Assessment of facial landmarks for bone asymmetry in geometric morphometric studies: A review

Background Anthropometrical points or landmarks are key for facial shape analysis using geometric morphometrics. In the early 1990s Fred Bookstein created a classification system with landmarks type 1., type 2., type 3. based on anatomical landmark homology. However, thirty years later, a uniform referece index of landmarks that can be used for assessing facial asymmetry still does not exist. The criteria for selecting landmark points are not fully defined, which makes classification of homology and and landmarks prone to arbitrary interpretations. A systematic review of literature indicates that authors of studies do not provide explanation for choosing exact points. Most of them also do not provide a clear definition of landmarks and landmarks classification according to homology. Objective The objective of this systematic review was to assess, in an evidence based manner, which landmarks inside the Bookstein groups of on facial hard tissues can be reccomended for facial asymmetry assessment using geometric mophometrics. Search methods An electronic search of 9 databases up to March 2020 by two reviewers was conducted to identify relevant articles. Selection criteria Prospective randomized, non-randomized controlled trials and cross sectional studies reporting on facial asymmetry using three-dimensional images and geometric morphometric methods. The reasons for assessing facial asymmetry were not considered. Data collection and analysis The 23 selected studies were categorized according to the number and specification of the research subjects, types of three-dimensional images, number of landmarks, and the craniofacial region of interest. All landmarks were extracted with the following data: name, abbreviation, and the author’s definition of the location. Results The craniofacial region is divided into neurobasic cranial part, ethmomaxillar part and the mandible. Assessment of neurobasic cranial asymmetry was conducted in 6 studies and 45 different landmarks were recorded, of which 11 were medial and 34 bilateral. Bregma and Lambda occur most frequently and according to homology both belong to type 1 landmarks. Assessment of ethmomaxillary asymmetry was conducted in 21 studies and 68 different landmarks were recorded, of which 16 were medial and 52 bilateral. Nasion and Jugale occur most frequently and according to homology Nasion belongs to type 1 landmarks and Jugale to type 2 landmarks.  Conclusion The selection and definition of craniofacial hard tissue landmarks is one of the most important tasks in the design of morphometric studies, and thus for the purpose of assessing facial asymmetry. The review provides an extensive cross-section of possible landmarks with the definition of the location as well as the possible location variation. The list of these landmarks should be observed through the classification of landmarks according to their homology, as well as possible variations of the classification.