MRI-based stereolithographic models of the temporomandibular joint: technical innovation Gerhard Undt, 1 Klaus Wild, 1 Gerd Reuther, 2 Rolf Ewers 1 1 Department of Oral and Maxillofacial Surgery, University of Vienna, Vienna, Austria (Head: R. Ewers, DDS, MD, PhD); 2 Department of Radiology, Private Clinic Doebling, Vienna, Austria (Head: G. Reuther, MD, PhD) SUMMARY. A new technique of manufacturing dual-colour stereolithographic models of hard and soft tissues of the temporomandibular joint (TMJ) is presented. Sagittal T1/PD weighted magnetic resonance (MR) images of joints with and without disc displacement were obtained in the closed and open mouth positions. Individual interactive contour identification of bony structures and the articular disc followed by binary interpolation provided the data for the generation of acrylic TMJ models. Three dimensional in vivo visualization of the articular disc in relation to bony structures in the closed and open mouth positions allows a new perception of normal and pathological TMJ anatomy. # 2000 European Association for Cranio-Maxillofacial Surgery INTRODUCTION First systematic attempts to depict the articular disc arthrographically were undertaken in the 1940s (Norgaard, 1944). In arthrotomography (Wilkes, 1978; Katzberg et al., 1979), both upper and lower joint spaces were filled with contrast medium followed by tomography. When comparing radio- graphic findings with anatomy of autopsy specimens, the diagnostic accuracy of single contrast arthrogra- phy for diagnosis of disc displacement was about 80% (Westesson & Bronstein, 1987; Tanimoto et al., 1989), while the diagnostic accuracy of computerized tomography (CT) in diagnosing the disc position is only about 70% (Westesson et al., 1987; Tanimoto et al., 1989). The most recent imaging technique, MRI, provides good definition of the soft tissues, is non- invasive, and has minimal radiologic risk (Katzberg et al., 1985). For the diagnosis of disc displacement, the MRI has been reported as up to 95% accurate, with a sensitivity of 0.9 and a specificity of 1.0 (Tasaki and Westesson, 1993). Some authors are even convinced that MRI techniques are superior to surgery for the demonstration of medial or lateral disc displacement (Westesson et al., 1987; Brooks & Westesson, 1993). Also for the assessment of osseous changes in the temporomandibular joint, modern MRI technology seems to be a reliable imaging modality. Tasaki and Westesson (1993) in their cadaver study report an accuracy of 93% (sensitivity 0.87 and specificity 1.0) for the diagnosis of osseous changes by MRI. Three-dimensional (3-D) visualization of anatomi- cal structures and the use of stereolithographic biomodels for preoperative planning have become a useful tool in modern oral and maxillofacial surgery. It has been shown that improved accuracy of diagnosis and the possibility of realistic and inter- active surgical simulation when using stereolitho- graphic models may reduce operating times significantly (D’Urso et al., 1999). Furthermore, the use of biomodels in addition to plain radiographs for patient education improves the level of informed consent when compared with the use of standard images alone. The possibility of 3-D visual reconstruction of osseous structures from CT scans (Marsh & Vannier, 1983; Hemmy et al., 1983; Cutting et al., 1986) has promoted the development of techniques to generate solid models based on CT-data by milling procedures (Brix et al., 1985; Lambrecht & Brix, 1989). The creation of models with the aid of stereolithography allowed the exact reproduction of cranial bony anatomy including closed cavities and generated compact, smooth and detailed surfaces (Mankovich et al, 1990; Stoker et al., 1992; Klein et al., 1992; Wolf et al., 1994). The high precision and accuracy of these acrylic models sufficient for clinical use has been proven (Barker et al., 1994; Bouyssie et al., 1997; Santler et al., 1998). The production of coloured stereolithographic models from CT-data to depict both soft and bony tissue in cleft lip and palate infants has also been described (Nakajima et al., 1995). While Nakajima uses coloured plaster to fill the empty space within the model representing the bony structures, other authors have generated dual- colour stereolithographic models by providing higher doses of laser light to the structures that need to be coloured, which makes the transparent resin turn deep red (Kermer et al., 1998; Sato et al., 1998). In our study, this recently developed technique is applied for generating dual-colour stereolithographic models of the temporomandibular joint based on MRI data. While the disc is represented by red acrylic resin, the bony structures are depicted by transparent white resin, which allows a good 3-D reproduction of 258 Journal of Cranio-Maxillofacial Surgery (2000) 28, 258–263 # 2000 European Association for Cranio-Maxillofacial Surgery doi:10.1054/jcms.2000.0159, available online at http://www.idealibrary.com on