Some mechanical measuring devices are used to arrive at a more objective determination of occlusal wear. Examples of mechanical devices include the laboratory scale, stereo microscope, commercial profilometers, customized profilometers, a computerized three-dimensional measuring microscope and a laser profilometer. Regardless of the type of device, however, all seem to have a few insufficiencies in common; they result in high standard deviations as a result of inaccurate replicas, and they present repositioning problems and measuring device restrictions.¹³

Currently, digital mapping of tooth surfaces seems to be the most precise mechanical method for indirectly analyzing restoration wear. The methods of determining accuracy as well as the techniques themselves have varied.^13–16 Three techniques appear to achieve a high level of accuracy. The primary difference between the three techniques lies in the actual measuring device.

The primary benefits of these wear analysis methods are that they provide both quantitative and qualitative data, and they specify precisely how much composite wear has occurred and in which areas of the restoration the wear has taken place.

The three methods of digital mapping include the system developed by Clinical Research Associates, or CRA,¹⁷ the Minnesota system^18,19 and the three-dimensional laser digitizing method.^18–20 The third method uses a high-precision laser rather than a modified laboratory microscope (as is used by CRA) or a computer-driven stylus (Minnesota system).¹⁴

The three-dimensional laser digitizing method is used in larger clinical studies.^21,22 With this method, the computer-driven laser scans a stone cast of the restoration being assessed and creates a three-dimensional computerized surface model of the tooth. A standard computer algorithm is used to superimpose follow-up images over baseline images and to calculate the precise location and amount of composite wear over the total restored surface. The automatic matching software program (Laser Scan 3D), along with the three-dimensional data acquisition, can detect wear within 10 µm.²² Because the process is quick and highly accurate, complex analyses of three-dimensional wear can be conducted on a large number of samples.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Seventy-four patients ranging in age from 20 to 60 years, whose treatment plan called for Class II restorations, were selected from the patient pool at Tufts Dental Clinic, Boston. Thirty-nine patients received Pertac (ESPE America) resin-based composite restorations and 35 patients received TPH (Dentsply L.D. Caulk) resin-based composite restorations. A total of 100 posterior restorations were placed in contact and occlusion with natural dentition.⁹ Cavity preparation and insertion were performed by the same operator.

Polyether impressions (Impregum, ESPE America) were taken at baseline and stone casts were made (Die-Keen, Heraeus Kulzer Inc.). Patients were examined at recall appointments after six months, 12 months and 24 months. To test the potential comparability of the composites, two clinical posterior composite studies were performed to demonstrate that the composite material is an independent variable and should not play a role in the wear analysis process. From the two studies, a total of 21 models were chosen in which enamel areas outside the occlusal contact points were present.

The models were analyzed using the following:

– a human evaluator indirect cast comparison method, or ICCM, to measure vertical height loss (using the Vivadent scale) (Figure 1);

– the three-dimensional laser digitizing ICCM to measure volume loss;

– the normalized three-dimensional laser digitizing ICCM to measure vertical height loss.

View larger version (76K): [in this window] [in a new window]   Figure 1. Vertical height loss using the human evaluator indirect cast comparison method, or ICCM, vs. the normalized three-dimensional laser digitizing ICCM.

Human evaluator ICCM. After each stone cast was poured, two of us (R.P., G.K.) who had undergone special calibration training in evaluating wear models independently assessed each cast for wear using the set of 18 previously calibrated standard models that vary by 25-µm increments. The scale measures the amount of generalized wear based on the interface between the tooth and the restoration. Restorations that exhibited a degree of wear between those of two of the standard cast models were assigned the lower values.

On a random basis, both evaluators individually graded each of the 21 test models three times against the baseline stone cast model and then against the Vivadent casts. To determine a wear measurement for each model, we averaged together their readings.

Magnifying loupes with a x2 magnification and a pointed lighting source were used during the evaluation for better determination of restoration abrasion.

Three-dimensional laser digitizing ICCM and normalized three-dimensional laser digitizing ICCM. To gain a more objective reading of wear, we performed a computerized analysis in two ways (that is, three-dimensional laser digitizing and normalized three-dimensional laser digitizing) to establish two separate measurements (volume loss and vertical height loss).

For the laser-digitizing portions of the study, the same stone models from each of the 21 patients were measured. Some of these casts represented wear in patient restorations after six months in vivo, while others represented wear after a longer period (up to 31 months [although the study period was 24 months, some patients were unable to attend the 24-month recall appointment and were examined later]). These follow-up casts from each patient were compared with the baseline stone casts.

The three-dimensional laser digitizing unit, developed at the University of Munich, Germany (by Laser Scan 3D), is the only computerized wear analysis system currently available that uses a laser sensor. The system’s lateral resolution is 25 µm and the vertical resolution is less than 5 µm. The wear sensor uses a triangulation angle of 30 degrees.²¹

To use this unit, the operator seats each specimen in a holder in the machine. The specimens are then scanned with the laser. These scans are processed by the automatic matching software to create a computerized model of the tooth and restoration. Total three-dimensional volume for each model is calculated as well as volume for designated areas of the models. Before being scanned, the stone models are scored at nonessential places to serve as reference points for the superimposition of the computerized images of the tooth surface.

By superimposing images of the baseline and follow-up models, we can detect wear differences. After the images are scanned into the computer, the baseline model images are compared with their corresponding follow-up images by aligning the score marks and using software matching; total three-dimensional volume loss, in cubic micrometers, then can be determined mathematically using algorithms.²² The vertical height loss is calculated by using a normalized three-dimensional laser digitizing ICCM. In addition, the total three-dimensional volume of the restoration must be differentiated from the total three-dimensional volume of the tooth.

We could not directly compare wear measurements made using the Vivadent scale (average height loss in micrometers) with measurements made conventionally with the three-dimensional laser digitizing unit (volume loss measured over the total restoration surface in cubic micrometers). For this reason, an adjustment was made to the digitizing unit that allowed it to measure test specimens in the same fashion as the Vivadent scale—essentially, we incorporated the Vivadent scale into the digitizing method.

Ten line tracings of the three-dimensional data ("cuts") were made through the stone casts to measure vertical height loss at different points at the margins of the restorations. These measurements were averaged to derive a quantitative amount of wear that could be compared with the measurements recorded by the human evaluators using the Vivadent models.

Therefore, to compare the methods of the human evaluator ICCM with those of the three-dimensional laser digitizing ICCM (Figure 2), a two-step data conversion was needed. We converted the continuous scale obtained from the three-dimensional laser digitizing ICCM setup to a 25-µm interval scale, and normalized the volumetric data according to the projected area of the tooth and to the ratio of the projected restoration. This resulted in the independent mean height loss per pixel. A normalization formula was applied to these measurements to convert them to micrometers and to allow a logically valid data comparison. The mathematical formula used is as follows:

View larger version (101K): [in this window] [in a new window]   Figure 2. Photograph of indirect reference casts (Vivadent).

Therefore, height (mm) = three-dimensional volume (mm³)/area (mm²).

According to Lang and colleagues,²³ the unit is expressed as mm³/mm² to identify it as a normalized volume.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
During the 31-month study period, the evaluator measured less than 25 µm of wear in each of the 21 case restorations. By contrast, the results obtained with the three-dimensional laser digitizing ICCM and the normalized three-dimensional laser digitizing ICCM were much higher, revealing up to 300 µm of wear.

In 19 (90 percent) of 21 cases, the human evaluators did not observe as much wear in the stone cast models as the three-dimensional laser digitizing method detected. Specifically, the human evaluators’ ICCM rated 16 cases as exhibiting 0 µm of wear and five cases as exhibiting 25 µm of wear. The normalized three-dimensional laser digitizing ICCM rated wear as follows: one case, 0 µm; three cases, 25 µm; three cases, 50 µm; seven cases, 75 µm; two cases, 100 µm; one case, 125 µm; two cases, 150 µm; zero cases, 175 µm; one case, 200 µm; zero cases, 225 to 275 µm; and one case, 300 µm.

We performed a statistical evaluation to determine whether a difference existed between wear measurements reported by the evaluators and those derived from the three-dimensional digitizing process. Comparisons of the means were performed with the Wilcoxon test, the correlations were calculated with Kendall and any agreement between the ICCMs was calculated with Cohen’s .

Po represents observed concordance and Pe represents concordance by chance. The statistic describes the extent to which the recorded degree of agreement would have improved on chance. In other words, it compares the amount of agreement recorded with that which would have been recorded had the evaluation been done at random.

Results of the comparisons between the normalized three-dimensional laser digitizing ICCM and the human evaluator ICCM, between the three-dimensional ICCM and the human evaluator ICCM, and between the three-dimensional ICCM and the normalized three-dimensional ICCM are found in the table.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

One possible reason for these reported low-wear ratings by evaluators is that the human eye has a maximum resolution power of 100 µm when an object is just 25 cm (about 10 inches) away from the viewer.²⁶ Use of the normalized three-dimensional laser digitizing ICCM resulted in scores ranging from 0 to 300 µm. Under ideal conditions, the mean values for the two test methods should be the same; thus, the mean difference should lie at the 0 point of the Y-axis when graphically depicted. In this study, the mean of all the values from the two methods is 83 µm, which is almost as large as four evaluation categories (0 to 24 µm, 25 to 49 µm, 50 to 74 µm, 75 to 99 µm), while the deviation in lower amounts of wear was less than that in higher amounts of wear.

The normalized three-dimensional laser digitizing ICCM evaluation agrees in principle with the Vivadent method, which involves the use of 18 standardized first-molar models to evaluate restoration wear.

Varying shapes of the reference casts may also explain why low-wear ratings were recorded by evaluators. In a comparison of Leinfelder’s scale with the M-L scale, Taylor and colleagues²⁷ reported lower values for the Leinfelder scale in comparable dimensions. They argued that the simple geometric form of the M-L scale may be in contrast to the irregular Leinfelder tooth form, and thus could contribute to the complex psychological nature of this evaluation process.

The three-dimensional ICCM, on the other hand, offers the advantage of eliminating such subjectivity from clinical studies of resin-based composite wear. It also offers the advantage of measuring composite restoration wear over the total surface area. Qualitative data also are provided and reveal the precise location and nature of tooth and restoration wear.

There is a fundamental difference between the human evaluator ICCM and three-dimensional ICCM procedures, which requires the data from the three-dimensional ICCM (which measures volume loss) to be converted to normalized data from the three-dimensional ICCM (which measures vertical height loss) in regard to the area corresponding, in principle, to a linear abrasion measurement. The high probability of correlation between the human evaluator ICCM and the normalized three-dimensional laser digitizing ICCM certainly points to a strong connection between the examined data. To document a strong agreement, however, the regression straight line of the values would have to agree with the straight-line y = x (line of equality). According to Bulman and Osborn,²⁸ the statistic of probability corresponds with the highest reliability for comparison of test methods. The intervals are 25 µm, and a weighted statistic was calculated to allow for the adjacent categories in the wear evaluation.

The mean value comparisons, the correlation coefficient and statistic values do not result in agreement between the human evaluator ICCM and the three-dimensional ICCM. Our results indicated that none of the combinations of comparison approximated a straight line (y = x); therefore, it is not possible to directly compare the measurement data from these wear analysis methods.

Our results showed that the mean amount of restoration wear determined by the normalized three-dimensional ICCM was two evaluation categories (at least 50 µm) higher than the mean amount determined by the human evaluator ICCM. Even when we allow for systematic differences in the statistic, the degree of agreement between the two methods with a value of 0.32 would be too negligible to be worthy of consideration.

The high abrasion values in the three-dimensional ICCM analysis suggest that it might not be possible to differentiate restoration edge roughness or fragmentation from wear.

The cost (ranging from about $75 to $300 for a set of models) as well as the need to create a new standard reference for wear judgment might be the dominant obstacles to accepting the three-dimensional evaluation method.

	CONCLUSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSION REFERENCES

 
Systematic differences among the various ICCMs highlight the unsolved problems of subjective evaluations of restoration wear. Even in a metrical simulation of subjective analysis of abrasion, the vertical height loss and volume wear results cannot be compared. In practice, this indicates that a new data pool for the comparison of various materials needs to be constructed, since the existing reference casts used in the human evaluator ICCMs are not comparable.