Comparison of particle-size distributions determined by optical scanning and by sieving in the assessment of masticatory performance

Aim of this study was to introduce a feasible and valid technique for the assessment of masticatory performance that is comparable to the standard sieving method. Twenty-one chewing samples (Optosil) comminuted by healthy dentate adults were analysed with a sieving and scanning method. Scanning was performed using a conventional flatbed scanner (1200 dpi). All scanned images underwent image analysis (ImageJ), which yielded descriptive parameters such as area, best-fitting ellipse for each particle. Of the 2D-image, a volume was estimated for each particle, which was converted into a weight. To receive a discrete distribution of particle sizes comparable to sieving, five chewing samples were used to calculate a size-dependent area-volume-conversion factor. The sieving procedure was carried out with a stack of 10 sieves, and the retained particles per sieve were weighed. The cumulated weights yielded by either method were curve-fitted with the RosinRammler distribution to determine the median particle size x50. The RosinRammler distributions for sieving and scanning resemble each other. The distributions show a high correlation (0.9191.0, n = 21, P < 0.01, Pearsons correlation coefficient). The median particle sizes vary between 3.83 and 4.77 mm (mean: 4.31) for scanning and 3.53 and 4.55 mm (mean: 4.21) for sieving. On average, scanning overestimates the x50 values by 2.4%. A modified BlandAltman plot reveals that 95% of the x50 values fall within 10% of the average x50. The scanning method is a valid, simple and feasible method to determine masticatory performance.