If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Meniscus tear is a known risk factor for osteoarthritis (OA). Quantitative assessment of meniscus degeneration, prior to surface break-down, is important to identification of early disease potentially amenable to therapeutic interventions. This work examines the diagnostic potential of ultrashort echo time-enhanced T2∗ (UTE-T2∗) mapping to detect human meniscus degeneration in vitro and in vivo in subjects at risk of developing OA.
UTE-T2∗ maps of 16 human cadaver menisci were compared to histological evaluations of meniscal structural integrity and clinical magnetic resonance imaging (MRI) assessment by a musculoskeletal radiologist. In vivo UTE-T2∗ maps were compared in 10 asymptomatic subjects and 25 ACL-injured patients with and without concomitant meniscal tear.
In vitro, UTE-T2∗ values tended to be lower in histologically and clinically normal meniscus tissue and higher in torn or degenerate tissue. UTE-T2∗ map heterogeneity reflected collagen disorganization. In vivo, asymptomatic meniscus UTE-T2∗ values were repeatable within 9% (root-mean-square average coefficient of variation). Posteromedial meniscus UTE-T2∗ values in ACL-injured subjects with clinically diagnosed medial meniscus tear (n=10) were 87% higher than asymptomatics (n=10, P<0.001). Posteromedial menisci UTE-T2∗ values of ACL-injured subjects without concomitant medial meniscal tear (n=15) were 33% higher than asymptomatics (P=0.001). Posterolateral menisci UTE-T2∗ values also varied significantly with degree of joint pathology (P=0.001).
Significant elevations of UTE-T2∗ values in the menisci of ACL-injured subjects without clinical evidence of subsurface meniscal abnormality suggest that UTE-T2∗ mapping is sensitive to sub-clinical meniscus degeneration. Further study is needed to determine whether elevated subsurface meniscus UTE-T2∗ values predict progression of meniscal degeneration and development of OA.
. The ability to non-invasively detect molecular changes that indicate loss of meniscal integrity in knees at risk for developing OA, particularly those suffering traumatic injury, is important to identifying disease states that may be amenable to therapeutic interventions.
Magnetic resonance imaging (MRI) and physical examination remain the primary modes for diagnosis of gross meniscal pathology, with current technologies permitting sensitivities and specificities generally better than 80% and 90%, respectively, as confirmed by arthroscopy
. Conventional MRI evaluation of the meniscus focuses on two aspects of tissue status: integrity of the surfaces as indications of tear and gross tissue loss, and relative signal intensity within the meniscus interior as indication of tissue hydration. The sensitivities and specificities of MRI and physical exam to meniscal degeneration prior to break-down of the surface are more difficult to assess
have been investigated as tools to assess the biochemical status of the meniscus with the goal of detecting microscopic alterations before gross damage occurs. Quantitative T1rho measurements detect differences between healthy controls and Anterior Cruciate Ligament-injured subjects within 3 months of injury in the posterior lateral meniscus, but the physiologic basis for this difference is not yet understood
. However, the biophysical basis for this observation also needs more investigation. A study by Rausher et al found that meniscus T2, but not T1rho, differentiated healthy controls from subjects with mild or severe OA after adjustment for age
. Another study by Zarins et al found that while both T2 and T1rho discriminated intact from torn menisci, neither metric discriminated healthy menisci from intact menisci with intra-substance abnormality
T2 is usually measured with spin echo sequences that, in order to satisfy safety restrictions, use long echo times (TE∼10 ms) and, consequently, only capture long-T2 relaxations. T2∗, the effective T2 shortened by phase dispersion from both main and local magnetic field inhomogeneities, captures fast-T2 relaxations (T2∗<10 ms) that reflect spin–spin interactions of protons bound to collagen and the degree of collagen fibril-alignment
. Ultrashort echo time-enhanced T2∗ (UTE-T2∗) mapping is a novel quantitative technique with the potential to measure short-T2∗ relaxations from joint tissues that are not well captured by standard T2 mapping
The aim of the current work is to examine the diagnostic potential of UTE-T2∗ mapping to detect meniscus degeneration. UTE-T2∗ of meniscus is evaluated in vitro, compared to histology as the gold standard, and in vivo, in ACL-injured human patients with meniscal abnormalities related to ACL tear.
Patients and methods
Human meniscus specimens
Sixteen human meniscus specimens (eight medial, eight lateral) were obtained from six intact cadaver knee joints (LifeLegacy Foundation, Phoenix, AZ, USA) and two tibial plateau explants with menisci attached (Musculoskeletal Transplant Foundation, Edison, NJ, USA). Menisci sourced from intact knee joints varied in health by gross inspection. Menisci sourced from explants were from specimens originally harvested for meniscal allograft transplantation.
MRI of meniscus specimens
All MRI images were acquired on a clinical 3T MRI scanner (MAGNETOM Trio Tim, Siemens, Erlangen, Germany) with an eight-channel knee coil (Invivo Inc., Gainesville, FL, USA). Intact cadaver joints (n=6) were scanned with the NIH-sponsored Osteoarthritis Initiative (OAI) sequences and scanner (www.oai.ucsf.edu) described in detail by Peterfy et al
. The OAI images of intact cadaver joints were reviewed by a musculoskeletal radiologist who assessed each joint for meniscus degeneration and/or tear. Low signal in all pulse sequences indicated normal meniscus. Abnormal signal within the meniscus was considered secondary to degeneration: Grade 1 – minimal, linear signal within the meniscus; Grade 2 – globular signal replacing most of the meniscus; Grade 3 – abnormal signal extending to the femoral, tibial or inner margin of the meniscus. Abnormal signal extending to the meniscus surface in more than two consecutive slices indicated tear. UTE-T2∗ mapping images were acquired on all meniscal specimens.
Following MRI, menisci were dissected from intact joints (n=12 menisci from six knees) and from tibial plateaus (n=4 menisci from two explants) and fixed in 10% neutral buffered formalin. Gross sections (1 cm wide) were removed from the posterior medial and lateral horns to coincide with the MRI sections evaluated (Fig. 1). Histologic sections were then processed, paraffin-embedded, vertically sectioned (5 μm thickness) parallel to the sagittally oriented MR images, and stained with hematoxylin/eosin (HE), alcian blue (AB), Safranin-O (Saf-O), and picro-sirius red (PSR) using standard techniques
. PLM analysis of the collagen network was performed using a Nikon Eclipse TE2000-U polarized light microscope (Nikon, Chiyoda-ku, Tokyo) with two orthogonal polarizers. PSR-stained sections were placed between the polarizers and rotated so that the polarizers were aligned 45° against the superior meniscus surface. HE, AL, Saf-O and PSR/PLM images were graded using a simplified version of the scoring system devised by Pauli et al
Ten asymptomatic human subjects with no known or suspected knee pathology and 25 subjects recruited from the clinical practices of three surgeons for ACL tear participated in these studies. All subjects gave informed consent and all studies were approved by the institutional review board.
Asymptomatic subjects (n=10) underwent MR imaging for UTE-T2∗ mapping of the left knee, one time daily for three consecutive days. Subjects were scanned at same time (±1 h) each day. One asymptomatic subject was scanned on only two consecutive days. ACL-injured subjects (n=25) underwent imaging of the injured knee one time within 4 weeks prior to ACL-repair surgery. At the time of surgery, meniscus status was assessed with arthroscopic examination and palpation. Status of menisci in ACL-injured subjects (tear/no tear; degree of degeneration) was determined from surgical and/or clinical MRI reports (clinical MRI report was unavailable in five cases).
UTE-T2∗ imaging and mapping
UTE-T2∗ mapping images for all tissue specimens and human subjects were acquired with a three-dimensional (3D) sequence acquisition-weighted stack of spirals (AWSOS)
. Briefly, the single-echo AWSOS acquisition was repeated 11 times at 11 different echo times (0.6, 1, 2, 3, 4, 5, 7, 10, 20, 30, 40 ms). Other acquisition parameters were: 140 mm field of view (FOV), 2 mm slice thickness, 52–60 slices, 24 in-plane spirals, 11.52 ms spiral readout time, 5 μs data sampling interval. For in vitro imaging (cadaver knee joints and explants), FA/TR was 30/100 ms with 512 matrix and scantime of 5.12 min per TE image, for a total scantime of 56 min. For in vivo imaging, FA/TR was 30/80 ms with 256 matrix and scantime of 1.92 min per TE image for a total scantime of 22 min. UTE-T2∗ acquisitions were sagittally oriented and centered on the femorotibial joint.
Regions of interest (ROIs) were manually segmented, by one individual with 10 years prior segmenting experience, from a single section from each of the medial and lateral compartments of each knee or explant. In order to reduce variability across subjects with respect to hoop-fiber alignment, a mid-sagittal slice was selected so that the cross-section of meniscus captured in the imaging plane was orthogonal to the hoop fibers [Fig. 1(a)]. The posterior horns were chosen for evaluation because they are in the zone of injury during ACL tear. Menisci were segmented by manually outlining the posterior aspects of the medial or lateral meniscus seen on an AWSOS image collected at TE time of 7 ms in which the contrast between menisci and surrounding tissues was strongest [Fig. 1(b)]. The mean and standard deviations (SDs) of UTE-T2∗ values observed in each meniscus ROI were calculated and recorded. Regional differences were assessed by segmenting smaller sub-regions within the posteromedial meniscus [Fig. 1(c)]. Sub-ROIs included the “inner” portion of the meniscus, consisting of approximately 40% of all pixels seen in the posteromedial meniscus section and the “outer” portion, consisting of approximately 25% of pixels. Inner and outer sub-regions were segmented to roughly correspond to the red and white zones delineated by Gatehouse et al
. Regional differences of the posterolateral meniscus were not evaluated because inner/outer segmentation was more difficult laterally preventing consistent segmentation across subjects.
In vivo UTE-T2∗ mapping repeatability assessment
Three-day intersession UTE-T2∗ repeatability was evaluated in two ways. First, relative inter-subject intersession reproducibility across all 10 asymptomatic subjects was expressed by the root-mean-square average coefficients of variation (RMSA-CV) for each ROI. RMSA-CV was determined by √((∑CV2)/n) where intra-subject CV was calculated by dividing the SD of a subjects' UTE-T2∗ values from Day 1, Day 2, and Day 3 by the mean of the subjects' UTE-T2∗ values from Day 1, Day 2, and Day 3 for each ROI, where n was the number of subjects. Second, absolute intersession precision was expressed as median of the intra-subject SDs for each ROI.
In vitro and in vivo, medial and lateral meniscus UTE-T2∗ values were evaluated separately. In vitro, non-parametric Spearman's rho assessed correlations between meniscal specimens' UTE-T2∗ values and histopathic scores. In addition, specimens were binned into two groups by histologic score (‘less degenerate’ scores 0–4 vs ‘more degenerate’ scores 5–9) and average UTE-T2∗ values calculated for each group of medial or lateral menisci, respectively. Non-parametric Mann–Whitney tests assessed pairwise differences between groups. Human subject UTE-T2∗ values were binned into three groups: asymptomatic, ACL-injured without clinical evidence of subsurface meniscal abnormality, and ACL-injured with tear to the meniscus. Non-parametric Kruskal–Wallis statistics examined differences across these groups and Mann–Whitney tests assessed pairwise differences. Regional differences in UTE-T2∗ values were assessed by non-parametric Wilcoxon signed rank tests. All statistical analyses were performed using Excel (Microsoft, Seattle, WA, USA) and SPSS (SPSS Inc., Chicago, IL, USA). Statistical significance was accepted for P<0.05.
In vitro histologic and clinical MRI assessments
Table II summarizes clinical MRI assessment, bulk mean UTE-T2∗ value and histopathic score of each meniscus sample examined. In vitro, mean UTE-T2∗ values ranged 6–13 ms. Areas of histologically confirmed tear showed focally higher values in the region of the tear (Fig. 2). Non-parametric correlation of unbinned UTE-T2∗ values and histopathic scores shows UTE-T2∗ and histology are not significantly correlated in either medial or lateral menisci (Spearman's rho P=0.62, 0.54; P=0.10, 0.16, respectively). UTE-T2∗ values binned by histologic grade (scores 0–4 vs 5–9) were not found to vary significantly with histopathic degeneration in this small sample size (medial n=8, Mann–Whitney P=0.13; lateral n=8, Mann–Whitney P=0.29). Medially, more degenerated samples (scores 5–9, n=6) showed a mean UTE-T2∗ value of 9.7 [95% confidence interval (CI)=8.1–11.3] ms, while less degenerated samples (scores 0–4, n=2) had an average UTE-T2∗ value of 6.5 (95%CI=5.5–7.5) ms. Laterally, the difference was smaller: 8.3 (6.4–10.1) vs 9.8 (7.2–12.3) ms, n=4, 4.
Table IIUTE-T2∗ values, clinical MRI and histologic evaluations of human menisci specimens
Comparison across all three in vitro measures of meniscus status was performed by grossly binning each of the metrics: mean UTE-T2∗ value (≤8 ms or ≥10 ms); histopathic score (≤3 or ≥5); clinical MRI assessment by a musculoskeletal radiologist (tear/no tear, degenerate/normal signal). Five specimens demonstrated relatively low mean UTE-T2∗ values (≤8 ms), relatively normal histopathic scores (≤3) and showed no clinical signs of tear or degeneration [Fig. 3(a) ]. Five other specimens demonstrated relatively high mean UTE-T2∗ values (≥10 ms), relatively degenerate histopathic scores (≥5), and evidence of tear and/or degenerate signal in corresponding clinical MRI images of posterior meniscus [Fig. 3(c)]. Five of the remaining specimens demonstrated highly heterogeneous UTE-T2∗ maps, relatively high histopathic scores (≥5) and grade I–III degeneration on the clinical MRIs [Fig. 3(b)]. In the last specimen, UTE-T2∗ was high (11 ms), but histology (score 3) did not support the clinical diagnosis of grade II degeneration.
Meniscus UTE-T2∗ mapping repeatability in vivo
Example in vivo UTE-T2∗ maps from a representative asymptomatic subject of the repeatability analysis are shown in Fig. 4. UTE-T2∗ values in asymptomatic subjects' menisci (n=10, mean age=27 (95%CI=25–29) years, mean body mass index (BMI)=24 (95%CI=22–26), five females, 10 left) exhibited relative intersession precision errors of 9% (RMSA-CV) corresponding to absolute precision errors of 1.0 ms.
Regional UTE-T2∗ variations
No differences were detected between the “inner” and “outer” aspects of the posteromedial menisci of asymptomatic subjects (n=10, Wilcoxon signed rank test, P=0.17) or ACL-injured subjects with (n=10, Wilcoxon signed rank test, P=0.68) or without concomitant medial meniscus pathology (n=15, Wilcoxon signed rank test, P=0.69). The mean age and BMI of ACL-injured subjects were 29 (95%CI=25–32) years and 28 (95%CI=26–30), respectively (n=25, 15 females, 16 left).
Meniscus UTE-T2∗ quantitative comparisons
UTE-T2∗ voxel values, calculated with a mono-exponential T2-fit, observed in menisci of asymptomatic subjects ranged approximately 6.2–17.1 ms (tenth, ninetieth percentiles, respectively). The average UTE-T2∗ value across the entirety of the posteromedial meniscus observed among the 10 asymptomatic subjects was 9.8 (95%CI=8.9–10.7) ms. Compared to asymptomatic subjects, the UTE-T2∗ maps of subjects with ACL-injury appeared more heterogeneous and demonstrated higher UTE-T2∗ values, as shown in Fig. 5. Quantitatively, meniscus UTE-T2∗ values varied significantly with degree of joint pathology (Kruskal–Wallis, P<0.0001), as shown in Fig. 5d. UTE-T2∗ menisci values in subjects with clinically diagnosed medial meniscus tear (n=10, mean=18.3 (95%CI=15.0–21.6) ms) were significantly greater (87% higher, Mann–Whitney test, P<0.0001) than the menisci of asymptomatic subjects. Menisci UTE-T2∗ values of ACL-injured subjects without clinical evidence of subsurface medial meniscus abnormality (n=15, mean=13.1 (95%CI=11.7–14.4) ms) were 33% higher (Mann–Whitney, P=0.001) than menisci UTE-T2∗ values of asymptomatic subjects. There was also a significant difference between ACL-injured subject groups: those with concomitant medial meniscus tear had significantly higher posteromedial UTE-T2∗ values than ACL-injured subjects without clinical evidence of subsurface meniscus pathology (Mann–Whitney, P=0.01).
In lateral menisci, UTE-T2∗ values of the posterior aspect varied significantly with degree of joint pathology (Kruskal–Wallis, P=0.001), as shown in Fig. 5e. UTE-T2∗ menisci values in subjects with clinically diagnosed lateral meniscus tear (n=6, mean=16.7 (95%CI=11.3–22.1) ms) were 77% higher (Mann–Whitney, P=0.001) than the menisci of asymptomatic subjects (n=10, mean=9.4 (95%CI=8.6–10.2) ms). Menisci UTE-T2∗ values of ACL-injured subjects without lateral meniscus abnormality (n=19, mean=11.6 (95%CI=10.5–12.7) ms) were 23% higher (Mann–Whitney, P=0.005) than menisci UTE-T2∗ values of asymptomatic subjects. There was also a significant difference between ACL-injured groups: those with concomitant lateral meniscus tear had significantly higher posterolateral UTE-T2∗ values than ACL-injured subjects without clinical evidence of subsurface meniscus pathology (Mann–Whitney, P=0.019).
In vivo 3D UTE-T2∗ mapping of the meniscus is feasible, repeatable, and achievable in scantimes tolerated by subjects with knee pain. This study demonstrates that non-invasive UTE-T2∗ mapping is sensitive to meniscus tears and can be used to quantitatively characterize meniscus degeneration in patients. Significant elevations of UTE-T2∗ values in the posterior horns of menisci of ACL-injured subjects without clinical evidence of subsurface meniscal abnormality suggest that UTE-T2∗mapping is sensitive to sub-clinical meniscus degeneration in knees known to be at risk for developing OA. This finding is not unexpected given that the posterior horn of the medial meniscus is a secondary restraint to anterior tibial translation and the posterior horns of both menisci are in the pathway of ACL-injury.
The histopathic and clinical status of meniscus tissue is reflected by the magnitude and spatial distribution of UTE-T2∗ values in the tissue. In vitro, mean UTE-T2∗ values were relatively low (≤8 ms) in samples exhibiting no or only mild degeneration by histology and without clinical MRI evidence of tear or degeneration. Specimens exhibiting both histopathic and clinical MRI evidence of more severe degeneration had relatively high mean UTE-T2∗ values (≥10 ms). In 30% of the cases examined, however, relatively low mean UTE-T2∗ values (≤8 ms) were seen in meniscus tissue found to be degenerate both clinically and histologically. In each of these cases, UTE-T2∗ maps demonstrated highly heterogeneous distributions of UTE-T2∗ values with focal regions of elevated UTE-T2∗ values corresponding to clinically diagnosed regions of degeneration. These findings highlight the inadequacy of relying on a mean UTE-T2∗ value that is calculated by averaging across all UTE-T2∗ voxel values within an ROI. Measures of UTE-T2∗ texture, like heterogeneity, entropy or measures of order
, may increase the sensitivity of UTE-T2∗ to different degrees of tissue degeneration and improve its ability to characterize the molecular integrity of the tissue.
The meniscus histology grading system used in this study was adapted from the system developed by Pauli et al to ‘quantify zonal degeneration . . . for an improved assessment of progressive changes with aging and disease’
. A zonal analysis of spatially matched MRI and histology images may help improve the correlation between UTE-T2∗ values and histopathic scores. However, because of the difference in scale between histology (micrometers) and the MRI images (millimeters) in this study, the images represented different quantities of tissue. Therefore, a finer zonal analysis of tissue sub-regions was not pursued.
The time-course for UTE-T2∗ change in the meniscus following ACL-injury or due to OA has not to our knowledge been reported and has the potential to be highly variable between patients. Among asymptomatics, however, meniscus UTE-T2∗ maps were repeatable with intersession precision error of 9% RMSA-CV, corresponding to an absolute error within 1 ms. This is likely a conservative estimate of UTE-T2∗ repeatability because RMSA-CV emphasizes values not near to center of the distribution and is sensitive to small changes in UTE-T2∗ mean. The repeatability error reported here is in-line with intercession repeatability reports from other qMRI measures of meniscus physiology in the literature, including T1rho (3–13% CV)
. Intersession UTE-T2∗ repeatability of 9% RMSA-CV measured in asymptomatics, a clinically stable population, helps to establish the difference between day-to-day fluctuations in the UTE-T2∗ metric from potentially clinically relevant changes observed in menisci of diseased or injured knees. It is important to point out that while UTE-T2∗ repeatability may be worse in an ACL-injured population, the 9% repeatability achieved here is likely more than sufficient to detect the 87% increase in UTE-T2∗ values between asymptomatics and subjects with torn menisci and the 33% increase between asymptomatics and ACL-surgery subjects without concomitant meniscus tear. Previously, Tsai et al reported the average T2 value of the posterior horn of the medial meniscus in normal young adults to be 9.65 ms
. The UTE-T2∗ values measured in menisci of asymptomatic subjects in the current work (9.8 ms) are in excellent agreement with this prior value.
An extensive macroscopic and histopathic analyses of 214 human menisci recently reported by Pauli et al concluded that ‘degeneration of the menisci initiates within the substance of the tissue rather than the surface’
. The results of the current work suggest that UTE-T2∗ mapping is sensitive to early degenerative changes within the substance of the meniscus. Specifically, ACL-injured subjects with intact menisci that do not exhibit evidence of degeneration detectable by gross inspection, arthroscopic palpation or clinical MRI evaluation, have significantly elevated meniscus UTE-T2∗ values compared to healthy controls. In light of these findings, it is reasonable to conclude that non-invasive MRI UTE-T2∗ mapping of subsurface meniscus structural integrity is sensitive to sub-clinical meniscus degeneration in a population known to be at risk for meniscal degeneration. Longitudinal evaluation is needed to determine the clinical implications of subsurface degeneration detected by UTE-T2∗, and whether UTE-T2∗ mapping of meniscus can be used to predict progressive meniscal degeneration or the development of broader joint pathology.
The UTE-T2∗ value reported in this work is not a pure measure of any single component of T2∗ relaxation in meniscus. Rather, the UTE-T2∗ metric, calculated from a mono-exponential fit routine, represents a weighted combination of all T2∗ decay components present in the same voxel. An ex vivo examination of multi-component UTE-T2∗-fitting demonstrated that up to four types of T2∗ decay types could be detected in articular cartilage
. Therefore, it is expected that a larger proportion of total T2∗ signal from meniscus voxels will emanate from short-T2 components compared to articular cartilage. A preliminary exploration of multi-exponential analyses of UTE-T2∗ data from normal human menisci found that a bi-exponential model provided a better fit to the image data did a mono-exponential model
. Further, 46% of total UTE-T2∗ signal in meniscus was found to be due to short-T2 components in the tissue (mean short-T2∗ relaxation time of 1.54 ms), and 53% of total signal was due to long-T2∗ components (mean long-T2∗ relaxation time of 13.6 ms)
. The relative values and distributions of long and short-T2∗ components in torn and degenerate menisci have yet to be rigorously evaluated.
In this work, degenerative menisci demonstrated higher UTE-T2∗ values than menisci of asymptomatic subjects, and only 10% of pixels in asymptomatic menisci had values less than 6.2 ms. Strictly speaking, a UTE (i.e., echo time <1 ms) is not required to study meniscus T2∗ relaxation. However, the inclusion of a UTE with high signal at 0.6 ms facilitates both capture and curve-fitting of rapidly decaying ‘short’ (i.e., <6 ms) T2∗ signals providing increased sensitivity to subtle differences between meniscus regions that may help to detect earlier changes to meniscus health. The choice of the 11 echoes acquired in this work was determined by Monte Carlo simulations
for covering both short- and long-components of T2∗ relaxation. Further work is needed to optimize the number of echoes needed to adequately detect subtle UTE-T2∗ differences across tissue regions and across patients while also reducing total scantime.
UTE-T2∗ mapping is a novel tool for the detection and quantification of subsurface meniscus degeneration. The ability to diagnose and quantitatively stage meniscus status, prior the surface break-down is important to identifying disease states potentially amenable to interventions to delay or prevent the onset of OA. Further study is needed to determine whether elevated subsurface meniscus UTE-T2∗ values, particularly in the absence of clinically diagnosed abnormality, predict progression of meniscal degeneration and development of OA.
Ashley Williams contributed to conception and design of the study, collection and assembly of data, analysis and interpretation of data, drafting, critical revision and final approval of the article. Dr. Yongxian Qian contributed to acquisition of data, revising the article for critical intellectual content, and final approval of the article. Dr. Sara Golla contributed to analysis and interpretation of data, revising the article for critical intellectual content, and final approval of the article. Dr. Constance Chu is responsible for the integrity of the work as a whole, and she obtained the funding for this study. Dr. Chu also contributed to conception and design of the study, provision of patients and study materials, acquisition and interpretation of study data, revising the article for critically important intellectual content, final approval of the article.
Funding for this work was provided by the NIH [ RO1 AR052784 (CR Chu) and P60 AR054731 (CR Chu/K Kwoh)]. The study sponsors had no involvement with collection, analysis or interpretation of data, nor any involvement with writing this manuscript or the decision to submit it to Osteoarthritis and Cartilage.
Conflict of interest
The authors of this work have no conflicts of interest to report relevant to this work.
The authors would like to acknowledge the contributions of Michele Mulkeen for her excellent histology preparations and Stephen Bruno for his tireless coordination of study subjects and scanner time.
Meniscal tear as an osteoarthritis risk factor in a largely non-osteoarthritic cohort: a cross-sectional study.