The Clinical Significance of Dark Cartilage Lesions Identified on MRI

Transcription

1 Musculoskeletal Imaging Original Research Markhardt and Kijowski Dark artilage Lesions on MRI Musculoskeletal Imaging Original Research. Keegan Markhardt 1 Richard Kijowski Markhardt K, Kijowski R Keywords: anisotropy, black line sign, cartilage, MRI DOI: /JR Received January 20, 2015; accepted after revision June 1, Department of Radiology, University of Wisconsin linical Science enter, E3/311, 600 Highland ve, Madison, WI ddress correspondence to. K. Markhardt (keegan.markhardt@gmail.com). JR 2015; 205: X/15/ merican Roentgen Ray Society The linical Significance of Dark artilage Lesions Identified on MRI OJETIVE. The purpose of this study was to determine the clinical significance of foci of low signal intensity in morphologically normal cartilage. MTERILS ND METHODS. This retrospective study included 887 patients who underwent 898 knee MRI examinations performed within 6 months of arthroscopic knee surgery. musculoskeletal radiologist reviewed all MRI examinations for the presence of foci of low signal in cartilage where there was no visible morphologic abnormality, referred to as dark cartilage lesions. The surgical reports of all patients were reviewed for the presence of cartilage degeneration at arthroscopy. Logistic regression was used to model the probability of dark cartilage lesions corresponding to cartilage degeneration at arthroscopy as a function of patient age. RESULTS. In the 5388 articular surfaces assessed on MRI, 142 dark cartilage lesions were identified. The proportion of dark cartilage lesions corresponding to cartilage degeneration at arthroscopy was 52.0% (13 of 25) in the patella, 57.1% (28 of 49) in the trochlea, 90.9% (10 of 11) in the medial femoral condyle, 50.0% (two of four) in the lateral femoral condyle, 80.0% (four of five) in the medial tibial plateau, and 70.8% (34 of 48) in the lateral tibial plateau. There was a direct correlation (R 2 = 0.89) between patient age and the likelihood that a dark cartilage lesion would correspond to cartilage degeneration at arthroscopy. ONLUSION. Dark cartilage lesions may be found on every articular surface of the knee joint and may be a sign of otherwise occult cartilage degeneration. I n clinical practice, the articular cartilage is typically assessed using MRI sequences that evaluate the morphologic appearance of cartilage [1]. However, currently available morphologic cartilage imaging techniques are limited by their relatively low sensitivity for detecting early cartilage degeneration. The reported sensitivity of both 2D and 3D sequences for detecting surgically confirmed cartilage softening, fibrillation, and superficial partial-thickness defects within the knee ranges between 44% and 75% [2 6]. This relatively low sensitivity is primarily attributed to suboptimal spatial resolution. Moreover, an in-plane spatial resolution of 0.3 mm is required in order to distinguish superficial morphologic changes of cartilage, such as fibrillation and pitting, from the smooth surface of normal cartilage, which is beyond the resolution of most MRI sequences used in clinical practice [7]. Identification of changes in the signal intensity of articular cartilage may also be used to detect cartilage degeneration, and this is especially helpful when cartilage morphologically appears to be normal on MRI. These signal intensity changes, which are thought to primarily reflect changes in cartilage T2 relaxation time, are most easily identified using quantitative T2 mapping techniques but can also be detected on sequences routinely used for clinical cartilage assessment, especially T2-weighted fast spin-echo images [8]. Most investigators have reported an increase in the T2 relaxation time and increased signal intensity on T2-weighted images of degenerative cartilage, attributed to multiple factors including increased water content, decreased macromolecular content, and disruption of collagen matrix ultrastructure [9 12]. onversely, in vitro and in vivo studies have also shown that cartilage degeneration may be associated with a decrease in T2 relaxation time and decreased signal intensity on T2-weighted images [13 19], which may be due to factors such as magnetization transfer and disruption of tissue anisotropy [20]. However, the presence of low signal intensity in cartilage is not widely JR:205, December

2 Markhardt and Kijowski recognized as an MRI finding of cartilage degeneration in clinical practice. Stephens and associates [13] recently identified what has been termed the cartilage black line sign of the central femoral trochlea of the knee. This lesion was defined as linear low signal intensity on T2-weighted images oriented perpendicular to the subchondral bone and located in an area of the cartilage that was morphologically normal on MRI and was found to correlate with cartilage fissuring at arthroscopy. Subsequent studies by Wissman and associates [16, 21] have attempted to further define the clinical significance of this lesion. Earlier studies by roderick and associates [14] and König and associates [15] found that areas of low signal intensity within the articular cartilage of the knee may correlate with areas of cartilage degeneration at arthroscopy and biopsy, respectively. These areas of abnormal signal intensity were not restricted to the femoral trochlea. lthough these studies support the hypothesis that low signal intensity within cartilage may be an MRI finding of cartilage degeneration, their results are difficult to apply to current clinical practice because they used T1-weighted and non fat-suppressed T2-weighted sequences with relatively low spatial resolution. Using more modern imaging sequences, iswal and associates [22] found areas of heterogeneous change in signal intensity, containing both increased and decreased signal intensity, to be reflective of cartilage degeneration at arthroscopy and a risk factor for progression of cartilage loss over time. However, they did not assess areas within cartilage that had decreased signal intensity without associated increased signal intensity. To our knowledge, no large study using modern imaging sequences and arthroscopic correlation has been performed to fully assess the clinical implications of these focal low-signal-intensity cartilage lesions. Therefore, the goal of this study was to determine the clinical significance of foci of low signal intensity in morphologically normal cartilage by correlating MRI findings with arthroscopic findings on all articular surfaces of the knee in a large number of symptomatic patients. TLE 1: Routine 3.0-T Knee MRI Examination Protocol Imaging Parameter Materials and Methods Study Group This retrospective study was performed in compliance with HIP regulations, with the approval of our institutional review board and with a waiver of informed consent. database of all consecutive arthroscopic knee surgeries performed at the University of Wisconsin Hospital and linics over a 7-year period between June 1, 2007, and June 1, 2014, was reviewed to identify patients who had an MRI examination of the knee performed on the same 3.0-T scanner within a 6-month period before arthroscopic knee surgery. Patients were excluded from the study if they had osteochondritis dissecans, history of a cartilage repair procedure, or evidence of chondrocalcinosis on knee radiographs. No patient was excluded on the basis of any other factor including age, sex, history of prior knee surgery other than a cartilage repair procedure, severity of cartilage degeneration, or MR image quality. The study group consisted of 887 patients meeting these criteria, including 511 male (58%) and 376 female (42%) patients between 12 and 81 years old (mean and median ages, 40.9 and 42 years, respectively). These 887 patients underwent a total of 898 MRI examinations and arthroscopic surgeries (bilateral imaging studies and procedures were performed on the knees of 11 patients). MRI Examinations ll knee MRI examinations were performed on the same 3.0-T MR scanner (Signa HDx, version 14.0, GE Healthcare) using an 8-channel phased-array extremity coil. ll MRI examinations consisted of an axial frequency-selective fatsuppressed T2-weighted fast spin-echo sequence, a coronal proton density weighted fast spin-echo sequence, a coronal frequency-selective fat-suppressed proton density weighted fast spin-echo sequence, a sagittal proton density weighted fast spin-echo sequence, and a sagittal frequency-selective fat-suppressed T2-weighted fast spin-echo sequence. The imaging parameters of the sequences performed during the MRI examinations are summarized in Table 1. Sequence xial T2 FS oronal PD oronal PD FS Sagittal PD Sagittal T2 FS Repetition time (ms) TE time (ms) Matrix size FOV (cm) Slice thickness (mm) andwidth (khz) Echo-train length Signal averages Scan time (min:s) 3:30 3:25 3:26 3:26 3:16 Note FS = fat-suppressed, PD = proton density weighted fast spin-echo sequence, T2 = T2-weighted fast spin-echo sequence. Review of MRI Examinations The MRI examinations performed on the 898 knees of the 887 patients in the study group were reviewed by a fellowship-trained musculoskeletal radiologist with 5 years of clinical experience. For assessment of interobserver agreement, 100 randomly chosen MRI examinations of the knee were also reviewed independently by a second fellowship-trained musculoskeletal radiologist with 12 years of clinical experience. The radiologists reviewed the MRI examinations on a PS workstation (Horizon Medical Imaging, McKesson) without knowledge of the arthroscopic findings. ll sequences were reviewed to determine the presence or absence of low signal intensity in the middle zone of cartilage on each articular surface that was morphologically normal. Specifically, the presence of low signal intensity in the middle zone of cartilage was assessed only on those articular surfaces of the knee where there was no appreciable cartilage fissure, fibrillation, or defect. Foci of low signal intensity in the middle zone, referred to as a dark cartilage lesion in the present study, were defined as areas with a conspicuous decrease in the normal signal intensity evident on all sequences, which could not be accounted for by the normal, gradual changes in shading of signal intensity related to the regiondependent differences in tissue anisotropy [23]. Only areas of low signal intensity that were similar to the signal intensity of normal tendon or ligament and seen in two imaging planes were included. More subtle areas of low signal intensity were not considered to be dark cartilage lesions. The two musculoskeletal radiologists then reviewed in consensus all MRI examinations in which a dark cartilage lesion was identified on an articular surface of the knee by one or both radiologists. Only those dark cartilage lesions that were definitively thought to be present by both ra JR:205, December 2015

3 Dark artilage Lesions on MRI Fig. 1 Drawing shows number of dark cartilage lesions classified by location using International artilage Repair Society injury evaluation system. (dapted with permission from [40]) diologists in the consensus review were included in the analysis. During the consensus review, the radiologists classified the dark cartilage lesion as linear or nonlinear using the criteria described by Wissman and associates [16, 21], who defined linear cartilage black line sign lesions for the femoral trochlea as those that measured less than 2 mm in thickness and extended for two or more slices, which would be greater than 3 mm in length on the basis of the sequence parameters used. The radiologists also assessed the various dark cartilage lesions in terms of the following characteristics: 1, location on a standardized knee map provided by the International artilage Repair Society; 2, visibility, comparing proton density weighted and fat-suppressed T2-weighted sequences; 3, presence or absence of extension to the articular surface; 4, orientation of a linear dark cartilage lesion relative to the articular surface (i.e., perpendicular or oblique); 5, length and width in the dimension parallel to the articular surface, measured using electronic calibers on the PS workstation and rounded to the nearest millimeter; 6, presence or absence of adjacent or subjacent increased signal intensity in cartilage on T2-weighted images; and 7, presence or absence of subjacent subchondral bone marrow edema. rthroscopic Knee Surgeries rthroscopic surgery was performed on 898 knees of the 887 patients in the study group within the 6-month period after MRI examination. The time interval from MRI examination to surgery ranged from 1 to 181 days (mean and median interval, 57.5 and 43 days, respectively). ll arthroscopic knee surgeries were performed by one of four experienced orthopedic surgeons at our institution who specialized in sports medicine and had between 11 and 31 years of clinical experience. The surgeons at our institution routinely inspect all articular surfaces both visually and with a surgical probe at arthroscopy and describe the grade of cartilage on each articular surface in the surgical report by use of a Noyes classification system (grade 0, normal; grade 1, near normal, with either softening [1] or superficial cartilage fibrillation [1]; grade 2, superficial partial-thickness cartilage defect [< 50% of the total thickness of the articular surface]; grade 2, deep partial-thickness cartilage defect 50% of the total thickness of the articular surface]; and grade 3, full-thickness cartilage defect) [24]. The surgical reports of all patients were retrospectively reviewed by a fellowship-trained musculoskeletal radiologist who was blinded to the MRI findings, and the grade of cartilage lesion on each articular surface of all 898 knees was recorded. Statistical nalysis ll statistical tests and graphics were obtained using R software (version 3.1.0, R Foundation). Statistical significance was defined as a two-sided p value less than There was no adjustment of p values for multiple testing. ohen s unweighted kappa statistic was used to assess interobserver agreement between radiologists for determining the presence and absence of dark cartilage lesions on each articular surface of the knee in 100 randomly selected patients. The Pearson chi-square test was used to compare sex and knee sidedness of patients with and without dark cartilage lesions. n unpaired t test that assumed unequal variance was used to compare the age of patients with and without dark cartilage lesions. The ochran-rmitage test was used to determine the association between the presence of dark cartilage lesions within the knee and the number of articular surfaces with cartilage degeneration at arthroscopy. The proportion of dark cartilage lesions corresponding to cartilage degeneration at arthroscopy were computed for each age quinquennium (5-year age group) and plotted. Least squares regression was used to model the proportion of true-positive dark cartilage lesions as a function of the age group midpoint. Residuals were examined to assess possible violations of model assumptions. Logistic regression was used to model the probability of dark cartilage lesions corresponding to cartilage degeneration at arthroscopy as a function of patient age. Odds ratios and 95% Is were calculated. The Fisher exact test was used to determine the association between surgical findings and the following characteristics of the various dark cartilage lesions: 1, linear or nonlinear classification; 2, presence or absence of extension to the articular surface; 3, orientation of a linear dark cartilage lesion relative to the articular surface (i.e., perpendicular or oblique); 4, presence or absence of adjacent or subjacent increased signal intensity in cartilage on T2-weighted images; and 5, presence or absence of subjacent subchondral bone marrow edema. n unpaired Welch t test that assumed unequal variance was used to determine the association between surgical findings and the cross-sectional area of the dark cartilage lesion. Results The 887 patients in the study group underwent a total of 898 MRI examinations and arthroscopic surgeries. Surgical indications for arthroscopy were anterior cruciate ligament tear in 30% of patients, medial meniscal tear in 66% of patients, lateral meniscal tear in 23% of patients, and various other indications in 6% of patients. The number of articular surfaces of the 898 knees in the 887 patients that were found to have cartilage de- JR:205, December

4 Markhardt and Kijowski TLE 2: Summary of 898 Dark artilage Lesions Sorted by Shape and ompartment: Number Detected and Number Showing Positive orrelation to Degeneration at rthroscopy Dark artilage Lesions ny Shape Linear Nonlinear rticular Surface No. Detected (Frequency [%]) generation at arthroscopy was as follows: 487 (54.2%) for the patella, 336 (37.4%) for the femoral trochlea, 399 (44.4%) for the medial femoral condyle, 178 (19.8%) for the lateral femoral condyle, 241 (26.8%) for the medial tibial plateau, and 288 (32.1%) for the lateral tibial plateau. Of the 5388 articular surfaces of the knee assessed on MRI in the 887 patients in the study group, a total of 142 dark cartilage lesions were identified in 131 patients. One hundred twenty-one knees had a single lesion, nine knees had lesions on two articular surfaces, and one knee had lesions on three articular surfaces. Sixty-nine dark cartilage lesions (48.6%) were present in the right knee, and 73 dark cartilage lesions (51.4%) in the left knee. There were 82 male (62.6%) and 49 female (37.4%) patients with dark cartilage lesions (mean and median age, 40.5 years and 41 years, respectively; age range, years). There was no statistically significant difference between patients with and without dark cartilage lesions with regard to age (p = 0.91), sex (p = 0.28), or knee sidedness (p = 0.89). s shown in Table 2 and Figure 1, dark cartilage lesions were identified on every articular surface of the knee, with lesions most frequently seen on the femoral trochlea and lateral femoral condyle. The overall average frequency of dark cartilage lesions across all compartments was 2.6%. Of the 142 dark cartilage lesions, 51 had a linear shape, and 91 had a nonlinear shape. Depending on the articular surface, the frequency of linear dark cartilage lesions ranged from 0.0% to 3.1%, and the frequency of nonlinear dark cartilage lesions ranged from 0.2% to 4.7%. s shown in Table 3, there was no substantial change in the frequency of dark cartilage lesions with No. With Positive orrelation (PPV [%]) No. Detected (Frequency [%]) No. With Positive orrelation (PPV [%]) increasing patient age. There was a trend (p = 0.07) toward a higher frequency of dark cartilage lesions in the knees of patients with lower numbers of articular surfaces with cartilage degeneration identified at arthroscopy. There was high interobserver agreement between radiologists for determining the presence or absence of dark cartilage lesions on each articular surface of the knee in 100 randomly chosen patients, with a ohen s unweighted kappa measure of The dark cartilage lesions were evident on all sequences but were more conspicuous on the fat-suppressed proton density weighted and T2-weighted fast spin-echo images than on the non fat-suppressed proton density weighted fast spin-echo images. Patellar dark cartilage lesions were best seen on the sagittal images; trochlear dark cartilage lesions were best seen on the axial and coronal images; and dark cartilage lesions on the femoral condyles and tibia plateau were best seen on the coronal images. Dark cartilage lesions were commonly seen in certain locations on each articular surface of the knee as depicted in Figure 1. In particular, 48 of the 49 dark cartilage lesions on the femoral trochlea were located in the central portion of the articular surface. The majority of linear dark cartilage lesions was oriented perpendicular to (73%) and contacted (91%) the articular surface. The surface area of dark cartilage lesions was typically larger in the lateral tibial plateau (mean surface area, 70.3 mm 2 ), but dark cartilage lesions were otherwise of similar size on the remaining articular surfaces (mean surface area, mm 2 ). s shown in Table 2, combining all articular surfaces of the knee, 34 of 51 linear dark cartilage lesions (66.7%) corresponded to No. Detected (Frequency [%]) No. With Positive orrelation (PPV [%]) Patella 25 (2.8) 13 (52.0) 8 (0.9) 5 (62.5) 17 (1.9) 8 (47.1) Trochlea 49 (5.5) 28 (57.1) 28 (3.1) 17 (60.7) 21 (2.3) 11 (52.4) Medial femoral condyle 11 (1.2) 10 (90.9) 7 (0.8) 7 (100) 4 (0.4) 3 (75.0) Lateral femoral condyle 4 (0.4) 2 (50.0) 2 (0.2) 2 (100) 2 (0.2) 0 (0) Medial tibial plateau 5 (0.6) 4 (80.0) 0 (0) 0 (N) 5 (0.6) 4 (80.0) Lateral tibial plateau 48 (5.3) 34 (70.8) 6 (0.7) 3 (50.0) 42 (4.7) 31 (73.8) Total 142 (2.6) 91 (64.1) 51 (0.9) 34 (66.7) 91 (1.9) 57 (62.6) Note PPV = positive predictive value, N = not applicable. cartilage degeneration at arthroscopy, and 57 of 91 nonlinear dark cartilage lesions (62.6%) corresponded to cartilage degeneration at arthroscopy. For the femoral trochlea, 17 of 28 linear dark cartilage lesions (60.7%) corresponded to cartilage degeneration at arthroscopy, and 11 of 21 nonlinear dark cartilage lesions (52.4%) corresponded to cartilage degeneration at arthroscopy. Older patient age was the main feature that could significantly differentiate between true-positive and falsepositive dark cartilage lesions (p < 0.01; odds ratio, 1.06 [95% I, ]). s shown in Table 3, the likelihood that a dark cartilage lesion would correspond to an area of cartilage degeneration at arthroscopy was found to increase with patient age in a linear manner (R 2 = 0.89), with the proportion of true-positive dark cartilage lesions rising from 20% for individuals younger than 20 years old to 100% for individuals older than 65 years old. s shown in Table 4, most dark cartilage lesions were found to correspond to superficial and deep partial-thickness cartilage defects at arthroscopy, with smaller numbers of dark cartilage lesions corresponding to cartilage softening and fibrillation. Figures 2 6 show examples of dark cartilage lesions on MRI. Some characteristics of dark cartilage lesions were useful for differentiating between true-positive and falsepositive lesions. The extension of a linear lesion (p = 0.03) but not of a nonlinear lesion (p = 0.15) to the articular surface could be used to significantly differentiate between true-positive and false-positive dark cartilage lesions. In particular, all three linear dark cartilage lesions that did not extend to the articular surface did not correspond to cartilage degeneration at arthroscopy, in JR:205, December 2015

5 Dark artilage Lesions on MRI TLE 3: Summary of Dark artilage Lesions Sorted by Shape and Patient ge: Number Detected and Number Showing Positive orrelation to Degeneration at rthroscopy Patient ge Range (y) No. of rticular Surfaces ssessed Dark artilage Lesions ll rticular Surfaces ombined Femoral Trochlea Frequency (%) a ny Shape Linear Nonlinear Linear Nonlinear /5 (20) 0/2 (0) 1/3 (33) 0/2 (0) 0/1 (0) /26 (42) 6/14 (43) 5/12 (42) 3/9 (33) 2/6 (33) /15 (40) 3/7 (43) 3/8 (38) 3/5 (60) 1/2 (50) /11 (55) 4/4 (100) 2/7 (29) 1/1 (100) 1/3 (33) /14 (71) 5/6 (83) 5/8 (63) 2/2 (100) 1/2 (50) /13 (69) 3/3 (100) 6/10 (60) 2/2 (100) 0/0 (N) /13 (77) 3/4 (75) 7/9 (78) 3/3 (100) 3/3 (100) /13 (62) 1/2 (50) 7/11 (64) 1/2 (50) 2/3 (67) /13 (92) 4/4 (100) 8/9 (89) 2/2 (100) 1/1 (100) /9 (89) 2/2 (100) 6/7 (86) 0/0 (N) 0/0 (N) /7 (100) 1/1 (100) 6/6 (100) 0/0 (N) 0/0 (N) > /3 (100) 2/2 (100) 1/1 (100) 0/0 (N) 0/0 (N) Total /142 (64) 34/51 (67) 57/91 (63) 17/28 (61) 11/21 (52) Note Except where otherwise indicated, data are given as ratio of no. of lesions with positive correlation to no. of lesions detected, with positive predictive value (%) given in parentheses. N = not applicable. a cross all articular compartments. TLE 4: Summary of Dark artilage Lesions onfirmed at rthroscopy Sorted by the Noyes lassification System for rthroscopic Grade of artilage Degeneration Grade ny Shape Linear Nonlinear cluding one located on the femoral trochlea. djacent or subjacent increased signal intensity in cartilage on T2-weighted images could also be used to significantly differentiate (p < 0.01) between true-positive and false-positive lesions, with 87.1% (27 of 31) of mixed dark and bright lesions corresponding to cartilage degeneration at arthroscopy, compared with 57.7% (64 of 111 cases) of isolated dark cartilage lesions. There was no significant difference between the proportion of linear and nonlinear dark cartilage lesion that corresponded to cartilage degeneration at arthroscopy for the femoral trochlea (p = 0.71) or for all articular surfaces combined (p = 0.72). In addition, the orientation of a linear dark cartilage lesion relative to the articular surface (p = 0.17) and the crosssectional area of the dark cartilage lesion (p = 0.60) could not be used to significantly differentiate between true-positive and false-positive lesions. Finally, subjacent subchondral bone marrow edema was seen in only 12 of 142 dark cartilage lesions (8.5%), and this was not a helpful finding because only 58.3% (seven of 12) of these lesions corresponded to cartilage degeneration at arthroscopy, which was not significantly different (p = 0.76) from the 64.6% (84 of 130) of lesions without subjacent subchondral bone marrow edema that corresponded to cartilage degeneration at arthroscopy. Discussion Our study expands on the work of Stephens and associates [13] and Wissman and associates [16, 21], who investigated linear dark cartilage lesions in the central femoral trochlea, termed the cartilage black line sign. Wissman and associates defined this linear dark cartilage lesion as measuring less than 2 mm in thickness and two or more slices in length and extending through the entire thickness of the cartilage. Using this definition, we found 27 lesions with the cartilage black line sign lesions in the central femoral Fig year-old man who underwent arthroscopic partial medial meniscectomy. and, xial () and sagittal () fat-suppressed T2-weighted fast spin-echo MRI sequences show nonlinear dark cartilage lesion in patella (arrows)., Grade 2 cartilage fibrillation was seen at arthroscopy (arrow) corresponding to dark cartilage lesion at MRI. JR:205, December

6 Markhardt and Kijowski Fig year-old man who underwent arthroscopic partial medial meniscectomy. and, xial fat-suppressed T2-weighted fast spin-echo () and coronal fat-suppressed proton density weighted fast spin-echo () MRI sequences show linear dark cartilage lesion in central femoral trochlea (arrows)., Grade 2 cartilage fissuring was seen at arthroscopy (arrow) corresponding to dark cartilage lesion at MRI. trochlea. Wissman and associates retrospectively reviewed 1300 consecutive knee MRI examinations for these linear dark cartilage lesions in the femoral trochlea and reported a slightly lower frequency of this lesion (1.9%) than we found in the present study (3.1%). Stephens and associates [13] found that both of two patients with a cartilage black line sign showed evidence of cartilage degeneration at arthroscopy, whereas Wissman and associates [16] found that one of eight patients showed evidence of cartilage degeneration at arthroscopy. We have improved on the arthroscopic correlation sample size of these studies, and our findings more strongly support an association with cartilage degeneration, with 17 of 27 lesions (63.0%) corresponding to cartilage degeneration at arthroscopy. One linear dark cartilage lesion in the central femoral trochlea in our study did not contact the articular surface and therefore did not meet the criteria for the cartilage black line sign. This case was negative on arthroscopy, which further supports the proposed definition of this lesion. Our study also found a similar association with cartilage degeneration for nonlinear dark cartilage lesions on the femoral trochlea that did not meet the morphologic criteria of the cartilage black line sign, with 11 of 21 nonlinear dark cartilage lesions (52.4%) corresponding to cartilage degeneration at arthroscopy. Therefore, the morphologic appearance of a focal dark cartilage lesion in the central femoral trochlea does not substantially change the likelihood that the lesion corresponds to cartilage degeneration at arthroscopy. Wissman and associates [21] also found progression of lesions with the cartilage black line sign over time. onsistent with these findings, we found that patient age was the main feature that could differentiate between true-positive and false-positive dark cartilage lesions, with older patients being more likely to have positive correlation at arthroscopy. This relationship was true both for cartilage black line sign lesions and for dark cartilage lesions in general. We also found a trend toward a higher frequency of dark cartilage lesions in the knees of patients who had fewer articular surfaces with cartilage degeneration at arthroscopy. Together, these findings suggest that dark cartilage lesions may precede morphologic evidence of cartilage degeneration on MRI. Fig year-old woman who underwent arthroscopic partial medial meniscectomy. and, oronal fat-suppressed proton density weighted fast spin-echo () and sagittal fat-suppressed T2-weighted fast spin-echo () MRI sequences show obliquely oriented linear dark cartilage lesion in central aspect of medial femoral condyle (arrows)., Grade 2 unstable cartilage flap was seen at arthroscopy (arrow) corresponding to dark cartilage lesion at MRI JR:205, December 2015

7 Dark artilage Lesions on MRI Fig year-old woman who underwent arthroscopic partial medial and lateral meniscectomy. and, oronal fat-suppressed proton density weighted fast spin-echo () and sagittal fat-suppressed T2-weighted fast spin-echo () MRI sequences show nonlinear dark cartilage lesion in lateral tibial plateau (arrows). Note associated increased signal within deep zone of cartilage adjacent to dark cartilage lesion on T2-weighted image (arrowhead, )., Grade 2 cartilage degeneration was seen at arthroscopy (arrow) corresponding to dark cartilage lesion at MRI. Our study also investigated dark cartilage lesions on the other articular surfaces of the knee and dark cartilage lesions that were nonlinear. lthough dark cartilage lesions were rare on some articular surfaces, correlation with cartilage degeneration was found on every articular surface. Furthermore, nonlinear dark cartilage lesions were found to have a similar likelihood of corresponding to cartilage degeneration at arthroscopy as linear dark cartilage lesions. Looking beyond the knee joint, a recent study by Pfirrmann and associates evaluated 44 patients with hip pain and found that linear areas of low signal intensity within the acetabular cartilage was a useful sign for detecting surgically confirmed cartilage delamination [25]. Figure 4 shows a similar linear area of low signal intensity within the cartilage of the medial femoral condyle that corresponded to a flap of delaminated cartilage at arthroscopy. Therefore, a low-signal-intensity lesion within articular cartilage can be an MRI finding of degeneration not only irrespective of its morphologic appearance or location within the knee but likely also irrespective of the joint. Overall, 64.1% of dark cartilage lesions corresponded to areas of cartilage degeneration at arthroscopy, and the proportion of arthroscopically positive dark cartilage lesions directly increased with patient age. ategorization by morphologic appearance (linear or nonlinear), obliquity (perpendicular or oblique), or size could not be successfully used to distinguish between dark cartilage lesions that corresponded to normal and abnormal cartilage at arthroscopy. However, adjacent increased signal intensity in cartilage on T2-weighted images was a feature that could significantly differentiate between true-positive and false-positive lesions, with 87.1% of mixed dark and bright lesions corresponding to cartilage degeneration at arthroscopy, compared with 57.7% of isolated dark cartilage lesions. iswal and associates [22] also found mixed dark and bright areas within cartilage to be both reflective of cartilage degeneration at arthroscopy and a risk factor for progression of cartilage loss over time. Dark cartilage lesions in our study were seen on all sequences of the routine MRI ex- Fig year-old woman who underwent arthroscopic partial lateral meniscectomy. and, xial fat-suppressed T2-weighted fast spin-echo () and coronal fat-suppressed proton density weighted fast spin-echo () MRI sequences show linear dark cartilage lesion in central femoral trochlea (arrows)., Normal cartilage was seen at arthroscopy (arrow). JR:205, December

8 Markhardt and Kijowski aminations but were more conspicuous on the fat-suppressed proton density weighted and T2-weighted fast spin-echo images than the non fat-suppressed proton density weighted fast spin-echo images. The influence of T2 relaxation time on the appearance of normal and degenerative cartilage has been extensively described in the literature [9, 12, 23, 26 29]. lthough proton density weighted fast spin-echo images have a much shorter effective TE than do T2-weighted images, the proton density weighted contrast of cartilage is thought to primarily reflect variations in its T2 relaxation time, especially with the addition of fat suppression [8]. The cartilage degeneration that was detected as areas of low signal intensity within morphologically normal cartilage in our study corresponded to areas of cartilage softening, fibrillation, and partial-thickness flaps and defects, which were likely beyond the spatial resolution of morphologic imaging. lthough the precise manner in which degeneration causes decreased signal intensity within articular cartilage is not known, the most likely causes are alterations in tissue anisotropy and magnetization transfer effects [20]. Highly organized collagenous structures such as cartilage have significant directional organization or anisotropy, which generally promote rapid T2 decay. However, as the structural orientation is tilted away from the direction of the main magnetic field ( 0 ), there is prolongation of T2 decay with maximal increase in T2 signal observed at 54.7 from 0 the magic angle. The organization of the collagen ultrastructure of cartilage results in a depth-dependent effect of anisotropy, with more pronounced anisotropic effects in the deep zone, owing to more perpendicularly oriented collagen fibers, and less anisotropic effects in the middle zone, where only a small portion of the curved collagen fibers are in an orientation that promotes rapid T2 decay [23]. Destruction of the extracellular architecture in degenerative cartilage could lead to loss of regional magic angle effect, resulting in a foci of relatively low signal intensity [30, 31]. lternatively, flattening of the normally curving collagen fibers in the middle zone of cartilage owing to macromolecular degeneration could lead to a more linear fiber orientation, resulting in a localized area of more rapid T2 decay [32, 33]. lteration of tissue anisotropy in degenerative cartilage could explain why dark cartilage lesions in our study tended to be located in certain areas on each articular surface. Each articular surface has a unique anisotropy pattern [8, 27, 34, 35], and these factors are dependent on normal local tissue orientation and relative orientation to 0 [23]. nother possible cause for low signal intensity within degenerative cartilage is the magnetization transfer effect, which refers to dipole-dipole interactions between bound protons in complex macromolecules and mobile water protons [36, 37]. ollagen macromolecules are the dominant component of articular cartilage that contributes to the magnetization transfer effect [38]. Imaging sequences that use a large number of off-resonance radiofrequency pulses, such as multislice fast spin-echo or turbo spin-echo sequences, cause more selective saturation of collagen bound protons. natural equilibrium then occurs where the mobile water protons give up their magnetization to the bound protons in collagen, thereby decreasing the measured signal from water. Fragmented collagen in degenerative cartilage is thought to provide more efficient magnetization transfer because a greater surface area is available to interact with both loosely bound bulk water in the macromolecular matrix and adjacent synovial fluid [39]. In this way, superficial fibrillation, small cartilage fissures, and other areas of early cartilage degeneration may appear more conspicuous on fast spin-echo and turbo spin-echo images owing to an edge of decreased water signal caused by increased magnetization exchange between mobile water protons and fragmented collagen macromolecules. Our study had several limitations. One limitation was that all MRI examinations were performed on a 3.0-T scanner. However, we are confident that our findings are also applicable to 1.5-T MRI, given that, to our knowledge, all prior investigations of dark cartilage lesions have used 1.5-T scanners [13, 16, 21]. nother limitation of our study was the presence of selection bias owing to the retrospective nature of the analysis. We attempted to reduce the effects of selection bias by evaluating a large number of patients with a wide range of ages and not discriminating by surgical indication. further limitation was that the information regarding the presence or absence of cartilage degeneration on each articular surface of the knee joint at arthroscopy was obtained by retrospective review of surgical reports. Thus, it is possible that cartilage lesions could have been missed at arthroscopy and that errors could have been made when describing the arthroscopic findings in the surgical report. Moreover, precise correlation between the location of the dark cartilage lesion on MRI and the location of the arthroscopic area of cartilage degeneration on a particular articular surface was not always possible. We attempted to mitigate this limitation by correlating with arthroscopic images when available and only including a dark cartilage lesion when there was no morphologic cartilage lesion seen otherwise on MRI. Finally, a potential weakness of our statistical analysis was that it did not account for clustering of dark cartilage lesions within subjects. However, this was uncommon, and only a small proportion of patients were found to have more than one cartilage lesion within the knee. In conclusion, our study has shown that areas of low signal intensity within cartilage may be seen on all six articular surfaces of the knee. significant proportion of these dark cartilage lesions were found to correspond to areas of cartilage degeneration at arthroscopy, indicating that these foci of low signal intensity in morphologically normal cartilage can be reported as areas of possible cartilage degeneration, especially in older individuals. It should be noted that to appropriately identify these dark cartilage lesions, the radiologist needs to be familiar with the normal, gradual changes in shading of signal intensity related to the joint-specific pattern of cartilage ultrastructure. cknowledgment We thank lejandro Muñoz del Rio for expert assistance with statistical analysis. References 1. Gold GE, Hargreaves, Stevens KJ, eaulieu F. dvanced magnetic resonance imaging of articular cartilage. Orthop lin North m 2006; 37: redella M, Tirman PF, Peterfy G, et al. ccuracy of T2-weighted fast spin-echo MR imaging with fat saturation in detecting cartilage defects in the knee: comparison with arthroscopy in 130 patients. JR 1999; 172: Gold GE, Fuller SE, Hargreaves, Stevens KJ, eaulieu F. Driven equilibrium magnetic resonance imaging of articular cartilage: initial clinical experience. J Magn Reson Imaging 2005; 21: Kijowski R, lankenbaker DG, Woods M, Shinki K, De Smet, Reeder S. 3.0-T evaluation of knee cartilage by using three-dimensional IDEL GRSS imaging: comparison with fast spin-echo imaging. Radiology 2010; 255: JR:205, December 2015

9 Dark artilage Lesions on MRI 5. Kijowski R, Davis KW, Woods M, et al. Knee lear cleft: the black line of the trochlear trough. Trattnig S. Transverse relaxation mechanisms in ar- joint: comprehensive assessment with 3D isotro- Skeletal Radiol 2012; 41: ticular cartilage. J Magn Reson 2004; 169: pic resolution fast spin-echo MR imaging diag- 17. Hodler J, erthiaume MJ, Schweitzer ME, Resn- 30. Dunham J, hambers MG, Jasani MK, itensky L, nostic performance compared with that of con- ick D. Knee joint hyaline cartilage defects: a com- hayen J. hanges in the orientation of proteogly- ventional MR imaging at 3.0 T. Radiology 2009; parative study of MR and anatomic sections. J cans during the early development of natural mu- 252: Mohr. The value of water-excitation 3D FLSH and fat-saturated PDw TSE MR imaging for detecting and grading articular cartilage lesions of the knee. Skeletal Radiol 2003; 32: Rubenstein JD, Li JG, Majumdar S, Henkelman RM. Image resolution and signal-to-noise ratio requirements for MR imaging of degenerative cartilage. JR 1997; 169: Goodwin DW. MRI appearance of normal articular cartilage. Magn Reson Imaging lin N m 2011; 19: Mlynárik V, Trattnig S, Huber M, Zembsch, Imhof H. The role of relaxation times in monitoring proteoglycan depletion in articular cartilage. J Magn Reson Imaging 1999; 10: Nieminen MT, Rieppo J, Toyras J, et al. T2 relaxation reveals spatial collagen architecture in articular cartilage: a comparative quantitative MRI and polarized light microscopic study. Magn Reson Med 2001; 46: Watrin, Ruaud JP, Olivier PT, et al. T2 mapping of rat patellar cartilage. Radiology 2001; 219: Liess, Lüsse S, Karger N, Heller M, Glüer. Detection of changes in cartilage water content using MRI T2-mapping in vivo. Osteoarthritis artilage 2002; 10: Stephens T, Diduch DR, alin JI, Gaskin M. The cartilage black line sign: an unexpected MRI appearance of deep cartilage fissuring in three patients. Skeletal Radiol 2011; 40: roderick LS, Turner D, Renfrew DL, Schnitzer TJ, Huff JP, Harris. Severity of articular cartilage abnormality in patients with osteoarthritis: evaluation with fast spin-echo MR vs arthroscopy. JR 1994; 162: König H, Sauter R, Deimling M, Vogt M. artilage disorders: comparison of spin-echo, HESS, and FLSH sequence MR images. Radiology 1987; 164: Wissman RD, Ingalls J, Nepute J, et al. The troch- omput ssist Tomogr 1992; 16: Menezes NM, Gray ML, Hartke JR, urstein D. T 2 and T 1ρ MRI in articular cartilage systems. Magn Reson Med 2004; 51: urstein D, Gray ML. Is MRI fulfilling its promise for molecular imaging of cartilage in arthritis? Osteoarthritis artilage 2006; 14: Markhardt K, hang EY. Hypointense signal lesions of the articular cartilage: a review of current concepts. lin Imaging 2014; 38: Wissman RD, England E, Mehta K, et al. The trochlear cleft: initial experience in elite athletes. J omput ssist Tomogr 2014; 38: iswal S, Hastie T, ndriacchi TP, ergman G, Dillingham MF, Lang P. Risk factors for progressive cartilage loss in the knee: a longitudinal magnetic resonance imaging study in forty-three patients. rthritis Rheum 2002; 46: Xia Y. Magic-angle effect in magnetic resonance imaging of articular cartilage: a review. Invest Radiol 2000; 35: Noyes FR, Stabler L. system for grading articular cartilage lesions at arthroscopy. m J Sports Med 1989; 17: Pfirrmann W, Duc SR, Zanetti M, Dora, Hodler J. MR arthrography of acetabular cartilage delamination in femoroacetabular cam impingement. Radiology 2008; 249: Xia Y. Heterogeneity of cartilage laminae in MR imaging. J Magn Reson Imaging 2000; 11: Goodwin DW, Wadghiri YZ, Zhu H, Vinton J, Smith ED, Dunn JF. Macroscopic structure of articular cartilage of the tibial plateau: influence of a characteristic matrix architecture on MRI appearance. JR 2004; 182: Mlynárik V, Degrassi, Toffanin R, Vittur F, ova M, Pozzi-Mucelli RS. Investigation of laminar appearance of articular cartilage by means of magnetic resonance microscopy. Magn Reson Imaging 1996; 14: Mlynárik V, Szomolányi P, Toffanin R, Vittur F, rine osteoarthritis. J Orthop Res 1990; 8: Dunham J, Shackleton DR, Nahir M, et al. ltered orientation of glycosaminoglycans and cellular changes in the tibial cartilage in the first two weeks of experimental canine osteoarthritis. J Orthop Res 1985; 3: lhadlaq H, Xia Y. The structural adaptations in compressed articular cartilage by microscopic MRI (µmri) T 2 anisotropy. Osteoarthritis artilage 2004; 12: lhadlaq H, Xia Y. Modifications of orientational dependence of microscopic magnetic resonance imaging T 2 anisotropy in compressed articular cartilage. J Magn Reson Imaging 2005; 22: lark JM. Variation of collagen fiber alignment in a joint surface: a scanning electron microscope study of the tibial plateau in dog, rabbit, and man. J Orthop Res 1991; 9: Goodwin DW, Zhu H, Dunn JF. In vitro MR imaging of hyaline cartilage: correlation with scanning electron microscopy. JR 2000; 174: Wolff SD, alaban RS. Magnetization transfer contrast (MT) and tissue water proton relaxation in vivo. Magn Reson Med 1989; 10: Henkelman RM, Stanisz GJ, Graham SJ. Magnetization transfer in MRI: a review. NMR iomed 2001; 14: Kim DK, eckler TL, Hascall V, alabro, alaban RS. nalysis of water-macromolecule proton magnetization transfer in articular cartilage. Magn Reson Med 1993; 29: Mosher TJ, Pruett SW. Magnetic resonance imaging of superficial cartilage lesions: role of contrast in lesion detection. J Magn Reson Imaging 1999; 10: International artilage Repair Society (IRS). IRS artilage Injury Evaluation Package. www. cartilage.org/_files/contentmanagement/irs_ evaluation.pdf. Developed January 27 30, ccessed June 17, 2015 JR:205, December