Maile, Howard;
(2023)
Keratoconus detection and personalised progression modelling using computational methods.
Masters thesis (M.Phil), UCL (University College London).
Preview |
Text
Howard_Maile_2023.pdf - Accepted Version Download (26MB) | Preview |
Abstract
Keratoconus is a common corneal disorder in young adults characterised by bilateral thinning and distortion of the cornea that disrupts this refractive system leading to blurred vision. Untreated, it can progress to cause blindness and is a leading cause of visual loss in young adults globally. Early detection, understanding disease progression and administering optimal treatment are thus important areas of research. Initially, the existing literature on the algorithmic detection of subclinical keratoconus was surveyed and critically evaluated which resulted in a comprehensive systematic review as well as the identification of avenues for further research. This led to the exploration of using convolutional neural networks (CNNs) applied to raw Anterior Segment Optical Coherence Tomography (AS-OCT) images. The results showed that raw images can be used to classify disease severity (early versus late stage keratoconus) with validation accuracy of 96.5% and provide the groundwork for further research. Subsequent chapters deal with the problem of predicting and analysing keratoconus progression. A prognostic model was developed to predict keratoconus progression to requiring corneal cross-linking (CXL) with explained variation of 33%. Given a series of corneal parameters at the first appointment, the model generates a personalised time-to-event curve which could aid patient empowerment, triage and service provision. Finally, the problem of defining progression after the CXL operation was investigated. Rather than using fixed thresholds which do not account for the heteroscedasticity within corneal measurements, a variation of Bland Altman analysis was used to generate adaptive thresholds for Kmax, Front K2 and Back K2 which, when combined, constitute a new definition of post CXL progression. This novel method was compared and contrasted with existing definitions of progression using the Kaplan-Meier estimator and could provide a more reliable estimate of keratoconus progression after CXL.
| Type: | Thesis (Masters) |
|---|---|
| Qualification: | M.Phil |
| Title: | Keratoconus detection and personalised progression modelling using computational methods |
| Open access status: | An open access version is available from UCL Discovery |
| Language: | English |
| Additional information: | Copyright © The Author 2023. Original content in this thesis is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) Licence (https://creativecommons.org/licenses/by-nc/4.0/). Any third-party copyright material present remains the property of its respective owner(s) and is licensed under its existing terms. Access may initially be restricted at the author’s request. |
| UCL classification: | UCL UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Brain Sciences UCL > Provost and Vice Provost Offices > School of Life and Medical Sciences > Faculty of Brain Sciences > Institute of Ophthalmology |
| URI: | https://discovery.ucl.ac.uk/id/eprint/10173920 |
Archive Staff Only
 |
View Item |
