The differential diagnosis of dental wear lesions affects their clinical management. We hypothesized that surface texture parameters can differentiate simulated erosion, abrasion, and erosion-abrasion lesions on human enamel and dentin. This in vitro study comprised 2 parts (both factorial 4 × 2), with 4 lesion types (erosion, abrasion, erosion-abrasion, and sound [no lesion; control]) and 2 substrates (enamel and dentin). Flattened/polished dental specimens were used in part 1, whereas natural dental surfaces were used in part 2. Testing surfaces were evaluated in blind conditions, using average surface roughness (Sa) and the following scale-sensitive fractal analysis parameters: area-scale fractal complexity (Asfc), exact proportion length-scale anisotropy of relief (eplsar), scale of maximum complexity (Smc), and textural fill volume (Tfv). Two-way analyses of variance, followed by Fisher’s protected least significant difference tests (α = 0.05), were used to evaluate the effects of lesion and substrate. Classification trees were constructed to verify the strength of potential associations of the tested parameters. In part 1, Asfc, Sa, and Tfv were able to differentiate erosion and erosion-abrasion lesions from the sound (no lesion) control in both substrates; only Asfc differentiated erosion and erosion-abrasion enamel lesions (all P < 0.05). The best association of parameters correctly classified up to 84% and 94% of the lesions on enamel and dentin, respectively. In part 2, only Asfc differentiated erosion and erosion-abrasion lesions from the sound (no lesion) control in both substrates, whereas eplsar was able to differentiate erosion from erosion-abrasion (all P < 0.05). The association of parameters correctly classified up to 81% and 91% of the lesions in enamel and dentin, respectively. Asfc, Sa, and Tfv were able to differentiate erosion and erosion-abrasion lesions, despite their complicated surface textures. The association of parameters improved the differentiation of lesions for both enamel and dentin in polished or natural surfaces.
Read full abstract