Quantitative evaluation of the accuracy of digital subtraction radiography in serial reproducibility of periapical intraoral images

Dental radiography is mandatory for sufficient diagnosis and pre-treatment patient evaluation. Intraoral radiographs are indispensably utilized both at the decision making process for patient treatment plan construction as well as in postoperative evaluation and follow up visits. Digital radiographic imaging has been introduced as an evolutionary alternative to film-based radiography. Digital subtraction radiography uses two digital images taken under different time points and with specially created computer programs these two images are superimposed upon one another. The resulted image (subtraction image) does not include the common grey scale values and represents / reflects the differences between the initial images. The accuracy of digital subtraction radiography depends on the ability to reproduce the same X-ray projection geometry in serial roentgenograms that are supposed to be subtracted and finally compared. Techniques and holding devices for obtaining serially identical radiographs have been proposed in order to achieve accurate alignment with parallelism between the fixture axis and the film or sensor. Third generation subtraction radiography systems are considered more accurate as they process and adjust different projection geometries so as to produce corrected baseline images before their superimposition and subsequent subtraction. The aim of the present in vitro investigation was to access the accuracy of IRAS/C version 07.00.00 Academic edition computer program in creating digital subtraction images between paired intraoral periapical radiographs. Two screw-type implants were inserted in the right buccal segment referring to the sites of the first and the second molars in a human cadaver mandible. An alignment set up system was constructed and used in order to reproduce identical x-ray conditions with the variable of horizontal angulations being intentionally altered. 69 A total of 91 periapical radiographs of the implants were obtained with the first one being the control while the remaining 90 were divided to nine groups of 10 images taken at various horizontal angulation projections. The horizontal angulation differences between the control baseline X-ray and the test group X-rays were set on 2, 3, 4, 5, 6, 7, 8, 9 and 10 degrees respectively. All X-rays were digitized and processed using the IRAS/C program. Subtraction images were created by subtracting the control image from the test ones appling a 6 point superimposition set up by the examiner with the correlation value between 0.9-1. There were no statistically significant differences between the control X-ray and the X-rays obtained with 2-5 degrees horizontal deviations. On the contrary, X-rays with horizontal variance of 6-10 degrees gave statistically significant different images compared to the baseline control image. The results of the present study demonstrate that serial X-rays taken with different horizontal angulations in central beam while all the other parameters being stable, vary significantly when compared through subtraction and their difference is in accordance to the degree increase. Major prerequisite for intraoral radiographic reproducibility is the application of identical conditions in projection geometry. Digital subtraction radiography is accurate and reliable only in standardized in vitro conditions that are difficult to reproduce in the clinical practic