Evaluation of bone formation in calcium phosphate scaffolds with μCT-method validation using SEM

Calcium phosphate scaffolds are frequently used as bone graft substitutes. Bone formation within these scaffolds is commonly assessed using μCT. However, the similar x-ray attenuation of bone and CaP complicates phase separation in μCT images. Current methods rely on manual thresholding, which lacks standardization and may introduce bias. This gap motivated the development of a more objective thresholding approach. Prior research has shown that manual segmentation can lead to significant variability. No prior work had resolved how to systematically reduce thresholding errors. This study addresses the need for a standardized μCT analysis method. It aims to improve the accuracy of bone formation evaluation in CaP scaffolds.

The study aimed to develop a reliable method for evaluating bone formation in CaP scaffolds using μCT. Bone formation is typically assessed using μCT, but the similarity between CaP and bone in x-ray attenuation complicates accurate segmentation. Manual thresholding is commonly used but suffers from subjectivity. This study sought to minimize thresholding errors by using SEM as a reference. The canine animal model was used to generate μCT and SEM data. The goal was to correlate μCT area fractions with SEM-derived values. The study aimed to provide a standardized, objective thresholding method. This approach would improve the accuracy of bone formation measurements.

The study used μCT to evaluate bone formation in CaP scaffolds. Two scaffold architectures were tested: foamed and robocast. μCT datasets were obtained from a 12-week canine animal model. SEM images were acquired as reference data for threshold validation. μCT datasets were registered to corresponding SEM images. Global thresholds were determined by correlating area fractions in μCT and SEM images. Manual thresholding was also performed using two different methods. The accuracy of each thresholding approach was compared using area fraction errors.

Using SEM as a reference reduced thresholding errors to less than 3%. Manual thresholding resulted in average errors up to 17%. Thresholds based on a single SEM reference had lower errors than manual methods. The SEM-based method provided more consistent results across scaffold types. Foamed and robocast scaffolds showed similar bone formation trends. The correlation between μCT and SEM data was strong. Area fractions measured by μCT were validated against SEM-derived values. The study demonstrated that using SEM references improves thresholding accuracy.

The study proposed a method to reduce thresholding errors in μCT evaluation of bone formation. Using SEM as a reference improved the accuracy of μCT segmentation. Manual thresholding was found to be less reliable than the SEM-based approach. The method is applicable to both foamed and robocast CaP scaffolds. The authors suggest that this approach could standardize μCT analysis in bone graft studies. The method reduces subjectivity in threshold selection. It allows for more reliable quantification of bone formation. The results support the use of SEM-based thresholding in μCT studies.