- Email: [email protected]
The diagnostic performance of non-contrast 3-Tesla magnetic resonance imaging (3-T MRI) versus 1.5-Tesla magnetic resonance arthrography (1.5-T MRA) in femoro-acetabular impingement
European Journal of Radiology 88 (2017) 109–116
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
The diagnostic performance of non-contrast 3-Tesla magnetic resonance imaging (3-T MRI) versus 1.5-Tesla magnetic resonance arthrography (1.5-T MRA) in femoro-acetabular impingement Ana M. Crespo-Rodríguez a,∗ , Jose C. De Lucas-Villarrubia b , Miguel Pastrana-Ledesma a , Ana Hualde-Juvera a , Santiago Méndez-Alonso a , Mario Padron c a
Department of Radiology, Hospital Universitario Puerta de Hierro Majadahonda, c/ Joaquín Rodrigo 2, Majadahonda 28222, Madrid, Spain Department of Orthopaedics and Traumatology at the Hospital Universitario Puerta de Hierro Majadahonda, c/ Joaquín Rodrigo 2, Majadahonda 28222, Madrid, Spain c Department of Radiology, Clínica Cemtro, Avda Ventisquero de la Condesa 42 Madrid 28035, Madrid, Spain b
a r t i c l e
i n f o
Article history: Received 13 September 2016 Received in revised form 17 November 2016 Accepted 26 December 2016 Keywords: Sensitivitity Specificity Arthrography Magnetic resonance imaging (MRI) Hip Femoro-acetabular impingement
a b s t r a c t Objective: The aim of this study was to evaluate the diagnostic accuracy of 3-T non-contrast MRI versus 1.5-T MRA for assessing labrum and articular cartilage lesions in patients with clinical suspicion of femoroacetabular impingement (FAI). Subjects and methods: Fifty patients (thirty men and twenty women, mean age 42.5 years) underwent 1.5-T MRA, 3-T MRI and arthroscopy on the same hip. An optimized high-resolution proton density spin echo pulse sequence was included in the 3-T non-contrast MRI protocol. Results: The 3-T non-contrast MRI identified forty-two of the forty-three arthroscopically proven tears at the labral-chondral transitional zone (sensitivity, 97.7%; specificity, 100%; positive predictive value (PPV), 100%; negative predictive value (NPV), 87.5%; accuracy 98%). With 1.5-T MRA, forty-four tears were diagnosed. However, there was one false positive (sensitivity, 100%; specificity, 85.7%; PPV, 97.7%; NPV, 100%; accuracy 98%). Agreement between arthroscopy and MRI, whether 3-T non-contrast MRI or 1.5-T MRA, as to the degree of chondral lesion in the acetabulum was reached in half of the patients and in the femur in 76% of patients. Conclusion: Non-invasive assessment of the hip is possible with 3-T MR magnet. 3-T non-contrast MRI could replace MRA as the workhorse technique for assessing hip internal damage. MRA would then be reserved for young adults with a strong clinical suspicion of FAI but normal findings on 3-T non-contrast MRI. When compared with 1.5-T MRA, optimized sequences with 3-T non-contrast MRI help in detecting normal variants and in diagnosing articular cartilage lesions. © 2017 Elsevier Ireland Ltd. All rights reserved.
1. Introduction
∗ Corresponding author. Present address: Department of Radiology, Hospital Clínico San Carlos, c/ Profesor Martín Lagos s/n, Madrid 28040, Madrid, Spain. E-mail addresses: [email protected] (A.M. Crespo-Rodríguez), [email protected] (J.C. De Lucas-Villarrubia), [email protected] (M. Pastrana-Ledesma), [email protected] (A. Hualde-Juvera), [email protected] (S. Méndez-Alonso), [email protected] (M. Padron). URLs: http://[email protected] (M. Pastrana-Ledesma), http://ana. [email protected] (A. Hualde-Juvera), http://[email protected] (S. Méndez-Alonso). http://dx.doi.org/10.1016/j.ejrad.2016.12.031 0720-048X/© 2017 Elsevier Ireland Ltd. All rights reserved.
Femoro-acetabular impingement (FAI) occurs when there is a conflict between the proximal femur and the acetabular rim [1]. Cam and pincer FAI can be differentiated on the basis of a predominance of either a femoral or an acetabular abnormality [2]. These changes result in repetitive microtrauma between the femur and the acetabular rim, leading to degeneration and tearing of the labral-chondral transitional zone (LCTZ), particularly in the anterosuperior region [3]. Lesions occur at the attachment of the labrum to the acetabulum, initially leaving the body of the labrum intact [4]. Beck et al. [5] and Schmid et al. [6] noted that in patients with FAI, adjacent acetabular cartilage damage was found in all patients with labral abnormalities. As a consequence, FAI has
110
A.M. Crespo-Rodríguez et al. / European Journal of Radiology 88 (2017) 109–116
become increasingly recognized as a potential cause of premature onset of osteoarthritis of the hip, especially in young (20–45 years) and active individuals [1,3]. FAI remains a controversial disease in terms of true incidence, clinical relevance, diagnosis and management [7]. Recently, an international consensus statement on FAI syndrome, the Warwick Agreement, has been published [8]. The syndrome has stimulated radiologists to re-evaluate the use of different hip MRI examinations, because of major improvements in resolution and technique in the past ten years. Several publications have advocated the use of direct MRA for the detection of labral tears because of its higher sensitivity as compared with non-contrast MRI at 1.5 T [9]. The enhanced signal-noise ratio (SNR), spatial resolution, and contrastto-noise ratio provided by the higher field strength would make 3-T MRI the ideal method for assessing articular cartilage and small structures such as the acetabular labrum [10]. The challenge for 3-T MRI is to reach as good a diagnostic performance as 1.5-T MRA in the assessment of acetabular labral abnormalities without the need for intra-articular contrast material and to improve the detection of cartilage abnormalities [10]. Besides image quality, several factors need to be considered when choosing between 3-T non-contrast MRI and 1.5-T MRA. These factors include the preferences of the radiologist and referring orthopaedic surgeon, patients’ reluctance to undergo an invasive procedure, risk, time, cost, labour, scheduling and coordination of either fluoroscopy or ecography and MR. The purpose of this study was to evaluate in retrospect the diagnostic performance of 3-T non-contrast MRI as compared with 1.5-TMRA in diagnosing LCTZ abnormalities in patients with FAI, surgical findings acting as the reference standard.
2. Material and methods 2.1. Subjects The retrospective study was performed with approval from our institutional review board, and with a waiver of informed consent. Criteria for inclusion were clinical suspicions of FAI, radiological findings of FAI (either cam or pincer) and 3-T non-contrast MRI, 1.5-T MRA and arthroscopy on the same hip. Criteria for exclusion criteria were previous surgery on the hip or other causes of groin pain ruled out on 3.0-T non-contrast MRI. The population studied consisted of consecutive series of patients with FAI who underwent hip arthroscopy between January 2012 and December 2014 in the Hospital Universitario Puerta de Hierro Majadahonda, Madrid. Imaging examinations were performed between October 2010 and February 2013. 3-T non-contrast MRI was performed before 1.5-T MRA, with a mean delay of three weeks. In this period of time 190 MRAs were done, most of them because of FAI.
2.2. Imaging examinations Non-contrast 3-T MRI was performed with a 3-T MR magnet (Achieva Dual Quasar 3-T Philips, Eindhoven, Netherlands) with the body phased array coil 6 Channel-Torso-coil. Our institution is a general hospital equipped with two 1.5-T and one 3-T MR magnets. Our standard hip protocol used to have similar imaging sequences in any magnet looking forward homogeneity. However, with the addition of a 3-T clinical MR magnet to our department, we took advantage of the increased SNR, spatial resolution, and contrastto-noise ratio afforded by higher field strength. Proton density (PD) weighted spin echo pulse sequences offer greater SNR than T1 or T2 [11]. A high acquisition matrix improves spatial resolution. In the particular case of the hip, we decided to introduce fat-suppressed
highresolution-PD (HR-PD) weighted spin echo pulse sequences into clinical practice (Table 1). Direct MRA of the hip was a two-step procedure. First, intraarticular injection was performed using an echogenic tip Chiba ® needle (Cook ) under ultrasound guidance. The 20cc of contrast ® solution contained 0.2cc of gadoterate meglumine (Dotarem ), 3cc ® of 2% mepivacaine and 17cc of sterile saline. The second step was to perform MR on patients, a 5 kg weight being used to apply traction, within thirty minutes of the injection of contrast medium. These MR examinations were performed with a 1.5-T MR magnet (Achieva Nova 1.5-T Philips, Eindhoven, Netherlands) with the cardiac body 5 Channel-Cardiac-coil. A three dimensional T1 weighted gradient echo sequence, specifically T1 High-Res Isotropic Volume Excitation (THRIVE) allowing reconstruction in any plane, was included in the protocol (Table 1). 2.3. Criteria for interpretations MR studies were analysed by three musculoskeletal radiologists with four (AHJ), six (AMCR), and twenty (MPL) years of experience, who were blinded to each others’ readings, of the original report and of the surgical results. The final interpretation was based on a majority consensus. Labral appearance was evaluated as normal or degenerative. Besides labral enlargement, intra-substance areas of intermediate signal intensity and irregular margins [12], intra-substance or flap tears were considered to be degenerative changes. They were characterized by high signal intensity on MRI or contrast medium in MRA being noted within the substance of the labrum extending to the articular surface. In LCTZ the evaluation included the assessment both of separation or tearing of the labrum and of adjacent cartilage damage. The presence of a tear at the LCTZ was defined as a cleft noted through the synovial fluid signal intensity on MRI [4] or contrast medium on MRA [12]. This chondro-labral separation has been termed a watershed lesion [3], a basal tear [10] or a longitudinal peripheral tear [13]. None classification system for hip labral tears was used as the Lage arthroscopic classification [13] does not correlate well with the Czerny MRA [14] or an MRA modification of the Lage classification [15]. The presence of a labral or paralabral cyst was also considered an indirect sign of tearing [16], similar to parameniscal cyst in the knee. The articular cartilage was graded on MRI by means of a modified version of Outerbridge’s classification system [17]. In the modified ranking, Grade 0 indicated intact cartilage. Grade 1 signified chondral softening or blistering with an intact surface, whilst Grade 2 indicated shallow superficial ulceration, fibrillation, or fissuring involving less than 50% of the depth of the articular surface. Grade 3 pointed to deep ulceration, fibrillation, fissuring or a chondral flap involving 50% or more of the depth of the articular cartilage without exposure of sub-chondral bone, whilst Grade 4 denoted full-thickness chondral wear with the exposure of sub-chondral bone. Apart from the assessment of femoral and acetabular articular surfaces, special attention was paid to cartilage within the LCTZ. 2.4. Arthroscopic evaluation Surgery was performed by an orthopaedic surgeon (JCDLV) with eight years of experience in hip arthroscopy and more than 300 procedures performed. This orthopaedic surgeon was aware of the MR imaging findings relating to patients at the time of arthroscopy. He provided dictated operative reports about all patients who underwent hip arthroscopy. Each report included an assessment of the integrity of the labrum, of the LCTZ and of articular carti-
A.M. Crespo-Rodríguez et al. / European Journal of Radiology 88 (2017) 109–116
111
Table 1 Protocol of 3-T non-contrast MRI and 1.5-T direct MR arthrography (MRA) of the hip. It is specified pulse sequence, plane, RT or repetition time (msec), ET or echo time (ms), matrix size, section thickness (mm), gap, echo train length, field of view (FOV), inversion time (IT), flip angle and fat suppression. Pulse sequences included are T1, 3D WATer Selective Fluid (3D-WATSf), STIR, Proton Density- High Resolution (PD-HR) and T1 High-Res Isotropic Volume Excitation (THRIVE). Fat suppression (FS) techniques detailed are SPectral Attenuated Inversion Recovery (SPAIR), Spectral Presaturation with Inversion Recovery (SPIR) and PRinciple Of Selective Excitation Technique (ProSet). Pulse sequence
Plane
RT
ET
Matrix size
Section thickness
Gap
Echo train length
FOV
IT
Flip angle
Fat supression
Hip 3-T non-contrast MRI
T1 STIR STIR 3D-WATSf PD-HR
coronal coronal axial sagittal coronal sagittal
633 9335 10107 20 3084
20 60 60 4,1 27
265 × 265 280 × 280 296 × 203 308 × 217 372 × 270
3 mm 3 mm 3 mm 1,18 × 1,18 × 1 3,5 mm
0,8 0,8 0,1 – 0,5
3 13 13 – 13
350 × 396 350 × 396 380 × 295 365 × 256 160 × 160
– 230 230 – –
– – – 35 –
– – – ProSet SPAIR
Hip 1.5-T direct MRA
THRIVE TSE T1
axial axial oblique coronal sagittal
3,6 500–600
8 18
224 × 185 160 × 160
1×1x 1 3
– 0,3
– 3
230 × 189 160 × 160
– –
7◦ –
SPAIR –
500–600 20
18 7
160 × 160 200 × 150
3 1,2 × 1,2 × 1,2
0,3 –
3 –
160 × 160 230 × 230 × 90
– –
– 50◦
SPIR ProSet
TSE T1 3D-WATSf
lage, together with a description of the treatment administered. Arthroscopic findings were used as the reference standard. 2.5. Statistical analysis The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of the different diagnostic techniques were calculated with a 95% confidence interval (CI). Cohen’s Kappa Index (k) was used to measure agreement between the different diagnostic techniques in assessing LCTZ pathology. This agreement was classified along the lines recommended by Landis and Koch. Thus, Kappa values between 0.0 and 0.20 indicated a slight agreement, from 0.21 to 0.40 fair agreement, between 0.41 and 0.60 moderate agreement, from 0.61 to 0.80 substantial agreement, and between 0.81 and 1.0 almost perfect agreement. 3. Results The study group consisted of fifty patients, thirty male and twenty female. The average age of the group was 42.5 years, the age range ran from 16 to 58. A history of chronic hip or groin pain and a clinical examination with findings consistent with FAI were present in all cases. In accordance with radiographic findings, these patients with FAI were classified as having pincer impingement (thirty-eight patients) or cam impingement (twelve patients), while deformities due to arthrosis were also noted in twenty-one patients (42%). The mean delay between the 1.5-T MRA and hip arthroscopy was eight months. In our series most of patients underwent hip arthroscopy before the maximum waiting time detailed on the Health System. Some patients delayed the intervention for personal reasons. A review of the patients’ clinical notes revealed no episodes of hip trauma between the time of the MRA examinations and the time of arthroscopy. No adverse events arose from performing the 3-T non-contrast MRI, the 1.5-T MRA or the hip arthroscopy. In the fifty patients with FAI who underwent hip arthroscopy, forty-three were found to have damage, involving forty-three tears at the LCTZ (Fig. 1) and two intra-substantial labral tears or flaps. Seven patients were found to have a paralabral cyst (Fig. 2). One sub-labral sulcus, as a normal variant, was found. Seven hips were normal in respect of the labrum and the LCTZ. In arthroscopic evaluations of labral appearance, thirty-six patients were identified as having labral degenerative changes, including two of intra-substance labral tears. The retrospective review of imaging examinations yielded a total of seven discrepant cases with 3-T non-contrast MRI and twelve with 1.5-T MRA. Four cases were read as negative for labral
abnormality when 3-T non-contrast MRI was used, and five when 1.5-T MRA was used, where the operative report was positive (false negatives). Three cases were read as positive with 3-T non-contrast MRI and seven with 1.5-T MRA in which the arthroscopic operation reports were negative (false positives). In the forty-three cases that had arthroscopically demonstrated tears at the LCTZ (Fig. 1), the tears were detected by 1.5-T MRA in every instance, whilst 3-T non-contrast MRI detected forty-two out of forty-three (97.7%). With 1.5-TMRA there was one false positive. The results of evaluation for labral appearance and for tears at the LCTZ from 3-T non-contrast MRI and 1.5-T direct MRA of the hip, using hip arthroscopy as a reference standard, are presented in the Table 2. Statistical indicators diagnostic performance of both imaging techniques are presented in Table 3. For cartilage within the LCTZ, agreement as to the grade between arthroscopy and MRI, whether 3-T non-contrast MRI or 1.5-TMRA, was reached in twenty-five out of the fifty patients (Fig. 3). In concordance with our previous observations [18], MRI was better than arthroscopy for diagnosis of chondral lesions of Grade 4. However, special difficulty was found, either with 3-T non-contrast MRI or 1.5-T MRA compared to hip arthroscopy, in the morphological characterization of low grade cartilage lesions (Fig. 4). In evaluations based on arthroscopy, eight patients were found to have normal cartilage at the LCTZ, thirteen were identified as having chondral lesions of Grade 1, twelve of Grade 2, eleven of Grade 3 and six of Grade 4. On 3-T non-contrast MRI, only three patients were identified as having Grade 3 chondral lesions and fourteen Grade 4. This is a reasonable level of agreement (k = 0.38, 95%CI 0.1–0.55). For acetabular cartilage, agreement between arthroscopy and MRI, whether using 3-T non-contrast MRI or 1.5-T MRA, as to the grade was achieved in about 56% of the patients. Arthroscopic evaluation showed twenty-two patients to have normal cartilage at the acetabular surface, seven to have chondral lesions of Grade 1, seven Grade 2, six Grade 3 and nine Grade 4.On 3-T non-contrast MRI, four patients were identified as having Grade 3 chondral lesions and twenty Grade 4. This is a reasonable level of agreement (k = 0.39, 95% CI0.12 to 0.68). For femoral articular cartilage, agreement between arthroscopy and MRI, whether using 3-T non-contrast MRI or 1.5-T MRA, was seen in thirty-nine patients (76%). 4. Discussion Hip arthroscopy is considered to be the reference standard for assessing intra-articular lesions of the hip joint, and it facilitates proceeding to therapeutic intervention. There is a considerable
112
A.M. Crespo-Rodríguez et al. / European Journal of Radiology 88 (2017) 109–116
Fig. 1. Full-thickness cartilage lesion in a fifty-two-year old male with pincer FAI. (a) On 3-T non-contrast MR, HR-PD weighted image (wi) with FS on a coronal view illustrates the plane of separation at the LCTZ (long arrow) with fraying and low signal intensity of cartilage (short arrow). (b) HR-PD wi with FS on a sagittal view demonstrates the hyperintense signal of the articular fluid extending through a complete plane of separation at the labral–chondral transitional zone (LCTZ), from the articular surface to the subchondral line (long arrow). (c) Sagittal T1 wi fat-suppressed MRA shows contrast media filling the full thickness cartilage lesion (long arrow). Intra-articular injection of contrast solution enlarges the joint and perilabral recess (curved arrow). Also a small artifact due to a small intra-articular air bubble is visible (dashed arrow). (d) Arthroscopic image shows a small delamination cartilage tear at the LCTZ (black arrow). Table 2 Comparison between 3-T non-contrast MRI and 1.5-T MRA of the hip and arthroscopy, using arthroscopy as standard of reference, for depiction of lesions at the acetabular labrum and the labral-chondral transitional zone. Hip arthroscopy Acetabular labrum
Labral-chondral transitional zone
Degenerative
Normal
Total
Tear
Normal
Total
3-T non-contrast MRI
Positive Negative Total
32 4 36
3 11 14
35 15 50
42 1 43
0 7 7
42 8 50
1.5-T direct MRA
Positive Negative Total
31 5 36
7 7 14
38 12 50
43 0 43
1 6 7
44 6 50
Table 3 Statistical indicators of the comparison between hip 3-T non-contrast MRI and 1.5-T MRA with hip arthroscopy for diagnosis of labral and labral-chondral transitional zone pathology, using hip arthroscopy as the reference standard. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), accuracy and also Cohen’s kappa index were calculated with a 95% confidence interval (CI).
Sensitivity Specificity PPV NPV Accuracy Kappa coefficient
Acetabular labrum abnormalities% (CI 95%)
Tears at the labral-chondral transitional zone% (CI 95%)
3T non-contrast MRI
1.5-T MRA
3T non-contrast MRI
1.5-T MRA
88.9 (77.2–100) 78.6 (53.5–100) 91.4 (80.7–100) 73.3 (47.6–99) 86 (75.4–96,6) 0.66 (0.42−0.89)
86.1 (73.4–98.8) 50 (20.2–79.7) 81.6 (67.9–95.2) 58.3 (26.3–90.4) 76 (63.16–88.8) 0.37 (0.09–0.66)
97.7 (92–100) 100 (92.8–100) 100 (98.8–100) 87.5 (58.3–100) 98 (93.12–100) 0.92 (0.76–1)
100 (98.8–100) 85.7 (52.6–100) 97.7 (92.2–100) 100 (91.7–100) 98 (93.1–100) 0.91 (0.73–1)
A.M. Crespo-Rodríguez et al. / European Journal of Radiology 88 (2017) 109–116
113
Fig. 2. Posterior paralabral cyst in a thirty-five-year-old male with a pincer FAI. (a) Sagittal WATS and (b) sagital PD-HR wi show fluid signal intensity next to the posterior acetabular margin (long arrow). (c) Sagittal reformatted THRIVE MRA image demonstrates a cystic lesion (long arrow) full-filled of contrast media between the labrum (curved arrow) and the posterior acetabulum. (d) Arthroscopic image shows the posterior paralabral cyst adyacent to the labrum (black arrow).
interest in different techniques of hip MRI as alternative, less invasive, methods that can establish diagnoses of disorders of the hip [19], provide guidance to surgeons as to the likely site of labral and chondral abnormalities and possibly enable surgeons to decide pre-operatively on the appropriate surgical technique. Direct MRA is the preferred method because it is better at enlarging the joint and providing images of small internal structures, allowing a better evaluation of diseases or conditions within the joint. Direct MRA at 1.5-T has been shown to be a reliable method for defining and detecting acetabular labral tears [9]. A systematic review of the literature and meta-analysis found a pooled sensitivity of approximately 87% and a pooled specificity about 64% for labral tears [6,9,14,20–25]. Tables 4 and 5 summarize Smith et al. [9] and add later relevant studies [18,26,27], all of them with more than fifty patients. However, sensitivity of direct MRA in the detection of cartilage damage is only moderate [6]. Indirect MRA is less invasive than the direct technique, does not require fluoroscopy or a physician to perform the injection, is relatively easy to schedule and has improved contrast resolution relative to non-contrast MRI. However, the disadvantages of indirect MRA, as compared with direct MRA, include the lack of capsular distention, higher contrast load, and less reliability in rendering the joint opaque. Although we do not have experience with this technique, few reports in the literature indicate favourable sensitivity
and specificity values with indirect MRA, whether with 1.5-T [28] or with 3-T magnets [29]. The same meta-analysis also examined the diagnostic performance of 1.5-T non-contrast MRI for acetabular labral tears and chondral abnormalities (Table 5), finding a pooled sensitivity of 66% and a pooled specificity of 79% [9]. Recently, two retrospective studies, both with high resolution MRI protocols, have reported a better diagnostic performance of 1.5-T non-contrast MRI for acetabular labral tears (pooled sensitivity of 96%) and chondral abnormalities[4,30]. To our knowledge, only two retrospective studies have compared the diagnostic performance of non-contrast MRI and MRA at a 3-T magnetic field for detection of acetabular labral tears and chondral defects in the same patient population [26,27] with divergent results. On the one hand, Tian et al. [26] reported that 3.0-T non-contrast MRI was not effective enough for evaluating the acetabular labral tears, with a relatively low sensitivity and specificity (Table 5), leading to the conclusion that 3.0-T MRA of the hip is a reliable evaluation modality for diagnosing acetabular labral tears, with a superior diagnostic performance (Table 4). On the other hand, Magee [27] reported that 3.0-T non-contrast MRI was near equivalent to 3.0-T MRA (Table 4 and 5) for detection of acetabular labral tears. Only in one patient out of forty-three, 3.0-T MRA allowed for detection of an acetabular labral tear not seen on 3.0-T non-contrast MR examination.
114
A.M. Crespo-Rodríguez et al. / European Journal of Radiology 88 (2017) 109–116
Fig. 3. Full-thickness cartilage lesion in a forty-six-year-old female with pincer FAI. (a) Coronal T1-wi of MRA shows a small full-thickness cartilage lesion that seems to be an acetabular cartilage defect (long arrow); acetabular labrum with normal signal intensity (dashed arrow) and perilabral recess (curved line) are also identified. On 3-T non-contrast MRI, (b) HR-PD wi with FS enables differentiation of a cartilage delamination tear: A hyperintense line extending from a disrupted cartilage surface deep to the acetabular cartilage layers is visible (long arrow). (c) HR-PD wi with FS on a sagittal view demonstrates a full-thickness cartilage delamination tear (long arrow). (d) The artrhoscopic image shows a small delamination tear at the acetabular cartilage (black arrow).
Table 4 Sensitivity and specificity for the diagnostic accuracy of direct MR arthrography (MRA) of the hip for detecting acetabular labral tears in different studies, modified from Smith et al. [9]. Magnetic field strength, tesla (T), patients, true positive (TP), false positive (FP), false negative (FN), true negative (TN), sensitivity and specificity. Study
Magnetic field streght
Patients
TP
FP
FN
TN
Sensitivity
Specificity
Czerny et al 1999 [14] Schmid et al 2003 [6] Keeney et al 2004 [19] Byrd and Jones 2004 [20] Chan et al 2005 [21] Freedman et al 2006 [22] Toomayan et al 2006 [23] Studler et al 2008 [24] Ziegert et al 2009[25] Tian et al 2014 [26] Magee 2015 [27] Crespo-Rodríguez et al 2015 [18]
1.0-T 1.0-T 1.5-T 1.5-T 1.5-T 1.5-T 1.5-T 1.5-T 1.5-T 3.0-T 3.0-T 1.5-T
40 40 102 40 17 24 30 57 144 34 43 51
30 14 66 23 16 22 22 43 140 19 38 43
2 5 5 7 1 1 0 6 0 2 1 1
3 4 27 9 0 1 2 1 4 2 4 0
5 17 4 1 0 0 6 7 0 11 0 7
91% 78% 71% 72% 100% 96% 92% 98% 97% 90,5% 90% 100%
71% 77% 44% 13% 0% 100% 54% 84,6% 87,5%
Table 5 Sensitivity and specificity for the diagnostic accuracy of non-contrast MRI of the hip for detecting acetabular labral tears in different studies, modified from Smith et al. [9]. Magnetic field strength, tesla (T), patients, true positive (TP), false positive (FP), false negative (FN), true negative (TN), sensitivity and specificity. Study
Magnetic field streght
Patients
TP
FP
FN
TN
Sensitivity
Specificity
Byrd and Jones 2004 [20] Mintz et al 2005 [30] Tian et al 2014 [26] Magee 2015 [27] Crespo-Rodríguezet al 2016
1.5-T 1.5-T 3.0-T 3.0-T 3.0-T
40 92 90 43 50
7 86 36 39 42
4 2 7 1 0
25 3 23 3 1
4 1 24 0 7
22% 97% 61.2% 93% 97.7%
50% 33% 77.4% 100%
A.M. Crespo-Rodríguez et al. / European Journal of Radiology 88 (2017) 109–116
115
Fig. 4. Small “pocket” cartilage lesion in a 40-year-old man with cam FAI in the left hip. (a) Sagittal and (b) coronal HR-PD wi with FS at a 3-T magnet show hiperintensity of the articular fluid within the small cleft at the LCTZ (long arrows), an abnormal signal intensity within the labrum related to degeneration (dashed arrow) and hypointensity foci within de acetabular cartilage (short arrow). (c) Sagittal reformatted THRIVE image of MRA shows contrast media full-filling a small cleft at the LCTZ (long arrow) and extending to the labrum (dashed arrow). (d) Arthroscopic image shows a small “pocket” cartilage lesion (black arrow) immediately adjacent to attachment of the labrum (L) to the acetabular cartilage (C).
In our opinion, moving to higher field strengths (specifically 3T) improves the SNR, which should then be able to be traded off for higher image resolution. Overall, the data being reported here show an excellent diagnostic performance for 3-T non-contrast MRI. If positive imaging findings are yielded by 3-T MRI, 1.5-T MRA does not provide any additional information. With regard to normal variations in the acetabular labrum and the difficulty of distinguishing them from labral tears, 3-T non-contrast MRI can provide valuable additional information for a differential diagnosis. It can accurately identify cartilage abnormalities next to the cleft within the LCTZ in labral tears, while in sub-labral sulcus a cleft within the LCTZ is seen without there being any cartilage abnormality. The issue that remains is whether MRA must be retained in daily practice for cases with negative findings on 3-T non-contrast MRI but a strong suspicion of FAI. Some authors consider that a negative imaging study, even at 3-T, does not completely exclude major intra-articular pathology that can be identified and treated arthroscopically [18,27]. However, it is possible to avoid performing an invasive procedure in at least 80% of patients with positive findings. Hence, our imaging protocol for patients with hip pain and clinical suspicion of FAI is first radiographs, followed by 3-T non-contrast MRI of the hip. In young patients (under 45 years old) with a clinical strong suspicion of FAI and normal findings on 3-T MRI would it proceed to MRA. The next question to be addressed would be whether
3-T MRA should be performed on this reduced group of patients. In the rest of patients, an individualized re-evaluation of clinical settings, diagnostic tools and treatment options is recommended. A number of limitations in this study must be acknowledged. The first is the small number of patients involved. The second, selection of patients with radiological signs of FAI and the high mean age of the group can explain the high rate of LCTZ injury. The third is that no control group was included in the study, because healthy individuals do not undergo hip arthroscopy. The fourth is that the surgeon was aware of the MRI results and this information was used in pre-operative planning. The last, radiologists knew that all patients underwent arthroscopy at the time they reviewed the images for this paper. In conclusion, non-invasive assessment of the hip is possible with 3-T MR. MRA can be replaced by 3-T non-contrast MRI as the workhorse technique for the assessment of internal disorders of the hip. MRA would be reserved for young adults with a strong clinical suspicion of FAI, but negative findings on 3-T non-contrast MRI. Optimized sequences on 3-T non-contrast MRI add value in detecting normal variations, such as sub-labral sulcus, and in diagnosing articular cartilage lesions in the hip, as compared with 1.5-T MRA.
116
A.M. Crespo-Rodríguez et al. / European Journal of Radiology 88 (2017) 109–116
Conflict of interest The authors and author’s institutions have no conflicts of interest. Acknowledgement The authors gratefully acknowledge Luis Ramos González, PhD MD, former Chairman of the Department of Radiology at the Hospital Universitario Puerta de Hierro Majadahonda, Madrid, Spain, for his support in this scientific paper and for encouraging everyone to pursue excellence in the patient care. The authors gratefully acknowledge Isabel Millán Santos, PhD for her contribution to statistical analysis of this paper. References [1] R. Ganz, J. Parvizi, M. Beck, M. Leunig, H. Notzli, K.A. Siebenrock, Femoroacetabular impingement: a cause for osteoarthritis of the hip, Clin. Orthop. Relat. Res. 417 (2003) 112–120, http://dx.doi.org/10.1097/01.blo. 0000096804.78689.c2. [2] C.W. Pfirrmann, B. Mengiardi, C. Dora, F. Kalberer, M. Zanetti, J. Hodler, Cam and pincer femoroacetabular impingement: characteristic MR arthrographic findings in 50 patients, Radiology 240 (2006) 778–785, http://dx.doi.org/10. 1148/radiol.2403050767. [3] J.C. McCarthy, P.C. Noble, M.R. Schuck, J. Wright, J. Lee, The Otto E. Aufranc Award: the role of labral lesions to development of early degenerative hip disease, Clin. Orthop. Relat. Res. 393 (2001) 25–37. [4] S.L. James, K. Ali, F. Malara, D. Young, J. O’Donnell, D.A. Connell, MRI findings of femoroacetabular impingement, Am. J. Roentgenol. 187 (2006) 1412–1419, http://dx.doi.org/10.2214/AJR.05.1415. [5] M. Beck, M. Leunig, J. Parvizi, V. Boutier, D. Wyss, R. Ganz, Anterior femoroacetabular impingement: part II. Midterm results of surgical treatment, Clin. Orthop. Relat. Res. 429 (2004) 170–177. [6] M.R. Schmid, H.P. Nötzli, M. Zanetti, T.F. Wyss, J. Hodler, Cartilage lesions in the hip: diagnostic effectiveness of MR arthrography, Radiology 226 (2003) 382–386, http://dx.doi.org/10.1148/radiol.2262020019. [7] K.A. Jung, C. Restrepo, M. Hellman, H. AbdelSalam, W. Morrison, J. Parvizi, The prevalence of cam-type femoroacetabular deformity in asymptomatic adults, J. Bone Joint Surg. Br. 93 (2011) 1303–1307, http://dx.doi.org/10.1302/0301620X.93B10.26433. [8] D.R. Griffin, E.J. Dickenson, J. O’Donnell, R. Agricola, T. Awan, M. Beck, et al., The Warwick Agreement on femoroacetabular impingement syndrome (FAI syndrome): an international consensus statement, Br. J. Sports Med. 50 (2016) 1169–1176, http://dx.doi.org/10.1136/bjsports-2016-096743. [9] T. Smith, G. Hilton, A.P. Toms, S.T. Donell, C.B. Hing, The diagnostic accuracy of acetabular labral tears using magnetic resonance imaging and magnetic resonance arthrography: a meta-analysis, Eur. Radiol. 21 (2011) 863–874, http://dx.doi.org/10.1007/s00330-010-1956-7. [10] P. Robinson, Conventional 3-T MRI and 1.5-T MR arthrography of femoroacetabular impingement, Am. J. Roentgenol. 199 (2012) 509–515, http://dx.doi.org/10.2214/AJR.12.8672. [11] A. Mohr, F.W. Roemer, H.K. Genant, C. Liess, Using fatsaturated proton density-weighted MR imaging to evaluate articular cartilage, Am. J. Roentgenol. 181 (2003) 280–281, http://dx.doi.org/10.2214/ajr.181.1. 1810280a. [12] C.A. Petersilge, M.A. Haque, W.J. Petersilge, J.S. Lewin, J.M. Lieberman, R. Buly, Acetabular labral tears: evaluation with MR arthrography, Radiology 200 (1996) 231–235, http://dx.doi.org/10.1148/radiology.200.1.8657917.
[13] L.A. Lage, J.V. Patel, R.N. Villar, The acetabular labral tear: an arthroscopic classification, Arthroscopy 12 (1996) 269–272, http://dx.doi.org/10.1016/ S0749-8063(96)90057-2. [14] C. Czerny, S. Hofmann, M. Urban, C. Tschauner, A. Neuhold, M. Pretterklieber, M.P. Recht, J. Kramer, MR arthrography of the adult acetabular capsular-labral complex: correlation with surgery and anatomy, Am. J. Roentgenol. 173 (1999) 345–349, http://dx.doi.org/10.2214/ajr.173.2.10430132. [15] D.G. Blankenbaker, A.A. De Smet, J.S. Keene, J.P. Fine, Classification and localization of acetabular labral tears, Skeletal Radiol. 36 (2007) 391–397, http://dx.doi.org/10.1007/s00256-006-0240-z. [16] D.A. McCall, M.R. Safran, MRI and arthroscopy correlations of the hip: a case-based approach, Instr. Course Lect. 61 (2012) 327–344. [17] R.E. Outerbridge, The etiology of chondromalacia patellae, J. Bone Joint Surg. Br. 43 (1961) 752–757. [18] A.M. Crespo Rodríguez, J.C. De Lucas Villarrubia, M.A. Pastrana Ledesma, Millán Santos I, M. Padrón, Diagnosis of lesions of the acetabular labrum, of the labral-chondral transition zone, and of the cartilage in femoroacetabular impingement: correlation between direct magnetic resonance arthrography and hip arthroscopy, Radiologia 57 (2015) 131–141. [19] J.A. Keeney, M.W. Peelle, J. Jackson, D. Rubin, W.J. Maloney, J.C. Clohisy, Magnetic resonance arthrography versus arthroscopy in the evaluation of articular hip pathology, Clin. Orthop. Relat. Res. 429 (2004) 163–169. [20] J.W. Byrd, K.S. Jones, Diagnostic accuracy of clinical assessment, magnetic resonance imaging, magnetic resonance arthrography, and intra-articular injection in hip arthroscopy patients, Am. J. Sports Med. 32 (2004) 1668–1674. [21] Y.S. Chan, L.C. Lien, H.L. Hsu, Y.L. Wan, M.S. Lee, K.Y. Hsu, C.H. Shih, Evaluating hip labral tears using magnetic resonance arthrography: a prospective study comparing hip arthroscopy and magnetic resonance arthrography diagnosis, Arthroscopy 21 (2005) 1250, http://dx.doi.org/10.1016/j.arthro.2005.07.007. [22] B.A. Freedman, B.K. Potter, P.A. Dinauer, J.R. Giuliani, T.R. Kuklo, K.P. Murphy, Prognostic value of magnetic resonance arthrography for Czerny stage II and III acetabular labral tears, Arthroscopy 22 (2006) 742–747, http://dx.doi.org/ 10.1016/j.arthro.2006.03.014. [23] G.A. Toomayan, W.R. Holman, N.M. Major, S.M. Kozlowicz, T.P. Vail, Sensitivity of MR arthrography in the evaluation of acetabular labral tears, Am. J. Roentgenol. 186 (2006) 449–453, http://dx.doi.org/10.2214/AJR.04.1809. [24] U. Studler, F. Kalberer, M. Leunig, M. Zanetti, J. Hodler, C. Dora, C.W. Pfirrmann, MR arthrography of the hip: differentiation between an anterior sublabral recess as a normal variant and a labral tear, Radiology 249 (2008) 947–954, http://dx.doi.org/10.1148/radiol.2492080137. [25] A.J. Ziegert, D.G. Blankenbaker, A.A. De Smet, J.S. Keene, K. Shinki, J.P. Fine, Comparison of standard hip MR arthrographic imaging planes and sequences for detection of arthroscopically proven labral tear, Am. J. Roentgenol. 192 (2009) 1397–1400, http://dx.doi.org/10.2214/AJR.08.1609. [26] C.Y. Tian, J.Q. Wang, Z.Z. Zheng, AH Ren, 3.0T conventional hip MR and hip MR arthrography for the acetabular labral tears confirmed by arthroscopy, Eur. J. Radiol. 83 (2014) 1822–1827, http://dx.doi.org/10.1016/j.ejrad.2014.05.034. [27] T. Magee, Comparison of 3. 0-T MR vs 3. 0-T MR arthrography of the hip for detection of acetabular labral tears and chondral defects in the same patient population, Br. J. Radiol. 88 (2015) 20140817, http://dx.doi.org/10.1259/bjr. 20140817. [28] M.B. Zlatkin, D. Pevsner, T.G. Sanders, C.R. Hancock, C.E. Ceballos, M.F. Herrera, Acetabular labral tears and cartilage lesions of the hip: indirect MR arthrographic correlation with arthroscopy—a preliminary study, Am. J. Roentgenol. 194 (2010) 709–714, http://dx.doi.org/10.2214/AJR.07.3669. [29] C.N. Petchprapa, L.D. Rybak, K.S. Dunham, R. Lattanzi, M.P. Recht, Labral and cartilage abnormalities in young patients with hip pain: accuracy of 3-Tesla indirect MR arthrography, Skeletal Radiol. 44 (2015) 97–105, http://dx.doi. org/10.1007/s00256-014-2013-4. [30] D.N. Mintz, T. Hooper, D. Connell, R. Buly, D.E. Padgett, H.G. Potter, Magnetic resonance imaging of the hip: detection of labral and chondral abnormalities using noncontrast imaging, Arthroscopy 21 (2005) 385–393, http://dx.doi. org/10.1016/j.arthro.2004.12.011.
Contents lists available at ScienceDirect
European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad
The diagnostic performance of non-contrast 3-Tesla magnetic resonance imaging (3-T MRI) versus 1.5-Tesla magnetic resonance arthrography (1.5-T MRA) in femoro-acetabular impingement Ana M. Crespo-Rodríguez a,∗ , Jose C. De Lucas-Villarrubia b , Miguel Pastrana-Ledesma a , Ana Hualde-Juvera a , Santiago Méndez-Alonso a , Mario Padron c a
Department of Radiology, Hospital Universitario Puerta de Hierro Majadahonda, c/ Joaquín Rodrigo 2, Majadahonda 28222, Madrid, Spain Department of Orthopaedics and Traumatology at the Hospital Universitario Puerta de Hierro Majadahonda, c/ Joaquín Rodrigo 2, Majadahonda 28222, Madrid, Spain c Department of Radiology, Clínica Cemtro, Avda Ventisquero de la Condesa 42 Madrid 28035, Madrid, Spain b
a r t i c l e
i n f o
Article history: Received 13 September 2016 Received in revised form 17 November 2016 Accepted 26 December 2016 Keywords: Sensitivitity Specificity Arthrography Magnetic resonance imaging (MRI) Hip Femoro-acetabular impingement
a b s t r a c t Objective: The aim of this study was to evaluate the diagnostic accuracy of 3-T non-contrast MRI versus 1.5-T MRA for assessing labrum and articular cartilage lesions in patients with clinical suspicion of femoroacetabular impingement (FAI). Subjects and methods: Fifty patients (thirty men and twenty women, mean age 42.5 years) underwent 1.5-T MRA, 3-T MRI and arthroscopy on the same hip. An optimized high-resolution proton density spin echo pulse sequence was included in the 3-T non-contrast MRI protocol. Results: The 3-T non-contrast MRI identified forty-two of the forty-three arthroscopically proven tears at the labral-chondral transitional zone (sensitivity, 97.7%; specificity, 100%; positive predictive value (PPV), 100%; negative predictive value (NPV), 87.5%; accuracy 98%). With 1.5-T MRA, forty-four tears were diagnosed. However, there was one false positive (sensitivity, 100%; specificity, 85.7%; PPV, 97.7%; NPV, 100%; accuracy 98%). Agreement between arthroscopy and MRI, whether 3-T non-contrast MRI or 1.5-T MRA, as to the degree of chondral lesion in the acetabulum was reached in half of the patients and in the femur in 76% of patients. Conclusion: Non-invasive assessment of the hip is possible with 3-T MR magnet. 3-T non-contrast MRI could replace MRA as the workhorse technique for assessing hip internal damage. MRA would then be reserved for young adults with a strong clinical suspicion of FAI but normal findings on 3-T non-contrast MRI. When compared with 1.5-T MRA, optimized sequences with 3-T non-contrast MRI help in detecting normal variants and in diagnosing articular cartilage lesions. © 2017 Elsevier Ireland Ltd. All rights reserved.
1. Introduction
∗ Corresponding author. Present address: Department of Radiology, Hospital Clínico San Carlos, c/ Profesor Martín Lagos s/n, Madrid 28040, Madrid, Spain. E-mail addresses: [email protected] (A.M. Crespo-Rodríguez), [email protected] (J.C. De Lucas-Villarrubia), [email protected] (M. Pastrana-Ledesma), [email protected] (A. Hualde-Juvera), [email protected] (S. Méndez-Alonso), [email protected] (M. Padron). URLs: http://[email protected] (M. Pastrana-Ledesma), http://ana. [email protected] (A. Hualde-Juvera), http://[email protected] (S. Méndez-Alonso). http://dx.doi.org/10.1016/j.ejrad.2016.12.031 0720-048X/© 2017 Elsevier Ireland Ltd. All rights reserved.
Femoro-acetabular impingement (FAI) occurs when there is a conflict between the proximal femur and the acetabular rim [1]. Cam and pincer FAI can be differentiated on the basis of a predominance of either a femoral or an acetabular abnormality [2]. These changes result in repetitive microtrauma between the femur and the acetabular rim, leading to degeneration and tearing of the labral-chondral transitional zone (LCTZ), particularly in the anterosuperior region [3]. Lesions occur at the attachment of the labrum to the acetabulum, initially leaving the body of the labrum intact [4]. Beck et al. [5] and Schmid et al. [6] noted that in patients with FAI, adjacent acetabular cartilage damage was found in all patients with labral abnormalities. As a consequence, FAI has
110
A.M. Crespo-Rodríguez et al. / European Journal of Radiology 88 (2017) 109–116
become increasingly recognized as a potential cause of premature onset of osteoarthritis of the hip, especially in young (20–45 years) and active individuals [1,3]. FAI remains a controversial disease in terms of true incidence, clinical relevance, diagnosis and management [7]. Recently, an international consensus statement on FAI syndrome, the Warwick Agreement, has been published [8]. The syndrome has stimulated radiologists to re-evaluate the use of different hip MRI examinations, because of major improvements in resolution and technique in the past ten years. Several publications have advocated the use of direct MRA for the detection of labral tears because of its higher sensitivity as compared with non-contrast MRI at 1.5 T [9]. The enhanced signal-noise ratio (SNR), spatial resolution, and contrastto-noise ratio provided by the higher field strength would make 3-T MRI the ideal method for assessing articular cartilage and small structures such as the acetabular labrum [10]. The challenge for 3-T MRI is to reach as good a diagnostic performance as 1.5-T MRA in the assessment of acetabular labral abnormalities without the need for intra-articular contrast material and to improve the detection of cartilage abnormalities [10]. Besides image quality, several factors need to be considered when choosing between 3-T non-contrast MRI and 1.5-T MRA. These factors include the preferences of the radiologist and referring orthopaedic surgeon, patients’ reluctance to undergo an invasive procedure, risk, time, cost, labour, scheduling and coordination of either fluoroscopy or ecography and MR. The purpose of this study was to evaluate in retrospect the diagnostic performance of 3-T non-contrast MRI as compared with 1.5-TMRA in diagnosing LCTZ abnormalities in patients with FAI, surgical findings acting as the reference standard.
2. Material and methods 2.1. Subjects The retrospective study was performed with approval from our institutional review board, and with a waiver of informed consent. Criteria for inclusion were clinical suspicions of FAI, radiological findings of FAI (either cam or pincer) and 3-T non-contrast MRI, 1.5-T MRA and arthroscopy on the same hip. Criteria for exclusion criteria were previous surgery on the hip or other causes of groin pain ruled out on 3.0-T non-contrast MRI. The population studied consisted of consecutive series of patients with FAI who underwent hip arthroscopy between January 2012 and December 2014 in the Hospital Universitario Puerta de Hierro Majadahonda, Madrid. Imaging examinations were performed between October 2010 and February 2013. 3-T non-contrast MRI was performed before 1.5-T MRA, with a mean delay of three weeks. In this period of time 190 MRAs were done, most of them because of FAI.
2.2. Imaging examinations Non-contrast 3-T MRI was performed with a 3-T MR magnet (Achieva Dual Quasar 3-T Philips, Eindhoven, Netherlands) with the body phased array coil 6 Channel-Torso-coil. Our institution is a general hospital equipped with two 1.5-T and one 3-T MR magnets. Our standard hip protocol used to have similar imaging sequences in any magnet looking forward homogeneity. However, with the addition of a 3-T clinical MR magnet to our department, we took advantage of the increased SNR, spatial resolution, and contrastto-noise ratio afforded by higher field strength. Proton density (PD) weighted spin echo pulse sequences offer greater SNR than T1 or T2 [11]. A high acquisition matrix improves spatial resolution. In the particular case of the hip, we decided to introduce fat-suppressed
highresolution-PD (HR-PD) weighted spin echo pulse sequences into clinical practice (Table 1). Direct MRA of the hip was a two-step procedure. First, intraarticular injection was performed using an echogenic tip Chiba ® needle (Cook ) under ultrasound guidance. The 20cc of contrast ® solution contained 0.2cc of gadoterate meglumine (Dotarem ), 3cc ® of 2% mepivacaine and 17cc of sterile saline. The second step was to perform MR on patients, a 5 kg weight being used to apply traction, within thirty minutes of the injection of contrast medium. These MR examinations were performed with a 1.5-T MR magnet (Achieva Nova 1.5-T Philips, Eindhoven, Netherlands) with the cardiac body 5 Channel-Cardiac-coil. A three dimensional T1 weighted gradient echo sequence, specifically T1 High-Res Isotropic Volume Excitation (THRIVE) allowing reconstruction in any plane, was included in the protocol (Table 1). 2.3. Criteria for interpretations MR studies were analysed by three musculoskeletal radiologists with four (AHJ), six (AMCR), and twenty (MPL) years of experience, who were blinded to each others’ readings, of the original report and of the surgical results. The final interpretation was based on a majority consensus. Labral appearance was evaluated as normal or degenerative. Besides labral enlargement, intra-substance areas of intermediate signal intensity and irregular margins [12], intra-substance or flap tears were considered to be degenerative changes. They were characterized by high signal intensity on MRI or contrast medium in MRA being noted within the substance of the labrum extending to the articular surface. In LCTZ the evaluation included the assessment both of separation or tearing of the labrum and of adjacent cartilage damage. The presence of a tear at the LCTZ was defined as a cleft noted through the synovial fluid signal intensity on MRI [4] or contrast medium on MRA [12]. This chondro-labral separation has been termed a watershed lesion [3], a basal tear [10] or a longitudinal peripheral tear [13]. None classification system for hip labral tears was used as the Lage arthroscopic classification [13] does not correlate well with the Czerny MRA [14] or an MRA modification of the Lage classification [15]. The presence of a labral or paralabral cyst was also considered an indirect sign of tearing [16], similar to parameniscal cyst in the knee. The articular cartilage was graded on MRI by means of a modified version of Outerbridge’s classification system [17]. In the modified ranking, Grade 0 indicated intact cartilage. Grade 1 signified chondral softening or blistering with an intact surface, whilst Grade 2 indicated shallow superficial ulceration, fibrillation, or fissuring involving less than 50% of the depth of the articular surface. Grade 3 pointed to deep ulceration, fibrillation, fissuring or a chondral flap involving 50% or more of the depth of the articular cartilage without exposure of sub-chondral bone, whilst Grade 4 denoted full-thickness chondral wear with the exposure of sub-chondral bone. Apart from the assessment of femoral and acetabular articular surfaces, special attention was paid to cartilage within the LCTZ. 2.4. Arthroscopic evaluation Surgery was performed by an orthopaedic surgeon (JCDLV) with eight years of experience in hip arthroscopy and more than 300 procedures performed. This orthopaedic surgeon was aware of the MR imaging findings relating to patients at the time of arthroscopy. He provided dictated operative reports about all patients who underwent hip arthroscopy. Each report included an assessment of the integrity of the labrum, of the LCTZ and of articular carti-
A.M. Crespo-Rodríguez et al. / European Journal of Radiology 88 (2017) 109–116
111
Table 1 Protocol of 3-T non-contrast MRI and 1.5-T direct MR arthrography (MRA) of the hip. It is specified pulse sequence, plane, RT or repetition time (msec), ET or echo time (ms), matrix size, section thickness (mm), gap, echo train length, field of view (FOV), inversion time (IT), flip angle and fat suppression. Pulse sequences included are T1, 3D WATer Selective Fluid (3D-WATSf), STIR, Proton Density- High Resolution (PD-HR) and T1 High-Res Isotropic Volume Excitation (THRIVE). Fat suppression (FS) techniques detailed are SPectral Attenuated Inversion Recovery (SPAIR), Spectral Presaturation with Inversion Recovery (SPIR) and PRinciple Of Selective Excitation Technique (ProSet). Pulse sequence
Plane
RT
ET
Matrix size
Section thickness
Gap
Echo train length
FOV
IT
Flip angle
Fat supression
Hip 3-T non-contrast MRI
T1 STIR STIR 3D-WATSf PD-HR
coronal coronal axial sagittal coronal sagittal
633 9335 10107 20 3084
20 60 60 4,1 27
265 × 265 280 × 280 296 × 203 308 × 217 372 × 270
3 mm 3 mm 3 mm 1,18 × 1,18 × 1 3,5 mm
0,8 0,8 0,1 – 0,5
3 13 13 – 13
350 × 396 350 × 396 380 × 295 365 × 256 160 × 160
– 230 230 – –
– – – 35 –
– – – ProSet SPAIR
Hip 1.5-T direct MRA
THRIVE TSE T1
axial axial oblique coronal sagittal
3,6 500–600
8 18
224 × 185 160 × 160
1×1x 1 3
– 0,3
– 3
230 × 189 160 × 160
– –
7◦ –
SPAIR –
500–600 20
18 7
160 × 160 200 × 150
3 1,2 × 1,2 × 1,2
0,3 –
3 –
160 × 160 230 × 230 × 90
– –
– 50◦
SPIR ProSet
TSE T1 3D-WATSf
lage, together with a description of the treatment administered. Arthroscopic findings were used as the reference standard. 2.5. Statistical analysis The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of the different diagnostic techniques were calculated with a 95% confidence interval (CI). Cohen’s Kappa Index (k) was used to measure agreement between the different diagnostic techniques in assessing LCTZ pathology. This agreement was classified along the lines recommended by Landis and Koch. Thus, Kappa values between 0.0 and 0.20 indicated a slight agreement, from 0.21 to 0.40 fair agreement, between 0.41 and 0.60 moderate agreement, from 0.61 to 0.80 substantial agreement, and between 0.81 and 1.0 almost perfect agreement. 3. Results The study group consisted of fifty patients, thirty male and twenty female. The average age of the group was 42.5 years, the age range ran from 16 to 58. A history of chronic hip or groin pain and a clinical examination with findings consistent with FAI were present in all cases. In accordance with radiographic findings, these patients with FAI were classified as having pincer impingement (thirty-eight patients) or cam impingement (twelve patients), while deformities due to arthrosis were also noted in twenty-one patients (42%). The mean delay between the 1.5-T MRA and hip arthroscopy was eight months. In our series most of patients underwent hip arthroscopy before the maximum waiting time detailed on the Health System. Some patients delayed the intervention for personal reasons. A review of the patients’ clinical notes revealed no episodes of hip trauma between the time of the MRA examinations and the time of arthroscopy. No adverse events arose from performing the 3-T non-contrast MRI, the 1.5-T MRA or the hip arthroscopy. In the fifty patients with FAI who underwent hip arthroscopy, forty-three were found to have damage, involving forty-three tears at the LCTZ (Fig. 1) and two intra-substantial labral tears or flaps. Seven patients were found to have a paralabral cyst (Fig. 2). One sub-labral sulcus, as a normal variant, was found. Seven hips were normal in respect of the labrum and the LCTZ. In arthroscopic evaluations of labral appearance, thirty-six patients were identified as having labral degenerative changes, including two of intra-substance labral tears. The retrospective review of imaging examinations yielded a total of seven discrepant cases with 3-T non-contrast MRI and twelve with 1.5-T MRA. Four cases were read as negative for labral
abnormality when 3-T non-contrast MRI was used, and five when 1.5-T MRA was used, where the operative report was positive (false negatives). Three cases were read as positive with 3-T non-contrast MRI and seven with 1.5-T MRA in which the arthroscopic operation reports were negative (false positives). In the forty-three cases that had arthroscopically demonstrated tears at the LCTZ (Fig. 1), the tears were detected by 1.5-T MRA in every instance, whilst 3-T non-contrast MRI detected forty-two out of forty-three (97.7%). With 1.5-TMRA there was one false positive. The results of evaluation for labral appearance and for tears at the LCTZ from 3-T non-contrast MRI and 1.5-T direct MRA of the hip, using hip arthroscopy as a reference standard, are presented in the Table 2. Statistical indicators diagnostic performance of both imaging techniques are presented in Table 3. For cartilage within the LCTZ, agreement as to the grade between arthroscopy and MRI, whether 3-T non-contrast MRI or 1.5-TMRA, was reached in twenty-five out of the fifty patients (Fig. 3). In concordance with our previous observations [18], MRI was better than arthroscopy for diagnosis of chondral lesions of Grade 4. However, special difficulty was found, either with 3-T non-contrast MRI or 1.5-T MRA compared to hip arthroscopy, in the morphological characterization of low grade cartilage lesions (Fig. 4). In evaluations based on arthroscopy, eight patients were found to have normal cartilage at the LCTZ, thirteen were identified as having chondral lesions of Grade 1, twelve of Grade 2, eleven of Grade 3 and six of Grade 4. On 3-T non-contrast MRI, only three patients were identified as having Grade 3 chondral lesions and fourteen Grade 4. This is a reasonable level of agreement (k = 0.38, 95%CI 0.1–0.55). For acetabular cartilage, agreement between arthroscopy and MRI, whether using 3-T non-contrast MRI or 1.5-T MRA, as to the grade was achieved in about 56% of the patients. Arthroscopic evaluation showed twenty-two patients to have normal cartilage at the acetabular surface, seven to have chondral lesions of Grade 1, seven Grade 2, six Grade 3 and nine Grade 4.On 3-T non-contrast MRI, four patients were identified as having Grade 3 chondral lesions and twenty Grade 4. This is a reasonable level of agreement (k = 0.39, 95% CI0.12 to 0.68). For femoral articular cartilage, agreement between arthroscopy and MRI, whether using 3-T non-contrast MRI or 1.5-T MRA, was seen in thirty-nine patients (76%). 4. Discussion Hip arthroscopy is considered to be the reference standard for assessing intra-articular lesions of the hip joint, and it facilitates proceeding to therapeutic intervention. There is a considerable
112
A.M. Crespo-Rodríguez et al. / European Journal of Radiology 88 (2017) 109–116
Fig. 1. Full-thickness cartilage lesion in a fifty-two-year old male with pincer FAI. (a) On 3-T non-contrast MR, HR-PD weighted image (wi) with FS on a coronal view illustrates the plane of separation at the LCTZ (long arrow) with fraying and low signal intensity of cartilage (short arrow). (b) HR-PD wi with FS on a sagittal view demonstrates the hyperintense signal of the articular fluid extending through a complete plane of separation at the labral–chondral transitional zone (LCTZ), from the articular surface to the subchondral line (long arrow). (c) Sagittal T1 wi fat-suppressed MRA shows contrast media filling the full thickness cartilage lesion (long arrow). Intra-articular injection of contrast solution enlarges the joint and perilabral recess (curved arrow). Also a small artifact due to a small intra-articular air bubble is visible (dashed arrow). (d) Arthroscopic image shows a small delamination cartilage tear at the LCTZ (black arrow). Table 2 Comparison between 3-T non-contrast MRI and 1.5-T MRA of the hip and arthroscopy, using arthroscopy as standard of reference, for depiction of lesions at the acetabular labrum and the labral-chondral transitional zone. Hip arthroscopy Acetabular labrum
Labral-chondral transitional zone
Degenerative
Normal
Total
Tear
Normal
Total
3-T non-contrast MRI
Positive Negative Total
32 4 36
3 11 14
35 15 50
42 1 43
0 7 7
42 8 50
1.5-T direct MRA
Positive Negative Total
31 5 36
7 7 14
38 12 50
43 0 43
1 6 7
44 6 50
Table 3 Statistical indicators of the comparison between hip 3-T non-contrast MRI and 1.5-T MRA with hip arthroscopy for diagnosis of labral and labral-chondral transitional zone pathology, using hip arthroscopy as the reference standard. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), accuracy and also Cohen’s kappa index were calculated with a 95% confidence interval (CI).
Sensitivity Specificity PPV NPV Accuracy Kappa coefficient
Acetabular labrum abnormalities% (CI 95%)
Tears at the labral-chondral transitional zone% (CI 95%)
3T non-contrast MRI
1.5-T MRA
3T non-contrast MRI
1.5-T MRA
88.9 (77.2–100) 78.6 (53.5–100) 91.4 (80.7–100) 73.3 (47.6–99) 86 (75.4–96,6) 0.66 (0.42−0.89)
86.1 (73.4–98.8) 50 (20.2–79.7) 81.6 (67.9–95.2) 58.3 (26.3–90.4) 76 (63.16–88.8) 0.37 (0.09–0.66)
97.7 (92–100) 100 (92.8–100) 100 (98.8–100) 87.5 (58.3–100) 98 (93.12–100) 0.92 (0.76–1)
100 (98.8–100) 85.7 (52.6–100) 97.7 (92.2–100) 100 (91.7–100) 98 (93.1–100) 0.91 (0.73–1)
A.M. Crespo-Rodríguez et al. / European Journal of Radiology 88 (2017) 109–116
113
Fig. 2. Posterior paralabral cyst in a thirty-five-year-old male with a pincer FAI. (a) Sagittal WATS and (b) sagital PD-HR wi show fluid signal intensity next to the posterior acetabular margin (long arrow). (c) Sagittal reformatted THRIVE MRA image demonstrates a cystic lesion (long arrow) full-filled of contrast media between the labrum (curved arrow) and the posterior acetabulum. (d) Arthroscopic image shows the posterior paralabral cyst adyacent to the labrum (black arrow).
interest in different techniques of hip MRI as alternative, less invasive, methods that can establish diagnoses of disorders of the hip [19], provide guidance to surgeons as to the likely site of labral and chondral abnormalities and possibly enable surgeons to decide pre-operatively on the appropriate surgical technique. Direct MRA is the preferred method because it is better at enlarging the joint and providing images of small internal structures, allowing a better evaluation of diseases or conditions within the joint. Direct MRA at 1.5-T has been shown to be a reliable method for defining and detecting acetabular labral tears [9]. A systematic review of the literature and meta-analysis found a pooled sensitivity of approximately 87% and a pooled specificity about 64% for labral tears [6,9,14,20–25]. Tables 4 and 5 summarize Smith et al. [9] and add later relevant studies [18,26,27], all of them with more than fifty patients. However, sensitivity of direct MRA in the detection of cartilage damage is only moderate [6]. Indirect MRA is less invasive than the direct technique, does not require fluoroscopy or a physician to perform the injection, is relatively easy to schedule and has improved contrast resolution relative to non-contrast MRI. However, the disadvantages of indirect MRA, as compared with direct MRA, include the lack of capsular distention, higher contrast load, and less reliability in rendering the joint opaque. Although we do not have experience with this technique, few reports in the literature indicate favourable sensitivity
and specificity values with indirect MRA, whether with 1.5-T [28] or with 3-T magnets [29]. The same meta-analysis also examined the diagnostic performance of 1.5-T non-contrast MRI for acetabular labral tears and chondral abnormalities (Table 5), finding a pooled sensitivity of 66% and a pooled specificity of 79% [9]. Recently, two retrospective studies, both with high resolution MRI protocols, have reported a better diagnostic performance of 1.5-T non-contrast MRI for acetabular labral tears (pooled sensitivity of 96%) and chondral abnormalities[4,30]. To our knowledge, only two retrospective studies have compared the diagnostic performance of non-contrast MRI and MRA at a 3-T magnetic field for detection of acetabular labral tears and chondral defects in the same patient population [26,27] with divergent results. On the one hand, Tian et al. [26] reported that 3.0-T non-contrast MRI was not effective enough for evaluating the acetabular labral tears, with a relatively low sensitivity and specificity (Table 5), leading to the conclusion that 3.0-T MRA of the hip is a reliable evaluation modality for diagnosing acetabular labral tears, with a superior diagnostic performance (Table 4). On the other hand, Magee [27] reported that 3.0-T non-contrast MRI was near equivalent to 3.0-T MRA (Table 4 and 5) for detection of acetabular labral tears. Only in one patient out of forty-three, 3.0-T MRA allowed for detection of an acetabular labral tear not seen on 3.0-T non-contrast MR examination.
114
A.M. Crespo-Rodríguez et al. / European Journal of Radiology 88 (2017) 109–116
Fig. 3. Full-thickness cartilage lesion in a forty-six-year-old female with pincer FAI. (a) Coronal T1-wi of MRA shows a small full-thickness cartilage lesion that seems to be an acetabular cartilage defect (long arrow); acetabular labrum with normal signal intensity (dashed arrow) and perilabral recess (curved line) are also identified. On 3-T non-contrast MRI, (b) HR-PD wi with FS enables differentiation of a cartilage delamination tear: A hyperintense line extending from a disrupted cartilage surface deep to the acetabular cartilage layers is visible (long arrow). (c) HR-PD wi with FS on a sagittal view demonstrates a full-thickness cartilage delamination tear (long arrow). (d) The artrhoscopic image shows a small delamination tear at the acetabular cartilage (black arrow).
Table 4 Sensitivity and specificity for the diagnostic accuracy of direct MR arthrography (MRA) of the hip for detecting acetabular labral tears in different studies, modified from Smith et al. [9]. Magnetic field strength, tesla (T), patients, true positive (TP), false positive (FP), false negative (FN), true negative (TN), sensitivity and specificity. Study
Magnetic field streght
Patients
TP
FP
FN
TN
Sensitivity
Specificity
Czerny et al 1999 [14] Schmid et al 2003 [6] Keeney et al 2004 [19] Byrd and Jones 2004 [20] Chan et al 2005 [21] Freedman et al 2006 [22] Toomayan et al 2006 [23] Studler et al 2008 [24] Ziegert et al 2009[25] Tian et al 2014 [26] Magee 2015 [27] Crespo-Rodríguez et al 2015 [18]
1.0-T 1.0-T 1.5-T 1.5-T 1.5-T 1.5-T 1.5-T 1.5-T 1.5-T 3.0-T 3.0-T 1.5-T
40 40 102 40 17 24 30 57 144 34 43 51
30 14 66 23 16 22 22 43 140 19 38 43
2 5 5 7 1 1 0 6 0 2 1 1
3 4 27 9 0 1 2 1 4 2 4 0
5 17 4 1 0 0 6 7 0 11 0 7
91% 78% 71% 72% 100% 96% 92% 98% 97% 90,5% 90% 100%
71% 77% 44% 13% 0% 100% 54% 84,6% 87,5%
Table 5 Sensitivity and specificity for the diagnostic accuracy of non-contrast MRI of the hip for detecting acetabular labral tears in different studies, modified from Smith et al. [9]. Magnetic field strength, tesla (T), patients, true positive (TP), false positive (FP), false negative (FN), true negative (TN), sensitivity and specificity. Study
Magnetic field streght
Patients
TP
FP
FN
TN
Sensitivity
Specificity
Byrd and Jones 2004 [20] Mintz et al 2005 [30] Tian et al 2014 [26] Magee 2015 [27] Crespo-Rodríguezet al 2016
1.5-T 1.5-T 3.0-T 3.0-T 3.0-T
40 92 90 43 50
7 86 36 39 42
4 2 7 1 0
25 3 23 3 1
4 1 24 0 7
22% 97% 61.2% 93% 97.7%
50% 33% 77.4% 100%
A.M. Crespo-Rodríguez et al. / European Journal of Radiology 88 (2017) 109–116
115
Fig. 4. Small “pocket” cartilage lesion in a 40-year-old man with cam FAI in the left hip. (a) Sagittal and (b) coronal HR-PD wi with FS at a 3-T magnet show hiperintensity of the articular fluid within the small cleft at the LCTZ (long arrows), an abnormal signal intensity within the labrum related to degeneration (dashed arrow) and hypointensity foci within de acetabular cartilage (short arrow). (c) Sagittal reformatted THRIVE image of MRA shows contrast media full-filling a small cleft at the LCTZ (long arrow) and extending to the labrum (dashed arrow). (d) Arthroscopic image shows a small “pocket” cartilage lesion (black arrow) immediately adjacent to attachment of the labrum (L) to the acetabular cartilage (C).
In our opinion, moving to higher field strengths (specifically 3T) improves the SNR, which should then be able to be traded off for higher image resolution. Overall, the data being reported here show an excellent diagnostic performance for 3-T non-contrast MRI. If positive imaging findings are yielded by 3-T MRI, 1.5-T MRA does not provide any additional information. With regard to normal variations in the acetabular labrum and the difficulty of distinguishing them from labral tears, 3-T non-contrast MRI can provide valuable additional information for a differential diagnosis. It can accurately identify cartilage abnormalities next to the cleft within the LCTZ in labral tears, while in sub-labral sulcus a cleft within the LCTZ is seen without there being any cartilage abnormality. The issue that remains is whether MRA must be retained in daily practice for cases with negative findings on 3-T non-contrast MRI but a strong suspicion of FAI. Some authors consider that a negative imaging study, even at 3-T, does not completely exclude major intra-articular pathology that can be identified and treated arthroscopically [18,27]. However, it is possible to avoid performing an invasive procedure in at least 80% of patients with positive findings. Hence, our imaging protocol for patients with hip pain and clinical suspicion of FAI is first radiographs, followed by 3-T non-contrast MRI of the hip. In young patients (under 45 years old) with a clinical strong suspicion of FAI and normal findings on 3-T MRI would it proceed to MRA. The next question to be addressed would be whether
3-T MRA should be performed on this reduced group of patients. In the rest of patients, an individualized re-evaluation of clinical settings, diagnostic tools and treatment options is recommended. A number of limitations in this study must be acknowledged. The first is the small number of patients involved. The second, selection of patients with radiological signs of FAI and the high mean age of the group can explain the high rate of LCTZ injury. The third is that no control group was included in the study, because healthy individuals do not undergo hip arthroscopy. The fourth is that the surgeon was aware of the MRI results and this information was used in pre-operative planning. The last, radiologists knew that all patients underwent arthroscopy at the time they reviewed the images for this paper. In conclusion, non-invasive assessment of the hip is possible with 3-T MR. MRA can be replaced by 3-T non-contrast MRI as the workhorse technique for the assessment of internal disorders of the hip. MRA would be reserved for young adults with a strong clinical suspicion of FAI, but negative findings on 3-T non-contrast MRI. Optimized sequences on 3-T non-contrast MRI add value in detecting normal variations, such as sub-labral sulcus, and in diagnosing articular cartilage lesions in the hip, as compared with 1.5-T MRA.
116
A.M. Crespo-Rodríguez et al. / European Journal of Radiology 88 (2017) 109–116
Conflict of interest The authors and author’s institutions have no conflicts of interest. Acknowledgement The authors gratefully acknowledge Luis Ramos González, PhD MD, former Chairman of the Department of Radiology at the Hospital Universitario Puerta de Hierro Majadahonda, Madrid, Spain, for his support in this scientific paper and for encouraging everyone to pursue excellence in the patient care. The authors gratefully acknowledge Isabel Millán Santos, PhD for her contribution to statistical analysis of this paper. References [1] R. Ganz, J. Parvizi, M. Beck, M. Leunig, H. Notzli, K.A. Siebenrock, Femoroacetabular impingement: a cause for osteoarthritis of the hip, Clin. Orthop. Relat. Res. 417 (2003) 112–120, http://dx.doi.org/10.1097/01.blo. 0000096804.78689.c2. [2] C.W. Pfirrmann, B. Mengiardi, C. Dora, F. Kalberer, M. Zanetti, J. Hodler, Cam and pincer femoroacetabular impingement: characteristic MR arthrographic findings in 50 patients, Radiology 240 (2006) 778–785, http://dx.doi.org/10. 1148/radiol.2403050767. [3] J.C. McCarthy, P.C. Noble, M.R. Schuck, J. Wright, J. Lee, The Otto E. Aufranc Award: the role of labral lesions to development of early degenerative hip disease, Clin. Orthop. Relat. Res. 393 (2001) 25–37. [4] S.L. James, K. Ali, F. Malara, D. Young, J. O’Donnell, D.A. Connell, MRI findings of femoroacetabular impingement, Am. J. Roentgenol. 187 (2006) 1412–1419, http://dx.doi.org/10.2214/AJR.05.1415. [5] M. Beck, M. Leunig, J. Parvizi, V. Boutier, D. Wyss, R. Ganz, Anterior femoroacetabular impingement: part II. Midterm results of surgical treatment, Clin. Orthop. Relat. Res. 429 (2004) 170–177. [6] M.R. Schmid, H.P. Nötzli, M. Zanetti, T.F. Wyss, J. Hodler, Cartilage lesions in the hip: diagnostic effectiveness of MR arthrography, Radiology 226 (2003) 382–386, http://dx.doi.org/10.1148/radiol.2262020019. [7] K.A. Jung, C. Restrepo, M. Hellman, H. AbdelSalam, W. Morrison, J. Parvizi, The prevalence of cam-type femoroacetabular deformity in asymptomatic adults, J. Bone Joint Surg. Br. 93 (2011) 1303–1307, http://dx.doi.org/10.1302/0301620X.93B10.26433. [8] D.R. Griffin, E.J. Dickenson, J. O’Donnell, R. Agricola, T. Awan, M. Beck, et al., The Warwick Agreement on femoroacetabular impingement syndrome (FAI syndrome): an international consensus statement, Br. J. Sports Med. 50 (2016) 1169–1176, http://dx.doi.org/10.1136/bjsports-2016-096743. [9] T. Smith, G. Hilton, A.P. Toms, S.T. Donell, C.B. Hing, The diagnostic accuracy of acetabular labral tears using magnetic resonance imaging and magnetic resonance arthrography: a meta-analysis, Eur. Radiol. 21 (2011) 863–874, http://dx.doi.org/10.1007/s00330-010-1956-7. [10] P. Robinson, Conventional 3-T MRI and 1.5-T MR arthrography of femoroacetabular impingement, Am. J. Roentgenol. 199 (2012) 509–515, http://dx.doi.org/10.2214/AJR.12.8672. [11] A. Mohr, F.W. Roemer, H.K. Genant, C. Liess, Using fatsaturated proton density-weighted MR imaging to evaluate articular cartilage, Am. J. Roentgenol. 181 (2003) 280–281, http://dx.doi.org/10.2214/ajr.181.1. 1810280a. [12] C.A. Petersilge, M.A. Haque, W.J. Petersilge, J.S. Lewin, J.M. Lieberman, R. Buly, Acetabular labral tears: evaluation with MR arthrography, Radiology 200 (1996) 231–235, http://dx.doi.org/10.1148/radiology.200.1.8657917.
[13] L.A. Lage, J.V. Patel, R.N. Villar, The acetabular labral tear: an arthroscopic classification, Arthroscopy 12 (1996) 269–272, http://dx.doi.org/10.1016/ S0749-8063(96)90057-2. [14] C. Czerny, S. Hofmann, M. Urban, C. Tschauner, A. Neuhold, M. Pretterklieber, M.P. Recht, J. Kramer, MR arthrography of the adult acetabular capsular-labral complex: correlation with surgery and anatomy, Am. J. Roentgenol. 173 (1999) 345–349, http://dx.doi.org/10.2214/ajr.173.2.10430132. [15] D.G. Blankenbaker, A.A. De Smet, J.S. Keene, J.P. Fine, Classification and localization of acetabular labral tears, Skeletal Radiol. 36 (2007) 391–397, http://dx.doi.org/10.1007/s00256-006-0240-z. [16] D.A. McCall, M.R. Safran, MRI and arthroscopy correlations of the hip: a case-based approach, Instr. Course Lect. 61 (2012) 327–344. [17] R.E. Outerbridge, The etiology of chondromalacia patellae, J. Bone Joint Surg. Br. 43 (1961) 752–757. [18] A.M. Crespo Rodríguez, J.C. De Lucas Villarrubia, M.A. Pastrana Ledesma, Millán Santos I, M. Padrón, Diagnosis of lesions of the acetabular labrum, of the labral-chondral transition zone, and of the cartilage in femoroacetabular impingement: correlation between direct magnetic resonance arthrography and hip arthroscopy, Radiologia 57 (2015) 131–141. [19] J.A. Keeney, M.W. Peelle, J. Jackson, D. Rubin, W.J. Maloney, J.C. Clohisy, Magnetic resonance arthrography versus arthroscopy in the evaluation of articular hip pathology, Clin. Orthop. Relat. Res. 429 (2004) 163–169. [20] J.W. Byrd, K.S. Jones, Diagnostic accuracy of clinical assessment, magnetic resonance imaging, magnetic resonance arthrography, and intra-articular injection in hip arthroscopy patients, Am. J. Sports Med. 32 (2004) 1668–1674. [21] Y.S. Chan, L.C. Lien, H.L. Hsu, Y.L. Wan, M.S. Lee, K.Y. Hsu, C.H. Shih, Evaluating hip labral tears using magnetic resonance arthrography: a prospective study comparing hip arthroscopy and magnetic resonance arthrography diagnosis, Arthroscopy 21 (2005) 1250, http://dx.doi.org/10.1016/j.arthro.2005.07.007. [22] B.A. Freedman, B.K. Potter, P.A. Dinauer, J.R. Giuliani, T.R. Kuklo, K.P. Murphy, Prognostic value of magnetic resonance arthrography for Czerny stage II and III acetabular labral tears, Arthroscopy 22 (2006) 742–747, http://dx.doi.org/ 10.1016/j.arthro.2006.03.014. [23] G.A. Toomayan, W.R. Holman, N.M. Major, S.M. Kozlowicz, T.P. Vail, Sensitivity of MR arthrography in the evaluation of acetabular labral tears, Am. J. Roentgenol. 186 (2006) 449–453, http://dx.doi.org/10.2214/AJR.04.1809. [24] U. Studler, F. Kalberer, M. Leunig, M. Zanetti, J. Hodler, C. Dora, C.W. Pfirrmann, MR arthrography of the hip: differentiation between an anterior sublabral recess as a normal variant and a labral tear, Radiology 249 (2008) 947–954, http://dx.doi.org/10.1148/radiol.2492080137. [25] A.J. Ziegert, D.G. Blankenbaker, A.A. De Smet, J.S. Keene, K. Shinki, J.P. Fine, Comparison of standard hip MR arthrographic imaging planes and sequences for detection of arthroscopically proven labral tear, Am. J. Roentgenol. 192 (2009) 1397–1400, http://dx.doi.org/10.2214/AJR.08.1609. [26] C.Y. Tian, J.Q. Wang, Z.Z. Zheng, AH Ren, 3.0T conventional hip MR and hip MR arthrography for the acetabular labral tears confirmed by arthroscopy, Eur. J. Radiol. 83 (2014) 1822–1827, http://dx.doi.org/10.1016/j.ejrad.2014.05.034. [27] T. Magee, Comparison of 3. 0-T MR vs 3. 0-T MR arthrography of the hip for detection of acetabular labral tears and chondral defects in the same patient population, Br. J. Radiol. 88 (2015) 20140817, http://dx.doi.org/10.1259/bjr. 20140817. [28] M.B. Zlatkin, D. Pevsner, T.G. Sanders, C.R. Hancock, C.E. Ceballos, M.F. Herrera, Acetabular labral tears and cartilage lesions of the hip: indirect MR arthrographic correlation with arthroscopy—a preliminary study, Am. J. Roentgenol. 194 (2010) 709–714, http://dx.doi.org/10.2214/AJR.07.3669. [29] C.N. Petchprapa, L.D. Rybak, K.S. Dunham, R. Lattanzi, M.P. Recht, Labral and cartilage abnormalities in young patients with hip pain: accuracy of 3-Tesla indirect MR arthrography, Skeletal Radiol. 44 (2015) 97–105, http://dx.doi. org/10.1007/s00256-014-2013-4. [30] D.N. Mintz, T. Hooper, D. Connell, R. Buly, D.E. Padgett, H.G. Potter, Magnetic resonance imaging of the hip: detection of labral and chondral abnormalities using noncontrast imaging, Arthroscopy 21 (2005) 385–393, http://dx.doi. org/10.1016/j.arthro.2004.12.011.