Abstract
INTRODUCTION
Graves’ disease (GD) is an autoimmune disease affecting the thyroid gland, orbital soft tissues and subcutaneous tissues of the extremities. Ophthalmic signs are clinically evident in 25-50% of patients with GD and 3-10% of cases develop severe disease.
This study discusses the current understanding of the pathophysiology of GD and the associated ophthalmopathy. It summarises the clinical features of the condition and examines the numerous scoring systems used to measure and monitor the ophthalmic features of the disease.
The three main radiological techniques used to image the orbit, ultrasound, computed tomography (CT) and magnetic resonance imaging (MRI) are discussed.
Although CT is regarded as the radiological investigation of choice for imaging orbit details subtle changes related to Graves’ ophthalmopathy (GO) can be missed. To accurately calculate extraocular muscle volume we have developed a computer software package, the Volume Estimation Tool. This software allows outlining of structures from a CT image and then calculates the volume of the structure.
The aim of this study is to determine if extraocular muscle volume and orbital volume estimates from CT imaging correlate with clinical signs and scores of disease activity or severity.
METHODS
Study Design
This is a retrospective study comparing the orbital CT scan findings in patients with GD with their clinical activity, as close as possible to the time of the scan, and their disease severity.
The initial hope had been to use high definition 3 dimensional ultrasound to calculate muscle volume. This proved inaccurate and non-reproducible for calculating muscle volumes.
We therefore, worked in conjunction with the Christchurch Hospital Medical Physics department to develop a computer programme to estimate muscle volume from coronal slices of CT scans. This volume estimation tool was used to calculate the volume of muscles and soft tissue orbits in a series of patients with clinical evidence of GO. The same muscle and orbital volumes were calculated for a set of matched controls without GO.
In addition, methods of measuring disease state in patients with GO were compared. Disease activity and severity were calculated from the clinical records of the patients with GD as close as possible to time of the CT scan. The activity of the GO and the severity of the disease were then correlated with the muscle and orbital volumes to see what relationships there are. The clinical symptoms and signs of GO were then related to muscle volume to determine which have the greatest impact on clinical activity and disease severity.
Participants
A record of all patients with thyroid eye disease was compiled by correlating records of patients seen in both the thyroid clinic, Nuclear Medicine Department and the ophthalmology clinic at Christchurch Hospital. From this cohort, we used radiology department records to obtain a subset of patients for whom head/neck CT scans had been ordered. A control group of patients with no history of GD and CT scans with normal radiology reports from the same period were selected.
Outcome Measures
Clinical activity and disease severity were calculated by different methods and compared. The activity and severity were also compared to the estimated muscle and orbital volumes to determine any possible relationships.
RESULTS
In patients with GO muscle volume may be a more sensitive indicator of muscle enlargement than subjective viewing of the CT images looking for enlargement of muscle bellies. Muscle volume for patients with GO was significantly higher than for patients with no GO (p<0.0001). Patients with GO also have larger orbital volumes than patients without GO (p=0.003).
Muscle volume in patients with GO also correlates positively with clinical activity (p<0.0001) and disease severity (p<0.0001).
When individual clinical features are examined there is a weak correlation between muscle volume and proptosis (p=0.05), but no association with lid retraction and muscle volume (p=0.12). Neither is there a correlation between horizontal or vertical muscle restriction and muscle volume (p=0.56 and p=0.44 respectively).
CONCLUSIONS
The results show that accurate calculation of extraocular muscle volume using this software package is possible. We confirm that calculating extraocular muscle volume in patients with GO gives additional useful information for assessing and managing patients with GO.
Muscle volume correlates positively with clinical activity and disease severity. Patients with more active and more severe disease have greater changes in muscle volume than those without.
For 2 patients with optic nerve compromise there was a high percentage of orbital volume taken up by muscle. This suggests this ratio, rather than muscle volume alone, may predict risk of optic nerve damage.
However, patients with very active disease or severe residual disease do not have an increased risk of restrictive eye movements despite having larger total muscle volumes.