This is the largest reported MRI study (n
32) of patients with LGMD2I due to the common mutation in FKRP
, and the first longitudinal natural history study to use quantitative MRI in LGMD2I. Quantitative fat imaging by the Dixon method demonstrated significant increases in fat replacement in 9 muscles across a 12 month period, while conventional physical testing of the patients did not demonstrate significant differences: indeed the median six minute walk distance appeared to increase, though not significantly. Qualitative grading of T1-weighted MRI images by a six-point scale was also shown to be insensitive. This study demonstrates the power of quantitative MRI to provide reliable objective outcome measures in a slowly progressing disease, as would be required in a clinical trial of therapy. It also provides new information about the rate of natural progression of pathology in individual muscle groups.
In we have highlighted potential target muscles for the longitudinal assessment of pathological changes that were identified by our study. Targeting specific muscles for analysis in a longitudinal trial would maximize the power of a large therapeutic study. Muscles suitable for longitudinal analysis in LGMD2I were the medial gastrocnemius muscle in the lower leg and the vastus lateralis, gracilis and rectus femoris muscle in the thigh. These were identified not only because of fat fraction progression over 12 months, but because they demonstrated a wide range of fat fractions in patients at both the early and advanced stage of the disease and the ROIs were easy to delineate on follow-up. Unsuitable muscles included the tibialis anterior, which was only mildly affected at baseline and did not progress, and the biceps femoris long head which was already heavily fat-infiltrated at baseline and therefore unlikely to respond to therapy. Consistently well-positioned patients were essential for the delineation of muscles and the ROI placements at follow-up to enable reliable and uniform analysis of muscle pathology.
This work demonstrated that qualitative scoring of T1w images was not discriminating enough to detect progression of muscle pathology in patients with LGMD2I over a 12-month period, compared to the quantitative fat imaging method, which demonstrated significant increases in 9 out of the 14 muscles analyzed. The qualitative scoring is subjective and there are difficulties in assigning grades to some muscles due to the imprecise boundaries of the scoring system, especially between grades 2a, 2b and 3, leading to potential inconsistencies in scoring the T1w images. The interobserver repeatability of the T1w scores showed that whilst there was overall agreement in the scoring of the qualitative grades (65% of assessments), there was only 39% agreement in scoring grade 2b. The quantitative analysis of the images demonstrated good interobserver repeatability (inter-observer coefficient 1.43%) and was an objective measure of fat content in the individual muscles.
As shown in , the functional testing and physical strength measurements did not show any significant difference from baseline to follow up, apart from the FVC measurements: while these measurements are clinically important, their relationship to lower limb pathology is known to be indirect. The finding of decreased FVC can have diagnostic value in distinguishing LGMD2I patients from other LGMDs with normal respiratory function: it is important that lung function is monitored in any longitudinal trial, since respiratory care (including non-invasive ventilation) is an important treatment approach for LGMD2I patients. There were no significant correlations between the change in FVC and the change in fat fraction of any muscle group. Therefore, whilst functional assessments and strength tests are important as they provide clinically relevant information on the level of ability in these patients, objectively there was no statistical difference detected over 12 months in the tests. The median six minute walk distance, often used as the primary end-point of clinical trials of therapy, actually appeared to increase after 12 months demonstrating its unsuitability for use in slow progressing diseases. Learning effects from test repetition may affect these types of measurement.
Future work could include both asymptomatic and paediatric patients with the common FKRP mutation, in order to further define the natural history of the disease. Limitations of the study included the analysis of the quantitative Dixon scans at only one level. Ideally, one would like to assess the ROIs at all levels of potential target muscles and analyze the variability within muscles. This may give further information regarding the process of muscle damage occurring in this condition. In addition, while the technique can assess the changes in fat and water composition of the tissue, it cannot measure macromolecular species which may not be MR-visible. This means that other pathological processes which occur in LGMD2I, such as the development of fibrosis, cannot be assessed by this technique.
This study has shown that quantitative MR imaging by the Dixon technique provides an objective measurement of the muscle fat fraction that detects disease progression that cannot be identified by conventional functional testing, including measures such as the 6MWD presently used as primary trial end-points. In a slowly progressing neuromuscular condition, such as LGMD2I, quantitative MRI is able to demonstrate significant increases in muscle pathology in 9/14 of the muscles studied. Quantitative MRI could in the future be considered as a primary end-point of choice in the longitudinal monitoring of these patients in clinical trials.