Assessment and diagnosis of dry eye disease (DED) is a challenging task. The conventional ways of diagnosing DED are problematic due to their invasiveness, poor test reliability and significant test duration. Previously, ocular surface thermography has been shown to be able to detect early inflammation and dry eye. However, its diagnostic ability and ocular temperature metrics that can best diagnose DED are not clear. The objectives of this thesis were manyfold. First, the prevalence of dry eye in Singapore population was investigated as a helpful basis for the rest of the project. A cross-sectional dry eye survey was carried out using McMonnies dry eye questionnaire. Members of the public were interviewed at 46 (out of 62) selected mass rapid transit stations in Singapore and its vicinity. 1004 questionnaires were collected from participants aged 15 - 83 years and various ethnicity. Prevalence of symptomatic dry eye (SDE) was found to be 12.3% (about 0.5 million Singaporeans). Risk factors associated with SDE were found to be age, gender, ethnicity, hypertension and contact lens wear. Smoking was not associated with SDE.The main part of this thesis sought to evaluate the efficacy of ocular thermography in diagnosing DED. A new infrared detector (NEC Thermo Tracer TH 9260) with relatively high resolution was used. Inter-image, inter-occasion and inter-examiner repeatability was first studied on 21 healthy and 15 DED subjects. Ocular surface marking and ocular surface temperature (OST) acquisition was performed with a novel 'diamond' method using a custom-designed OST analysis V2 software. Ten out of the twelve tested OST indices were shown to be highly repeatable for three studied time points: 0 s, 5 s and 10 s. They were temperatures of the geometric center of the cornea (GCC), mean temperature (MOST) of the region of interest (ROI), maximum (MaxT) and minimum (MinT) temperatures of the ROI, extreme temporal (T1) and nasal conjunctiva (T4), mid temporal (CT) and nasal conjunctiva (CN) and temporal (LT) and nasal limbal (LN). Another 62 DED and 63 age- and sex-matched controls were then recruited and the ten static and dynamic OST indices were evaluated. Static measures were study of absolute OST at t = 0 s, 5 s and 10 s after eye opening. Dynamic measures were study of mean change and net change in OST over 10 s of sustained eye opening. Static measures on eight OST indices (GCC, MOST, MinT, MaxT, T4, CT, LT and LN) at t = 0 s, 5 s and 10 s and dynamic measures on two OST indices (T4 at 3 s onward and MaxT at 5 s onward) were found to be valuable in detecting DED. The temperature metrics (static and dynamic) were identified for further investigation. Thereafter, the diagnostic ability of the temperature metrics were evaluated singly and as combinations in terms of their area under the curve (AUC), Youden index and discrimination power. Receiver operating characteristic curves were plotted for each metric. Best detectors for DED were found to be the T4 temperature metrics: particularly T4-5 and T4-10 (i.e. absolute temperature of the extreme nasal conjunctiva at 5 s and 10 s). Values of T4-5 of < 34.8 °C were found to give sensitivity and specificity of 87.1% and 50.8% respectively and values of T4-10 of < 34.6 °C were found to give sensitivity and specificity of 77.6% and 61.9% respectively. The two temperature metrics had highest Youden index as compared to other metrics and were shown to be useful in view of AUC > 70% but of limited performance in view of their discrimination power. Nevertheless, measuring T4-5 and T4-10 was found to be comparable to other conventional methods for DED. T4-10 was better than T4-5 in view of higher AUC and Youden index. None of the tested dynamic metrics was good detector for DED and combining metrics were not able to increase the diagnostic ability. The last part of this thesis was to validate the effectiveness of some common conventional dry eye tests, to study their correlation with T4 temperature metrics and derive the best composite/combined tests for DED. Sixty two DED patients and 82 controls were studied. The conventional clinical tests examined were: symptom evaluation using McMonnies dry eye questionnaire (Mscore) and symptom count (Scount), fluorescein break-up time (FBUT) and corneal epithelial staining (CES), non-invasive break-up time (NIBUT) and tear meniscus height (TMH). Mscore and Scount was the best detectors for DED, followed by FBUT and CES. Discordance between signs and symptoms for DED was further confirmed. Combining CES with T4-10 (series) can be future objective tests for DED. Further research is warranted, particularly to (1) validate the ability of T4-10 as a stand-alone test for DED and (2) work out an algorithm and validate the diagnostic ability of the recommended combined test (CES and T4-10) using newly recruited subjects.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:713628 |
| Date | January 2017 |
| Creators | Tan, Li Li |
| Contributors | Morgan, Philip |
| Publisher | University of Manchester |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | https://www.research.manchester.ac.uk/portal/en/theses/assessment-of-dry-eyes-using-ocular-surface-thermography(77e912eb-6a15-4ea1-8b2c-7ed9958591ec).html |
Page generated in 0.002 seconds