with use of three-dimensional volume, isotropic voxel acquisitions with postprocessing reformation5. Most commonly, three oblique planes, relative to the acetabulum or femur, are acquired. From true axial plane images, the oblique coronal and oblique sagittal planes are prescribed, respectively, perpendicular and parallel to the acetabular fossa4. The oblique axial plane is most commonly prescribed from the oblique coronal plane images, parallel to the femoral neck long axis. Radial plane imaging has also become common, allowing for circumferential assessment of the femoral head-neck junction5,7-12.
Performing MRI with the intra-articular application of gadolinium contrast agent, direct magnetic resonance arthrography (MRA) accentuates the joint space and facilitates evaluation of the labrum, hyaline cartilage, and the ligamentum teres by improving the contrast resolution13-16. MRA also has the advantage of the distension effect, which further separates the labral, osteochondral, and capsular structures, improving the spatial resolution. Indirect MRA involves the intravenous injection of gadolinium and is thus less invasive than direct MRA. However, it requires a delay of thirty to ninety minutes between injection and MRI acquisition to allow for diffusion of the contrast agent into the joint space, thus improving contrast resolution15. Additionally, in indirect MRA, normal and pathological extra-articular soft tissues may also be enhanced, making the intra-articular abnormalities less conspicuous5.
There are various classes of MRI sequences that can be utilized in different combinations to comprehensively acquire images of the hip. Structures seen on MRI can be characterized by their signal intensity, with low signal intensity referring to dark or blacker, high signal intensity referring to bright or whiter, and intermediate signal intensity referring to gray. Each structure has a characteristic signal profile that is variable depending on the sequence used:
• Fluid: There is a low signal intensity on T1 sequences and a high signal intensity on fluid-sensitive sequences (T2, proton density, short tau inversion recovery [STIR]), accentuated when fat suppression is applied.
• Fat: There is a high signal intensity on T1, proton density, and T2 sequences and a low signal intensity when fat suppression is applied.
• Bone: Compact bone and trabeculae have a diffusely low signal intensity across all sequence classes.
• Bone marrow: The signal intensity varies depending on the balance between yellow and red marrow; the yellow marrow signal follows that of fat, and the red marrow, with its greater cellularity and fluid content, has a signal intensity in between the signal intensities of fat and fluid.
• Hyaline cartilage: Typically, there is an intermediate signal intensity across all sequence classes.
• Fibrocartilage, ligaments, or capsule: There is a diffusely low signal intensity across all sequence classes.
A standard protocol for routine, non-contrast MRI of the hip joint should include a T1-weighted sequence without fat suppression for assessment of the anatomy, joint alignment, fractures, and bone marrow abnormalities. Sequences that are sensitive to fluid (T2, proton density with fat suppression, or STIR) allow for the detection of abnormal edema within the soft tissues and bone and abnormal fluid-dominant structures such as cysts or regional fluid collection5. Fluid-sensitive sequences additionally improve the contrast between joint fluid and adjacent structures5. High-resolution MRI without a contrast agent has been shown to be adequate for the assessment of labral and chondral abnormalities17,18. For contrast MRA, T1-weighted sequences with or without fat suppression should be employed, in addition to fluid-sensitive sequences. With indirect MRA, fat suppression is strongly recommended to improve the spatial resolution5.
Anatomic Structures Visualized on MRI
The normal labrum typically manifests as a well-defined, sharply marginated, triangular low-signal-intensity structure on all sequence classes5. However, a rounded or flattened shape is not uncommon19-21. The labrum is fixed to the cartilage structure and the acetabulum structure, with the interfaces referred to as the chondrolabral junction and the acetabular-labral junction, respectively5,22. MRA is the standard method for labrum assessment with high sensitivity (69% to 100%) and specificity (64% to 96%) for the detection of labral tears23-29. Non-contrast MRI with optimized protocols can reach a sensitivity of 96% to 97%, a specificity of 33%, and an accuracy of 93% to 95% for the detection of labral tears17. When degenerate, the labrum may be enlarged or globular or may demonstrate a higher internal signal intensity and irregularity along its surface5,30.
Labral tears are demonstrated by contrast solution or fluid signal intensity within the substance of the labrum extending to the capsular or articular surface and are most commonly seen in the anterosuperior quadrant18 (Fig. 1). It should be noted that labral tears may not always manifest with fluid-type, high signal intensity. This is due to the fact the tear defect can contain synovitis or granulation tissue. As well, volume-averaging can occur between the very fine tear cleft (high signal) and the adjacent, more normal labral tissue (low signal). The resultant tear may appear to be of lower signal intensity, less than that of the joint fluid23. Several normal anatomic structures and signal patterns may mimic a labral tear. These include an increased signal at the chondrolabral junction, a cleft between the transverse ligament and the anteroinferior labrum, a posteroinferior sublabral sulcus, and cartilage undercutting of the labrum5,17,24,31,32. Labral detachments result in accumulation of contrast at the base of the labrum at the chondrolabral or acetabulolabral junctions. Labral tears or detachments may result in the formation of adjacent fluid-filled structures referred to as paralabral cysts5. Although they may theoretically fill with gadolinium, most will not and thus are most conspicuous on fluid-sensitive sequences33.
MRI17,18 and MRA13,30,34-36 are standard modalities for the evaluation of articular cartilage status. Standard sequences include T1 and proton density, with or without fat suppression. To detect cartilage damage in the hip, the sensitivity should be between 64% and 93% and specificity should be between 63% and 93%17,18,30.
Normal articular cartilage typically has an intermediate signal intensity on routine MRI sequences and has a slightly higher signal intensity when fat suppression is applied37. There can be signal heterogeneity (focal, laminar, or striated) within the substance of normal cartilage related to variations in collagen fiber orientation38.
Chondral damage can manifest with changes in signal intensity or morphology. In the earliest stages of degeneration, there will be increased water manifested by increased signal intensity on fluid-sensitive sequences. Morphological change includes alteration in contour with surface fibrillation and fissuring or in thickness, either partial or full29,38. In addition, chondropathy can be focal or diffuse. On MRI, areas with chondral loss will be identified by fluid filling the clefts or areas of defect and are typically best appreciated on fluid-sensitive sequences such as proton density or T2, with fat suppression applied.
Focal detachment of the cartilage from the subchondral bone is referred to as delamination. Delamination is seen on MRI as a focal change of signal intensity within the cartilage in most sequences as a hypointense line paralleling the articular surface7,35,39. Delamination in combination with a fissure or a defect forms a chondral flap and is seen on MRI as an interposition of fluid between the cartilage and the acetabular bone35,39-42 (Fig. 2). In MRA, sensitivity has been shown to be between 22% and 97% and specificity has been shown to be between 84% and 100%39-41. In a study published with a low sensitivity of 22% (six of twenty-seven hips), two-thirds of the delaminations were seen retrospectively41; therefore, MRA of hips with suspected delamination has to be analyzed very carefully. Subchondral bone changes due to cartilage damage can occur, including marrow edema or cysts with increased fluid signal and fibrosis or trabecular sclerosis, which manifest with reduced signal intensity38.
Recently, more advanced MRI protocols have been introduced that allow for interrogation of the hyaline cartilage at the biochemical level. They allow for detection of the microscopic changes in the extracellular matrix. Early degenerative changes of cartilage result in a decreased organization of collagen, a loss of proteoglycan, and an increase of water content43,44.
Delayed gadolinium-enhanced magnetic resonance imaging of cartilage (dGEMRIC) uses the indirect MRA technique to quantify the concentration of proteoglycan45,46. T1ρ is a sequence that is also sensitive to proteoglycan predominantly but does not require any contrast injection47,48. T2 mapping is another non-contrast technique but is more sensitive to collagen orientation and integrity49-51. Currently, these techniques are not standard in clinical use.
The synovial membrane is a thin layer between the joint capsule and the joint cavity. In a healthy hip, the synovium is too thin to be directly visualized by MRI as it blends in with the capsule. An effusion manifests as a distension of the capsule by the increased fluid within the joint. According to Mitchell et al., grade-1 joint effusion is defined as an asymmetric finding of fluid within the hip joint, grade-2 joint effusion is defined as fluid surrounding the femoral neck, and grade-3 joint effusion is defined as when the capsular recesses are distended52. The signal of joint fluid follows that of water: a low signal intensity on T1 sequences and a high signal intensity on fluid-sensitive sequences. However, complex effusions related to an inflammatory exudate, infection, or hemorrhage may have a slightly higher signal intensity on T1 sequences and a slightly lower signal intensity on the fluid-sensitive sequences.
Synovitis reflects inflammation in the joint and is manifested by diffuse or focal thickening of the synovial membrane. The thickening can be smooth or nodular, and there may be fronds or septations that can be seen extending into the joint fluid. Synovitis is most accurately detected with use of the T1 sequence with fat suppression after the administration of gadolinium53,54.
This is a benign metaplasia of the synovium, resulting in variably calcified cartilaginous nodules. These nodules can separate from the synovium and can become free bodies in the joint space55. Even when pathological findings of synovial chondromatosis sometimes can be detected in a conventional radiograph, MRI is the gold standard for diagnosis and staging. Findings on MRI include multiple lobulated, fairly uniformly sized, intra-articular structures of variable signal intensity bathed by a fluid signal56 (Fig. 3) and, in long-standing cases, synovial thickening and bone remodeling and erosion57. Extension was visible in the iliopsoas in 40% of the hips (six of fifteen patients) and in the obturator externus bursae in 71% of the hips (five of seven patients)57,58.
Pigmented Villonodular Synovitis
This rare disease is a benign proliferation of the synovia with hemosiderin pigmentation59. Diffuse or nodular synovial thickening with a low signal intensity in T2 sequences and a heterogeneous low to intermediate signal intensity in T1 sequences can be found60. The hemosiderin leads to a pathognomonic blooming artifact (enlargement of the area with low signal intensity) in gradient-echo sequences59,60. Frequent additional MRI features are osseous erosions and subchondral cysts in the acetabulum or the femoral head61.
Undisplaced hip fractures (occult fractures) can be difficult to diagnose on conventional radiographs. Delayed diagnosis increases the risk for adverse events such as dislocation, nonunion, osteonecrosis, the need for a prosthetic replacement, and increased health costs and mortality62-64. In 1% to 10% of all hip fractures, further diagnostic imaging is required to confirm or rule out a fracture (1% in a study with 590 patients, 2.5% in a study with 1353 patients, and 10% in a study with 235 patients)65-67. MRI is recommended for all patients with severe hip pain after trauma and normal conventional radiographs68-72. A T1 sequence has been recommended to detect fractures combined with additional T2 or STIR sequences to assess the surrounding bone marrow and soft tissues64. The fracture manifests as appearing linear or band-like, with a low T1 signal intensity focus and a high surrounding signal intensity on fluid-sensitive sequences73. There is a high sensitivity in both MRI (99% to 100%) and bone scintigraphy (94% to 98%), but there is a higher specificity in MRI (98% to 100%) compared with bone scintigraphy (94% to 95%); MRI also has a higher spatial resolution74-77. Computed tomography (CT) is not yet well studied as a diagnostic tool for occult fracture66, as a study has showed low sensitivity (83%)67.
An insufficiency fracture results from the repetitive stress of everyday activities and can be found in the femoral neck and supra-acetabular region. It is described as a process of repetitive bone remodeling under submaximal stress; fracture is the end stage78. Conventional radiography is often insufficient for accurate diagnosis, with a sensitivity of only 10% to 15% at initial presentation, rising to 30% to 70% after three weeks75,79. MRI is the imaging modality of choice, with a sensitivity of 99% to 100% and a specificity of 86%75,80. Insufficiency fracture manifests on MRI as a line of low signal intensity in T1 sequences, surrounded by an area of high signal intensity in T2 sequences, indicating bone marrow or soft-tissue edema81. Compared with nuclear scintigraphy, which also has a high sensitivity but a lower specificity, MRI can additionally characterize the fracture in terms of orientation, exact location, and extent of the fracture line and periosteal edema82. Another advantage of MRI is the ability to distinguish between abnormal fracture and insufficiency fracture83. Sensitivity in CT (69%) is substantially lower than that in MRI (99%)75.
Osteonecrosis of the femoral head can result in pain, collapse of the articular surface, or predisposition to osteoarthritis. Osteonecrosis can be detected in the very early stages with MRI, manifesting as subchondral edema84. MRI is the modality of choice for this application, with a sensitivity of 97% and a specificity of 98%85.
Eventually, a well-defined area forms, with a thin, peripheral low signal and a linear boundary on T1 images, with both internal and external marrow edema. In a later stage of osteonecrosis, after the start of osseous repair, the double-line sign can be detected86. This describes a rim of low signal intensity around an area of high signal intensity in T1 and T2 sequences. In the end stage, osseous collapse and sclerosis can be detected as a low T1 signal intensity and a variable T2 signal intensity85.
Transient Osteoporosis of the Hip
This rare, self-limiting disease is characterized by pain associated with loss of bone and occurs most often in the hip in middle-aged men87,88. MRI demonstrates diffuse bone marrow edema involving the femoral head, starting in the subchondral region and extending distally across the femoral neck region89 (Fig. 5). Transient osteoporosis of the hip does not progress to osteonecrosis but resolves within four to ten months.
Bone Marrow Edema
Bone marrow edema, or bone marrow lesion, is purely a description of an MRI pattern. The pattern is equal to an increase of fluid in the tissue, but not all of the underlying pathological conditions correspond with an increase in fluid content90. Bone marrow edema is defined as a diffuse area in bone marrow with a decrease of signal intensity in a T1 sequence and an increase of signal intensity in T2 and STIR sequences84,87. The intravenous application of a contrast agent leads to an increase of T1 signal intensity, indicating hypervascularity91. Bone marrow edema occurs in various diseases such as osteoarthritis, infection, osteonecrosis, neoplasm, sickle cell anemia, and osteoporosis, as well as after trauma and as an isolated finding (transient bone marrow edema syndrome)84,87,92-95. No specific histological association with bone marrow edema could be found96. The progression of bone marrow edema corresponds with cartilage loss97.
Infection of the hip may occur in the joint space (septic arthritis) or in the marrow space (osteomyelitis). Septic arthritis is manifested by joint effusion, cartilage and bone erosion, subchondral marrow edema on both bone surfaces of the articulation, and periarticular soft-tissue edema (Fig. 6). Osteomyelitis demonstrates areas of intense marrow edema, variable periosteal and soft-tissue edema, bone destruction and resorption, cloaca, or sinus tract98. In both forms of infection, the administration of intravenous gadolinium contrast agent will make areas of active synovial, marrow, and soft-tissue inflammation more conspicuous on T1 sequences with fat suppression98,99. To detect osteomyelitis on MRI, the sensitivity is 90% to 98% and the specificity is 79% to 89%99,100.
Femoroacetabular impingement is a common pathomorphology of the hip and is described as a cause for primary hip osteoarthritis8,42,101. Acetabular overcoverage (pincer type) or femoral-head asphericity (cam type) leads to pathological interaction between the acetabulum and the femoral neck during motion and subsequently to cartilage damage and osteoarthritis42.
The shape of the acetabulum can be described by its coverage, depth, and orientation. The overcoverage found in pincer-type femoroacetabular impingement can be due to acetabular retroversion102, global overcoverage (a center edge angle of >40°)103, or protrusio (femoral head touching or crossing ilioischial line)42. In the normal hip, the acetabular fossa opening is directed anteriorly and thus is anteverted. To quantify the acetabular version, a transverse plane of the pelvis including both hips on MRI or CT is required. Tönnis and Heinecke and Reynolds et al. defined acetabular version on CT as the angle between a sagittal plane through the midfemoral head at the level of the maximal diameter and the line connecting the outer rim of the acetabulum (equatorial edge line)104,105. In hips with a retroverted acetabulum, this angle will be negative; a positive angle indicates an acetabular anteversion. A normal acetabulum is anteverted by 20° to 23°, with a range from 15° to 25°105-107. This method is validated for MRI, showing similar results by measuring the bone or labral rim of the acetabulum108 (Fig. 7).
In cam-type femoroacetabular impingement, bone and/or cartilage excess located on the femoral head-neck junction leads to decreased femoral head-neck offset9,12,42,109. The cam deformity is quantified by the alpha angle13,110. The alpha angle is traditionally measured on an oblique axial slice and is defined as the angle between a line through the center of the narrowest part of the femoral neck and the center of the femoral head and another line through the center of the femoral head and the anterior beginning of the femoral head asphericity110. To evaluate the entire femoral head-neck junction circumferentially, radial imaging can be used. Depending on where the alpha angle is measured on the femoral head-neck junction, different threshold values have to be used12. On the basis of normative data and comparisons with symptomatic individuals with cam femoroacetabular impingement, a value of >50° at the 3 o’clock position and a value of >60° at the 1:30 position are considered26,111 (Fig. 8).
Developmental dysplasia of the hip leads to early osteoarthritis due to an undercoverage of the acetabulum and subsequently abnormal chronic shear and increased load. On MRI, associated abnormalities such as labral hypertrophy, degeneration, or tear and cartilage degeneration can be detected112 (Fig. 1).
Source of Funding: There was no source of external funding for this study.
Investigation performed at The Ottawa Hospital, Ottawa, Ontario, Canada
Disclosure: None of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of any aspect of this work. None of the authors, or their institution(s), have had any financial relationship, in the thirty-six months prior to submission of this work, with any entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. Also, no author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.
- Copyright © 2015 by The Journal of Bone and Joint Surgery, Incorporated