View 2019 Abstracts

The use of texture analysis software is not new, and has been explored in several medical specialties in attempts to aid radiological or even microbiological diagnosis of a variety of diseases(1-3). Within Ophthalmology texture analysis software has shown potential in the ability to classify orbital tumours(4).

Thyroid associated ophthalmopathy is the most common cause of orbital inflammation and proptosis in the adult population within the western population(5), affecting over 400,000 patients in the UK(6). Identifying active thyroid associated ophthalmopathy and prompt treatment is crucial in preventing detrimental impact to structures of the eye. We know through previous works of Potts et al that comparison between the inflamed recti and adjacent temporalis muscle can be used as a surrogate marker of thyroid associated ophthalmopathy activity (7).

We aim to explore the use of texture analysis software in identifying active thyroid associated ophthalmopathy on MRI scans within a heterogeneous retrospective cohort.

This single centre retrospective study used a heterogeneous sample of patients who presented from 2018-2019 with active thyroid associated ophthalmopathy requiring pulsed intravenous steroid therapy. MRI images of these patients were retrospectively obtained and fully anonymised. Texture analysis software MaZda was then used to compare the signal intensity ratio (SIR) between the medial rectus muscle and the adjacent temporalis muscle as reference (7).

The texture classification and signal intensity determined by MaZda between the medial rectus and temporalis muscle was then compared using Microsoft Excel, IBM SPSS, and MaZda software.

Preliminary results show that the MaZda texture analysis software is easy to use once calibrated in assisting the interpretation of active thyroid associated ophthalmopathy patient’s MRI scans. Initial analysis shows distinctly different signal intensity and pixel characteristics between inflamed medial rectus muscle and the adjacent temporalis muscle in patients with active disease.

The results of this study show promise for the use of texture analysis software in providing quantitative data in support of the radiological diagnosis of active thyroid associated ophthalmopathy. We hope this proof of concept study drives further research interest and plan to follow up with larger direct cohort comparison studies to analyse real world specificity and sensitivity.

References:

1. Muehlematter UJ, Mannil M, Becker AS, Vokinger KN, Finkenstaedt T, Osterhoff G, et al. Vertebral body insufficiency fractures: detection of vertebrae at risk on standard CT images using texture analysis and machine learning. Eur Radiol. 2018.
2. Hui TCH, Chuah TK, Low HM, Tan CH. Predicting early recurrence of hepatocellular carcinoma with texture analysis of preoperative MRI: a radiomics study. Clin Radiol. 2018;73(12):1056.e11-.e16.
3. Holli-Helenius K, Salminen A, Rinta-Kiikka I, Koskivuo I, Bruck N, Bostrom P, et al. MRI texture analysis in differentiating luminal A and luminal B breast cancer molecular subtypes – a feasibility study. BMC Med Imaging. 2017;17(1):69.
4. Jyothi S, Mir S, Salvage D, Vize C. Non-invasive Classification of Orbital Tissue Pathology Based on Texture Analysis Parameters from Magnetic Resonance Images. European Ophthalmic Review. 2018;12(2).
5. Zoumalan CI, Cockerham KP, Turbin RE, Volpe NJ, Kazim M, Douglas RS, et al. Efficacy of corticosteroids and external beam radiation in the management of moderate to severe thyroid eye disease. J Neuroophthalmol. 2007;27(3):205-14.
6. Lazarus JH. Epidemiology of Graves’ orbitopathy (GO) and relationship with thyroid disease. Best Pract Res Clin Endocrinol Metab. 2012;26(3):273-9.
7. Mayer E, Herdman G, Burnett C, Kabala J, Goddard P, Potts MJ. Serial STIR magnetic resonance imaging correlates with clinical score of activity in thyroid disease. Eye (Lond). 2001;15(Pt 3):313-8.