Abstract

The primary purpose was to answer the following question: What is the location and pattern of necrosis and associated chondrolabral lesions and can they be accurately detected on traction MR arthrography compared with intra-operative findings in patients undergoing hip preservation surgery for femoral head necrosis (FHN)? Retrospective, diagnostic case series on 23 patients (23 hips; mean age 29 ± 6 years) with diagnosis of FHN undergoing open/arthroscopic joint preserving surgery for FHN and pre-operative traction MR arthrography of the hip. A MR-compatible device for weight-adapted application of leg traction (15–23 kg) was used and coronal, sagittal and radial images were acquired. Location and pattern of necrosis and chondrolabral lesions was assessed by two readers and compared with intra-operative findings to calculate diagnostic accuracy of traction MR arthrography. On MRI all 23 (100%) hips showed central FHN, most frequently antero-superiorly (22/23, 96%) where a high prevalence of femoral cartilage damage was detected (18/23, 78%), with delamination being the most common (16/23, 70%) damage pattern. Intra-operative inspection showed central femoral head cartilage damage most frequently located antero-superiorly (18/23, 78%) with femoral cartilage delamination being most common (14/23, 61%). Traction MR arthrography enabled detection of femoral cartilage damage with a sensitivity/specificity of 95%/75% for reader 1 and 89%/75% for reader 2. To conclude, femoral cartilage damage occurs at the zone of necrosis and can be accurately detected using traction MR arthrography of the hip which may be helpful for surgical decision making in young patients with FHN.

INTRODUCTION

Femoral head necrosis (FHN) is a major cause of osteoarthritis of the hip joint [1]. In young patients, efforts are made to halt the course of osteoarthritis. Two distinct surgical principles have been applied traditionally to treat FHN. Core decompression [2, 3] and vascularized bone grafting [3, 4] aim to restore femoral head blood supply. Rotational [5, 6] or angular [7] femoral osteotomies are performed to position healthy bone into the main weight-bearing areas to prevent collapse and allow healing of the necrotic area. Historically these procedures have been performed from an extra-articular approach. More recently, open [8–14] and arthroscopic [15–17] procedures for surgical treatment of FHN have been described which offer direct access to the joint for bone grafting and chondrolabral repair [18].

To date, the optimal surgical treatment for FHN remains controversial [1]. This may be related to the fact that surgical planning often relies on imaging studies which allow staging of the disease but do not provide further information on location and extent of intra-articular damage [19]. More specifically the use of standard imaging MRI planes does not enable an exact localization of necrosis [19]. In contrast radial images which are oriented perpendicular to the femoral neck could potentially allow an accurate anatomic allocation of FHN, comparable to imaging of the cam deformity [20]. Currently, a detailed pre-operative image analysis of intra-articular damage in FHN is lacking [19], despite its detrimental effect on outcome of joint preserving surgery for FHN [21]. Such information could be valuable for planning of chondrolabral repair surgery [18]. MRI of the hip, even after intra-articular contrast injection often fails to detect hip cartilage damage [22, 23]. Typically, the tight coaptation of the hip prevents contrast accumulation which makes differentiation of cartilage layers difficult [22]. Recently application of leg traction during MR arthrography of the hip was introduced to overcome this problem and has shown promising first results in detection of intra-articular lesions in FAI [24–26]. Therefore, it was hypothesized that traction MR arthrography with radial imaging could be useful in the pre-operative characterization of FHN and associated chondrolabral lesions.

The primary purpose was to answer the following question: What is the location and pattern of necrosis and associated chondrolabral lesions and can they be accurately detected on traction MR arthrography compared with intra-operative findings in patients undergoing hip preservation surgery for FHN?

MATERIALS AND METHODS Patients

After IRB-approval a retrospective, diagnostic study on patients with FHN diagnosed was performed between January 2016 and June 2019 at a tertiary hip preservation centre. Pre-operative diagnosis of FHN was based on the presence of a geographical low signal band on MRI according to the revised Association Research Circulation Osseous (ARCO) [27]. The institutional database was searched for patients undergoing open or arthroscopic reconstructive hip surgery for treatment of FHN with complete intra-articular inspection. Out of 27 patients (27 hips), 23 patients (23 hips) had pre-operative traction MR arthrography of the hip according to the institutional routine protocol and were included in the study.

Mean age was 29 ± 6 years (11 men, 45%; Table I). Most frequently FNH occurred idiopathic (14/23, 61%). The majority of hips (16/23; 70%) were diagnosed with advanced ARCO Stage III [27]. Nine hips (9/23, 39%) had Tönnis Grade 0, and 14 hips (14/23, 61%) had Tönnis Grade 1 of radiographic osteoarthritis [28] (Table I).

Table I.

Demographic and radiographic characteristics

Patients (hips) 23 (23) Age (years) 29 ± 6 Men (%) 11 (45) Body mass index (kg/m2) 25 ± 3 Tönnis grade of osteoarthritis, hips (%) Grade 0: 9 (39) Grade 1: 14 (61) LCE angle (°) 28 ± 9 Hip dysplasia (LCE < 25°), hips (%) 7 (30) Deep hip (LCE > 39°), hips (%) 1 (4) Retroversion (positive COS, PWS and ISS signs), hips (%) 3 (12) Alpha angle (°) 74 ± 12 Cam deformity (α > 60°), hips (%) 21 (91) Aetiology, hips (%) 14 (61) idiopathic 6 (26) systemic/toxic 3 (13) post-traumatic 2019 revised ARCO criteria (0–IV), hips (%) I (normal X-ray, MRI abnormal): 2 (9) II (X-ray abnormal, MRI abnormal): 5 (22) IIIA (subchondral fracture, depression ≤ 2 mm): 8 (35) IIIB (subchondral fracture, depression > 2 mm): 8 (35) Previous surgery, hips (%) 2 (9) core decompression Surgical approaches, hips (%) 8 (35) surgical hip dislocation 6 (26) surgical hip dislocation, femoral osteotomy 6 (26) hip arthroscopy, core decompression 3 (13) surgical hip dislocation, periacetabular osteotomy Femoral osteochondroplasty, hips (%) 21 (91) Chondrolabral repair, hips (%) 7 (30) curettage of necrosis, bone grafting and cartilage suture 3 (13) curettage of necrosis, bone grafting, AMIC 3 (13) subchondral drilling femoral head 1 (4) curettage of necrosis and osteochondral autograft 5 (22) subchondral drilling acetabulum 9 (39) labrum refixation Patients (hips) 23 (23) Age (years) 29 ± 6 Men (%) 11 (45) Body mass index (kg/m2) 25 ± 3 Tönnis grade of osteoarthritis, hips (%) Grade 0: 9 (39) Grade 1: 14 (61) LCE angle (°) 28 ± 9 Hip dysplasia (LCE < 25°), hips (%) 7 (30) Deep hip (LCE > 39°), hips (%) 1 (4) Retroversion (positive COS, PWS and ISS signs), hips (%) 3 (12) Alpha angle (°) 74 ± 12 Cam deformity (α > 60°), hips (%) 21 (91) Aetiology, hips (%) 14 (61) idiopathic 6 (26) systemic/toxic 3 (13) post-traumatic 2019 revised ARCO criteria (0–IV), hips (%) I (normal X-ray, MRI abnormal): 2 (9) II (X-ray abnormal, MRI abnormal): 5 (22) IIIA (subchondral fracture, depression ≤ 2 mm): 8 (35) IIIB (subchondral fracture, depression > 2 mm): 8 (35) Previous surgery, hips (%) 2 (9) core decompression Surgical approaches, hips (%) 8 (35) surgical hip dislocation 6 (26) surgical hip dislocation, femoral osteotomy 6 (26) hip arthroscopy, core decompression 3 (13) surgical hip dislocation, periacetabular osteotomy Femoral osteochondroplasty, hips (%) 21 (91) Chondrolabral repair, hips (%) 7 (30) curettage of necrosis, bone grafting and cartilage suture 3 (13) curettage of necrosis, bone grafting, AMIC 3 (13) subchondral drilling femoral head 1 (4) curettage of necrosis and osteochondral autograft 5 (22) subchondral drilling acetabulum 9 (39) labrum refixation Table I.

Demographic and radiographic characteristics

Patients (hips) 23 (23) Age (years) 29 ± 6 Men (%) 11 (45) Body mass index (kg/m2) 25 ± 3 Tönnis grade of osteoarthritis, hips (%) Grade 0: 9 (39) Grade 1: 14 (61) LCE angle (°) 28 ± 9 Hip dysplasia (LCE < 25°), hips (%) 7 (30) Deep hip (LCE > 39°), hips (%) 1 (4) Retroversion (positive COS, PWS and ISS signs), hips (%) 3 (12) Alpha angle (°) 74 ± 12 Cam deformity (α > 60°), hips (%) 21 (91) Aetiology, hips (%) 14 (61) idiopathic 6 (26) systemic/toxic 3 (13) post-traumatic 2019 revised ARCO criteria (0–IV), hips (%) I (normal X-ray, MRI abnormal): 2 (9) II (X-ray abnormal, MRI abnormal): 5 (22) IIIA (subchondral fracture, depression ≤ 2 mm): 8 (35) IIIB (subchondral fracture, depression > 2 mm): 8 (35) Previous surgery, hips (%) 2 (9) core decompression Surgical approaches, hips (%) 8 (35) surgical hip dislocation 6 (26) surgical hip dislocation, femoral osteotomy 6 (26) hip arthroscopy, core decompression 3 (13) surgical hip dislocation, periacetabular osteotomy Femoral osteochondroplasty, hips (%) 21 (91) Chondrolabral repair, hips (%) 7 (30) curettage of necrosis, bone grafting and cartilage suture 3 (13) curettage of necrosis, bone grafting, AMIC 3 (13) subchondral drilling femoral head 1 (4) curettage of necrosis and osteochondral autograft 5 (22) subchondral drilling acetabulum 9 (39) labrum refixation Patients (hips) 23 (23) Age (years) 29 ± 6 Men (%) 11 (45) Body mass index (kg/m2) 25 ± 3 Tönnis grade of osteoarthritis, hips (%) Grade 0: 9 (39) Grade 1: 14 (61) LCE angle (°) 28 ± 9 Hip dysplasia (LCE < 25°), hips (%) 7 (30) Deep hip (LCE > 39°), hips (%) 1 (4) Retroversion (positive COS, PWS and ISS signs), hips (%) 3 (12) Alpha angle (°) 74 ± 12 Cam deformity (α > 60°), hips (%) 21 (91) Aetiology, hips (%) 14 (61) idiopathic 6 (26) systemic/toxic 3 (13) post-traumatic 2019 revised ARCO criteria (0–IV), hips (%) I (normal X-ray, MRI abnormal): 2 (9) II (X-ray abnormal, MRI abnormal): 5 (22) IIIA (subchondral fracture, depression ≤ 2 mm): 8 (35) IIIB (subchondral fracture, depression > 2 mm): 8 (35) Previous surgery, hips (%) 2 (9) core decompression Surgical approaches, hips (%) 8 (35) surgical hip dislocation 6 (26) surgical hip dislocation, femoral osteotomy 6 (26) hip arthroscopy, core decompression 3 (13) surgical hip dislocation, periacetabular osteotomy Femoral osteochondroplasty, hips (%) 21 (91) Chondrolabral repair, hips (%) 7 (30) curettage of necrosis, bone grafting and cartilage suture 3 (13) curettage of necrosis, bone grafting, AMIC 3 (13) subchondral drilling femoral head 1 (4) curettage of necrosis and osteochondral autograft 5 (22) subchondral drilling acetabulum 9 (39) labrum refixation  Diagnostic imaging

All patients underwent standardized supine radiographic imaging of the pelvis including antero-posterior (AP) pelvis views and cross-table lateral views [29]. All patients underwent traction MR arthrography following fluoroscopy guided injection. Intra-articular needle position was confirmed with 1–2 ml of iodinated contrast agent (iopamidol, 200 mg/ml; Iopamiro 200; Bracco) and 15–20 ml of diluted MR contrast agent (gadoteric acid, 0.0025 mmol/ml; Guerbet) were injected. MRI was performed at 3T (Skyra, Siemens Medical Solutions) using a large flexible body coil. An MR-compatible traction device (trac-view, Menges Medical) was used which includes a pulley, a cable connected to the weight and a supporting plate for the contralateral foot to prevent tilting of the pelvis [30, 31]. Traction load was adjusted depending on patient’s weight: for patients weighing <60/60–80/>80 kg were used 15/18/23 kg, respectively. Coronal and sagittal proton-density (PD) weighted turbo spin-echo images without fat-saturation were acquired: repetition time (TR)/echo time (TE), 2600/11 ms, slice thickness of 2 mm, 170 × 170 mm field of view, a matrix size of 269 × 384, acquisition time (AT) of 3–4 min. Radial PD-weighted turbo spin-echo images without fat-saturation were oriented around the femoral neck axis [20, 32]: TR/TE, 2100/14 ms, slice thickness of 4 mm, 170 × 170 mm field of view, a matrix size of 403 × 448, AT of 5 min for 12 slices.

Surgical treatment

Twenty-three patients (23 hips) underwent joint preserving surgery, which was performed by three experienced hip surgeons (Table I). In six hips (6/23, 26%) hip arthroscopy was performed including core decompression. In 17 hips (17/23, 74%), surgical hip dislocation was performed with direct treatment of the necrotic lesion at the femoral head performed in 14 hips (14/23, 61%). Most commonly curettage of the necrotic bone was performed, followed by impaction bone grafting and suturing of the femoral head cartilage (7/23, 30%) (Fig. 1) or autologous matrix-assisted chondrogenesis (3/23, 13%) [10]. Labral refixation was performed in nine hips (9/23, 39%) and subchondral drilling of the acetabulum was performed in five hips (5/23, 22%). Weight-bearing of the necrotic area was reduced with a concomitant femoral osteotomy in four hips [10]: three hips (3/23, 13%) inter-trochanteric varus-derotation osteotomies, one hip (1/23, 4%) posterior rotational femoral neck osteotomy. In two hips (8%), subtrochanteric derotation osteotomies were performed for correction of excessively high femoral version [33, 34]. Femoral osteochondroplasty was performed in 21 hips (21/23, 91%). In all patients, presence, location and damage pattern of intra-articular lesions was recorded intra-operatively according to the grading of Beck et al. [35]. For comparison with imaging findings surgical records were reviewed by an orthopaedic surgeon and femoroacetabular cartilage was graded: intact/delamination/thinning/defect. The acetabular labrum was graded as intact/chondrolabral separation/complex tear.

Fig. 1.

(A) AP pelvis radiograph of a 43-year-old man with FHN ARCO Stage IIIB who underwent a ‘trapdoor procedure’ for osteochondral repair via a surgical hip dislocation. (B) Coronal PD-w MR arthrogram which was performed in an external institution shows no obvious femoral cartilage lesion. (C) MR arthrography was repeated with traction. Joint distraction was achieved, contrast agent entered the central compartment of the hip and a large centrally located cartilage flap became visible (arrowhead). (D) Extensive femoral cartilage delamination was confirmed at surgery. The cartilage flap was held aside (asterisk), to perform a curettage of the necrotic bone until bleeding was observed. The defect was filled with an autologous bone graft. (E) Final result after re-suturing of the cartilage flap (asterisk).

Fig. 1.

(A) AP pelvis radiograph of a 43-year-old man with FHN ARCO Stage IIIB who underwent a ‘trapdoor procedure’ for osteochondral repair via a surgical hip dislocation. (B) Coronal PD-w MR arthrogram which was performed in an external institution shows no obvious femoral cartilage lesion. (C) MR arthrography was repeated with traction. Joint distraction was achieved, contrast agent entered the central compartment of the hip and a large centrally located cartilage flap became visible (arrowhead). (D) Extensive femoral cartilage delamination was confirmed at surgery. The cartilage flap was held aside (asterisk), to perform a curettage of the necrotic bone until bleeding was observed. The defect was filled with an autologous bone graft. (E) Final result after re-suturing of the cartilage flap (asterisk).

Image analysis

Radiographs were assessed by one senior orthopaedic surgeon for staging of FHN according to the 2019 revised ARCO grading [27]. The lateral-centre edge angle was measured, presence of acetabular retroversion (cross-over sign, posterior-wall sign and ischial-spine sign) was assessed [36] and the Tönnis grade of osteoarthritis [28] was evaluated. MRI analysis was performed independently from each other, blinded to the initial report and the surgical report by a radiology resident and a senior radiologist (7 and 19 years of experience).

Radial images were used to assess the location of FHN at each half-hour position of the clock-face. Image criterion for FHN was defined as the presence of a geographical, hypointense line in the femoral head and was recorded dichotomously (yes/no) [19]. On each radial image a line perpendicular to the femoral neck axis was drawn for definition of central (medial to the perpendicular line) or peripheral (lateral to the perpendicular line) FHN (Fig. 2). Anatomic allocation of FNH was performed using a previously described approach (see below) [20].

Fig. 2.

Differentiation between peripheral and central location of FHN and femoral cartilage damage is shown on a radial image. This is based on a line which is drawn perpendicular to the femoral neck axis. If the lesion is medial/lateral to this line it is referred to as being located central/peripheral, respectively. In this hip peripheral and central FHN (dashed line) was present at the 2 o’clock and at the 8 o’clock position.

Fig. 2.

Differentiation between peripheral and central location of FHN and femoral cartilage damage is shown on a radial image. This is based on a line which is drawn perpendicular to the femoral neck axis. If the lesion is medial/lateral to this line it is referred to as being located central/peripheral, respectively. In this hip peripheral and central FHN (dashed line) was present at the 2 o’clock and at the 8 o’clock position.

Coronal-, sagittal- and radial- traction MR images were used to assess chondrolabral lesions. Definite anatomic allocation of lesions to the corresponding clock-face positions was performed on the 12 radial images using a previously validated technique [20, 32]: Each image represents a ‘half-hour’ position and can be allocated according anatomic landmarks. For necrosis and femoral cartilage damage the greater trochanter served as landmark for the femoral 12 o’clock position. For acetabular cartilage and labrum lesions the acetabular teardrop served as landmark for the acetabular 6 o’clock position. Accordingly, the acetabulum and femur were divided into four quadrants: 12–2:30 o’clock = antero-superior; 3–5:30 o’clock = antero-inferior; 6–8:30 o’clock = postero-inferior and 9–11:30 o’clock = postero-superior.

Femoral and acetabular cartilage lesions were graded using previously described classifications: delamination (subchondral contrast accumulation), thinning (partial-thickness loss of cartilage), defect (full-thickness loss of cartilage; Fig. 3) [24, 37]. Femoral cartilage lesions were further divided into peripheral and centrally located lesions. Labrum lesions were graded as chondrolabral separation (contrast extending into the labral base) or complex tear (contrast extension into the labral base and within the labral substance) [24, 37].

Fig. 3.

Schematic drawings of femoral cartilage damage and corresponding appearance of femoral cartilage lesions on traction MR arthrography. (A) Delamination: subchondral contrast accumulation (arrowhead). (B) Thinning: partial-thickness loss of cartilage (arrowheads). (C) Defect: full-thickness loss of cartilage (arrowhead).

Fig. 3.

Schematic drawings of femoral cartilage damage and corresponding appearance of femoral cartilage lesions on traction MR arthrography. (A) Delamination: subchondral contrast accumulation (arrowhead). (B) Thinning: partial-thickness loss of cartilage (arrowheads). (C) Defect: full-thickness loss of cartilage (arrowhead).

Statistical analysis

Descriptive statistics including mean and SD were calculated. Inter-observer reproducibility and intra-observer reliability were assessed using Cohen’s kappa: < 0 poor agreement; 0.0–0.2, slight agreement; 0.21–0.4 fair agreement; 0.41–0.6, moderate agreement; 0.61–0.8, substantial agreement; 0.81–1, almost perfect agreement [38]. Sensitivity, specificity, accuracy, positive predictive value and negative predictive value of traction MR arthrography for detecting intra-articular lesions was calculated. SPSS version 25.0 (IBM, Armonk, New York, USA) was used for statistical analysis.

RESULTS

On MRI all 23 (100%) hips showed central FHN, which was most frequently located in the antero-superior (22/23, 96%) quadrant. Peripheral FHN was present in 8/23 (35%) hips, most frequently located in the antero-superior and -inferior quadrants (each 3/34, 13%; Fig. 4).

Fig. 4.

Location of necrosis (left image), femoral cartilage damage assessed on MRI (middle image) and intra-operatively (right image) are shown from an antero-posterior (images on top) and lateral perspective (images at bottom). PS indicates the postero-superior quadrant. Necrosis was most commonly (96%) located at central antero-superior femoral head, the location where femoral cartilage damage was most commonly observed on MRI and intra-operatively (each 78%). Among the observed lesions femoral cartilage delamination was the most common damage pattern on MRI (70%) and at surgical inspection (61%).

Fig. 4.

Location of necrosis (left image), femoral cartilage damage assessed on MRI (middle image) and intra-operatively (right image) are shown from an antero-posterior (images on top) and lateral perspective (images at bottom). PS indicates the postero-superior quadrant. Necrosis was most commonly (96%) located at central antero-superior femoral head, the location where femoral cartilage damage was most commonly observed on MRI and intra-operatively (each 78%). Among the observed lesions femoral cartilage delamination was the most common damage pattern on MRI (70%) and at surgical inspection (61%).

On traction MR arthrography most hips (19/23, 83%) demonstrated central femoral cartilage damage, most frequently located antero-superiorly (18/23, 78%). Femoral cartilage delamination was the most commonly (16/23, 70%) observed damage pattern (Fig. 4). Corresponding to that, intra-operative inspection showed central femoral head cartilage damage in the majority of the hips (19/23, 83%) most frequently located antero-superiorly (18/23, 78%), with femoral cartilage delamination being most common (14/23, 61%; Fig. 4). Traction MR arthrography enabled detection of femoral cartilage damage with a sensitivity/specificity of 95%/75% for Reader 1 and 89%/75% for Reader 2 (Table II; Figs 1 and 5). MRI grading of femoral cartilage lesions was consistent with surgical grading in 70% (16/23) of the cases for reader 1 (κ = 0.427, P = 0.002) and in 61% (14/23) of the cases for reader 2 (κ = 0.403, P = 0.001; Table III).

Fig. 5.

AP pelvis radiograph of a 36-year-old woman with FHN ARCO Stage IIIA undergoing a surgical hip dislocation and autologous matrix-induced chondrogenesis. (B) Coronal PD-w MR arthrogram which was performed in an external institution shows no obvious femoral cartilage lesion. (C) Two months later MR arthrography was repeated with leg traction. Minimal joint distraction was achieved but extensive subchondral contrast accumulation (arrowheads) became visible indicating femoral cartilage delamination involving two thirds of the weight-bearing surface. (D, E) Delaminated femoral cartilage got completely detached during the dislocation manoeuvre and was found lying in the acetabulum (arrowheads) leaving a large cartilage femoral cartilage defect (asterisk).

Fig. 5.

AP pelvis radiograph of a 36-year-old woman with FHN ARCO Stage IIIA undergoing a surgical hip dislocation and autologous matrix-induced chondrogenesis. (B) Coronal PD-w MR arthrogram which was performed in an external institution shows no obvious femoral cartilage lesion. (C) Two months later MR arthrography was repeated with leg traction. Minimal joint distraction was achieved but extensive subchondral contrast accumulation (arrowheads) became visible indicating femoral cartilage delamination involving two thirds of the weight-bearing surface. (D, E) Delaminated femoral cartilage got completely detached during the dislocation manoeuvre and was found lying in the acetabulum (arrowheads) leaving a large cartilage femoral cartilage defect (asterisk).

Table II.

Diagnostic performance of traction MR arthrography in detection of chondrolabral lesions compared with surgery

Parameter . Femoral cartilage . Acetabular cartilage . Labrum . Reader 1 . Reader 2 . Reader 1 . Reader 2 . Reader 1 . Reader 2 . No. of true-positives, (no. of hips) 18 17 18 19 20 21 No. of false-negatives, (no of hips) 1 2 2 1 1 0 No. of true-negatives, (no. of hips) 3 3 2 1 2 2 No. of false-positives (no. of hips) 1 1 1 2 0 0 Sensitivity (%) 95 (74–100) 89 (67–99) 90 (68–99) 95 (75–100) 95 (76–100) 100 (84–100) Specificity (%) 75 (19–99) 75 (19–99) 67 (9–`99) 33 (1- 91) 100 (16–100) 100 (16–100) Positive predictive value (%) 95 (77–99) 94 (76–99) 95 (78–99) 90 (81–96) 100 (100–100) 100 (100–100) Negative predictive value (%) 75 (29–96) 60 (26–86) 50 (18–82) 50 (8–92) 67 (23–93) 100 (100–100) Accuracy (%) 91 (72–99) 87 (66–97) 87 (66–97) 87 (66–97) 96 (78–100) 100 (85–100) Inter-observer agreement (Cohen’s κ)* 0.588 0.417 0.633 Parameter . Femoral cartilage . Acetabular cartilage . Labrum . Reader 1 . Reader 2 . Reader 1 . Reader 2 . Reader 1 . Reader 2 . No. of true-positives, (no. of hips) 18 17 18 19 20 21 No. of false-negatives, (no of hips) 1 2 2 1 1 0 No. of true-negatives, (no. of hips) 3 3 2 1 2 2 No. of false-positives (no. of hips) 1 1 1 2 0 0 Sensitivity (%) 95 (74–100) 89 (67–99) 90 (68–99) 95 (75–100) 95 (76–100) 100 (84–100) Specificity (%) 75 (19–99) 75 (19–99) 67 (9–`99) 33 (1- 91) 100 (16–100) 100 (16–100) Positive predictive value (%) 95 (77–99) 94 (76–99) 95 (78–99) 90 (81–96) 100 (100–100) 100 (100–100) Negative predictive value (%) 75 (29–96) 60 (26–86) 50 (18–82) 50 (8–92) 67 (23–93) 100 (100–100) Accuracy (%) 91 (72–99) 87 (66–97) 87 (66–97) 87 (66–97) 96 (78–100) 100 (85–100) Inter-observer agreement (Cohen’s κ)* 0.588 0.417 0.633 Table II.

Diagnostic performance of traction MR arthrography in detection of chondrolabral lesions compared with surgery

Parameter . Femoral cartilage . Acetabular cartilage . Labrum . Reader 1 . Reader 2 . Reader 1 . Reader 2 . Reader 1 . Reader 2 . No. of true-positives, (no. of hips) 18 17 18 19 20 21 No. of false-negatives, (no of hips) 1 2 2 1 1 0 No. of true-negatives, (no. of hips) 3 3 2 1 2 2 No. of false-positives (no. of hips) 1 1 1 2 0 0 Sensitivity (%) 95 (74–100) 89 (67–99) 90 (68–99) 95 (75–100) 95 (76–100) 100 (84–100) Specificity (%) 75 (19–99) 75 (19–99) 67 (9–`99) 33 (1- 91) 100 (16–100) 100 (16–100) Positive predictive value (%) 95 (77–99) 94 (76–99) 95 (78–99) 90 (81–96)