Shade-Matching Ability of a Spectrophotometer in Comparison to Clinician

Objectives: To compare the shade-matching ability of a spectrophotometer to a clinician.
Methods: Twenty extracted, anterior teeth with varying shades A1-D4 were prepared with a class V preparation (approximately 6mm x 2mm). Ten samples were restored with a one-shade universal supra-nano composite, the other 10 used universal nano-hybrid composite shade IVA (Universal A shade) according to the manufacturer’s instructions. All teeth were covered entirely with Opalescence^® Boost 40% HP (Ultradent) and bleached per manufacturer’s instructions, two consecutive times. The teeth were placed in distilled water for 24 hours. L*a*b* and VITA shade were read with the CrystalEye Spectrophotometer (Olympus) over the restoration and two sites 1mm above and below. Color difference (△E) between the restoration and adjacent sites were calculated from the L*a*b*values. For the clinicians’ VITA shading, three clinicians were calibrated, while two clinicians shaded each time point. Each tooth was given a VITA shade value and a matching score (1=matches or 0=not matches). The spectrophotometer and clinician readings were performed at 24 hours post-operative, 24 hours, 1-week, 2-weeks, and 4-weeks post bleaching. Restorations were considered an acceptable match if △E < 3.7, according to the literature. VITA shades were considered exact match if the restoration and adjacent sites had an exact VITA shade value. Cohen's kappa was used to measure the agreement in shade-matching, and the McNemar test to compare the VITA shade value exact match between the restoration and adjacent sites recorded by the spectrophotometer versus clinician.
Results: Clinicians reported that 97% of the restoration and enamel comparisons matched well (matching score=1) while the spectrophotometer recorded an acceptable match (△E <3.7) in 49.5% of the comparisons. Kappa between the spectrophotometer and clinicians was 0.02. VITA shade value exact match from clinician is statistically different from spectrophotometer (p < 0.001).
Conclusions: The spectrophotometer had a significantly lower color match ability compared to clinicians.