Detection for non-cavitated fissure caries with optical coherence tomography / Erum Zain

Erum, Zain (2017) Detection for non-cavitated fissure caries with optical coherence tomography / Erum Zain. PhD thesis, University of Malaya.

	PDF (The Candidate's Agreement) Restricted to Repository staff only Download (1659Kb)
Preview	PDF (Thesis PhD) Download (5Mb) | Preview

Abstract

Early detection and preventive management of non-cavitated fissure caries (NCFC) are the standard of care today. The aim of this study was to explore the potential of an optical method, Optical Coherence Tomography (OCT), in detection of NCFC and discriminating its stages of severity on the occlusal surface. Objective: Firstly, to evaluate the accuracy of OCT in detecting sound fissure and NCFC using qualitative OCT analysis and to determine the lesion dimension agreement between OCT and PLM. Secondly, to optically characterize Ekstrand (Ek) histological codes 0, 1 and 2 and to identify OCT outcome measures that can differentiate between Ek codes. Thirdly, to validate the identified OCT outcome measures. Finally, compare the diagnostic accuracy of OCT outcome measures and ICDAS-II criteria. Methods: Premolars with sound and naturally occurring NCFC were scan bucco-lingually on the occlusal surface using Swept Source OCT (SS-OCT) and validated using Polarized Light Microscopy (PLM) and Micro-computed tomography (Micro-CT). Bespoke software developed with MATLAB (Mathworks Inc) was used for OCT data extraction and processing. Results: OCT qualitative assessment can detect NCFC with high sensitivity (SN) 0.98 and specificity (SP) 0.95. Agreement between OCT and PLM was observed for width of wall and height of slope lesion. Characterisation of OCT signals based on Ekstrand histological classification showed difference in rate of attenuation between Ek codes. Mean depth-resolved intensity attenuated in two phases, in phase I, Ek code 0 had bimodal pattern, where a sudden drop was seen at 30 μm, however, at 40 μm slow attenuation (1.43E+06 a.u./μm) was observed. Ek code 2 exhibited the fastest rate of attenuation (6.43E+06 a.u./μm). Thereafter, in phase II, relatively slower attenuation was observed for all three Ek codes. Area under the curve, partial (AUCP) and total (AUCT) were the outcome measures to detect sound fissure from NCFC and differentiated stages of NCFC respectively. AUCP at cut-off 3.038E+10 (a.u.) showed 0.86 SN and >0.99 SP with 0.95 area under receiver operating characteristic (AUROC) curves. AUCT at a cut-off value of 7.031E+10 (a.u.) was limited to SN 0.61 and SP 0.73 with 0.66 AUROC. Validation of AUCP showed no significant difference in detecting sound and NCFC when groove-fossa system was assessed overall compared to groove where both had accurate SN (> 0.99) and good SP (0.79). For differentiation, AUCT showed similar SN at different anatomical locations (0.80), however, SP (0.75) was highest at the occlusal fissure. Quantitative OCT analysis when compared to visual criteria ICDAS-II showed similar SN (> 0.99) for detection of NCFC, whereas higher SP (> 0.99 ) was observed for OCT. OCT was able to accurately delineate ICDAS-Radiological (R)–code 2 versus ICDAS-R-code 1 with high SN (> 0.99) and good SP (0.83) as compared to ICDAS-II criteria. Conclusion: The attenuation characteristics of backscattered intensity and the derived outcome measures AUCP and AUCT can significantly detect sound fissure from NCFC and differentiate stages of severity respectively and can be used for predicting Ek codes 0, 1 and 2 with good accuracy.

Actions (For repository staff only : Login required)