|Year : 2014 | Volume
| Issue : 1 | Page : 13-18
Assessment of bone-fill following regenerative periodontal therapy by image subtraction using commercially available software
Pavan Kumar Yellarthi1, Viswa Chandra Rampalli2, Naveen Anumala2, Rama Raju Devaraju3
1 Department of Oral Medicine and Radiology, GITAM Dental College & Hospital, Visakhapatnam, Andhra Pradesh, India
2 Department of Periodontics, SVS Institute of Dental Sciences, Mahabubnagar, Telangana, India
3 Departments of Oral Medicine and Radiology, SVS Institute of Dental Sciences, Mahabubnagar, Telangana, India
|Date of Submission||12-Jun-2014|
|Date of Acceptance||24-Aug-2014|
|Date of Web Publication||26-Sep-2014|
Rama Raju Devaraju
Department of Oral Medicine and Radiology, SVS Institute of Dental Sciences, Mahabubnagar, Telangana
Source of Support: This study received no external support and was
funded by the authors’ institutions., Conflict of Interest: None
| Abstract|| |
Context: There are few applicable methods and scant literature on using non-specific image analysis software to assess and quantify post-periodontal surgery bone-fill. Aims: The present study describes the method of digital subtraction radiography (DSR) and morphometric area analysis (MA) in assessing bone-fill following regenerative periodontal therapy by using two image processing tools. Settings and Design: For the bone-fill assessment, radiographs of 78 angular bone defects from 30 subjects who underwent periodontal flap surgery with hydroxyapatite bone graft placement were utilized. Materials and Methods: The initial radiographic image obtained at baseline was subtracted from the radiographic images taken at 12/24 weeks by using the commercially available image processing software. After digital subtraction, the digitized and excluded interdental bone was transferred to open source software for area calculation. Bone-fill calculated by using DSR and MA was compared to the bone-fill values obtained by using the 1 mm 2 counting grid placed over the digital sensor. Statistical Analysis: Intragroup comparison of bone-fill between various groups was performed using repeated measures analysis of variance (ANOVA) followed by multiple comparisons using Bonferroni correction. One-way ANOVA followed by the post hoc test was used for intragroup and intergroup comparison. Results: DSR and MA showed higher mean bone-fill levels over radiographic grid. Conclusion: Considering that the software aims to reduce the inaccuracy behind manual assessment of bone-fill, the overestimation of bone-fill as compared to a grid may be attributed to the enhanced sensitivity of the method.
Keywords: Alveolar bone loss, dental radiography, digital subtraction radiography, periodontitis
|How to cite this article:|
Yellarthi PK, Rampalli VC, Anumala N, Devaraju RR. Assessment of bone-fill following regenerative periodontal therapy by image subtraction using commercially available software
. J Indian Acad Oral Med Radiol 2014;26:13-8
|How to cite this URL:|
Yellarthi PK, Rampalli VC, Anumala N, Devaraju RR. Assessment of bone-fill following regenerative periodontal therapy by image subtraction using commercially available software
. J Indian Acad Oral Med Radiol [serial online] 2014 [cited 2020 Jun 2];26:13-8. Available from: http://www.jiaomr.in/text.asp?2014/26/1/13/141826
| Introduction|| |
One of the important endpoints of regenerative periodontal therapy is new bone formation.  However, the identification and quantification of new bone formation within the treated area remains a great challenge to periodontists in general. , Periodontics has traditionally relied on conventional radiography to detect and assess periodontal osseous regeneration.  However, conventional radiography has its own drawbacks when used to assess bone-fill after periodontal therapy; errors in the assessment of bone gain may occur because of image quality, angular distortions or the use of non-standardized images. ,,, Minimal amount of new bone formation is difficult to appreciate by the naked eye and hence digital subtraction radiography (DSR) facilitates qualitative and quantitative visualization of even minor density changes (<5%) in bone by enhancing the changed osseous components in the image. ,
The use of appropriate software in DSR is paramount as it relies on comparing the pixel values of two images of the same region taken at different time intervals and then by subtracting the first image from the second.  The cost and complexity of the software , and the additional training required have been cited as barriers to the adoption of DSR in a dental office. Secondary to the careful alignment of the detector and beam, the ability to optimize an image and extract measurement functions is the foundation for a myriad of softwares, commercial and open source, medical and non-medical, expensive and free. ,, Only small differences in diagnostic accuracy are found in subtraction images from low and high-cost equipment, , thus aiding the practitioner in adopting a more open approach to utilize software supporting DSR. Apart from identifying sites with bone-fill, the quantitative analysis of sites with bone-fill is of particular challenge. In the absence of computed tomography (CT), DSR or histomorphometry, bone-fill is generally quantified as percentage bone-fill (PBF) or as change in bone level as compared to preset landmarks such as the CEJ, bone contour or the root surface.  Bone-fill is difficult to quantitate accurately by morphometry in conventional and digital radiographs owing to various radiographic errors that may creep in. ,, Even in radiographs obtained by digital radiography, the choice of bone graft and the bone defect morphology may impact the quantitative assessment of bone-fill.  Radiolucent bone grafts show an increase in density that correlates with de novo bone-fill.  This increase in density is sometimes imperceptible when a radiopaque bone graft is used leading to inaccurate assessment of bone-fill even in digitized radiographs. , In these situations, the enhanced sensitivity of image subtraction holds enormous potential in assessing and quantifying bone-fill to the practicing periodontist. 
Specific image analysis software tools are increasingly being used to evaluate post-surgical bone-fill;  yet for a variety of reasons, ,, there are few applicable methods and consequently scant literature on using non-specific image analysis software to assess and quantify post-periodontal surgery bone-fill. The present study describes the method of DSR and MA in assessing bone-fill following regenerative periodontal therapy by using two image processing software. The usefulness of this method was also evaluated by comparing values obtained by DSR-MA to those obtained by using a more 'conventional' external 1 mm counting grid placed over the radiographic sensor.
| Materials and Methods|| |
Prior to the study, approval from the institutional Review Board (No.: SVS/2012/3/004) was obtained and informed consent was taken from all the subjects. For the bone-fill assessment, radiographs of 78 angular bone defects from 30 subjects (mean age 40.27 ± 9.66), who underwent periodontal flap surgery with hydroxyapatite bone graft (G-graft® , Surgiwear, Shahjahanpur, India) placement were utilized. Bone morphology was confirmed by transgingival probing and by direct visualization during periodontal flap surgery.
Standard digital radiographs were taken at baseline (before surgery), 12, and 24 weeks by paralleling/long-cone technique at pre-set parameters using a commercially available radiovisiography (RVG) system (Kodak RVG 5100® Digital Radiography System, Carestream Health, Rochester, USA). After the imaging plate was placed in the sensor holder for paralleling technique (XCP Kits for Digital Sensors® , BlueDent, Chennai, India), addition silicon impression material (Elite HD + Regular Body Normal Set® , Zhermack, Badia Polesine, Italy) was added around the biting surface and allowed to set. This arrangement ensured standardized alignment of the aiming device and the holder ensuring correct positioning of the collimator in subsequent radiographic images. Duplicate radiographic images at all the time intervals were taken by using a metallic counting grid (IOPA film grid® , BlueDent, Chennai, India). The bone-fill was measured in square millimeters by point-counting the number of squares over newly developing radioopacities at different follow up periods. Squares that were half filled or unclear were not counted.
Digital subtraction technique (DSR)
The initial radiographic image obtained at baseline was subtracted from the radiographic images taken at 12 and 24 weeks by using commercially available image processing software (Adobe Photoshop® 6.0, Adobe Systems, San Jose, USA). To reduce brightness and contrast variations, both images were adjusted by using the image > adjust > auto levels and auto contrast tool. Before digital subtraction, both radiographs were moved in appropriate directions as needed [Figure 1], to reduce geometric distortion. These images were then superimposed and subtracted by selecting the image > calculation > exclusion > new channel tools [Figure 2]. The remnant interdental image layer [Figure 3] was outlined by using the magnetic lasso tool [Figure 4] and the layer was copied and saved as a separate joint photographic expert group (JPEG) document at low compression [Figure 5].
|Figure 1: The initial radiographic image obtained at baseline was subtracted from the radiographic images taken at 12 and 24 weeks, in Adobe Photoshop® 6.0|
Click here to view
|Figure 2: Excluding the graft area through the image > calculation > exclusion > new channel tools|
Click here to view
Morphometric area analysis (MA)
After digital subtraction, the digitized and excluded interdental layer was transferred to open source software for area calculation (ImageJ, Research Services Branch, NIH, Bethesda, Maryland, USA). The layer was converted into a grayscale image, and the measurement scale was set to account for any magnification/reduction of the radiograph because of RVG [Figure 6]. The area of the layer was calculated (in mm 2 ) by initially enclosing the entire area with the rectangular selection tool and then by using Analyze > Analyze Particles tool [Figure 7].
Bone-fill calculated by using DSR and MA was compared to the values obtained by using the 1 mm 2 counting grid placed over the digital sensor. Site-specific intragroup comparison of bone-fill between various groups was performed using repeated measures analysis of variance (ANOVA) followed by multiple comparisons using Bonferroni correction. One-way ANOVA followed by the post hoc test was used for intragroup and intergroup comparisons. A P-value of <0.05 was considered statistically significant and P-value of <0.001 was considered as highly significant.
| Results|| |
Of the 78 defects, 23 were combined defects, 16 were one-wall defects, 18 were two-wall defects and 21 were three-wall defects. No angulation, exposure technique, or software processing errors were encountered during the study.
Both the radiographic methods showed a highly significant increase (p ≤ 0.001) in mean bone-fill [Table 1] in the subtracted radiographs at 12 and 24 weeks. However, the mean bone-fill levels were higher in the DSR-MA group with mean differences ranging from - 0.03 ± 0.00 to 0.91 ± 0.54 depending on the defect evaluated [Table 2].
|Table 1: Intra-group defect-wise comparison of bone-fill at different time intervals using repeated measures ANOVA. All clinical values are in mm2|
Click here to view
|Table 2: Pair-wise bone-fill comparison between DSR-MA and grid at different time intervals using ANOVA|
Click here to view
These mean differences for both the radiographic methods from baseline to 12 weeks, to baseline to 24 weeks were statistically highly significant ( p ≤ 0.001) in combined defects and was significant at 12 weeks (p ≤ 0.05) and highly significant at 24 weeks in two-wall defects [Table 2]. The mean difference was not significant in one-wall and three-wall defects at both the time intervals.
| Discussion|| |
In this study, image subtraction was done in commercially available image analysis software (Adobe Photoshop® 6.0) by modifying a method previously described by Carvalho. , This software was used to effect image subtraction evaluation of the radiographs from peri-apical surgeries,  forensic odontology,  and implantology.  Two recent studies on bone-fill following periodontal therapy have used this software primarily for density estimation  and bone-fill estimation.  While the former used Adobe Photoshop® for DSR, volumetric, and densitometric estimations, the latter used a commercial tool for DSR and a free tool for morphometric analysis. While Adobe Photoshop® can be used for volumetric analysis, ,, we observed that the procedure was more complicated in underexposed radiographs. Because the bone-fill area would be picked up by the 'lasso' tool, additional adjustments to the contrast and sharpness were necessary which we felt was beyond the ability of the built-in functions of the software. Hence different software was used for morphometric analysis.
ImageJ is a freeware medical imaging software program released by the National Institute of Health (NIH) for the analysis of scientific images. Interestingly, several recent studies ,, have utilized this software for volumetric analysis of bone-fill after periodontal regenerative procedures. However, the analysis was done by measuring linear distances on pre-determined landmarks and using relavent formulae to determine the percentage bone-fill or the net volume. However, ImageJ supports automated area measurement of a complex object through the 'analyze particles' plug-in obviating the need for linear measurements and formulae.  The tool is easy to use and is not susceptible to noise in the outline of bone-fill. The authors have not come across any study that has partly or wholly utilized this plug-in to estimate bone-fill after periodontal surgery.
Bone-fill calculated by using DSR-MA was compared to the values obtained by using the external metallic grid placed over the digital sensor. Counting grids are routinely used to analyze periodontal bone-fill , and both the radiographic methods showed a highly significant increase in mean bone-fill with DSR-MA showing higher mean bone-fill levels over counting grids. This is in agreement with the analyses and observations of previous studies which have indicated that conventional radiographic interpretation is more subjective and may underestimate bone-fill when compared to subtraction radiography. ,, When compared defect-wise, the mean difference for both the radiographic methods was significant in combined and two-wall defects and non-significant in one-wall and three-wall defects at both the time intervals. Radiographically, combined, two- and three-wall defects are difficult to distinguish due to superimposition of the external cortical bone and tooth whereas one-wall infrabony defect is the most easily distinguishable.  In three-wall defects, there is a loss of cancellous bone and the accuracy of subtraction or conventional radiography in this situation is more or less similar.  This reason coupled with the easy identifiability of the single wall by subtraction and conventional radiography  might have contributed to the almost similar bone-fill measurement. Combined and two-wall defects present with a corticocancellous bone loss. The higher bone-fill estimation by DSR-MA can be explained by the fact that DSR primarily evaluates change by subtracting images in a buccolingual direction. 
However, this methodology has some limitations. The software, while exhibiting good sensitivity to bone-fill, , are not specific for digital subtraction and can be affected by the competency of the evaluator especially during the demarcation of the excluded bone-fill slice. DSR-MA has been compared to a more conventional method without making any assumption on which method is the gold standard. While counting grids are popular in periodontics, the use of one method as a gold standard to judge the relative merits of the other in this study can be misleading as the accuracy of current radiographic methods of determining architecture of osseous defects and bone-fill are still subject to debate. 
In conclusion, the methodology and the software described in the study seem to be effective in identifying areas of bone-fill after regenerative periodontal surgery. Considering that the software aims to reduce the inaccuracy behind manual assessment of bone-fill, the overestimation of bone-fill as compared to a counting grid may be attributed to the enhanced sensitivity of the method. For the proposed method to gain acceptance, further validation on the accuracy and reproducibility of the software for subtraction radiography may be necessary.
| Acknowledgments|| |
The authors would like to thank J Rahul for his help on the use of image processing software.
| References|| |
|1.||Guimarães Mdo C, Passanezi E, Sant'Ana AC, Grechi SL, Taba Junior M. Digital subtraction radiographic analysis of the combination of bioabsorbable membrane and bovine morphogenetic protein pool in human periodontal infrabony defects. J Appl Oral Sci 2010;18:379-84. |
|2.||Hwang YJ, Fien MJ, Lee SS, Kim TI, Seol YJ, Lee YM, et al. Effect of scaling and root planing on alveolar bone as measured by subtraction radiography. J Periodontol 2008;79:1663-9. |
|3.||Bae KH, Han JS, Seol YJ, Butz F, Caton J, Rhyu IC. The biologic stability of alumina-zirconia implant abutments after 1 year of clinical service: A digital subtraction radiographic evaluation. Int J Periodontics Restorative Dent 2008;28:137-43. |
|4.||Danesh-Meyer MJ, Chen ST, Rams TE. Digital subtraction radiographic analysis of GTR in human intrabony defects. Int J Periodontics Restorative Dent 2002;22:441-9. |
|5.||Eickholz P, Hausmann E. Evidence for healing of Class II and Class III furcations 24 months after guided tissue regeneration therapy: Digital subtraction and clinical measurements. J Periodontol 1999;70:1490-500. |
|6.||Mol A. Image processing tools for dental applications. Dent Clin North Am 2000;44:299-318. |
|7.||van der Stelt PF. Better imaging: The advantages of digital radiography. J Am Dent Assoc 2008;139(Suppl):7S-13S. |
|8.||Levin M. Digital imaging: A work multiplier. Endodontic Therapy 2005;5:26. |
|9.||Möystad A, Svanaes DB, Larheim TA. Personal computer equipment for dental digital subtraction radiography vs. industrial computer equipment and conventional radiography. Scand J Dent Res 1992;100:117-22. |
|10.||Dehdashti AR, Binaghi S, Uske A, Regli L. Comparison of multislice computerized tomography angiography and digital subtraction angiography in the postoperative evaluation of patients with clipped aneurysms. J Neurosurg 2006;104:395-403. |
|11.||Machtei EE. Outcome variables for the study of periodontal regeneration. Ann Periodontol 1997;2:229-39. |
|12.||Kumar PG, Kumar JA, Anumala N, Reddy KP, Avula H, Hussain SN. Volumetric analysis of intrabony defects in aggressive periodontitis patients following use of a novel composite alloplast: A pilot study. Quintessence Int 2011;42:375-84. |
|13.||Williams AL, Gornet MF, Burkus JK. CT evaluation of lumbar interbody fusion: Current concepts. AJNR Am J Neuroradiol 2005;26:2057-66. |
|14.||Sciadini MF, Johnson KD. Evaluation of recombinant human bone morphogenetic protein-2 as a bone-graft substitute in a canine segmental defect model. J Orthop Res 2000;18:289-302. |
|15.||Carvalho FB, Gonçalves M, Tanomaru-Filho M. Evaluation of chronic periapical lesions by digital subtraction radiography by using Adobe Photoshop CS: A technical report. J Endod 2007;33:493-7. |
|16.||Carvalho FB, Gonçalves M, Guerreiro-Tanomaru JM, Tanomaru-Filho M. Evaluation of periapical changes following endodontic therapy: Digital subtraction technique compared with computerized morphometric analysis. Dentomaxillofac Radiol 2009;38:438-44. |
|17.||Berketa J, James H, Marino V. A pilot study in the recovery and recognition of non-osseointegrated dental implants following cremation. J Forensic Odontostomatol 2011;29:38-44. |
|18.||Kwon HJ, Lee DW, Park KH, Kim CK, Moon IS. Influence of the tooth- and implant-side marginal bone level on the interproximal papilla dimension in a single implant with a microthread, conical seal, and platform-switched design. J Periodontol 2009;80:1541-7. |
|19.||Jenabian N, Haghanifar S, Maboudi A, Bijani A. Clinical and radiographic evaluation of Bio-Gen with biocollagen compared with Bio-Gen with connective tissue in the treatment of class II furcation defects: A randomized clinical trial. J Appl Oral Sci 2013;21:422-9. |
|20.||Singh VP, Nayak DG, Uppoor AS, Shah D. Clinical and radiographic evaluation of Nano-crystalline hydroxyapatite bone graft (Sybograf) in combination with bioresorbable collagen membrane (Periocol) in periodontal intrabony defects. Dent Res J (Isfahan) 2012;9:60-7. |
|21.||Mahantesha, Shobha KS, Mani R, Deshpande A, Seshan H, Kranti K. Clinical and radiographic evaluation of demineralized bone matrix (grafton) as a bone graft material in the treatment of human periodontal intraosseous defects. J Indian Soc Periodontol 2013;17:495-502. |
|22.||Sowmya NK, Tarun Kumar AB, Mehta DS. Clinical evaluation of regenerative potential of type I collagen membrane along with xenogenic bone graft in the treatment of periodontal intrabony defects assessed with surgical re-entry and radiographic linear and densitometric analysis. J Indian Soc Periodontol 2010;14:23-9. |
|23.||Bethesda: Research Services Branch, National Institute of Health. Available from: http://www.rsbweb.nih.gov/ij/docs/guide/user-guide-A4booklet.pdf. [Last accessed on 2014 Jan 2]. |
|24.||Cury PR, Araujo NS, Bowie J, Sallum EA, Jeffcoat MK. Comparison between subtraction radiography and conventional radiographic interpretation during long-term evaluation of periodontal therapy in Class II furcation defects. J Periodontol 2004;75:1145-9. |
|25.||Wenzel A, Warrer K, Karring T. Digital subtraction radiography in assessing bone changes in periodontal defects following guided tissue regeneration. J Clin Periodontol 1992;19:208-13. |
|26.||Peters TG. Detection of periodontal defects by conventional and subtraction radiography. Conn Dent Stud J 1987;7:23-4. |
|27.||Gomes-Filho IS, Sarmento VA, de Castro MS, da Costa NP, da Cruz SS, Trindade SC, et al. Radiographic features of periodontal bone defects: Evaluation of digitized images. Dentomaxillofac Radiol 2007;36:256-62. |
|28.||Kullendorff B, Gröndahl K, Rohlin M, Nilsson M. Subtraction radiography of interradicular bone lesions. Acta Odontol Scand 1992;50:259-67. |
|29.||Eickholz P, Hausmann E. Evidence for healing of periodontal defects 5 years after conventional and regenerative therapy: Digital subtraction and bone level measurements. J Clin Periodontol 2002;29:922-8. |
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7]
[Table 1], [Table 2]