Method: Eighteen extracted molars were collected (IRB #13-02372-XM), cleaned, and stored in water. The teeth were fixed in stainless steel rings with crowns exposed. The stainless steel rings incorporated spherical references. The crown surfaces were scanned using an optical 3D scanner (COMET xS, Steinbichler). Mesio-occluso-distal slots (4-mm wide and 4-mm deep) were prepared. Fifteen minutes after cavity preparation the crowns were scanned again. Teeth were kept in water, except during the scanning. Using the reference spheres, the before and after preparation scans were precisely aligned by minimizing the root-mean-square difference (Cumulus software). Cusp flexure was determined by calculating the difference between the before and after scans across the buccal and lingual surfaces (CuspFlex software). T- and chi-square tests (significance level 0.05) were used to test the hypothesis that there was no preference in cusp flexure or directions.
Result: Cavity preparation caused bending in buccal and lingual walls. Cusp flexure (mean±standard deviation) was 2.56±1.95 µm and 3.08±2.86 µm for the buccal and lingual walls, respectively. The difference in flexure between buccal and lingual walls was not significant. Cavity walls bent either outward or inward. Of the 18 teeth, 8 showed crown expansion (both walls bent outward), 3 contraction (both walls bent inward), and 7 parallel (both walls bent in same direction). There was not enough evidence that either pattern of cusp movement was more likely than others.
Conclusion: Evidence of cuspal flexure caused by cutting a cavity in extracted teeth during in vitro testing was documented. No significant differences were found in the amount or direction of flexure between buccal or lingual cusps.