For many years, digital subtraction radiography in dentistry has been used to qualitatively assess changes in radiographic density. Numerous authors have demonstrated the ability of this method to improve diagnostic performance for the detection of approximal dental caries [3], periapical pathology [4] and periodontal disease [5-9].
More recently quantitative methods have been developed for comparing radiographs longitudinally. These methods provide for extremely reproducible and objective measures of radiographic density changes. These methods have been shown to be more sensitive than qualitative methods.[10-16]
Longitudinal radiographic analysis can therefore be defined as either qualitative or quantitative methods which characterize the changes in radiographic density that occur over time. By utilizing longitudinal radiographic analysis earlier diagnosis of disease may be possible; furthermore, assessment of therapeutic agents or surgical interventions may be more precisely evaluated for their effectiveness.
Until the development of this project these methods have been applied to basic research and limited clinical trials only.
The purpose of the Longitudinal Radiographic Analysis project is to place these methods of analysis directly into a clinical environment for basic patient care and widespread clinical research applications.
First the patient is referred to the Reproducible Radiology Clinic. The patient will be imaged using a cephalostat device with large source to object projection geometry and a rotating anode high frequency medical x-ray generator. This technique will ensure that consistent projection geometry will be achieved for each follow-up examination. All exposure parameters will be recorded for follow-up examinations. The radiographs will be processed using an automatic film processor. The reference radiograph will then be digitized and entered into the CAREİ system by the radiology technologist. Follow-up examinations will be performed by the same technician in the Reproducible Radiology Clinic. The follow-up radiographs will be spatially registered to the reference image using real-time subtraction and a micromanipulator. The radiograph will then be digitized and entered into the CAREİ system. Each subsequent follow-up examination will be entered into the CARE system creating a longitudinal series of images.
In order to perform a longitudinal radiographic analysis, the clinician uses an image analysis workstation to first query the patient/examination database for his patient. He may use any available workstation to access the data on the central image file server. Once the patient record is retrieved a list of longitudinal examinations will be displayed. The clinician selects the examination of interest and the entire series of digital images will be retrieved from the central file server and loaded into RAM. Next the user determines the type of longitudinal analysis he wants to use. He may select from subtraction, semi-quantitative or quantitative analyses. The user can select any of the images in the series to serve as the reference or "baseline" image for the analysis. Next the user selects any number of images in the series to be compared to the reference image as long as that image was acquired chronologically after the reference image.
The user is instructed to define the area of interest on the reference image which will be used in the gamma correction of the follow-up images. Dependent upon the analysis chosen the user will be directed to the appropriate action using dialogue boxes. The results of the longitudinal analysis will be displayed in a results window. The image or image data can be output to hardcopy if that is needed or output as digital data on floppy disk. (See flow diagrams on following pages).
The system is composed of three components: 1) image acquisition module, 2) image analysis workstation, and 3) central image file server. The acquisition module is designed for the registration, digitization and image management of longitudinally acquired radiographs. The image analysis workstation is designed to load a longitudinal study from a central file store and perform image analysis such as digital subtraction, semi-quantitative densitometric analysis, and quantitative densitometric analysis. The image file server serves as a central image storage system were images are stored from the acquisition module and retrieved by an image analysis workstation.
The software will be based on MSDOS 5.0 and Windows 3.0 operating system environment. The basic database structure will utilize Novell Btrieve and run on a Novell Netware 386 network. The program which will be called CAREİ (Computer Aided Radiographic Evaluation) is based on a modification of the RadWorksİ program for direct digital radiography.
The software designed for the acquisition module will provide for the data management of two relational structures, patient information and examination information, in a Btrieve database. The patient information will be composed of patient last name, middle initial, first name, chart number and date of birth. The examination information will be composed of date and time of baseline exam, referring doctor, referring service, image modality (ie. intraoral), exam type (ie. longitudinal), and exam area (ie. max. left). After the patient and examination information of a longitudinal exam is entered into the database. A reference image is acquired to which all subsequent follow-up radiographs will be registered.
An examination image file will consist of a file header of variable length composed of all database fields and the dates of all subsequent examinations followed by the binary image data of the reference and all follow-up images. Therefore the examination image file will increase in size as new follow-up examinations are added.
The software designed for the image analysis workstation will provide for access to the patient/examination database and central image archive. From the software, the database is queried for a specific patient. A list of all examinations will be displayed for that patient. Once the examination is selected the entire examination image file is retrieved from the central archive and loaded into RAM. The user can then determine the type of analysis which is to be performed. The software will support digital subtraction, image enhancement, and at least three different methods of quantitative densitometric analysis. The user will be allowed to select any image within the series to serve as the reference image. Then any number of follow-up images which are chronologically after the reference image can be selected for analysis. The resultant subtracted image or processed image will be displayed in a results windows. Quantitative image analysis results will also be displayed and can be printed out.
Novell Netware 386 version 3.1.1 will serve as the system software and file server software. The patient/examination database will reside on the central server. The database is based on Novell Btrieve and will be compatible with the RadWorksİ program for direct digital acquisition radiography systems. This will allow the two imaging systems to be merged at a later time.
2. Kundel, H. and G. Revesz, Lesion conspicuity, structured noise, and film reader error. AJR, 1976. 126: p. 1233-1238.
3. Grondahl, H.-G., et al., Statistical contrast enhancement of subtraction images for radiographic caries diagnosis. Oral Surg Oral Med Oral Pathol, 1982. 53: p. 219-223.
4. Grondahl, H., K. Grondahl, and R. Webber, A digital subtraction technique for dental radiography. Oral Surg, 1983. 55: p. 96-102.
5. Grondahl, H.-G. and K. Grondahl, Subtraction radiography for the diagnosis of peridontal bone lesions. Oral Surgery, 1983. 55: p. 208-213.
6. Webber, R., U. Ruttimann, and H. Gršhndahl, X-ray image subtraction as a basis for assessment of periodontal changes. J Perio Res, 1982. 17: p. 509-511.
7. Lurie, A., R. Greenberg, and K. Kornman, Subtraction radiology demonstrates crestal bone loss in experimentally induced marginal periodontitis. Oral Surg, 1983. 55: p. 537-542.
8. Rethman, M., et al., Diagnosis of bone lesions by subtraction radiography. J Perio Res, 1985. 56(6): p. 324-329.
9. Hausmann, E., et al., Usefulness of subtraction radiography in the evaluation of periodontal therapy. J Periodontol, 1985. (Spec iss): p. 4-8.
10. Vos, M., et al., Quantitative measurement of periodontal bone changes by digital subtraction. J Periodont Res, 1986. 21: p. 583-591.
11. Okano, T., et al., Quantitative evaluation of proximal bone lesions using digital subtraction radiography. Dentomaxillofac Radiol, 1988. 17: p. 99-103.
12. McHenry, K., et al., Methodological aspects and quantitative adjuncts to computerized subtraction radiography. J Periodont Res, 1987. 22: p. 125-132.
13. Heaven, T., A. Firestone, and F. Feagin, Quantitative radiographic measurement of dentinal lesions. J Dent Res, 1990. 69: p. 51-54.
14. Bragger, U., et al., Computer assisted densitometric image analysis for the quantitation of radiographic alveolar bone changes. J Periodont Res, 1987. 22: p. 227-229.
15. Bragger, D., et al., Computer-assisted densitometric image analysis in periodontal radiography. J Clin Periodontol, 1988. 15: p. 27-37.
16. Bragger, U., et al., Computer assisted densitometric image analysis in periodontal radiography. A methodological study. J Clin Periodontol, 1988. 15: p. 27-37.
dove@uthscsa.edu