[Protocol optimization of three-dimensional spiral CT of craniofacial bones]

To identify and optimize parameters determining the diagnostic quality of three-dimensional surface models derived from craniofacial spiral CT.

Three dry bone skull specimens were scanned with 1-slice, 4-slice, and 16-slice spiral CT. A total of 60 surface models were calculated with variation of several parameters of the CT scan and calculation of the primary data set and SSD reformations. Two observers evaluated the quality of the resulting models independently using a structured questionnaire.

Slice thickness was the only independent factor that influenced image quality. The quality of the 1-slice CT models was poor in all cases. The 4-slice CT models were rated as superior, but also showed relevant impairments. In contrast, the 16-slice CT models provided improved quality with only minor shortcomings.

Actual clinical applications of three-dimensional models for surgical planning or guidance require high quality of the underlying CT data sets. The slice thickness should not exceed 1 mm. CT examination should be performed using a 4-slice system, or preferably a 16-slice system with a dedicated post-processing protocol.