Established microleakage tests have their own disadvantages. In this study, 3D reconstruction method was tried to overcome these disadvantages.

Four types of microleakage tests were used and relationships among them were estimated: penetrated dye volume; marginal adaptability; degree of dye penetration and relative penetrated length to cavity wall.

Twenty-four Class V cavities were bulk filled with composite (Esthet X) following surface treatments: N group (no treatment); E group (etching only); T group (etching + Prime & Bond NT). 50% silver nitrate was used as a dye solution after thermocycling (5℃ & 55℃, 1,000 times). Teeth were serially ground with a thickness of 0.2 mm. Volume of dye penetration was estimated from a three-dimensionally reconstructed image with a software (3D-DOCTOR). Percentage of margin without gap was estimated from SEM and degree of dye penetration and the relative length of dye penetration to overall cavity wall were also estimated.

ANOVA and Scheffe test for dye volume, Kruskal-Wallis and Mann-Whitney test for marginal quality, Spearman's rho test for checking of relationships among methods were used.

The results were as follows:

1. Dye penetration could be seen from several directions, furthermore, its volumetric estimation was possible.

2. Reverse relationship was found between dye volume and marginal quality (r = -0.881 / p = 0.004).

3. Very low relationship was seen between dye volume and two-dimensional tests (degree of dye penetration and relative length). However, 2D evaluation methods showed high relationship (p = 0.002-0.054) each other.

4. Three times vertical section could be recommended as a 2D test.