The effect of region of interest variations on morphologic operations data and gray-level values extracted from digitized dental radiographs

• Published in: Oral surgery, oral medicine, oral pathology, oral radiology and endodontics oral medicine, oral pathology, oral radiology and endodontics, 1999-11, Vol.88 (5), p.636-639
• Main Authors: Shrout, M.K., Farley, B.A., Patt, S.M., Potter, B.J., Hildebolt, C.F., Pilgram, T.K., Yokoyama-Crothers, N., Dotson, M., Hauser, J., Cohen, S., Kardaris, E., Hanes, P.
• Format: Article
• Language: English
• Subjects: Alveolar Bone Loss - diagnostic imaging; Alveolar Bone Loss - pathology; Alveolar Process - diagnostic imaging; Alveolar Process - pathology; Biological and medical sciences; Dentistry; Humans; Investigative techniques, diagnostic techniques (general aspects); Medical sciences; Otorhinolaryngology. Stomatology. Orbit; Radiodiagnosis. Nmr imagery. Nmr spectrometry; Radiographic Image Enhancement; Radiography, Bitewing - methods; Radiography, Dental, Digital - methods; Statistics, Nonparametric
• ISSN: 1079-2104; 1528-395X
• DOI: 10.1016/S1079-2104(99)70098-3

Objective. To determine the correlations among morphologic operations (MO) values and the correlations among gray-level values for regions of interest (ROIs) placed at various locations on digital images of alveolar bone for 45 patients. Study Design. As part of a larger study, up to 7 vertical bite-wing radiographs were taken and digitized for each of 45 patients. Sets of 2 rectangular ROIs were placed on the digitized images of interdental alveolar bone at 4 locations for each patient. The ROIs (1 crestal and 1 apical) were placed between second premolars and first molars in all 4 dental quadrants. Gray-level values were measured, and MO analysis was performed on each ROI. Descriptive statistics were calculated and correlations determined. Results. Paired correlations (such as apical vs crestal, left vs right, maxillary apical vs mandibular apical) of MO values were weak ( r = 0.01–0.21), but corresponding correlations for gray-level values were relatively strong ( r = 0.60–0.92). Conclusion. MO values varied with ROI location considerably more than did gray-level values. Additionally, ROI size and shape apparently affected MO data. Accurate placement and documentation of ROIs appear to be critical considerations in analyses that use MOs. (Oral Surg Oral Med Oral Pathol Oral Radiol Endod 1999;88:636-9)