This article reports on the current status of the American Association of Orthodontists Foundation (AAOF) Craniofacial Growth Legacy Collection—an AAOF-supported multi-institutional project that uses the Internet and cloud computing to collect and share craniofacial images and data for orthodontic research and education. The project gives investigators and clinicians all over the world online access to longitudinal information on craniofacial development in untreated children with malocclusions of various types. It also is a unique source of control samples for testing the validity of consensually accepted beliefs about the effects of orthodontic treatment or of failure to treat.
With support from the American Association of Orthodontists Foundation (AAOF), 9 of the 12 known collections of longitudinal craniofacial growth records in the United States and Canada have joined to create a Web site of cephalometric images with an underlying numeric database, accessible through any standard Internet browser at www.AAOFLegacyCollection.org . Its function is to make representative materials from the participating collections readily available to clinicians, craniofacial investigators, students, and other interested persons.
The 9 participating collections represent the work of hundreds of investigators over many years. The materials they contain are literally irreplaceable. In addition to cephalograms of different types, the collections include intraoral radiographs, hand-wrist films, study casts, and written records charting the physical development through time of children of different ethnicities and growth patterns. Each participating collection is independent from the others and has pursued its own strategies of design and data collection. The product of these different and complementary strategies constitutes a rich record of longitudinal craniofacial development among children who never received orthodontic treatment. The intent of the Legacy Collection is to gather representative subsets of these records starting with lateral cephalograms from all available collections and to make the combined materials freely available via the Internet.
Each of the 9 participating collections is located at a major university or research institute. Each contains physical records of longitudinal craniofacial development in untreated growing children. These records were acquired during a brief historical window in time, approximately between 1930 and 1985. Before 1930, the technologic capacity for tracking and recording the growth and development of the internal structures of the human head through time did not exist. However, well before the end of the 20th century, the continued gathering of such records from untreated children was precluded by the recognition of the possibility of deleterious effects from the excessive use of ionizing radiation. Hence, longitudinal collections of this kind can never again be acquired. It therefore seems imperative to preserve the available original images in digital form before their loss or physical deterioration makes that impossible.
Out of concern for the preservation of this important research legacy, the National Institute of Dental Research in 1988 sponsored a survey of existing longitudinal records collections in the United States and Canada by Hunter et al. Their report identified 12 extant collections containing various combinations of x-ray cephalograms, plaster study casts of the teeth, other physical records, and demographic documentation. Taken together, these 12 collections have been the source of a large portion of the available published information on longitudinal craniofacial growth, including several hundred peer-reviewed articles in major scientific journals. Many of these publications are required reading for orthodontic and oral surgery residents in the United States and elsewhere, and they have contributed strongly to the belief systems that clinicians use in the delivery of orthodontic and maxillofacial surgical treatment throughout the world. But as valuable as these publications are, only a small portion of the irreplaceable information contained in the original collections has yet been analyzed.
The idea of preserving a substantial collection of these irreplaceable images and associated numeric data had been a dream of clinicians and craniofacial investigators for many years. In 2008, Mark Hans and his associates at Case Western Reserve University sought and received support from the AAOF to organize a meeting of representatives from interested institutions to investigate the potential for developing a shared virtual resource of longitudinal craniofacial growth records. This meeting led to the formation of a consortium among the separate collections to test the feasibility of constructing a sharable image base and database with continuing support from the AAOF.
Nine of the 12 collections identified in the study of Hunter et al collaborated in this consortium. The universities and institutes involved include Case Western Reserve University (Cleveland, Ohio), University of Michigan (Ann Arbor), University of Iowa (Iowa City), University of Oklahoma (Oklahoma City), Oregon Health Sciences University (Portland), University of the Pacific (San Francisco, Calif), University of Toronto (Toronto, Ontario, Canada), Forsyth Institute (Cambridge, Mass), and Wright State University (Dayton, Ohio). A valuable liaison among the AAOF, the American Association of Orthodontists, and the consortium was provided by Leslie Will and Carla Evans.
The initial phase of the AAOF Legacy Collection project was designed specifically as a test of feasibility of a collaborative project that could integrate contributions from the 9 collections. The Craniofacial Research Instrumentation Laboratory (CRIL) at the University of the Pacific was the designated site for the development of a prototype sharable database and Web site to which each participating collection would contribute materials for a representative subset of cases. The prototype system was constructed using components previously developed at the CRIL with the support of the National Institutes of Health and the AAOF. During an initial test period from June 2009 to December 2010, the 9 participating collections successfully demonstrated the feasibility of the collaboration. The AAOF then authorized a full-scale project aimed at the collection of longitudinal records for a representative set of subjects from each collection with an emphasis on lateral cephalograms. All samples were anonymized and met the institutional review board criteria of their respective institutions. The main data collection period lasted from January 2011 to December 2014. At that point, the AAOF Legacy Collection database contained records for 762 untreated subjects from the 9 collections. Tables I and II summarize the composition of the collection at that time. Further documentation may be found in the Statistics section under the “About” tab of the Web site itself.
|AAOF legacy collection contents|
|Number of collections||9|
|Number of cases||762|
|Number of images||16,024|
|Average number of images per case||21|
|Average age at first image (y)||5.6||5.6||5.6|
|Average age at last image (y)||18.2||18.7||18.5|
|Average range (y)||12.8||12.9||12.8|
Operation of the database and Web site
The AAOF Legacy Collection may be viewed from locations throughout the world using any standard Internet browser. At the beginning of our project, it was conceived of primarily as an archive, a relatively simple repository in which cases and images could be stored against degradation or loss. Fortunately, the development of modern database technology has made it possible to search for cases and images with similar properties across the entire collection. It is this ability to search for, aggregate, and prepare for analysis cases and images with properties of interest that is the main emphasis of this introductory report.
The Web site image display is arranged by the participating collection. In each such collection, cases are listed sequentially in a Detailed Collection Inventory. The attributes listed for each subject include case identification, sex, Angle class, image types available, age at each time point, number of time points, and whether conventional hard tissue landmarks have been identified. Since each of these attributes is treated by the database as an individual variable, case and image selections can be made on the basis of any available criterion or combination of criteria.
To demonstrate the operation of some of these capabilities, Figure 1 displays a representative lateral cephalogram of a patient. Like every other cephalogram in the collection, it is stored and identified in the database and can be displayed on the computer at full screen size. At the bottom of the figure is a menu of check boxes listing properties that can be varied to increase the usefulness of the image. The contrast or brightness of the image can be altered, it can be displayed side by side with any other image for comparison purposes, and a negative version can be displayed if desired. If hard tissue landmarks have been previously identified, they can be displayed overlaid on the image with or without labels. The “Previews” button controls the display of large thumbnails representing each available time point for that subject, permitting the user to get an otherwise unavailable sense of longitudinal changes through time for that particular case. This capability is illustrated in Figure 2 in a series of 12 lateral cephalograms for an untreated subject with an anterior open bite.