New detection devices such as Digital Imaging Fiber-Optic TransIllumination (DIFOTI) have been introduced to improve early detection of carious surfaces. One of the questions that we set out to answer was whether DIFOTI, which is based on light propagation just below the tooth surface, can be used to determine lesion depth. DIFOTI uses fiber-optic transillumination of safe visible light to image the tooth. Light delivered by a fiber-optic is collected on the other side of the tooth by a mirror system and fed to a digital electronic charge-coupled device. Then the acquired data are sent to a computer for analysis with dedicated algorithms, which produce digital images that can be viewed by the clinician and patient in real time or stored for later use.

An initial study indicated that DIFOTI had the potential to improve early detection,⁴ but the ability to measure lesion depth over time has not been demonstrated. We designed our study, in part, to determine if DIFOTI could help measure the depth of an early approximal lesion. In addition, a belief among faculty members at some dental schools at which Class II amalgam preparations still are required for dental licensure is that D-speed films detect early cavitated approximal lesions easier than do the faster F-speed films. The advantage of using the higher speed films is that patients are exposed to less ionizing radiation. Our study compared F-speed radiographic film with histologic lesion depth and cavitation on early approximal lesions. To our knowledge, F-speed film has not yet been calibrated in this manner, nor has DIFOTI been evaluated for measuring lesion depth.

Every two weeks, the demineralization process was interrupted only long enough to take digital images and bitewing radiographs.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
Sample preparation. We collected human teeth extracted at the University of the Pacific, School of Dentistry, Department of Oral and Maxillofacial Surgery. We immediately stored them in a saturated solution of thymol in double deionized water and sterilized them using gamma radiation.⁵ We selected the crowns of seven teeth (canines and premolars) with approximal surfaces that were visibly free of caries (no cavitations, white- or brown-spot lesions or any other signs of demineralization or remineralization). We used mild detergent and deionized water to clean surface debris and then lightly polished them with a micropolish (Alumina, No. 1 Alpha, Buehler, Lake Bluff, Ill.) and a rag wheel. We then coated the teeth twice with a clear acid-resistant varnish and allowed them to dry after each coat. (We had conducted pilot studies to confirm that the clear acid-resistant varnish would not affect the image quality of DIFOTI.) After the varnish was hardened completely, we used a dental curette to make an approximately 1-millimeter scratch through the hardened varnish on the mesial and distal surfaces of each of the seven teeth, creating 14 small "windows." Through these windows, acid could enter and contact the enamel surface during the artificial demineralization process.

Artificial caries demineralization of samples. We made a demineralization solution that simulates the in vivo demineralization process that occurs in the mouth according to the pH cyclic model developed by Featherstone and colleagues.⁶ The solution consisted of calcium phosphate (2 millimolars per liter) in an acetate (0.075 moles per liter) buffer. We adjusted the pH of this solution to 4.8. After we scratched the 14 windows through the varnish on the mesial and distal surfaces of the teeth, we subjected the teeth to the demineralization solution for 14 weeks at 37 C.

Imaging. Every two weeks, we interrupted the demineralization process only long enough to take DIFOTI images and bitewing radiographs. We took the DIFOTI images according to the manufacturers’ instructions to simulate the angle and appearance of a bitewing radiograph, as well as at an approximately–45–degree angle, similar to that of a bitewing radiograph. We simulated the angle of a bitewing-like view to test the hypothesis that the light would penetrate the tooth far enough for us to be able to measure the depth of the lesion. A 45-degree–angle view shows the surface directly.

We used F-speed radiographic film (Kodak Insight, Eastman Kodak, Rochester, N.Y.) to take radiographs at an angle and a distance that would simulate a clinical bitewing radiograph using a radiograph unit (Model 46-158800G2, General Electric Company Medical Systems, Chalfont St. Giles, England) at 15 milliamperes and 5 impulses. We developed the radiographic films using fresh solutions in an automatic developer (Velopex Extra-X, Medivance Instruments, London). We manually scanned the radiographs at a high resolution (1,200 dots per inch) and evaluated them using data analysis software (Image-Pro, Media Cybernetics, Silver Spring, Md.) to measure the depth of the lesion.

Preparation of thin sections and use of polarized light microscopy. At the conclusion of the 14-week period of controlled demineralization, we made thin sections approximately 80-mm thick longitudinally through the center of the lesion in each sample using a high-speed hard-tissue microtome (Scifab, Layayette, Colo.) with water cooling. We subjected the thin specimens to histologic analysis using the gold standard polarized light microscopy (PLM) with a polarizing microscope (BX-50, Olympus, Melville, N.Y.), as described in a 1991 review article by ten Bosch and Angmar-Mansson.⁷ We then entered the PLM images into the computer and magnified them to measure the depth of the lesion using the data analysis software.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
DIFOTI results. After the samples had been demineralized for 14 weeks, we were not able to identify the depth of a lesion on any of the samples using DIFOTI at the bitewing-like angles (Figure 1). The 45-degree DIFOTI images, however, did successfully identify surface demineralization successfully as early as two weeks into the demineralization process when we still had difficulty doing so visually (Figure 2).

View larger version (111K): [in this window] [in a new window]   Figure 1. Images comparing Digital Imaging Fiber-Optic Trans-Illumination (DIFOTI) (Electro-Optical Sciences, Irvington, N.Y.), F-speed radiographic film (X-ray), polarized light microscopy (PLM) and photography at 14 weeks. The PLM images show the intact surface layer and that most of the body of each lesion did not survive the thin-sectioning procedure, though the outer surface is seen as intact on the photographs.

View larger version (58K): [in this window] [in a new window]   Figure 2. Digital Imaging Fiber-Optic Trans-Illumination (DIFOTI) (Electro-Optical Sciences, Irvington, N.Y.) images at two weeks comparing bitewing-like view with the 45-degree view. Note that the surface demineralization appears dark as early as two weeks in the 45-degree view but not in the bitewing-like view.

PLM results. After the samples had been demineralized for 14 weeks, we obtained PLM images (Figure 1) of 13 of the 14 samples; one sample had been destroyed during the thin- sectioning procedure. The mean depth of a lesion was 0.58 mm (± SD = 0.19 mm), which on average was situated slightly more than one-half way (± SD = 0.57 mm) to the DEJ.

Surface cavitation results. After the samples had been demineralized for 14 weeks, we saw no signs of surface cavitation on the lesions when we examined the samples using visual magnification and tactile methods. We were able to see white-spot lesions on all of the samples.

Interpretation of data. We calculated the mean and standard deviation of the depths of the lesions by imaging technique group. An analysis of variance indicated that differences existed among the means for the DIFOTI, F-speed radiographic film and PLM groups at the P < .0001 significance level. Using the Tukey test, we found that the DIFOTI group mean was significantly different from those of the F-speed radiographic film group and the PLM group (P < .001), but the radiographic film group mean and the PLM group mean were not significantly different from one another (P > .05). When we conducted a linear regression analysis of depths measured by radiography and by PLM, we found they had a good linear relationship with an r²value of .625 (Figure 3). Although the mean depth determined by PLM was approximately 30 percent higher than that determined by the radiographs, we determined that this difference was not statistically significant, using a two-tailed t test (P > .10).

View larger version (48K): [in this window] [in a new window]   Figure 3. Radiograph depth (F-speed radiographic film) versus polarized light microscopy (PLM) depth. The solid line is the linear regression line (2= .625). mm: Millimeters.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

We evaluated the ability of DIFOTI and F-speed radiographic film to measure the depth of a lesion compared with the gold standard of PLM histologic analysis. We also evaluated the ability of DIFOTI and F-speed radiographic film to measure the depth of a lesion compared with surface cavitation. The results of our study demonstrated that DIFOTI could not determine the depth of a lesion, F-speed radiographic film could measure the depth of a lesion and correlated reasonably well to PLM histologic analysis, and none of the samples showed signs of surface cavitation. Although DIFOTI’s manufacturer does not claim that the device measures lesion depth, many clinicians continue to use it to make treatment decisions about whether to restore early proximal lesions.

Figure 2 illustrates that DIFOTI images taken at the 90-degree bitewing-like view failed to show the depth of a lesion in any of the samples. All images did, however, show signs of subsurface demineralization early (at the first two-week imaging period), as seen in 45-degree-view images compared with the bitewing-like view. Our study reinforces the manufacturer’s view that DIFOTI is a good tool for detecting early surface changes and for alerting clinicians to use clinical judgment in addition to other detection methods to determine treatment and preventive options.

The F-speed film radiographs were able to measure the depth of demineralization represented by the radiolucent areas on the film. Scanning a radiograph and evaluating the image using data analysis software is a way to measure the depth of a lesion more objectively. Therefore, in our study, we scanned the radiographic film at a high resolution after it was developed—a step normally not taken in clinical practice—as the purpose of our study was to measure the depth of a lesion and not the ability of the clinician to do so.

In our study, it was not difficult to evaluate the presence of surface cavitation using visual magnification and tactile methods because we made the lesions artificially in vitro and so they were easy to access directly. None of the samples showed signs of surface cavitation, including slight pitting.

Early enamel lesions with intact surface layers can be remineralized chemically.⁹ Thus, the clinical decision whether to chemically remineralize an approximal lesion or to surgically intervene should be based not on the histologic depth of a lesion, but rather on whether the surface is cavitated. If it is cavitated, not only can the lesion no longer be cleaned, but there is nothing stopping the invasion of the bacteria into the more porous dentin. Therefore, the lesion could, and perhaps should, be restored surgically.

It is possible to tell if a lesion is cavitated by looking at a bitewing radiograph. A study by Pitts and Rimmer¹⁰ correlated radiography to cavitation. They reported a 10 percent chance of cavitation when the lesion seen on a radiograph was in the inner one-half of the enamel and a 40 percent chance of cavitation when the lesion was in the outer one-half of the dentin. Thus, the clinician should restore the tooth surgically only when a lesion seen on a radiograph is all of the way through the enamel and clearly penetrates the DEJ showing a dental radiolucency (Figure 4). The trend for the standard of care for evaluating an approximal lesion is to use high-speed radiographic film such as F-speed to lessen the patient’s exposure to ionizing radiation.

View larger version (15K): [in this window] [in a new window]   Figure 4. Diagram showing enamel divided into halves (E1, E2) and the dentin into thirds (D1, D2, D3). It shows the radiographic depth thought to be the minimum depth that should trigger surgical treatment. DEJ: Dentino-enamel junction.

Our study validates the ability of F-speed radiographic film to detect early approximal lesions and demonstrates that it is well-correlated to histologic analysis. In many cases, the depth measured by radiographic film was almost equal to that determined by PLM histologic analysis. Though the mean histologic depth of a lesion was 30 percent deeper than that measured on a radiograph (which is consistent with a previous study of D-speed film by Kleier and colleagues¹¹), this difference was not statistically significant. The ability to accurately measure the depth of a lesion in our study may have been enhanced by the fact that we scanned the radiograph at high resolution and magnified the image using data analysis software. Although our study suggested that the F-speed radiographic film is well-correlated with histologic analysis, results may not be reproducible at chairside by the clinician without magnification. Perhaps digital radiography may have an advantage in this respect; to our knowledge, this technology has not yet been calibrated to histologic lesion depth. Each digital system may need to be calibrated separately.

Our study had results similar to those of Pitts and Rimmer¹⁰ in that no samples were cavitated, even though there were clear radiographic lesions in the enamel but none penetrating the dentin. Lesions in enamel that have an intact surface can be remineralized.^3,9 The lesions in the PLM images in Figure 1 appear cavitated only because the intact enamel layer and part of the body of the lesions were lost during the procedure we used to create the 80-µm thin section. We speculate that this is because we subjected the samples in our study to 14 weeks of demineralization without periods of remineralization in between.

	CONCLUSIONS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

 
The results of our study show that DIFOTI is able to detect surface demineralization at an early stage, but that it is not able to measure the depth of a lesion. DIFOTI should not be used for assessing the depth of approximal lesions in the same manner as should a clinical bitewing radiograph. Using F-speed radiographic film to assess the depth of a lesion was well-correlated to histologic evaluation by PLM, as evidenced by the regression analysis we conducted. All of the lesions, as assessed by F-speed radiography, were in the enamel only; none of the lesions penetrated the DEJ, and none of the samples were cavitated. Depth assessment and, even more importantly, the ability to predict cavitation are necessary when deciding whether to treat lesions by remineralization or by surgical restorative intervention.