Improved Cartilage Quality on Delayed Gadolinium-Enhanced MRI of Hip Cartilage after Subchondral Drilling of Acetabular Cartilage Flaps in Femoroacetabular Impingement Surgery at Minimum 5-Year Follow-Up

Abstract

Objective

To assess whether subchondral drilling of acetabular cartilage flaps during femoroacetabular impingement (FAI) surgery improves (1) acetabular dGEMRIC indices and (2) morphologic magnetic resonance imaging (MRI) scores, compared with hips in which no additional treatment of cartilage lesions had been performed; and (3) whether global dGEMRIC indices and MRI scores correlate.

Design

Prospective cohort study of consecutive patients with symptomatic FAI treated with open surgery between 2000 and 2007. Patients with subchondral drilling of acetabular cartilage flaps were allocated to the study group, those without drilling to the control group. All patients underwent indirect 3-T MR arthrography to assess cartilage quality by dGEMRIC indices and a semiquantitative morphologic MRI score at minimum 5 years after surgery. dGEMRIC indices and morphologic MRI scores were compared between and among groups using analysis of covariance/paired t tests.

Results

No significant difference was found between the global dGEMRIC indices of the study group (449 ± 147 ms, 95% CI 432-466 ms) and the control group (428 ± 143 ms, 95% CI 416-442 ms; P = 0.235). In regions with cartilage flaps, the study group showed higher dGEMRIC indices (472 ± 160 ms, 95% CI 433-510 ms) compared with the control group (390 ± 122 ms, 95% CI 367-413 ms; P < 0.001). No significant differences were found for the morphologic MRI scores. A strong inversely linear correlation between the dGEMRIC indices and the morphologic MRI scores (r = −0.727, P < 0.001) was observed.

Conclusions

Treatment of acetabular cartilage flaps with subchondral drilling leads to better cartilage quality in regions with cartilage flaps at minimum 5 years of follow-up.

Introduction

Morphological changes of the hip joint in cam-type femoroacetabular impingement (FAI) are a common cause of hip pain and a risk factor for the development of hip osteoarthritis.^1,2 Shear forces induced by the repetitive inclusion of the aspherical femoral head into the acetabulum typically lead to a carpet like delamination of the acetabular cartilage creating cartilage flaps and detachment of the labrum at its base. ³ Surgical treatment aims to improve the range of motion by restoring the normal shape of the femoral neck. ⁴ While the benefit of labrum repair is commonly accepted,^5,6 indications and optimal techniques for the repair of different stages of cartilage damage such as cartilage flaps in cartilage delamination are debated. ⁷ Among possible surgical techniques, piercing of the subchondral bone with an awl (microfracturing) or a drill (subchondral drilling) are cost-effective and easy to perform 1-stage procedures to treat cartilage lesions. ⁸ Treatment is based on the rationale of creating a bone marrow clot within the cartilage defect and promote differentiation of pluripotent mesenchymal stem cells into stable fibrocartilaginous repair tissue. ⁹ In the knee joint, good long-term results for treatment of small lesions in otherwise healthy joints have been reported after microfracturing.^10,11 In the hip joint, however, long-term outcome for treatment of focal cartilage damages, especially for cartilage debonding or cartilage flaps, is sparse. ¹²

In a previous study of FAI patients with acetabular cartilage flaps caused by FAI, a lower rate of conversion to total joint replacement following subchondral drilling of the cartilage flaps compared with slight debridement only was reported. Despite that, the clinical scores and the progression of osteoarthritis on conventional radiographs in the remaining patients with a preserved hip did not differ. ¹³ There are however obvious limitations of conventional radiography in detecting focal changes of cartilage quality. Magnetic resonance imaging (MRI) is more sensitive in detecting subtle morphologic patterns of early joint degeneration secondary to FAI using semiquantitative scores. ¹⁴ The morphologic MRI scores include evaluation and grading of damages to the labrum and cartilage, bonce cysts, as well as acetabular and femoral osteophytes. ¹⁴ Furthermore, newer histologically validated quantitative MRI techniques such as delayed gadolinium enhanced MRI of cartilage (dGEMRIC) provide a compositional analysis of cartilage quality and are increasingly used in the setting of FAI.^15-17 However, neither morphologic MRI scores nor dGEMRIC have been used to date to investigate the status of joint degeneration following subchondral drilling of cartilage flaps of the hip.

Thus, we asked whether subchondral drilling of acetabular cartilage flaps during open FAI surgery improves global and regional (1) acetabular dGEMRIC indices and (2) morphologic MRI scores at minimum 5 years of follow-up compared with hips in which no additional treatment of the cartilage lesions had been performed; and (3) whether there is a correlation between global dGEMRIC indices and morphologic MRI scores.

Methods

Study Design and Patients

After institutional review board approval, we performed a prospective, diagnostic, cohort study using biochemical and morphologic MRI to assess the effect of subchondral drilling for treatment of acetabular cartilage damage following written informed consent. Inclusion criteria were (1) all consecutive patients treated at our institution with open surgical treatment of symptomatic FAI between January 2000 and December 2007, (2) intraoperative visually verified acetabular cartilage flaps, (3) a complete set of pre- and postoperative anterior-posterior pelvis and cross-table lateral hip radiographs, (4) indirect MR arthrography, including a morphologic MRI protocol and the acquisition of 3-dimensional (3D) dGEMRIC maps for biochemical assessment of hip cartilage at last clinical examination at minimum 5 years after surgery. Patients in whom subchondral drilling had been performed were allocated to the study group (28 patients, 29 hips), those with no specific treatment for the cartilage damage to the control group (43 patients, 51 hips). 4 patients (5 hips) were lost to follow-up. At latest follow-up, none of the patients in the study group had converted to total hip arthroplasty, while 7 patients (8 hips) in the control group had undergone total hip arthroplasty at a mean of 6.1 years after surgery and were excluded. This left 36 patients (43 hips) with a preserved hip joint in the control group. dGEMRIC was performed in 33 patients (37 hips): 15 patients (15 hips) of the study group and 19 patients (22 hips) of the control group. One patient had hips treated in both groups ( Fig. 1 ).

Figure 1.

Flowchart showing the inclusion criteria and exclusion criteria.

Groups were comparable for age, sex, body mass index, size of cartilage flap, pre- and postoperative radiographic hip anatomy and Tönnis grade of osteoarthritis. ¹⁸ Time of follow-up was 10 ± 2 years (95% CI 9-11 years) for the control group versus 8 ± 1 years (95% CI 7-9 years) for the study group (P = 0.003) ( Tables 1 and 2 ). Demographics, clinical scores as well as pre- and postoperative radiological morphological parameters between the patients who agreed to the dGEMRIC scan and those who did not were compared to evaluate a possible selection bias ( Tables 3 and 4 ).

Table 1.

Demographic Characteristics of Study and Control Groups. ^a

Parameters	Study Group	Control Group	P
No. of patients (hips)	15 (15)	19 (22)	—
Age at surgery (years)	33 ± 7 (28-37)	31 ± 10 (27-35)	0.575
Females (%)	13 (−7 to 33)	41 (19-63)	0.141
Height (cm)	175 ± 7 (171-179)	176 ± 9 (172-180)	0.737
Weight (kg)	78 ± 16 (70-87)	75 ± 11 (70-80)	0.457
Body mass index (kg/m2)	26 ± 6 (23-29)	24 ± 4 (23-26)	0.379
Follow-up (years)	8 ± 1 (7-9)	10 ± 2 (9-11)	0.003
Size of cartilage flap (cm2)	5 ± 2 (4-6)	6 ± 2 (5-6)	0.313

^aValues are shown as means ± standard deviations with 95% confidence intervals in parentheses.

Table 2.

Radiographic Characteristics of the Patients in the Study Group and Control Group Before and After FAI Surgery. ^a

Parameters	Preoperative			Postoperative
	Study Group	Control Group	P	Study Group	Control Group	P
Alpha angle (deg)	74 ± 8 (66-80)	62 ± 16 (51-72)	0.052	39 ± 6 (35-42)	39 ± 5 (37-41)	0.932
Lateral center edge angle (deg)	30 ± 3 (27-32)	33 ± 8 (28-38)	0.165	27 ± 3 (25-28)	29 ± 5 (27-32)	0.179
Acetabular index (deg)	7 ± 3 (5-9)	7 ± 6 (3-10)	0.996	7 ± 3 (5-9)	6 ± 5 (4-8)	0.432
Hips with positive crossover sign (%)	55 (15096)	33 (2-64)	0.333	20 (−3 to 43)	27 (7-47)	0.624
Hips with positive ischial-spine sign (%)	22 (−11 to 56)	33 (2-65)	0.599	20 (−3 to 43)	50 (27-73)	0.068
Hips with Tönnis grade > 0 (%)	33 (6-60)	18 (0-36)	0.438	47 (18-75)	32 (11-53)	1
Hips with Tönnis grade > 1 (%)	—	—	—	7 (−7 to 20)	18 (1-35)	0.629
Hips with radiographic OA progression (%)	—	—	—	20 (−3 to 43)	27 (7-47)	0.711

FAI = femoroacetabular impingement; OA = osteoarthritis.

^aValues are shown as means ± standard deviations with 95% confidence intervals in parentheses.

Table 3.

Comparison of Demographic Data and Clinical data Between the Patients with and without a dGEMRIC Scan. ^a

Parameters	Subchondral Drilling (29 Hips)		P	No Subchondral Drilling (51 Hips)		P
dGEMRIC Scan at Follow-Up	Yes, Study Group	No, Study Group		Yes, Control Group	No, Control Group
No. of hips (% of overall group)	15 (52)	14 (48)	—	22 (43)	29 (57)	—
Age (years)	33 ± 7 (28-37)	33 ± 8 (28-38)	0.956	31 ± 10 (27-35)	32 ± 8 (29-35)	0.685
Percentage of females, %	13 (1-25)	21 (6-36)	0.600	41 (23-59)	31 (23-59)	0.557
Height (cm)	175 ± 7 (171-179)	179 ± 7 (174-183)	0.213	176 ± 9 (172-180)	177 9 (174-181)	0.550
Weight (kg)	78 ± 16 (70-87)	80 ± 14 (72-89)	0.730	75 ± 11 (70-80)	78 ± 14 (73-83)	0.450
Body mass index (kg/m2)	26 ± 6 (23-29)	25 ± 4 (23-27)	0.792	24 ± 4 (23-26)	24 ± 4 (23-26)	0.725
Follow-up (years)	8 ± 1 (7-9)	7± 3 (6-9)	0.460	10 ± 2 (9-11)	7 ± 5 (5-8)	0.004
HHS at latest follow-up (0-100 points)	82 ± 16 (73-91)	93 ± 9 (87-98)	0.076	83 ± 13 (77-90)	88 ± 13 (82-94)	0.260
Merlde d’Aubigne score (0-18 points)	16 ± 2 (15-17)	17 ± 1 (16-18)	0.110	16 ± 1 (15-17)	17 ± 2 (16-17)	0.148

dGEMRIC = delayed gadolinium enhanced MRI of cartilage; HHS = Harris Hip Score.

^aValues are shown as means ± standard deviations with 95% confidence intervals in parentheses.

Table 4.

Comparison of Radiographic Characteristics of the Patients With and Without a dGEMRIC Scan.^a

Parameters	Subchondral Drilling (29 Hips)		P	No Subchondral Drilling (51 Hips)		P
dGEMRIC Scan at Follow-Up	Yes, Study Group	No		Yes, Control Group	No
No. of hips (% of overall group)	15 (52)	14 (48)		22 (43)	29 (57)
Preoperatively
Alpha angle (deg)	74 ± 8 (66-80)	62 ± 12 (52-71)	0.041	62 ± 16 (51-72)	59 ± 13 (51-67)	0.648
Lateral center edge angle (deg)	30 ± 3 (27-32)	29 ± 4 (26-32)	0.766	33 ± 8 (28-38)	34 ± 8 (30-37)	0.883
Acetabular index (deg)	7 ± 3 (5-9)	6 ± 4 (3-9)	0.615	7 ± 6 (3-10)	4 ± 6 (1-7)	0.144
Hips with positive crossover sign (%)	55 (15-96)	64 (39-90)	0.714	33 (2-64)	61 (43-79)	0.136
Hips with positive ischial-spine sign (%)	22 (−11 to 56)	64 (39-90)	0.092	33 (2-65)	61 (43-79)	0.136
Hips with Tönnis grade >0 (%)	33 (6-60)	21 (0-42)	0.681	18 (0-36)	38 (20-56)	0.125
Hips with Tönnis grade >1 (%)	0	0	1	0	3 (−3 to 9)	1
Postoperatively
Alpha angle (deg)	39 ± 6 (35-42)	38 ± 6 (34-41)	0.803	39 ± 5 (37-41)	39 ± 5 (37-41)	0.860
Lateral center edge angle (deg)	27 ± 3 (25-28)	26 ± 3 (24-28)	0.414	29 ± 5 (27-32)	29 ± 7 (26-31)	0.695
Acetabular index (deg)	7 ± 3 (5-9)	7 ± 4 (4-9)	0.617	6 ± 5 (4-8)	6 ± 5 (4-7)	0.575
Hips with positive crossover sign (%)	20 (−3 to 43)	21 (0-42)	0.924	27 (7-47)	17 (3-31)	0.388
Hips with positive ischial-spine sign (%)	20 (−3 to 43)	64 (39-90)	0.016	50 (27-73)	45 (27-63)	0.714
Hips with Tönnis grade >0 (%)	47 (18-75)	44 (18-70)	0.916	32 (11-53)	53 (35-71)	0.177
Hips with Tönnis grade >1 (%)	7 (−7 to 20)	11 (−5 to 27)	1	18 (1-35)	32 (15-49)	0.319
Hips with radiographic OA progression (%)	20 (−3 to 43)	22 (0-44)	1	27 (7-47)	37 (19-55)	0.511

dGEMRIC = delayed gadolinium enhanced MRI of cartilage; OA, osteoarthritis.

Values are shown as means with standard deviation and 95% confidence intervals in parentheses.

Diagnosis and Surgical Treatment of Femoroacetabular Impingement

Diagnosis of FAI was based on patient history, a clinical examination including a positive anterior impingement test and radiographic findings of cam-type FAI. An alpha angle >50° on a cross-table lateral view was defined as cam morphology. ¹⁹ During the study period, surgical hip dislocation was the standard treatment for patients with FAI syndrome. Surgical hip dislocations were performed according to a previously described technique by surgeons experienced in hip preservation surgery. ²⁰ This included femoral osteochondroplasty for resection of the cam deformity and/or acetabular rim trimming and reattachment of the labrum. Evaluation of the acetabular cartilage lesions was performed according to the Beck classification and the extent and depth of the lesions were recorded. ²¹ Size of the cartilage flaps and distribution of cartilage lesions was comparable between the 2 groups and was calculated according to a previously described method ( Table 1 ). ¹³ At the time of surgery, there was no evidence defining the optimal treatment of acetabular cartilage flaps. Institutional standard treatment at that time was either minimal debridement of the cartilage flap and additional subchondral drilling of the delaminated area (study group) or retaining the cartilage flap with minimal debridement and no further treatment (control group). The decision to perform additional subchondral drilling was made intraoperatively. The cartilage flap was lifted and holes were drilled with a 1.5-mm drill bit in the subchondral bone 3- to 4-mm apart until bleeding was observed. The cartilage flap was positioned back in place, the femoral head was reduced and the osteotomy of the greater trochanter was fixed with cortical screws. The postoperative protocol was identical for all patients and included partial weightbearing for 6 weeks followed by a clinical and radiologic examination at our outpatient clinic. Physical therapy and full weightbearing was commenced after consolidation of the greater trochanter was confirmed. No patient underwent revision FAI surgery. At the latest follow-up patients underwent clinical examination by orthopedic residents who were not involved in the surgical treatment of the patients. Anterior-posterior pelvic views and cross-table lateral views were obtained in all patients for evaluation of signs of osteoarthritis and the anatomy of the acetabulum (lateral center edge angle, acetabular index, presence/absence of the cross-over and ischial-spine sign) and the proximal femur (alpha angle). ¹⁹ Radiographic parameters were comparable between the study group and the control group at baseline and postoperatively regarding the degree of osseous correction ( Table 2 ).

Magnetic Resonance Imaging and Image Analysis

All patients underwent indirect MR arthrography using intravenous injection of 0.2 mmol/kg Gd-DTPA²⁻ (Magnevist; Bayer, Leverkusen, Germany) for acquisition of morphologic and dGEMRIC images using the same 3-T unit (Siemens Trio; Siemens, Erlangen, Germany) with flexible surface body coils. After injection, patients were instructed to walk around for at least 15 minutes to facilitate diffusion of contrast agent into the cartilage. After further 10 to 30 minutes MRI was performed with a standard, multiplanar protocol of 2D coronal-, sagittal-, radial proton-density (PD) weighted turbo spin echo images without fat saturation. ²² Sequence parameters for the 2D radial PD-weighted images were as follows: repetition time (TR)/echo time (TE) 1500 ms/18 ms, 4 mm slice thickness, 16 × 16 cm field of view, matrix size of 448 × 317, acquisition time (AT) of 4:30 minutes. ²³ The dGEMRIC sequence was performed 45 to 70 minutes after injection of the contrast agent and included a standard dual-flip angle T1 volume interpolated breath-examination (VIBE) sequence ²⁴ : TR/TE, 15 ms/3.3 ms, flip angles of 4° and 24°, slice thickness of 0.8 mm, 16 × 16 cm field of view, matrix size of 192 × 192, in-plane resolution of 0.83 × 0.83 mm interpolated to 0.4 × 0.4 mm, AT of 8.46 minutes for 128 slices. ²⁵

The analysis of MR images was performed by one reader (FS) with extensive experience in MR imaging of the hip joint and who was not involved in the surgical management of the patients. For analysis of dGEMRIC indices, a commercially available software (Osirix Version 6.0; Geneva, Switzerland) was used. ²⁶ The 3D dGEMRIC volume was reformatted to obtain 12 radial images oriented perpendicular to the femoral neck axis, analogously to the orientation of the 2D PD-weighted radial images, which served as morphologic guidance. ²⁵ For anatomic allocation of radial MR images, a previously established and validated method was used. ²³ The midpoint of the acetabular notch corresponds to the 6 o’clock position with 12 o’clock being directly opposed to it. Three o’clock indicates anterior and 9 o’clock corresponds to posterior independently of the side of the hip ( Fig. 2 ). ²³ For measurement of dGEMRIC indices a previously described method was used^25,27: At each clock face position, regions of interest (ROIs) were manually placed and further divided into equally sized peripheral and central zones, resulting in a total of 20 dGEMRIC measurements resulting from 20 ROIs per hip joint( Fig. 3 ). The regions for the 5 and 6 o’clock positions were not assessed, as they are devoid from cartilage.

Figure 2.

Image on the right shows radial images oriented perpendicular to the femoral neck axis. Left image shows clock-face definitions based on anatomic landmarks and definitions of peripheral and central regions of interest for assessment of dGEMRIC indices. ps, postero-superior; pi, postero-inferior; as, antero-superior; ai: antero-inferior; AN, acetabular notch. Reprinted with permission from Schmaranzer et al. ²⁵

Figure 3.

Two representative examples illustrating the assessment of morphologic MRI scores and dGEMRIC indices. (A, B) A patient who had undergone surgical hip dislocation without subchondral drilling at the age of 25 years. (C, D) A patient in whom surgical hip dislocation and subchondral drilling had been performed at the age of 29 years. (A-D) Both patients had surgically confirmed cartilage flaps extending from 11 o’clock to 3’ o clock. The 12 o’clock position was identified on (A, C) radial proton density–weighted (PD-w) images and on the (B, D) radially reformatted dGEMRIC maps using the acetabular teardrop as anatomical landmark (dashed lines). (A) Ten years after surgery without subchondral drilling morphologic MRI damage score was 3 points (intralabral tear [arrow] =1 point, acetabular rim osteophyte [asterisk] = 1 point, focal cartilage delamination [arrowhead] = 1 point]), (B) peripheral (p) and central (c) dGEMRIC indices were 325 ms and 302 ms. (C) In the patient in which subchondral drilling had been performed morphologic damage score was similar with 3 points 9 years postoperatively (intralabral tear [arrow] = 1 point, acetabular rim bone cyst [asterisk] = 1 point, focal cartilage thinning [arrowhead] = 1point]) but (D) peripheral (p) and central (c) dGEMRIC indices were higher, 482 ms and 375 ms, respectively.

To obtain morphologic scores of degenerative hip disease, a previously validated semi-quantitative grading scheme for radial imaging was used, which had been specifically designed to quantify early, pre-arthritic joint degeneration in patients with FAI and hip dysplasia. ¹⁴ On each of the 10 “full-hour” positions of the acetabulum, the following parameters were assessed and graded on PD-weighted radial images: labrum damage, paralabral cysts, cartilage damage, acetabular center bone cyst, acetabular rim bone cysts, femoral bone cysts, acetabular rim osteophytes, and femoral osteophytes ( Table 5 , Fig. 3 ). This results in a possible MRI score of 10 points per radial image and clock face position and eventually resulted in a total possible range of MRI score of 0 (normal joint) to 100 points (severe degenerative changes) for the entire joint.

Table 5.

Morphologic Magnetic Resonance Imaging Score According to Stelzeneder et al. ^{14 a}

Parameters	Normal	Pathologic
Labrum damage	No (0)	Intralabral tear (1)	Full-thickness tear (2)
Paralabral cysts	No (0)	Yes (1)
Cartilage damage	No (0)	Focal defect (1)	Generalized defect (2) b
Acetabulum center bone cysts	No (0)	Yes (1)
Acetabulum rim bone cysts	No (0)	Yes (1)
Femoral bone cysts	No (0)	Yes (1)
Acetabular rim osteophytes	No (0)	Yes (1)
Femoral osteophytes	No (0)	Yes (1)

^aNormal and pathological parameters with allocated points in parentheses. The value of all parameters summed up to a possible score between 0 (normal hip) and 10 points per region of interest or 0 to 100 points per joint (severe degenerative changes), respectively.

^bFull-thickness defect or at least cartilage thinning or delamination >1 cm in length.

To evaluate if subchondral drilling of the acetabular cartilage flaps improved global acetabular dGEMRIC indices (study question 1) or morphologic MRI scores (study question 2), mean values of the entire joint were compared between the study group and the control group. For the regional comparison, surgical reports were reviewed for the positions of zones with intact cartilage and zones with cartilage flaps. The corresponding clock-face positions ( Fig. 2 ) were defined as “intact cartilage” or “cartilage flap.” The dGEMRIC indices and MRI scores of each individual clock-face position were then grouped accordingly. Mean values were compared between the study group and the control group.

To evaluate if global dGEMRIC indices and morphologic MRI scores correlate (study question three), the Spearman correlation coefficient was calculated.

Statistical Analysis and Visualization of Results

Normal distribution of continuous variables was tested with the Kolmogorov-Smirnov test. Parametric variables were compared with paired and unpaired Student’s t tests. Nonparametric variables were compared with the Wilcoxon rank-sum test for paired data and the Mann-Whitney U test for unpaired data. For comparison of dGEMRIC indices and morphologic MRI scores among the study groups, we performed an analysis of covariance to adjust for differences in follow up between surgery and the MRI scan. Categorical data was compared with the chi-square test. Spearman correlation coefficient (r) was used to assess correlation between dGEMRIC indices and morphologic MRI scores and was graded as follows: r < 0.2 very weak; 0.20 to 0.39 for weak; 0.40 to 0.59 for moderate; 0.60 to 0.79 for strong; and ≥0.8 for very strong correlation. ²⁸ Interobserver reliability of the dGEMRIC measurements and the morphologic MRI scores was determined using intraclass correlation coefficients (ICC) and Cohen’s kappa (κ) based a random sample of 46 clock-face positions which were analyzed by a second reader (MSH). Four months following the initial analysis the first reader repeated the assessment of dGEMRIC indices and morphologic MRI scores for evaluation of intraobserver reliability. The required sample was based on a power analysis of measurements on 46 clock-face positions to achieve an ICC >0.80 with an alpha error of 0.05 and 80% power. Statistical analysis was performed using WinSTAT (Version 2012.1, Bad Krozingen, Germany, 2012). A P value of < 0.05 was defined as statistically significant.

For a more intuitive topographical visualization of dGEMRIC indices, MRI scores and relative frequency of intraoperatively documented cartilage damage were visualized as interpolated surface color plots using MATLAB (The MathWorks, Inc., Natick, MA, USA) on planar maps. ²⁵ A color spectrum from blue to red was used. High dGEMRIC indices and low morphologic MRI scores were coded in blue (good cartilage quality) while low dGEMRIC indices and high morphologic MRI scores were coded in red (degenerated cartilage). To map the intraoperative evaluation of cartilage damage, regions with a high frequency cartilage flaps were colored in red, those regions without flaps were illustrated in blue.

Results

Delayed Gadolinium-Enhanced MRI of Cartilage

No significant difference was found between the global dGEMRIC indices of the study group (449 ± 147 ms, 95% CI 432-466 ms) and the control group (428 ± 143 ms, 95% CI 416-442 ms; P = 0.235). In regions with cartilage flaps, the study group showed higher dGEMRIC indices (472 ± 160 ms, 95% CI 433-510 ms) compared to regions with cartilage flaps in the control group (390 ± 122 ms, 95% CI 367-413 ms; P < 0.001) ( Table 6 , Fig. 4 ). No significant difference in dGEMRIC indices in regions with intact cartilage was found between the groups (study group 442 ± 143 ms, 95% CI 424-461 ms; control group 442 ± 147 ms, 95% CI 426-458 ms; P = 0.556). Within the control group, higher dGEMRIC scores were observed in regions with intact cartilage compared to regions with cartilage flaps (mean difference 50 ± 171 ms, 95% CI 17-82 ms; P < 0.001). No significant difference was observed in the study group (mean difference −29 ± 182 ms, 95% CI −70 to 20 ms; P = 0.306) ( Table 6 , Fig. 4 ).

Table 6.

Comparison of dGEMRIC Indices and Morphologic MRI Scores Between the Study Group and the Control Group. ^a

MRI Parameter	Region	Study Group	Control Group	P
dGEMRIC index (ms)	Global	449 ± 147 (432-466)	428 ± 143 (416-442)	0.235
	Cartilage flaps	472 ± 160 (433-510)	390 ± 122 b (367-413)	<0.001
	Intact cartilage	442 ± 143 (424-461)	442 ± 147 b (426-458)	0.556
Morphologic MRI score (0-100)	Global	18 ± 7 (14-22)	20 ± 10 (15-24)	0.453
	Cartilage flaps	10 ± 5 (8-13)	13 ± 6 b (10-16)	0.116
	Intact cartilage	8 ± 7 (4-12)	8 ± 7 b (5-11)	0.800

MRI = magnetic resonance imaging; dGEMRIC = delayed gadolinium enhanced MRI of cartilage.

^aValues are shown as means ± standard deviations with 95% confidence intervals in parentheses.

^bThere was a significant difference of the scores of regions with cartilage flaps and intact cartilage within the control group (P < 0.001 for the dGEMRIC index and P = 0.001 for the morphologic MRI score); total possible range of MRI score 0 (normal joint) to 100 points (severe degenerative change).

Figure 4.

Topographical distribution of intraoperatively documented acetabular cartilage damage (upper row), dGEMRIC indices and morphologic MRI damage scores of the control group and study group. Size of the cartilage flaps was comparable for both groups (see also Table 1). For both groups, the cartilage flaps were most commonly found in the anterior-superior quadrant of the acetabulum, while no damage was present posterior-inferiorly. dGEMRIC indices and morphologic MRI damage scores were divided in regions with cartilage flaps and regions with intact cartilage according to the intraoperative findings. Accordingly, topographical distribution of the color plots do not overlap between regions with cartilage flaps and intact cartilage. Cartilage quality as assessed with dGEMRIC in regions with cartilage flaps was better in the study group compared to the control group. More severe morphologic and biochemical cartilage damage in regions with cartilage flaps compared to regions with intact cartilage were observed in the control group.

Morphologic MRI Scores

No significant difference was found between the of global morphologic MRI scores of the study group (18 ± 7 points, 95% CI 14-22 points) and the control group (20 ± 10 points, 95% CI 15-24 points; P = 0.453). In regions with cartilage flaps, the study group showed no significant difference in morphologic MRI scores (10 ± 5 points, 95% CI 8-13 points) compared with the control group (13 ± 6 points, 95% CI 10-16 points; P = 0.116) ( Table 6 , Fig. 4 ). No significant difference in morphologic MRI scores between both groups was observed in regions with intact cartilage (study group 8 ± 7 points, 95% CI 4-12 points; control group 8 ± 7 points, 95% CI 5-11 points; P = 0.800). Within the control group, higher morphologic MRI scores were observed in regions with cartilage flaps compared to regions with intact cartilage (mean difference 6 ± 6 points, 95% CI 3-8 points; P = 0.001). No significant difference was observed in the study group (mean difference 1 ± 12 points, 95% CI −6 to 8 points; P = 0.062) ( Table 6 , Fig. 4 ). No significant difference between both groups was observed for any of the subcategories of the morphologic MRI score ( Table 7 ).

Table 7.

Results for Morphologic Magnetic Resonance Imaging (MRI) Scoring of Both Groups.

Morphologic MRI Score a	Study Group (15 Hips)		Control Group (22 Hips)		P
	Mean ± SD (95% CI)	No. of Hips (%)	Mean ± SD (95% CI)	No. of hips (%)
Labrum damage (0-20 points)	4 ± 2 (3-5)	15 (100)	4 ± 2 (3-5)	22 (100)	0.602
Paralabral cyst (0-10 points)	0 ± 0 (0-1)	3 (20)	1 ± 1 (0-1)	6 (27)	0.535
Cartilage damage (0-20 points)	5 ± 3 (3-7)	14 (93)	5 ± 3 (3-6)	22 (100)	0.861
Acetabular center bone cysts (0-10 points)	0 ± 0 (0-0)	3 (20)	0 ± 1 (0-1)	6 (27)	0.496
Acetabular rim bone cysts (0-10 points)	0 ± 1 (0-1)	5 (33)	1 ± 1 (0-1)	10 (45)	0.328
Femoral bone cysts (0-10 points)	0 ± 1 (0-1)	3 (20)	0 ± 1 (0-1)	3 (14)	0.682
Acetabular rim osteophytes (0-10 points)	6 ± 2 (5-7)	15 (100)	6 ± 2 (5-7)	22 (100)	0.423
Femoral osteophytes (0-10 points)	2 ± 2 (1-4)	11 (73)	3 ± 3 (1-4)	15 (68)	0.833
Total MRI score (0-100 points)	18 ± 7 (14-22)	—	20 ± 10 (15-24)	—	0.5

^aTotal possible range of MRI score 0 (normal joint) to 100 points (severe degenerative change).

A strong inversely linear correlation between the dGEMRIC indices and the morphologic MRI scores (r_S = −0.727, P < 0.001) was observed ( Fig. 5 ).

Figure 5.

Correlation between the global dGEMRIC indices and the global morphologic MRI scores. With increasing dGEMRIC values, the morphologic MRI scores decrease which corresponds to mild cartilage degeneration and vice versa for severe cartilage degeneration.

Interobserver Reliability

Interobserver reliability was high, both for dGEMRIC measurements (ICC 0.940, 95% CI 0.894-0.966; P < 0.001) and for morphologic MRI scores (κ = 0.825, P < 0.001). Intraobserver reliability was high, both for dGEMRIC measurements (ICC 0.931, 95% CI 0.879-0.961; P < 0.001) and for global morphologic MRI scores (κ = 0.782, P < 0.001). Interobserver reliability for subcategories of morphologic MRI scores ranged from substantial for labrum damage (κ = 0.612, P < 0.001) to high for paralabral cysts (κ = 1, P < 0.001) ( Table 8 ). Intraobserver reliability for subcategories of morphologic MRI scores ranged from moderate for femoral osteophytes (κ = 0.580, P < 0.001) to high for paralabral cysts (κ = 1, P < 0.001) ( Table 8 ). Inter- and intraobserver reliabilities of the subcategories of the morphologic MRI scores are summarized in Table 8 .

Table 8.

Inter- and Intraobserver Reliabilities of Morphologic MRI Subscores. ^a

Parameters	Interobserver Reliability (κ)	Intraobserver Reliability (κ)
Labrum damage	0.612	0.696
Paralabral cysts	1	1
Cartilage damage	0.715	0.634
Acetabulum center bone cysts	0.878	0.847
Acetabulum rim bone cysts	0.702	0.737
Femoral bone cysts	0.779	0.694
Acetabular rim osteophytes	0.628	0.747
Femoral osteophytes	0.767	0.580

^aP values were <0.001 for all κ values.

Discussion

Although commonly performed, the effect of subchondral drilling for treatment of cartilage flaps in the hip joint remains unclear. To our knowledge, this is the first study evaluating the radiological long-term outcome of this treatment on cartilage regeneration using biochemically sensitive (dGEMRIC) and morphologic MRI. Global dGEMRIC indices and morphologic MRI scores of the whole joint did not show any differences between the study group and the control group. By contrast, a more detailed analysis differentiating the joint in regions with and without cartilage flaps showed that subchondral drilling resulted in higher dGEMRIC indices at the defect zone at a minimum 5 years of follow-up compared with hips in which no subchondral drilling had been performed.

As osteoarthritis in the mostly young FAI patients generally progresses slowly, long term follow-up is necessary to evaluate the effectiveness of a specific treatment. ²⁹ In a systematic review on the arthroscopic management of acetabular cartilage defects, Marquez-Lara et al. ³⁰ included 12 studies in which treatment by means of debridement, microfracturing/ subchondral drilling and autologous chondrocyte transplantation was performed. All studies were retrospective, with an average short follow-up of 27.1 months. Outcome measures were limited to clinical scores and all treatment options were associated with comparable improvements, which does not allow to draw definite conclusion regarding the efficacy of these procedures. ³⁰ Moreover, invasive methods such as second look arthroscopy for obtaining biopsies are rarely performed due to ethical reasons. Philippon et al. ³¹ investigated the results of microfracturing for isolated cartilage defects in 9 patients who underwent second look hip arthroscopy at an average of 20 months. Eight of 9 patients had 95% to 100% coverage of the initially diagnosed lesions with normal appearing or slightly discolored articular cartilage. ³¹ Similar results were found by Karthikeyan et al. ³² who documented an average of 96% filling of the defect with macroscopically good quality repair tissue in 19 of 20 patients at second-look arthroscopy after 17 months.

By contrast a comprehensive assessment of morphologic MR using semiquantitative scores and biochemically sensitive MRI techniques such as dGEMRIC can potentially overcome these limitations. ³³ This could help avoid the otherwise inevitable second look operation for macroscopic or histologic evaluation of repair tissue. In a goat model, Watanabe et al. ³⁴ investigated the ability of dGEMRIC measurements to evaluate the quality of repair tissue after microfracturing in the knee joint. In the repair tissue, a correlation was observed between pre- and postcontrast T1 values and glycosaminoglycan concentration. ³⁴ However, to date there is only limited literature on postoperative MR imaging of the hip following FAI surgery and most studies evaluated the prevalence of intra-articular lesions at short-term in patients with residual pain.^35-37 In the current study we did not observe a difference in global dGEMRIC indices and morphologic MRI scores between the study group in which subchondral drilling of acetabular cartilage flaps had been performed and the control group. It has been shown that cartilage damage in FAI is typically focal and requires a topographical assessment of damage patterns on MRI. ¹⁴ To account for that we compared regions with and without surgically documented cartilage flaps between groups. This analysis showed better biochemical cartilage quality as assessed with dGEMRIC in regions with acetabular flaps in the study group (472 ms, 95% CI 433-510 ms) compared with the control group (390 ms, 95% CI 367-413 ms). By contrast morphologic MRI scores did not differ significantly between the groups. This may indicate that dGEMRIC is a more sensitive tool to assess cartilage quality than a semiquantitative morphologic assessment which takes the whole joint into account. In addition, dGEMRIC indices were higher by 50 ms (95% CI 17-82 ms) and morphologic MRI scores were lower by 6 points (95% CI 3-8 points) in regions with documented intact cartilage compared to regions with untreated cartilage flaps in the control group. This observed difference in dGEMRIC indices is within the previously reported range of 50 to 100 ms, which is clinically relevant as it corresponds to different morphologic stages of acetabular cartilage damage ³⁸ . By contrast no significant difference between regions with intact cartilage and cartilage flaps was observed in the study group. These findings demonstrate the potential of subchondral drilling to improve the quality of chondral repair tissue or prevent more rapid deterioration whereas degeneration progresses over time in areas of cartilage flaps without further treatment.

Overall, we observed a strong inverse correlation between global dGEMRIC indices and global morphologic MRI scores. This further underlines the validity of our results as more severe glycosaminoglycan depletion corresponding to lower dGEMRIC indices were associated with higher, more severe morphologic MRI scores.

There are several limitations to this study. First, the allocation of the patients to the study or control group was not randomized. At the time of surgery, there was no evidence how best to treat cartilage flaps and both techniques were randomly used. In addition, subchondral drilling or debridement does not pose any technical challenges and can be performed with standard instruments readily available in all operations. In reviewing the surgical reports, no indications for selection bias could be found. Second, no pre-operative dGEMRIC scans were available for comparison since the used 3D dual-flip angle gradient echo technique was not yet established at that time. ²⁴ This is an inherent limitation to a long-term follow-up study. Third the image assessment of the morphologic damage score was based on radial PD-weighted images without fat saturation at 10 clock-face positions, which limits the assessment of bone marrow lesions to some degree and did not adjust for the size of degenerative lesions except for differentiation between focal (<1 cm) and generalized cartilage defects (>1 cm). Fourth, patients who had progressed to total hip arthroplasty had inevitably to be excluded from follow-up MRI. Of the eligible patients with preserved joints, 15 patients (54%; 15 hips) of the initial study group and 19 patients (53%; 22 hips) of the initial control group agreed to undergo a follow-up dGEMRIC scan. Key parameters (demographics, clinical scores as well as pre- and postoperative radiological morphological parameters) between the patients that agreed to the dGEMRIC scan and those who did not were extensively compared to evaluate a possible selection bias. There were no differences except for higher preoperative alpha angle in the subset of the study group which agreed to undergo dGEMRIC scan at follow-up. This indicates that more patients with severe deformities were actually included in the study group and further minimizes the risk of a bias of ending up with patients with better outcome in the study group. Fifth, there was a longer follow-up in the control group compared to the study group. To overcome this limitation, we performed an analysis of covariance including time of follow up. Sixth, we did not include a B1 prescan to reduce flip-angle variations secondary to B1 field inhomogeneities which occur when using a dual-flip angle technique for dGEMRIC. ³⁹ However, this should not jeopardize our observations to a relevant degree as both groups were imaged with the same sequence parameters.

In conclusion, treatment of acetabular cartilage flaps with subchondral drilling leads to a better cartilage quality of the defect zone at minimum 5 years of follow-up.