Comparison between different screening strategies to determine the statistical shape model of the pelvises for implant design. (September 2019)

• Record Type: Journal Article
• Authors: Shih, Kao-Shang; Hsu, Chi-Pin; Liu, Che-Wei; Wang, Lu-Lin; Hou, Sheng-Mou; Lin, Shang-Chih
• Is Part Of: Computer methods and programs in biomedicine. Volume 178(2019)
• Page Start: 265
• Page End: 273
• Publication Date: 2019-09
• Subjects: Periarticular osteosynthesis -- Statistical shape -- Anatomic profile -- Statistical shape model -- Iterative closest point; Medicine -- Computer programs -- Periodicals; Biology -- Computer programs -- Periodicals; Computers -- Periodicals; Medicine -- Periodicals; Médecine -- Logiciels -- Périodiques; Biologie -- Logiciels -- Périodiques; Biology -- Computer programs; Medicine -- Computer programs; Periodicals; Electronic journals; 610.28
• DOI: 10.1016/j.cmpb.2019.06.028
• Languages: English
• ISSNs: 0169-2607

Abstract

Highlights: The four-staged method to calculate the statistical shape model of numerous pelvis. Size-screening strategies of the profile difference are used to exclude some bones with improper sizes. The four groups of SSM were compared in terms of profile difference. The iterative closest points algorithm was used to register the bones for final SSM calculation. Abstract: Background and objectives: The statistical shape model (SSM) of numerous bones has been used to determine the anatomical representative of the population- or race-specific design for periarticular implants. Whether to include size- and profile-mismatched bones in the SSM calculation is debatable. Therefore, the objective of this study was to characterize the screening strategies for the mismatched bones to improve the SSM calculation. Methods: The bone database used in this study consisted of 20 pelvises. A systematic four-staged SSM calculation was used to evaluate the accuracy of the predicted SSM shape among the four size- and profile-screening strategies. Additionally, the surface-smoothing effects on the SSM results were investigated. Two comparison indices were used in terms of profile difference and surface smoothness. Results: Significant variations in size and profile existed for the collected bones. By normalizing the aspect ratio of all bones, exclusion of the size-mismatched bones reduced the maximum and root mean square (RMS) error values of the profile difference by 18.9% and 17.5%, respectively. After further excluding the profile-improper bones, normalization reduced the RMS profile difference by 24.1% compared with the non-normalized strategy. Exclusion of the size-improper bones for non-normalized strategy would have reduced the RMS profile difference by 15.4%. After smoothness, the RMS profile difference of SSM was only 6.1% higher than that of the non-smoothness SSM. Conclusions: The four-stage calculation showed that the most favorable strategy was to normalize bones to the same aspect ratio and exclude improperly shaped bones. The model permitted inclusion of the original characteristics of the bones and preserved their shapes and excluded only significantly improper bones. After SSM calculation, the smoothed process provided satisfaction in quality with a statistically insignificant loss in bone morphology for population- or race-specific designs of implants.