Keywords
femoroacetabular impingement - CAM - pincer - magnetic resonance arthrography - diagnostic
accuracy
Introduction
Femoroacetabular impingement (FAI) is today considered one of the main cause of symptomatic
early hip arthritis in young patients.[1] FAI is induced by an incongruence between the femoral head and the acetabulum due
to an alteration of the head–neck junction (“CAM”), of the acetabular rim (“pincer”),
or a mix of both conditions.[2]
[3] Symptoms are represented by groin pain, irradiated usually to the thigh and buttock,
and often associated to clicks during hip rotations. Range of motion is usually limited
by the early contact between these bone alterations and forced passive motion activities
are symptomatic.
Some clinical tests (Flexion-Adduction-Internal Rotation (FADDIR), Flexion, ABduction,
and External Rotation (FABER) Ribas, and posterior impingement) are useful to have
a suspect, but the confirmation with imaging studies is crucial.[1]
[4]
[5] Conventional radiology is essential for the diagnosis of FAI by the use of various
specific projections. “True” anterior–posterior, Dunn at 45 and 90 degrees, frog leg,
and Lequesne “false profile” are some of the most used radiologic views.[6]
[7] Common findings evaluable by radiology are the “pistol grip” deformity, the crossover
sign, the over-coverage of the acetabular rim (“eight sign”), and the alterations
of the α and the center-edge angles.[8]
[9] Such findings are usually sufficient to plan surgical treatment of the various deformity.
However, alterations of intra-articular structures are not detected by standard radiology
and may need further treatment during surgery.
Hip arthroscopy is considered the gold standard of treatment for FAI; it is associated
with good outcomes and with lesser morbidity with respect to open surgical dislocation,
originally described as the most indicated procedure.[10]
[11]
Magnetic resonance imaging (MRI) and computed tomography (CT) are useful to add information
about other intra-articular lesions and to better visualize the profile of bones,
respectively.[7]
[12]
[13] Direct 3-T MRI may be useful alone in the detection of soft tissue alterations,
but it is not diffusely available.[14] The use of MR with intra-articular injection of gadolinium contrast (MR arthrography
[MRA]) is a valid diagnostic tool to evaluate cartilage, synovial membrane, labrum,
and ligamentum teres (LT). However, diagnostic accuracy of this exam is still under
debate.[15]
[16]
[17]
The aim of this study was to assess the diagnostic accuracy of MRA in the detection
of intra-articular lesions of the hip in patients affected by FAI by using arthroscopy
as reference standard. The hypothesis of the study was that MRA is a valid diagnostic
exam to detect intra-articular lesions in hips affected by FAI.
Methods
Twenty-nine consecutive hip arthroscopies performed between June 2012 and April 2016
in 24 patients affected by symptomatic FAI (five patients underwent a bilateral-staged
procedure) were considered for this study. The Institutional Review Board approved
the study and all patients were adequately informed about the diagnostic procedure,
treatment, and follow-up. Inclusion criteria were symptomatic FAI with persistent
pain and functional limitation without any other source of groin or irradiated pain.
Exclusion criteria were patients with advanced hip osteoarthritis (OA), and other
sources of groin or irradiated pain (discal hernia, inguinal disruption, muscular
tears). Eighteen patients were males and 6 females. The mean age was 38.3 years (range,
18–59 years), the mean body mass index was 20.4 (range, 20–22).
Each patient underwent a specific diagnostic protocol. The medical history was focused
on the acquisition of the gait ability during childhood, the results of the early
ultrasound screening performed few months after birth, and the presence of clicks
in the hip. The radiologic study consisted in the true anteroposterior standing, the
bilateral false profile, and the 45- and 90-degree Dunn views of both hips.[18]
[19]
All patients underwent an MRA with a 1.5-T machine (Siemens Medical Solution, MR Area,
Erlangen, Germany). No patients referred allergy or contraindications to arthrography.
All procedures were performed by the same radiologist (G.C.) and consisted of local
disinfection and anesthesia with 1% lidocaine followed by the intra-articular injection
of diluted gadolinium through an anterior ultrasound-guided portal using a probe cover
and sterile gel. The following sequences of the preliminary direct MRI were obtained:
short tau inversion recovery, T1-weighted and PD coronal, T1-weighted, and T2-weighted
axial with 3-mm-thick slices. After contrast injection, further sequences were obtained:
T1-weighted fat saturation MRA (Fat-SAT) axial, coronal and oblique sagittal, T1-weighted
Vibe 3D coronal sequences with MPR sagittal, axial and radial reconstructions with
2-mm-thick slice, and coronal density protonil (DP) Fat-SAT. The purpose of such sequences
was to evaluate any tissue in the joint (cartilage, labrum, capsula, synovial membrane
and its folds, LT, and bones). Any pincer lesion was evaluated on precontrast axial
oblique PD-weighted MR sequences.
All arthroscopies were performed by the same surgeon (C.C.) with patients on lateral
position on a standard surgical bed with a dedicated hip traction system and by the
use of an imaging intensifier. The first surgical step consisted of the evaluation
of the central compartment under traction. After traction release, the second step
consisted of peripheral compartment evaluation.
Radiological and arthroscopic evaluation were performed independently and blinded
each other. Acetabular and femoral chondral lesions were classified according to the
Outerbridge classification[20] and staged as low grade (grades 0–2) and high grade (grades 3 and 4). Other intra-articular
findings were labral degeneration, labral tears, synovitis, LT tears, CAM lesions,
pincer lesions, loose bodies, and osteophytes. All these findings were handled as
dichotomous variables and staged as “yes or no.”
All patients' data were recorded in a custom-made database. The statistical analysis
was performed using SPSS software v. 11.0 (SPPS Inc, Chicago, Illinois, United States).
Absolute per cent agreement, sensitivity, specificity, positive predictive value (PPV),
and negative predictive value (NPV) of MRA were calculated by the 2 × 2 table method
and by assuming the arthroscopic assessment as reference standard.
Results
Absolute per cent agreement between MRA and arthroscopy was observed in 79.3% of the
cases for acetabular chondral lesions, 65.5% for femoral chondral lesions, 89.6% for
labral degeneration, 79.3% for labral tears, 75.8% for synovitis, 89.6% for LT tears,
89.6% for loose bodies, and 82.7% for osteophytes. A full agreement (100%) was observed
for CAM lesions, whereas agreement for pincer lesions correspondence was 68.9% ([Figs. 1]
[2]
[3]
[4]).
Fig. 1 T1 TSE axial, TR: 500, pelvis; TE: 22 milliseconds. (A) Bilateral bumps (red arrows) in a patient with bilateral CAM-type FAI with capsular
thickening, more prominent on the left side. (B) Intraoperative aspect of the same large bump of the left femoral head before distraction.
Fig. 2 MPR mdc VIBE axial, left hip; TR: 19, TE: 5.93 milliseconds. (A) Fissuring and fragmentation of the labrum at the inferior–medial position (straight
red arrow) and a bump on the femoral head/neck junction (curved red arrow) in a patient
with CAM-type FAI. (B) Intraoperative finding, demonstrating the alterations as observed at MRA.
Fig. 3 T1 TSE mdc FS, left hip; TR: 525, TE: 29 milliseconds. Fissuring of the superior–external
labrum (straight red arrow) with very small extracapsular leak of contrast agent (curved
red arrow) and intact periosteum in patient with CAM-type FAI. (A) These features are very suggestive of a labral tear. (B) Arthroscopy exactly confirmed the finding.
Fig. 4 MPR mdc VIBE axial, left hip; TR: 19, TE: 5.93 milliseconds. Signal intensity alteration
of the superior–external labrum without fissuring and/or detachment (straight red
arrow). (A) Early sign of osteoarthritis of the hip with osteophytes in the inferior–medial
position (curved red arrow). (B) During arthroscopy, there was a full correspondence of the lesion and its position.
Sensitivity was high for all the variables investigated, except for labral tears,
pincer lesions, and loose bodies. Specificity was high for labral tears, LT tears,
CAM lesions, pincer lesions, and loose bodies, while it was moderate for all the other
variables. PPV was high only for labral degeneration and CAM lesions, whereas it was
moderate to poor for all the other variables. NPV was high for all the variables,
except pincer lesions ([Table 1]).
Table 1
Sensitivity, specificity, PPV, and NPV are shown for all the tested pathologic findings
|
Chondral lesions
(A)
|
Chondral lesions
(FH)
|
Labral tears
|
Labral degeneration
|
LT tears
|
CAM lesions
|
Pincer lesions
|
Loose bodies
|
OP
|
|
Sensitivity (%)
|
100
|
100
|
33
|
95
|
100
|
100
|
33
|
50
|
100
|
|
Specificity (%)
|
68.4
|
50
|
85
|
71
|
88
|
100
|
85
|
96
|
73.9
|
|
PPV (%)
|
72.7
|
47.3
|
20
|
91.3
|
57.1
|
100
|
50
|
66.6
|
50
|
|
NPV (%)
|
100
|
100
|
91.6
|
83.3
|
100
|
100
|
73.9
|
92.3
|
100
|
Abbreviations: A, acetabulum; FH, femoral head; LT, ligamentum teres; OP, osteophytes;
PPV, positive predictive value; NPV, negative predictive value.
Discussion
FAI is more frequent than believed until years ago, and it represents a common source
of groin pain,[21] which is often (but not exclusively) related to bone alterations of acetabulum and
proximal femur. While bone alterations are easy to detect, soft tissue evaluation
is not simple. Standard radiology is dramatically important, but CT and MRI may also
be useful. Specifically, MRI usually provides more information on the status of soft
tissues than CT[13] without irradiation of the patients. As for other joints, the use of an intra-articular
contrast medium improves the diagnostic accuracy of MR.[14]
[15] At MRA, CAM impingement appears on coronal and axial images as a lacking offset
between femoral head and neck, with a focal osseous bump at their junction. FAI is
often associated to fibrocystic changes at the femoral head–neck junction that can
be early detected on MRA as small cysts of varying diameter.[22]
[23] Pincer impingement may be related to focal or global acetabular over-coverage. On
MRA, the pincer morphology can be evaluated on axial images drawing a line between
the lateral edges of the acetabulum. This method allows a better accuracy in the assessment
of acetabular alterations with respect to the crossover sign on standard X-rays; the
latter evaluation may often overrate an acetabular retroversion due to the tilt and
the inclination of the pelvis on the X-ray tube.[14]
Labral degeneration, labral lesions, and chondral lesions are generally well detected
on MRA.[24]
[25]
[26]
[27]
[28] In a meta-analysis conducted by Smith et al,[24] 16 studies on MRA of the hip were evaluated demonstrating sensitivity of 87% and
specificity of 64% in the detection of labral tears. Such value of specificity was
quite similar to that reported in this study, confirming its good capacity in excluding
labral tears. However, poor PPV in our study is probably related to the presence of
anatomical variants of the labrum that frequently occurs very close to the lesions,
particularly in the anterior–superior zone of the acetabulum.[25] Poor sensitivity observed was mainly due to limited number of labral lesions found
during arthroscopy.
There is not a univocal interpretation regarding the detection of chondral lesions.
Ho et al[26] considered MRA useful to detect high-grade chondral lesions, while Mintz et al also
reported a good evaluation of grade-2 lesions.[27] In other studies, the exact grade of chondral damage has not clearly specified,
thus making data interpretation rather difficult.[28]
[29]
In our study, MRA showed sensitivity and NPV of 100% in the evaluation of both acetabular
and femoral chondral lesions; therefore, the presence of low-grade chondral lesions
at MRA is an optimal predictor of the absence of high-grade chondral lesions during
arthroscopy. Conversely, MRA showed suboptimal specificity and PPV in the evaluation
of high-grade acetabular lesions and poor specificity and PPV in the evaluation of
femoral lesions. Indeed, in our study, preoperative MRA often overestimated chondral
lesions observed successively during arthroscopy, particularly on the femoral side.
The incidence and the severity of acetabular lesions was higher than femoral ones,
as Li et al previously observed.[14] Overall, the results of this study for chondral lesions are similar to the recent
literature regarding this topic.[14] Therefore, MRA may have an important role in excluding advanced acetabular and femoral
lesions.[30]
MRA may be useful in the evaluation of LT tears. The low specificity and PPV could
be due to the presence of intraligament degeneration and lesions that cannot be visualized
during arthroscopy. Moreover, DP sequences have been introduced in the MRA study after
the first three procedures, to improve its visualization. However, LT tears were identified
in 13.8% of cases; this prevalence was similar to those reported in other studies,
ranging from 4 to 51%.[31]
[32]
[33]
[34]
Suboptimal specificity of MRA in the assessment of acetabular and femoral osteophytes
was probably due to poor visualization of the inferior–medial region of the hip joint
during arthroscopy.
In our study, MRA depicted very well CAM lesions, whereas it showed suboptimal specificity
and NPV, and poor sensitivity and PPV in the assessment of pincer lesions.
The overall rationale of MRA of the hip in FAI would be to allow a better planning
of arthroscopic procedures. If X-rays and CT may be sufficient to plan the osteoplasty
of bone alterations, MRA may indicate where it would be necessary for labral repair
or debridement, when to perform microfractures for chondral lesions, or finally when
is indicated debridement of LT. MRA showed a good correlation with preoperative X-ray
findings. However, in patients with no radiologic signs of OA, MRA often revealed
high-grade acetabular and femoral chondral lesions. A correlation between MRA and
X-rays for pincer lesion was difficult due to the different method of evaluation (“eight
sign” on standing position vs. acetabular over-coverage on supine position). In our
study, X-ray evaluation showed a superior correlation with arthroscopy than with MRA.
This study has some limitations. First, it represents a preliminary study. The number
of patients is limited and no power analysis has been conducted. Although many results
agree with those reported in the literature, some contrasting findings might be related
to the lack of experience of radiologists and no reliability analysis was conducted
for all the investigated variables. Finally, the introduction of the DP sequences
after the first three exams could have jeopardized the results for the assessment
of LT tears.
In conclusion, the diagnosis of FAI is rather simple, but the detection of soft tissues
alterations is still today challenging for the orthopaedic surgeon. An integration
between clinical and radiological information is mandatory, particularly regarding
MRA, which is of paramount importance for the planning of the surgical procedure.
However, a certain disagreement exists between instrumental and arthroscopic findings,
which should be eliminated by a proper technical accuracy of imaging exams and an
improved knowledge of the pathology by all dedicated specialists.
