ABSTRACT
OBJECTIVES: To develop a new method for assessing residual ridge resorption in the edentulous maxilla. METHODS: Defined experimental and reference areas in the maxilla were drawn on transparent film laid over a panoramic radiograph and digitized. Bone areas were measured with an integrated planimetry program and expressed as a ratio R. The effect of positioning errors on reliability of the method was investigated on dry skulls. The correlation between the change in ratio and actual bone loss was examined by progressively reducing the height of an artificial residual ridge on one skull. RESULTS: The coefficient of variation for the absolute ratio in different head positions was < 0.05 and its correlation coefficient of the change in R and the degree of resorption was r2 > or = 98.3%, P = 0.0001. CONCLUSIONS: Comparison of the experimental area with the reference area on serial panoramic radiographs appears suitable for the assessment of residual resorption in the maxilla.
Subject(s)
Bone Resorption/diagnostic imaging , Jaw, Edentulous/diagnostic imaging , Maxillary Diseases/diagnostic imaging , Radiographic Image Enhancement , Radiography, Panoramic , Cephalometry/methods , Humans , Image Processing, Computer-Assisted , Linear Models , Observer Variation , Reproducibility of Results , Statistics as TopicABSTRACT
Magnification factors (VF) in relation to various head positions were evaluated for a digital panoramic radiography machine (Orthophos DS, Sirona) by means of a dry skull phantom. Six metallic objects were attached to the facial layer of the mandible. A set of 170 digital radiographs (program 1) was produced, with defined positioning errors within the vertical and horizontal plane. Repeated digital measurements of the object images were performed by one reader, applying the mouse-driven measurement tool integrated in the proprietary software. In addition, the central layer of the radiographic device was experimentally determined and its location compared to that of the accurately positioned dry mandible. All objects were located on the outside of the central layer of minimum distortion and therefore placed towards the sensor. One- and two-way analysis of variance was applied for the statistical evaluation of the data, with VF as dependent variable and the positions as factors, respectively. Vertical inclinations of the skull affected horizontal VF only in the premolar region (p = 0.0069). However, when the skull was rotated within the horizontal plane, VF was significantly influenced in all anatomical regions of the mandible (p = 0.0001). This was particularly obvious in the premolar region, with a maximum variation of 10.3%. Fluctuations of VF were more pronounced in the premolar region than at the mandibular angle. We conclude from this study that, as expected, digital panoramic radiographs exhibit identical distortion effects as compared to conventional panoramic x-rays. Exact quantification of regional magnification requires (spherical) reference objects with known dimensions. Horizontal magnification is quite unreliable and far less reproducible than vertical magnification. Whenever several panoramic radiographs of one patient are to be compared quantitatively, accurate positioning of the patients' head is a necessary prerequisite.
Subject(s)
Image Processing, Computer-Assisted/instrumentation , Mandible/diagnostic imaging , Phantoms, Imaging , Radiography, Panoramic/instrumentation , Artifacts , Equipment Design , HumansABSTRACT
OBJECTIVES: To evaluate the precision and accuracy of digital measurements in digital panoramic radiography. METHODS: A series of 70 digital panoramic radiographs were obtained of a dry skull in seven different positions with metallic pins and spheres fixed to the mandible. Three replicate measurements were performed with the mouse-driven cursor by one reader at 1:1 and 2:1 magnification. Precision was assessed with the reliability index (R) and Malony/Rastogi test and the effect of magnification on accuracy by paired Wilcoxon test. RESULTS: Vertical measurements were less reproducible than horizontal measurements. There were significant differences in assessments between images at 1:1 and 2:1 magnification (P < 0.05). The maximum variation in mean difference was 0.4% of actual object length for pins and 1.2% for spheres. The difference did not exceed 0.1 mm. R was lower for 2:1 magnification and consistently lower for spheres compared with pins. CONCLUSION: The most reliable measurements were obtained of linear objects in the horizontal plane. Digital measurements are sufficiently accurate for clinical use.
