Abstract
Objectives
This study evaluated (1) the detection and assessment of non-cavitated occlusal carious lesions by spectral domain optical coherence tomography (SD-OCT) and (2) the impact of varying angle of incidence (AI) of probe light and refractive index matching (RIM).
Methods
Nine extracted human molars with 18 occlusal lesions (ICDAS code 2) were visually selected. 18 regions of interest (ROI) were imaged with SD-OCT under varying AI (0°, ±5°, ±10°, ±15°) and with/without application of glycerine at 0°. X-ray micro computed tomography (μCT) was used as a validation standard. μCT and OCT signals were categorized according to the lesion extent: 1-sound, 2-lesion limited to half of enamel, 3-lesion limited to enamel, 4-lesion into dentin. Agreement between both methods was assessed. Intra- and inter-examiner reproducibility analyses were conducted. Statistics: Cohen’s kappa coefficient (κ), Spearman’s rho correlation ( r s ) and Wilcoxon test (α = 0.05).
Results
Slight to moderate agreement (κ = 0.153) between μCT and OCT was obtained at an AI of 0° (Wilcoxon: p = 0.02). With variation of Al a substantial agreement (κ = 0.607) was observed (p = 0.74). Spearman’s correlation between both methods was 0.428 at 0°, 0.75 with varying AI and 0.573 with glycerine. Kappa values for intra-and inter-examiner analysis ranged between 0.81 and 0.88 and between 0.25 and 0.73, respectively.
Conclusion
Variation of AI improves the detectability of non-cavitated occlusal carious lesions. RIM can enhance signal-to-noise ratio.
Clinical relevance
OCT could provide additional diagnostic information in single and longitudinal assessments of occlusal carious lesions.
1
Introduction
In recent years, in many industrialized countries a significant decrease of caries has been observed both on children and adolescents as well as adults . At the same time, in this overall trend of caries decline, the percentage of occlusal caries is seen to be increasing . Especially patients with non-cavitated occlusal carious lesions were observed more frequently . Such lesions can possess an intact surface area, macroscopically appearing as sound enamel . Fluoridation can foster remineralization of the surface of the enamel carious lesion possibly masking subsurface progression into dentine which complicates diagnosis . The penetration of the lesion into the dentin causes a mechanical weak spot underneath the enamel, which can merge into a cavitation when external forces are exerted. This cavitation then acts as a retention area for bacterial biofilm. For that reason, cavitation often defines the threshold to invasive therapy to restore the tooth surface. Early diagnosis of carious lesions and their progression is therefore of great importance to prevent restorative treatment need and allow for non-invasive therapeutic options. For determination of effective treatment strategies accurate non-invasive imaging methods enabling surface and sub-surface assessment of early findings due to caries are desirable.
Optical coherence tomography (OCT) is an established diagnostic method in retina assessment in ophthalmology, and provides non-invasive three-dimensional imaging with high resolution. Various studies subsequently evaluated capability and suitability of OCT as a new caries diagnostic tool using low coherence interferometry . This tomographic method enables contact-free non-destructive real time imaging. The first dental use of OCT was by Colston et al. and Feldchtein et al. for imaging of the soft- and hard-tissue in the oral cavity . Numerous OCT systems have been described and also the potential of the technology to characterize specific material characteristics has been demonstrated, e.g. by using polarization sensitive OCT (PS-OCT) . As the image contrast can be enhanced and artefacts can be minimized , PS-OCT is advantageous over non-polarization sensitive systems for the assessment of carious lesions. OCT were evaluated to assess progression of carious lesions on smooth enamel surfaces in vitro and to detect and monitor artificial occlusal early enamel caries as well as occlusal enamel and dentin caries in vitro , ex vivo and in vivo . Another research group reported that OCT could be used for the measurement of the remineralization of early artificial caries and dentin . The fact that detection of carious lesions on occlusal surfaces is still rather challenging and no single, reliable method with high sensitivity and specificity is clinically available today was the motivation for this study.
The aim of this study was therefore to evaluate the suitability of non-polarization sensitive spectral domain (SD)-OCT for the detection and assessment of non-cavitated occlusal carious lesions and to explore improvements to the method by variation of incidence angle of probe light and refractive index matching. In comparison to the prior studies mentioned above, in this study natural occlusal carious lesions were investigated and additional techniques for enhancement of caries detection were evaluated. We hypothesize that SD-OCT is suited to serve as a caries diagnostic tool in this scenario and that the variation of incidence angle of probe light and refractive index matching enhance detectability of non-cavitated occlusal carious lesions.
2
Materials and methods
Nine intact, extracted human molars with 18 initial occlusal carious lesions (ICDAS code 2, approval of the Ethics Committee of the University of Leipzig, protocol No. 299-10-04102010) were visually selected and immersed in 0.5% chloramine solution at 4 °C immediately after extraction. The teeth were cleaned mechanically. Fig. 1 depicts the experiment schematically. Regions of interest (ROI) were marked by 2 drill holes each in mesio-distal direction (diamond bur, micro lance, pointed, 46 μm, 957 AM, Komet/Co. Gebr. Brasseler GmbH). The roots of the specimens were completely embedded in Stycast ® compound (Emerson & Cuming, Westerlo, Belgium) using a prepared box-shaped silicon block in order to reproduce the same occlusal position (constant inclination of the occlusal surface to the sample stage) during multiple investigations.
The ROI were imaged 2-dimensionally by spectral domain OCT (SD-OCT, Telesto SP5, Thorlabs GmbH, Dachau, Germany) under different angles of incidence (AI) of probe light (0°, ±5°, ±10°, ±15°) and with/without surficial application of glycerine gel at 0° (layer thickness: 1 mm, Airblock, Dentsply DeTrey GmbH, Germany). SD-OCT is a variant of Fourier-domain OCT, using a broadband, low-coherence light source. Technical specifications of the SD-OCT system were as follows: center wavelength 1310 nm ± 100 nm, sensitivity ≤ 106 dB, axial/lateral resolution <7.5 (air)/15 μm, field of view 9 mm × 9 mm × 3 mm (pixel size 512 × 512 × 512), A-scan average 1, spacing 6 μm. A goniometer (radius of 40 mm; travel of ±15°; Edmund optics Ltd, UK) was used to variate the AI to the occlusal surface. AI was limited to ±15°. It was shown in a pilot study that with angles above this value occlusal structures may block the OCT beam resulting in non-visibility of the complete occlusal surface.
The images obtained by OCT were analyzed using Image J version 1.45 (Wayne Rasband, National institutes of Health, Bethesda, Maryland USA). ROI were prepared by dabbing with a wet pad, so that the lesions would not be overdry during OCT scanning. After OCT imaging, X-ray micro computed tomography images were used as a validation standard (μCT; 100 kV, filter Al +Cu, rotation step 0.2°, averaging 6, pixel size 3.5 μm; Skyscan 1172-100-50, Bruker Micro CT, Kontich, Belgium). Carious lesions imaged by OCT and μCT were assessed and categorized according to the extent of the lesions [ Table 1 ]. Agreement between both methods as well as Inter- and intra-examiner reproducibility were assessed by unweighted kappa coefficients (κ), Spearman’s rho correlations ( r s ) and Wilcoxon test (α = 0.05). The main examiner with 6 years of experience in OCT diagnostics made three consecutive measurements. Two other calibrated observers (one person with OCT experience, the other person with no prior experience in image-based diagnostics) reassessed the OCT images. All measurements were performed blind using the same monitor and image processing software under identical room lighting conditions.
| Score | Extent of lesions |
|---|---|
| 1 | No lesion |
| 2 | Lesion limited to half of enamel |
| 3 | Lesion limited to enamel |
| 4 | Lesion into dentin |
2
Materials and methods
Nine intact, extracted human molars with 18 initial occlusal carious lesions (ICDAS code 2, approval of the Ethics Committee of the University of Leipzig, protocol No. 299-10-04102010) were visually selected and immersed in 0.5% chloramine solution at 4 °C immediately after extraction. The teeth were cleaned mechanically. Fig. 1 depicts the experiment schematically. Regions of interest (ROI) were marked by 2 drill holes each in mesio-distal direction (diamond bur, micro lance, pointed, 46 μm, 957 AM, Komet/Co. Gebr. Brasseler GmbH). The roots of the specimens were completely embedded in Stycast ® compound (Emerson & Cuming, Westerlo, Belgium) using a prepared box-shaped silicon block in order to reproduce the same occlusal position (constant inclination of the occlusal surface to the sample stage) during multiple investigations.
The ROI were imaged 2-dimensionally by spectral domain OCT (SD-OCT, Telesto SP5, Thorlabs GmbH, Dachau, Germany) under different angles of incidence (AI) of probe light (0°, ±5°, ±10°, ±15°) and with/without surficial application of glycerine gel at 0° (layer thickness: 1 mm, Airblock, Dentsply DeTrey GmbH, Germany). SD-OCT is a variant of Fourier-domain OCT, using a broadband, low-coherence light source. Technical specifications of the SD-OCT system were as follows: center wavelength 1310 nm ± 100 nm, sensitivity ≤ 106 dB, axial/lateral resolution <7.5 (air)/15 μm, field of view 9 mm × 9 mm × 3 mm (pixel size 512 × 512 × 512), A-scan average 1, spacing 6 μm. A goniometer (radius of 40 mm; travel of ±15°; Edmund optics Ltd, UK) was used to variate the AI to the occlusal surface. AI was limited to ±15°. It was shown in a pilot study that with angles above this value occlusal structures may block the OCT beam resulting in non-visibility of the complete occlusal surface.
The images obtained by OCT were analyzed using Image J version 1.45 (Wayne Rasband, National institutes of Health, Bethesda, Maryland USA). ROI were prepared by dabbing with a wet pad, so that the lesions would not be overdry during OCT scanning. After OCT imaging, X-ray micro computed tomography images were used as a validation standard (μCT; 100 kV, filter Al +Cu, rotation step 0.2°, averaging 6, pixel size 3.5 μm; Skyscan 1172-100-50, Bruker Micro CT, Kontich, Belgium). Carious lesions imaged by OCT and μCT were assessed and categorized according to the extent of the lesions [ Table 1 ]. Agreement between both methods as well as Inter- and intra-examiner reproducibility were assessed by unweighted kappa coefficients (κ), Spearman’s rho correlations ( r s ) and Wilcoxon test (α = 0.05). The main examiner with 6 years of experience in OCT diagnostics made three consecutive measurements. Two other calibrated observers (one person with OCT experience, the other person with no prior experience in image-based diagnostics) reassessed the OCT images. All measurements were performed blind using the same monitor and image processing software under identical room lighting conditions.
Stay updated, free dental videos. Join our Telegram channel
VIDEdental - Online dental courses
