Determining the fidelity of multiple casts made from a single polyvinylsiloxane impression

Statement of problem: Indirect restorations may not fit the patient’s teeth, or they may require excessive adjustment. If multiple casts from the same impression are equivalent, dentists can use new pristine casts to determine if fit problems have a laboratory or clinical cause. If the problem is from the laboratory, it may be correctable on a pristine cast. Purpose: The purpose of this in vitro study is to determine if the areas critical for indirect restoration fit are equivalent on 6 casts made over 29 days from a single polyvinylsiloxane impression. Material and Methods: Eight impressions were made of a reinforced plaster reference cast with a fixed partial denture preparation from teeth #3-#5. Forty-eight casts were fabricated from these impressions at 1 hour, 3 hours, 22 hours, 8 days, 15 days, and 29 days. Half of the casts were made with low-expansion mounting stone, the other half were made with high-expansion type 5 stone. All casts were scanned to create STL files. Geomagic Control X was used to compare critical areas of the experimental cast STL files to the reference cast STL file by making two 2-dimensional slices facial-lingually through crown preparations #3 and #5, followed by 2 more 2-dimensional slices mesial-distally through interproximal contact points and crown preparations. Comparison points on these 2-dimensional slices were made on mesial, distal, facial, and lingual preparation walls and margins, on occlusal surfaces, and on interproximal contact areas of teeth #2 and #6. These planes and comparison points were precisely transferable across multiple substitutions of different experimental cast STL files. Therefore, precise point deviations from the reference cast were made across multiple experimental casts. Control data were obtained by comparing the original reference cast STL file to a second reference cast STL file made after the reference cast was repositioned on the scan table to test scanner accuracy, by comparing an experimental cast to the reference cast both early and late in the comparison sequence to test Geomagic algorithm consistency, and by comparing a new reference cast STL file to the original reference cast STL file to test reference cast deterioration. Data were analyzed with Tukey HSD and Games-Howell post hoc tests. A power analysis on trial data showed that 8 impressions would be enough for significant results. Results: There were no statistical differences among all 48 casts when using comparison points p > 0.6. The 48 experimental casts’ comparison points average