Introduction: Conventional cephalometry is an inexpensive and well-established method for evaluating patients with dentofacial deformities. However, patients with major deformities and in particular asymmetric cases are difficult to evaluate by conventional cephalometry. Reliable and accurate evaluation in the orbital and midfacial region in craniofacial syndrome patients is difficult due to inherent geometric magnification, distortion and the superpositioning of the craniofacial structures on cephalograms. Both two- and three-dimensional computed tomography (CT) have been proposed to alleviate some of these difficulties. Aims and objectives: The aim of our study is to compare the reliability of anatomic cephalometric points obtained from the two modalities: Conventional posteroanterior cephalograms and 3D CT of patients with facial asymmetry, by comparison of intra- and interobserver variation of points recorded from frontal X-ray to those recorded from 3D CT. Materials and methods: The sample included nine patients (5 males and 4 females) with an age range of 14 to 21 years and a mean age of 17.11 years, whose treatment plan called for correction of facial asymmetry. All CT scans were measured twice by two investigators with 2 weeks separation for determination of intraobserver and interobserver variability. Similarly, all measurement points on the frontal cephalograms were traced twice with 2 weeks separation. The tracings were superimposed and the average distance between replicate points readings were used as a measure of intra- and interobserver reliability. Intra-and interobserver variations are calculated for each method and the data were imported directly into the statistical program, SPSS 10.0.1 for windows. Results: Intraobserver variations of points defined on 3D CT were small compared with frontal cephalograms. The intraobserver variations ranged from 0 (A1, B1) to 0.6 mm with the variations less than 0.5 mm for most of the points. Interobserver variations of points between first and second tracings defined on PA Ceph and 3D CT were less than 1.5 mm for all the points. A direct comparison of frontal cephalometry and 3D CT using intraobserver variation for the first observer showed that 3D CT was more reliable than frontal cephalometry for 11 points. Direct comparison of frontal cephalometry and 3D CT for the second observer showed a similar tendency with 14 points significantly more reliable for 3D CT. Comparison of frontal cephalometry and 3D CT using interobserver variation for the tracings showed that 3D CT was more reliable than frontal cephalometry. Summary and conclusion: The study has shown that 3D CT is consistently more accurate and reliable than conventional frontal cephalometry. Unfortunately, the cost of a computerized tomography examination is considerable; this means that, at this stage, its use by surgeons and orthodontists is a last resort.