Differential fluorescence system as method of caries detection - a comparative study.

The aim of the present in vitro study was to evaluate the usefulness of a new differential fluorescence system (DFS) for smooth surfaces caries detection, and the comparison of readings from DFS, QLF and spectral measurements with transverse microradiographic and histopathologic analysis, as standard references.

Material and methods

The material comprised of 56 teeth, extracted for orthodontic and periodontal reasons, with a fully intact coronal part and 6 teeth with clinically observed secondary caries. The coronal part of each tooth was cleaned with a brush, a rubber and distilled water. All organic debris was removed from the root part of the teeth mechanically. The teeth were stored in a thymol-containing 0,9 % saline solution, under refrigeration. Thirty-one premolars, 24 molars and 7 incisors were included in the study. The teeth were classified and the proximal surfaces were numbered from 1 to 110.

Macroscopic examination

The teeth were examined macroscopically under a dental unit light. The surfaces were observed and classified. Clinical inspection of 110 surfaces revealed that 24 of them were sound, 64 had incipient caries, 15 had more advanced carious lesions and 6 had clinically visible secondary caries.

Photographic documentation

All approximal surfaces were photographed by a digital camera (Nikon COOLPIX 990) using the parallel technique. The images of the teeth were captured and registered in a PC.

Differential fluorescence imaging system(DFS)

The DFS imaging systems consisted of a light source and two types of cameras (Fig. 1.). A semi-conducting laser with a blue light emission (λ=407nm, P=25mW, G=20mW/cm²) was used as a light source. One of the cameras used in the study was bl/wh SIT (sitcom intensified target) equipped with filters (sensitivity 1x10 ¯6 lux), and the second one was a bl/wh camera (sensitivity 0,02 lux). The imaging board simultaneously captured two images of each surface, at different (red-625nm and green-500nm) wavelengths. The images were than combined and processed with a specially developed algorithm. This allowed obtaining a better contrast between the demineralised area and sound tissue.

Fig. 1. Schematic setup of the DFS system

Measurements of Differential Fluorescence Spectra (SM)

The spectra of fluorescence registered on the examined surfaces were recorded and analysed. The detection system consisted of a monochromator and linear CCD detector. The fibre sensor comprised of the main fibre used as a laser light guiding and six concentrically spaced fibres for fluorescence light receiving. The spectrometer was installed on a computer card with software controlling measurements and spectra processing. All examined surfaces were scanned with the fibre sensor. The ratio of mean value of fluorescence intensity obtained for a demineralised area and the neighbouring healthy region of each surface was registered. If no differences were observed the mean value for the whole surface was recorded.

Qlf measurements

The QLF system used in the study consisted of a camera (Panasonic WV-KS 152), a light source-a Xenon lamp emitting a non-coherent, visible, blue green light, combined with a filter system blocking all reflected and back scattered light. The images of fluorescence of the teeth were captured, recorded, stored, processed and analysed with a computer programme (Inspector Research Systems BV, Amsterdam, The Netherlands). Two parameters were measured: average fluorescence loss (%), and area of the lesion (mm²).

Histopathologic and microradiographic (TMR) measurements

The teeth were then prepared for sectioning for histopathological and TMR measurements. All teeth were embedded in methylmetacrylate. Teeth with incipient enamel lesions on both approximal surfaces were selected. These were hemisectioned with a diamond saw, perpendicular to the buccal and occlusal surfaces, to obtain two separate samples, and were then embedded again. Tooth slices, 300µm thick, from the areas evaluated and marked on the photographed surfaces were sawn, perpendicular to the enamel surface using a water-cooled saw.

Micro-radiographic images of the slices as well as an image of an aluminium step wedge for calibration were captured and recorded on holographic film (Kodak SO-253) exposed to Ni-filtered radiation at 20kV and 55mA with the exposure time of 1min. The samples with lesions defined as a white spot were manually ground to the thickness of 100µm and examined with TMR. The TMR system consisted of a PC connected to a microscope densitometer. The images of these slices were taken with the exposure time of 20 sec. The films were developed according to a standard conditions and then examined under the microscope at 16х magnification. The penetration of the demineralised area in a pulpal direction was established for each sample, according to following scale: 0=sound surface, 1=caries limited to the outer half of enamel, 2=caries extending to the inner half of enamel but not to the dentino-enamel (DEJ) junction, 3=caries penetrating to the DEJ but limited to the outer part of the dentin, 4=caries extending to the inner part of the dentin.

Mineral content and depth profile were assessed with designed software (Inspector Research Systems BV, Amsterdam, The Netherlands). The integrated mineral loss (Δz) was estimated only for enamel caries.

Data analysis

Data analysis was done using Statistica 6.0 Program. The Spearman rank correlation test was used to estimate the correlation between fluorescence measurements (QLF and DFS) and the depth of the lesions determined from TMR and histological analysis. For the specimens with enamel carious lesions, the relationship with fluorescence measurements and mineral loss (Δz) was also investigated. Sensitivity and specificity of the both fluorescence methods were calculated for D₃caries level.

Results

The selected images of dental surfces obtained by DFS and QLF are presented in Fig 1.

DFS 18 A QLF 18 A

DFS 18 B QLF 18 B

DFS 14 A QLF 14A

Validity and diagnostic accuracy of DFS and QLF were evaluated. The Spearman rank correlation coefficients between lesion depth and QLF and DFS were 0. 83 and 0. 79 respectively. The Pearson rank correlation coefficients between enamel mineral loss, measured by TMR, and QLF and DFS readings were 0.89 and 0.77, respectively. Sensitivity and specificity for DFS with respect to dentinal caries were 0.75 and 0.92, and for QLF were 0. 89 and 1, with a cut off value of 12.5 %.

Sperman’s rank correlation (r) QLF and DFS to lesion depth (SEMI)

	Valid N	Spearman R	t(N-2)	p-level
QLF	99	0.83	14.52894	.000000
DFS	99	0.79	12.5406	.000000

Pearson correlation (r) QLF i DFS to mineral loss (TMR)

	r(X,Y)	r2	t	p	N
QLF & MIN_LOSS	0,89	0,800348	16,51038	0,000000	70
DFS & MIN_LOSS	0,77	0,586442	-9,81971	0,000000	70

Sensitivity and specificity

	DFS			QLF
	cut off	sens.	spec.	cut off	sens	spec
D 3	12,5%	0,75	0,92	25%	0,89	1

Conclusions: For smooth surface diagnosis, both methods- QLF and DFS are of equal merit in quantitative aspects. A closer correlation was found between QLF and mineral changes. The DFS system seems to be an appropriate diagnostic tool for caries detection and quantification. Further investigation should be undertaken for adaptation this system for clinical trial.

0366 The application of a differential fluorescence system in caries diagnosis

A. MIELCZAREK¹, M. KWASNY², X.-Q. SHI³, and J. MICHALAK¹, ¹ Medical University of Warsaw, Poland, ² Military University of Technology, Warsaw, Poland, ³ Karolinska Institutet, Huddinge, Sweden

Objectives: To evaluate a new differential fluorescence system (DFS) for detection and quantification of carious lesions in comparison with QLF. The imaging board of this system simultaneously captures two images of each surface, at different wavelengths (red-625nm and green-500nm). Images are then combined and processed with a special algorithm.

Methods: Material comprised of 56 extracted teeth. The proximal surfaces were classified and numbered from 1 to 110. The teeth were examined macroscopically and photographed by a digital camera. DFS and QLF were used for registration of fluorescence images. Teeth were then prepared for histological and microradiographical measurements (TMR). Mineral content loss and depth profile (LD) were assessed with software (Inspector Research Systems BV, Amsterdam, The Netherlands). Integrated mineral loss (Δz) was evaluated only for enamel caries.

Results: Validity and diagnostic accuracy of DFS and QLF were evaluated. The Spearman rank correlation coefficients between lesion depth and QLF and DFS were 0. 83 and 0. 79 respectively. The Pearson rank correlation coefficients between enamel mineral loss and DFS and QLF were 0.89 and 0.77, respectively. Sensitivity and specificity for DFS with respect to dentinal caries were 0.75 and 0.92, and for QLF were 0. 89 and 1, with a cut off value of 12.5 %.

Conclusions: For smooth surface diagnosis, both methods are of equal merit in quantitative aspects. A closer correlation was found between QLF and mineral changes. The DFS system seems to be an appropriate diagnostic tool for caries detection and quantification. Further investigation should be undertaken for the adaptation of this system for clinical trial.

IADR/AADR/CADR 82nd General Session (March 10-13, 2004)