Imaging Science in Dentistry 2020; 50: 31-6

https://doi.org/10.5624/isd.2020.50.1.31

Impact of radiotherapy on mandibular bone: A retrospective study of digital panoramic

radiographs

Luiz Felipe Palma 1,2,*, Ricardo Yudi Tateno 2

, Cíntia Maria Remondes 3
, Marcelo Marcucci ,
3

Arthur Rodriguez Gonzalez Cortes 2,4

1
Discipline of Descriptive and Topographic Anatomy, Department of Morphology and Genetics, Federal University of São Paulo, São Paulo, Brazil
2
MSc Dentistry Program, Ibirapuera University, São Paulo, Brazil
3
Stomatology and Oral and Maxillofacial Surgery Center, Heliópolis Hospital, São Paulo, Brazil
4
Department of Dental Surgery, Faculty of Dental Surgery, University of Malta, Msida, Malta

ABSTRACT

Purpose: The purpose of this study was to investigate the impact of radiotherapy on mandibular bone tissue in head
and neck cancer patients through an analysis of pixel intensity and fractal dimension values on digital panoramic
radiographs.
Materials and Methods: Thirty patients with radiographic records from before and after 3-dimensional (3D)
conformational radiotherapy were selected. A single examiner carried out digital analyses of pixel intensity values
and fractal dimensions, with the areas of interest unilaterally located in the right angle medullary region of the
mandible below the mandibular canal and posterior to the molar region.
Results: Statistically significant decreases were observed in the mean pixel intensity (P = 0.0368) and fractal dim­
ension (P = 0.0495) values after radiotherapy.
Conclusion: The results suggest that 3D conformational radiotherapy for head and neck cancer negatively affected
the trabecular microarchitecture and mandibular bone mass. (Imaging Sci Dent 2020; 50: 31-6)

KEY WORDS: ‌Radiotherapy; Panoramic Radiography; Mandible

mechanisms leading to osteoradionecrosis,3 but conclusive

Introduction evidence is not yet available. Histomorphometric analysis
The treatment of head and neck cancer is based on 3 mo- of biopsies has proven to be a valuable laboratory method
dalities, which can be combined or applied individually: to evaluate bone quality, while clinical methods include
radiotherapy, chemotherapy, and surgery. Radiotherapy, digital radiography, ultrasonography, and computed tomog-
in which ionizing radiation is used to eliminate or control raphy.4
the neoplasm, has been adopted as a standard treatment in Panoramic radiographs, especially digital panoramic ra-
the early stages of head and neck cancer. Advanced cases, diographs, are quick and easy to obtain, inexpensive, and
in contrast, commonly require chemotherapy and/or sur- comfortable for the patient, and they typically involve the
gery.1,2 administration of low doses of radiation. In addition, dig-
The deleterious effects of radiation on bone tissue have ital radiography has been considered a useful tool for the
been closely examined with the goal of elucidating the determination of systemic bone quality through the assess-
ment of mandibular bone mass and morphology.5 Despite
conflicting results reported in the literature and limitations
Received October 31, 2019; Revised December 31, 2019; Accepted January 22, 2020
*Correspondence to : Dr. Luiz Felipe Palma related to technical aspects and the reproducibility of the
MSc Dentistry Program, Ibirapuera University, Av. Interlagos, 1329 - Chácara Flora, method, digital panoramic radiography has been described
São Paulo, SP, 04661-100, Brazil
Tel) 55-11-99-100-8038, E-mail) luizfelipep@hotmail.com as capable of predicting osteopenia and osteoporosis after

Copyright ⓒ 2020 by Korean Academy of Oral and Maxillofacial Radiology

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0)
which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Imaging Science in Dentistry·pISSN 2233-7822 eISSN 2233-7830

- 31 -
Impact of radiotherapy on mandibular bone: A retrospective study of digital panoramic radiographs

correlating those conditions with systemic bone mineral lar pathologies (e.g., condensing osteitis, bone sclerosis,
density values obtained through dual-energy X-ray absorp- fibrous dysplasia, periapical lesions, or pathological frac-
tiometry.6 tures) in the areas of interest.
A digital image is composed of an array of small areas This study was approved by the Research Ethics Commit-
known as pixels, to which specific numeric values are tee at Ibirapuera University (CAAE 91440318.3.0000.5597)
assigned.5 Pixel intensity analysis of a digital panoramic and Heliópolis Hospital (CAEE 91440318.3.3001.5449)
radiograph can be based on a grayscale that ranges from and was in compliance with the Declaration of Helsinki.
0 (black) to 255 (white),7 in which pixel intensity is consid- Participants’ privacy and confidentiality of information were
ered an indirect measurement of local bone mass5. Fractal ensured in all phases of the study.
analysis is a mathematical model that describes and ana-
lyzes complex forms and structural patterns, numerically Digital panoramic radiography
expressed as fractal dimensions.5 Although several methods Digital panoramic radiographs were obtained using a
of fractal analysis have been described,8 the box-counting PaX-400C device (Vatech, Hwaseong, Korea) with an ex-
technique has been the most widely applied and is suited posure time of 14 s and an electric current of 8 mA. As
for binary images9 such as radiographs. This method can suggested by the manufacturer, male and female subjects
reveal complex interconnections of the trabecular bone, received peak kilovoltages of 68 kVp and 67 kVp, respec-
consequently indicating bone quality.10 tively. The images were stored digitally as 8-bit grayscale
Thus, the aim of the present study was to evaluate the bitmap (.bmp) files with a resolution of 2900 × 1532 pixels.
impact of radiotherapy on mandibular bone tissue in head
and neck cancer patients through an analysis of pixel in- Pixel intensity and fractal dimension analyses
tensity and fractal dimension values on digital panoramic The pixel intensity and fractal dimension analyses were
radiographs. adapted from the original studies by White and Rudolph11
and Tosoni et al.5 and were conducted using the image
processing software ImageJ 1.52a (National Institutes of
Materials and Methods Health, Bethesda, MD, USA).
Sample characteristics, eligibility criteria, and Initially, a fixed selection tool was used (100 × 100 pix-
ethical issues els) to determine the region of interest (ROI) on each ra-
This retrospective study was conducted using digital diograph. This region was located unilaterally in the right
panoramic radiographs from the electronic collection of angle region of the mandible (below the mandibular canal
the Stomatology and Oral and Maxillofacial Surgery Cen- and posterior to the molar region), parallel to the posterior
ter at Heliópolis Hospital, obtained from patients between margin of the mandibular ramus, and included only trabec-
2015 and 2018. To be considered eligible, individuals were ular bone (Fig. 1).
required to be at least 18 years of age, to have a history of The mean pixel intensity value of each ROI was deter-
3-dimensional (3D) conformational radiotherapy for head mined using the software histogram, and fractal dimension
and neck cancer (involving the cervicofacial and supracla- analysis was then performed. For this purpose, the image
vicular fossa fields), and to have panoramic radiographic of the ROI was duplicated. This copy of the image was
records available from before and after radiotherapy. As a blurred using a Gaussian filter with a 35-pixel radius to re-
local protocol, all initial radiographic images were taken tain only large density variations and was subtracted from
during the oral interventions administered in the prepara- the original image, generating a third image. A constant
tion for oncologic treatment, ranging from 1 to 3 weeks of 128 (corresponding to the half-value of the gray levels
prior to radiotherapy. on the 8-bit scale) was added at each pixel position on the
The following exclusion criteria were applied: systemic third image to ensure that variations in brightness corre-
diseases known to alter bone metabolism (e.g., osteopo- sponding to specific features (trabeculae and medullary
rosis, thyrotoxicosis, or hyperparathyroidism), use of any spaces) were visible. The image was then converted into a
type of medication with known effects on bone metabolism binary image at a brightness threshold value of 128 to fa-
(e.g., corticosteroids, antiangiogenic agents, or antiresorp- cilitate image segmentation into components that visually
tive agents such as bisphosphonates or denosumab), poor resembled trabeculae and medullary spaces. The image
positioning during the radiographic examination, poor- was then subjected to erosion and dilation to reduce noise
quality images, and the presence of extensive mandibu- and was inverted to reveal the trabeculae. Next, the image

- 32 -
 Luiz Felipe Palma et al

Fig. 1. Region of interest selection.

(A) (B) (C) (D) (E)

(F) (G) (H) (I)

Fig. 2. Fractal dimension analysis process. A. Region of interest. B. Result of blurring of (A). C. Result of subtracting (B) from (A). D. Re-
sult of adding 128 to each pixel position of (C). E. Result of conversion of (D) into a binary image. F. Result of erosion and dilation of (E). G.
Result of inversion of (F). H. Result of skeletonizing of (G). I. Result of superpositioning (H) over (A).

was skeletonized, or eroded until only the center pixel line to reduce memory bias, and evaluated using the intraclass
remained, and superimposed over the original image to correlation coefficient (ICC). Another trained observer car-
certify that the structures corresponded to the trabeculae. ried out analyses of 10% of the sample, and inter-examiner
Finally, the box-counting method was applied with box siz- reproducibility was determined using the ICC as well.
es of 2, 3, 4, 6, 8, 12, 16, 32, and 64 pixels to determine the The characteristics of the subjects and their oncological
fractal dimension (Fig. 2). treatment were analyzed descriptively. The data distribution
was assessed using the Shapiro-Wilk test, and the pixel in-
Statistical analysis tensity and fractal dimension values before and after radio-
The analyses were conducted by a trained observer, who therapy were analyzed using the Wilcoxon sign-rank test.
was a dentist with expertise in oral radiology. Intra-exam- The statistical analysis was carried out using BioEstat 5.3
iner reproducibility was assessed by reanalysis of 10% of software (Instituto de Desenvolvimento Sustentável Ma-
the sample, performed 2 weeks after the initial analysis mirauá, Amazonas, Brazil) with a 5% significance level.

- 33 -
Impact of radiotherapy on mandibular bone: A retrospective study of digital panoramic radiographs

Table 1. Demographic and clinical characteristics of the participants

Patients
Characteristic
n %

Sex Male 24 80%

Female 6 20%
Primary tumor location Pharynx 13 43.3%
Oral cavity 9 30%
Larynx 4 13.3%
Parotid gland 3 10%
Occult primary 1 3.3%
Total dose of radiation ≤59 Gy 1 3.3%
60 to 69 Gy 9 30%
70 Gy 20 66.7%
Time period between pre- 3 to 12 months 12 40%
and post-radiotherapy radiographs 12 to 24 months 13 43.3%
24 to 35 months 5 16.7%

Table 2. Mean pixel intensity and fractal dimension values on pre- and post-radiotherapy radiographs in head and neck cancer patients

Variable Pre-radiotherapy Post-radiotherapy P value

Pixel intensity Mean 132.5±31.3 114.0±32.3 P<0.05

Maximum 182.1 170.1
Minimum 70.2 59.2
Fractal dimension Mean 1.4±0.1 1.3±0.1 P<0.05
Maximum 1.5 1.5
Minimum 1.2 1.2

tively, constituting a statistically significant difference

Results (P = 0.0368). The mean fractal dimension values pre- and
Thirty individuals were selected for the study, resulting post-radiotherapy were 1.4±0.1 and 1.3±0.1, respective-
in a total of 60 digital panoramic radiographs. Most of ly, similarly presenting a statistically significant difference
these individuals were male (80%), had a primary tumor (P = 0.0495). Detailed information on the analyses is pro-
located in the pharynx (43.4%), and had received a total ra- vided in Table 2.
diation dose of 70 Gy (66.7%). The most common interval
separating the pre- and post-radiotherapy radiographs was
12 to 24 months (43.4%). More detailed information on the
Discussion
sample is provided in Table 1. Bone quality is determined by the tissue microarchitec-
The normal distribution was rejected for the pixel in- ture and the degree of mineralization. Low bone quality has
tensity data before (P = 0.088) and after radiotherapy been directly associated with the failure of dental implants,
(P = 0.3081) and for the fractal dimension data after radio- since primary stability and osseointegration are influenced
therapy (P = 0.3851). Intra- and extra-examiner reproduc- to a large extent by the characteristics of the surrounding
ibility were confirmed for both pixel intensity (ICC, 0.89; bone.4 Evidence has shown that patients with osteoporosis
P = 0.001) and fractal dimension (ICC, 0.81; P = 0.01) experience more rapid development and greater severity of
evaluations. periodontal disease, as well as a lower accumulated long-
The mean values of pixel intensity before and after ra- term survival rate of dental implants12 and a higher risk of
diotherapy were 132.5±31.3 and 114.0±32.3, respec- fractures, than individuals without osteoporosis.13

- 34 -
 Luiz Felipe Palma et al

The present study investigated the characteristics of atomical structures, such as the airways, tonsils, soft palate,
mandibular bone tissue on digital panoramic radiographs, tongue, and hyoid bone.5 To date, ROI dimensions have
which are commonly used imaging exams in dental prac- not been standardized. The protocol proposed by Tosoni et
tice. Scientific evidence supports the use of digital pan- al.5 (100 × 100 pixels) was satisfactory for all individuals
oramic radiography for the analysis of bone tissue, despite examined in our study and provided a suitable medullary
limitations inherent to the technique concerning tissue bone area for the analyses.
properties, such as software and image receptor configura- Radiotherapy is an important and efficient therapeutic
tion, parameters of the radiographic device (such as peak modality for head and neck cancer, although it also affects
kilovoltage, exposure time, and current), and overlap of non-neoplastic cells in or adjacent to the irradiated area.
tissues and structures.6 Dual-energy X-ray absorptiometry Its effects may occur during or immediately after irradi-
is a costly gold-standard exam for the diagnosis of osteo- ation (acute effects) or even months or years later (late
porosis. Panoramic radiographs can detect a reduction in effects).17,18 Our results demonstrated that 3D conforma-
mandibular mineral density at a lower cost, which allows tional radiotherapy significantly affected subjects’ trabec-
early systemic diagnosis and prompt referral to a special- ular microarchitecture and the mandibular bone mass. One
ized professional.14 hypothesis to explain this phenomenon is that intraosseous
A pixel is a single digital point with a variable intensity, vascularization decreases as a result of exposure to ionizing
constituting the smallest unit of an image that can be rep- radiation, with changes in the bone cell pool, progressive
resented or controlled. In a grayscale radiographic image, fibrosis, and local necrosis.3 These alterations manifest ra-
the value of each pixel provides information about inten- diologically as low bone mineral density, trabecular frac-
sity alone.15 For pixel intensity analysis for the indirect ture, and loss and destruction of cortical structures.4 Nev-
measurement of local bone mass, some authorities have ertheless, the various studies and models addressing this
suggested methods to normalize variations in noise and subject have presented highly heterogeneous methods and
image production that could influence the comparison be- samples. It is therefore very difficult to realistically com-
tween radiographs obtained with the same devices and pa- pare them and apply the results to humans.19
rameters, with these methods including density markers,5 To the best of our knowledge, this is the first study to
binarization, and skeletonization.15 Herein, the direct use of investigate the impact of radiotherapy on the mandibular
histograms in ImageJ seemed to be satisfactory, making the bone tissue of head and neck cancer patients using pixel in-
procedures even quicker and simpler. tensity and fractal dimension analyses of digital panoram-
Analysis of the fractal dimension is a low-cost, easily ap- ic radiographs. As limitations and suggestions for future
plicable, non-invasive, and accurate method for quantifying research, some points need to be taken into account. First,
the interface pattern of trabecular bone and bone marrow, further studies should consider using more representative
with higher values indicating a more complex structure.10 sample sizes. In addition, accurate data on the total radi-
It is a technique used to identify scale-invariant structures ation dose received by each patient in the ROI should be
that are not affected by the conditions under which radio- provided to determine the association between the amount
graphic images are produced.8,9 However, a range of dental of ionizing radiation received and the level of bone tissue
radiographic processing methods have been developed as damage. Furthermore, because this was a retrospective
adaptations of the original article by White and Rudolph,11 study, the interval between the pre- and post-radiotherapy
which was used as a reference in our study. radiographs was not standardized and ranged from 3 to 35
Another conflicting point in the literature relates to the months. Lastly, analyses of other mandibular areas, bilat-
mandibular site used to determine the ROI and its size. The eral assessment by more than 1 experienced examiner, and
anterior region of the mandible seems to produce relatively the use of other radiographic techniques (e.g., periapical
inaccurate images due to the overlap of the cervical spine radiography) could provide more accurate and detailed in-
and relatively high likelihood of blurring.16 In our study, formation.
we used the angle of the mandible below the mandibular In summary, based on the results of pixel intensity and
canal and posterior to the molar region to avoid interfer- fractal dimension analyses, this study suggests that 3D
ence caused by masticatory stress, as proposed by Oliveira conformational radiotherapy for head and neck cancer neg-
et al.,15 and to provide bone structure even in fully edentu- atively affected the trabecular microarchitecture and man-
lous aged individuals with extremely atrophic mandibles. dibular bone mass.
However, this area is subject to the overlapping of other an-

- 35 -
Impact of radiotherapy on mandibular bone: A retrospective study of digital panoramic radiographs

Conflicts of Interest: None quency values at implant recipient sites. Int J Oral Maxillofac
Implants 2016; 31: 55-62.
10. de Molon RS, de Paula WN, Spin-Neto R, Verzola MH, Tosoni
GM, Lia RC, et al. Correlation of fractal dimension with histo-
References
morphometry in maxillary sinus lifting using autogenous bone
1. Palma LF, Gonnelli FA, Marcucci M, Dias RS, Giordani AJ, graft. Braz Dent J 2015; 26: 11-8.
Segreto RA, et al. Impact of low-level laser therapy on hy- 11. White SC, Rudolph DJ. Alterations of the trabecular pattern of
posalivation, salivary pH, and quality of life in head and neck the jaws in patients with osteoporosis. Oral Surg Oral Med Oral
cancer patients post-radiotherapy. Lasers Med Sci 2017; 32: Pathol Oral Radiol Endod 1999; 88: 628-35.
827-32. 12. Chambrone L. Current status of the influence of osteoporosis
2. Palma LF, Gonnelli FA, Marcucci M, Giordani AJ, Dias RS, on periodontology and implant dentistry. Curr Opin Endocri-
Segreto RA, et al. A novel method to evaluate salivary flow nol Diabetes Obes 2016; 23: 435-9.
rates of head and neck cancer patients after radiotherapy: a pilot 13. Dervis E. Oral implications of osteoporosis. Oral Surg Oral
study. Braz J Otorhinolaryngol 2018; 84: 227-31. Med Oral Pathol Oral Radiol Endod 2005; 100: 349-56.
3. Michel G, Blery P, Pilet P, Guicheux J, Weiss P, Malard O, 14. Neves FS, Barros AS, Cerqueira GA, Cruz GA, Reis AA, Alves
et al. Micro-CT analysis of radiation-induced osteopenia and LB, et al. Assessment of fractal dimension and panoramic ra-
bone hypovascularization in rat. Calcif Tissue Int 2015; 97: diomorphometric indices in women with celiac disease. Oral
62-8. Radiol (in press).
4. Konermann A, Appel T, Wenghoefer M, Sirokay S, Dirk C, 15. Oliveira ML, Pedrosa EF, Cruz AD, Haiter-Neto F, Paula FJ,
Jäger A, et al. Impact of radiation history, gender and age on Watanabe PC. Relationship between bone mineral density and
bone quality in sites for orthodontic skeletal anchorage device trabecular bone pattern in postmenopausal osteoporotic Bra-
placement. Ann Anat 2015; 199: 67-72. zilian women. Clin Oral Investig 2013; 17: 1847-53.
5. Tosoni GM, Lurie AG, Cowan AE, Burleson JA. Pixel inten- 16. Hwang JJ, Lee JH, Han SS, Kim YH, Jeong HG, Choi YJ, et
sity and fractal analyses: detecting osteoporosis in perimeno- al. Strut analysis for osteoporosis detection model using den-
pausal and postmenopausal women by using digital panoram- tal panoramic radiography. Dentomaxillofac Radiol 2017; 46:
ic images. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 20170006.
2006; 102: 235-41. 17. Gonnelli FA, Palma LF, Giordani AJ, Deboni AL, Dias RS,
6. Camargo AJ, Cortes AR, Aoki EM, Baladi MG, Arita ES, Segreto RA, et al. Low-level laser therapy for the prevention
Watanabe PC. Diagnostic performance of fractal dimension of low salivary flow rate after radiotherapy and chemotherapy
and radiomorphometric indices from digital panoramic radio- in patients with head and neck cancer. Radiol Bras 2016; 49:
graphs for screening low bone mineral density. Braz J Oral 86-91.
Sci 2016; 15: 131-6. 18. Gonnelli FA, Palma LF, Giordani AJ, Deboni AL, Dias RS,
7. Halazonetis DJ. What do 8-bit and 12-bit grayscale mean and Segreto RA, et al. Low-level laser for mitigation of low sali-
which should I use when scanning? Am J Orthod Dentofacial vary flow rate in head and neck cancer patients undergoing ra-
Orthop 2005; 127: 387-8. diochemotherapy: a prospective longitudinal study. Photomed
8. Amer ME, Heo MS, Brooks SL, Benavides E. Anatomical vari- Laser Surg 2016; 34: 326-30.
ations of trabecular bone structure in intraoral radiographs using 19. Jegoux F, Malard O, Goyenvalle E, Aguado E, Daculsi G. Ra-
fractal and particles count analyses. Imaging Sci Dent 2012; 42: diation effects on bone healing and reconstruction: interpre-
5-12. tation of the literature. Oral Surg Oral Med Oral Pathol Oral
9. Suer BT, Yaman Z, Buyuksarac B. Correlation of fractal dimen- Radiol Endod 2010; 109: 173-84.
sion values with implant insertion torque and resonance fre-

- 36 -