Quantitative Mapping of Glenohumeral Cartilage in Asymptomatic Subjects Using 3.0 T Magnetic Resonance
Imaging
Carly A. Lockard1, Katharine J. Wilson1, Charles P. Ho1, Richard C. Shin1, J. Christoph Katthagen1, Peter J. Millett1,2
1
Steadman Philippon Research Institute, Vail, CO, 2The Steadman Clinic, Vail, CO
Disclosures: Carly A. Lockard (None), Katharine J. Wilson (None), Charles P. Ho (Steadman Philippon Research Institute Scientific Advisory Board
member, Rotation Medical consultant, *), Richard C. Shin (None), J. Christoph Katthagen (None), Peter J. Millett (Arthrex consultant and royalties, Game
Ready, VuMedi stock and stock options, Springer Publishing royalties, *). *(Steadman Philippon Research Institute Research Support from: Smith &
Nephew Endoscopy, Arthrex, Siemens Medical Solutions, USA, Ossur Americas, Vail Valley Medical Center)
ABSTRACT INTRODUCTION: Transverse relaxation time (T2) mapping is a quantitative magnetic resonance imaging technique that allows objective,
non-invasive assessment of tissue properties including water content and collagen organization. T2 mapping values have been correlated to factors
associated with cartilage degeneration, and may allow early detection and treatment of osteoarthritis. The objective of this study was to quantify T2 mapping
values in asymptomatic shoulders at 3.0 T for the entire mappable region of the glenohumeral cartilage in the coronal and sagittal planes using a clinically
feasible mapping methodology. The mapping protocol and normative baseline values could then be used for future comparison with symptomatic shoulders
with cartilage degeneration, as a diagnostic tool for early cartilage degeneration.
METHODS: This study was approved by the Institutional Review Board and all subjects provided informed consent. 21 asymptomatic volunteers
underwent unilateral shoulder scans including T2 mapping in the coronal and sagittal planes at 3.0 T. The T2 maps were manually segmented by two raters,
with one rater segmenting all images twice to assess intra-rater reliability. The humeral head cartilage was segmented in the sagittal and coronal planes,
whereas the glenoid cartilage was segmented only in the coronal plane due to its orientation parallel to the sagittal plane. The T2 means and medians for
each cartilage region and plane were calculated and analyzed. The cartilage coverage provided by each mapping plane was qualitatively assessed.
RESULTS SECTION: The coronal humeral head segmentation captured the medial-inferior to superior-lateral cartilage including limited portions of the
anterior and posterior regions, since curvature and partial volume averaging difficulties increase anteriorly and posteriorly in the coronal plane (shown for
one subject in Fig.1, a). The sagittal segmentation captured the superior to superior-lateral cartilage over a wider anterior-posterior area, but failed to capture
the medial-inferior portion of the cartilage due to curvature and partial volume averaging difficulties (Fig.1, b).The coronal segmentation of the glenoid
cartilage spanned the majority of the glenoid face (Fig.1, c).
The mean glenoid cartilage T2 was 37.86 ± 2.06 ms (mean ± standard deviation). The mean humeral head cartilage T2 was 40.79 ± 2.74 ms (coronal
plane) and 34.20 ± 2.46 ms (sagittal plane). The coronal plane humeral head cartilage T2 was significantly greater than the sagittal plane humeral head
cartilage T2 (p<0.05), perhaps due to the different regions included in each plane and different stresses experienced in those regions. Figure 1, d-f shows the
T2 distribution for one subject over the coronal and sagittal humeral head and coronal glenoid segmentations. The intra-rater and inter-rater ICC for each
segmented region (glenoid: 0.91 and 0.64; humeral head, coronal plane: 0.71 and 0.81; humeral head, sagittal plane: 0.55 and 0.49, respectively) showed
either “fair to good” or “excellent” agreement for the two raters.
DISCUSSION: Baseline T2 values for humeral head and glenoid cartilage were established, including previously unreported sagittal plane humeral head T2
values. Due to the humeral head’s curvature, two-dimensional (2D) T2 mapping in a single plane does not provide complete humeral head cartilage
information. The coronal plane is most appropriate for segmentation of the medial-inferior to superior-lateral region of the humeral head, while the sagittal
plane is most appropriate for segmentation of the superior-anterior to superior-posterior region. The shoulder joint would benefit from future 3D quantitative
mapping techniques, particularly in the humeral head cartilage, to avoid curvature and partial volume averaging difficulties. Limitations of this study include
difficulty in segmenting the thin cartilage, the potential for magic angle effect (artificially long T2 for cartilage oriented at 55 degrees to the primary
magnetic field), and partial volume averaging due to the chondral curvature through the mapping plane.
SIGNIFICANCE: In order for T2 mapping to be applied for early detection of cartilage degeneration in the clinical setting a feasible mapping methodology
and normative baseline must be established. This work establishes a T2 mapping protocol and baseline T2 values for shoulder cartilage with clinically
feasible scan times in the sagittal and coronal planes, and demonstrates the advantages and limitations of each plane in terms of cartilage coverage.
ACKNOWLEDGEMENTS: The authors thank Bill Brock for contributing his time and expertise for MRI scan acquisition and Grant Dornan for providing
statistical analysis.
IMAGES AND TABLES:
Figure 1: 3D volume rendering of humeral head cartilage segmented in the A) coronal and B) sagittal planes. C) 3D volume rendering of glenoid cartilage
segmented in the coronal plane. 3D reconstruction of D) coronal and E) sagittal segmentation of humeral head cartilage with T2 mapping color-scale
applied. F) 3D reconstruction of coronal segmentation of glenoid cartilage with T2 mapping color-scale applied.
ORS 2017 Annual Meeting Poster No.2234