The accuracy of stereolithography in planning craniofacial bone replacement.

Stereolithography can be used to produce physical models of the craniofacial skeleton from three-dimensional computed tomography (CT) data. The purpose of this study was to assess its accuracy for modeling osseous defects of the midface. Maxillary resections simulating unilateral maxillectomy (N = 3), bilateral maxillectomy (N = 3), and unilateral orbitomaxillectomy (N = 3) were performed as for sinus tumor resection on nine fresh cadaver skulls. Stereolithographic models (SLMs) were made from the specimen's CT data. The accuracy of SLMs was determined by comparing distances between key landmarks on the skulls and SLMs. Each SLM was grossly accurate with some loss of thin delicate structures. The mean differences in overall dimensions between the SLMs and skull specimens were 1.5 mm (range: 0-5.5 mm) for craniofacial measures, 1.2 mm (range: 0-4.8 mm) for skull base measures, 1.6 (range: 0-5.8 mm) for midface measures, 1.9 mm (range: 0-7.9 mm) for maxilla measures, and 1.5 mm (range: 0-5.7 mm) for orbital measures. The mean differences in defect dimensions were 1.9 mm (range: 0.1-5.7 mm) for unilateral maxillectomy, 0.8 mm (range: 0.2-1.5 mm) for bilateral maxillectomy, and 2.5 mm (range: 0.2-7.0 mm) for orbitomaxillectomy defects. Midface SLMs may be more prone to error than those of other craniofacial regions because of the presence of thin walls and small projections. Thus, one should consider designing midface bone replacements that are larger in critical dimensions than those predicted by preoperative modeling. These findings have important implications for the planning of current surgical methods as well as future applications of tissue-engineered bone replacement.