Abstract
PURPOSE
To assess the accuracy of three-dimensional models of the human temporal bone generated from CT data.
METHODS
Thin-section CT of a left human cadaveric temporal bone was performed using multiple-scan planes (axial, coronal, and sagittal) at 1.5-mm section thickness and 0.25-mm pixel size with an edge-enhancement two-dimensional algorithm. CT data were converted to toggle point format based upon a threshold value of 200 (approximately -830 HU) obtained from prior experimentation with a CT phantom. Selective laser sintering of polycarbonate powder was performed at a beam diameter of 0.060 inches (1.5 mm), 100 scan lines per inch, layer thickness of 0.010 inches (0.25 mm), and layer repeat factor of 4. The polycarbonate models were then scanned in the axial, coronal, and sagittal planes and compared with the original CT data. Anatomic dissection of the models was performed for further verification of the imaging findings.
RESULTS
Models of high anatomic accuracy were generated. Shortening by a factor of 0.67 along the Z axis secondary to the layer repeat factor of 4 resulted in distortion of the models. Distortion in the XY plane ranged from 0% to 20%. Differences in model accuracy based on the initial CT scan plane were observed. A significant amount of nonsintered or partially sintered polycarbonate resulted in intermediate density on the CT images.
CONCLUSIONS
Selective laser sintering can result in accurate modeling of detailed anatomic structures in the human temporal bone. Further investigation of materials and factors contributing to the accuracy of selective laser sintering in the manufacturing of high-resolution anatomic models is warranted.
Full Text
The Full Text of this article is available as a PDF (2.7 MB).