Hostname: page-component-74656f9884-rngvg Total loading time: 0 Render date: 2026-11-05T11:29:59.915Z Has data issue: false hasContentIssue false
  • English
  • Français

Racial differences in bone density variation in head and neck cancer patients with osteoradionecrosis: A pilot study

Published online by Cambridge University Press:  14 July 2026

Srivatsa Surya Vasudevan Open the ORCID record for Srivatsa Surya Vasudevan [Opens in a new window] ,
Abhijit Gundale ,
Eleni Mijalis ,
Amol Takalkar ,
Kent Burton ,
Tara Moore-Medlin ,
Jerry McLarty  and
Cherie-Ann O. Nathan
Srivatsa Surya Vasudevan
Affiliation:
Otolaryngology - Head and Neck Surgery, Louisiana State University Health Sciences Center Shreveport, USA
Abhijit Gundale
Affiliation:
Otolaryngology - Head and Neck Surgery, University of Florida College of Medicine - Jacksonville, USA
Eleni Mijalis
Affiliation:
Otolaryngology - Head and Neck Surgery, The University of Mississippi Medical Center , USA
Amol Takalkar
Affiliation:
Department of Radiology and Imaging Sciences, Emory University Hospital, USA
Kent Burton
Affiliation:
Otolaryngology - Head and Neck Surgery, Louisiana State University Health Sciences Center Shreveport, USA
Tara Moore-Medlin
Affiliation:
Otolaryngology - Head and Neck Surgery, Louisiana State University Health Sciences Center Shreveport, USA
Jerry McLarty
Affiliation:
Otolaryngology - Head and Neck Surgery, Feist-Weiller Cancer Center, USA
Cherie-Ann O. Nathan*
Affiliation:
Otolaryngology - Head and Neck Surgery, Louisiana State University Health Sciences Center Shreveport, USA
*
Corresponding author: Cherie-Ann O. Nathan; Email: cherieann.nathan@lsuhs.edu
Rights & Permissions [Opens in a new window]

Abstract

Objectives

This study aimed to assess pre- and post-radiation therapy mandibular bone densities in head and neck cancer patients with osteoradionecrosis and factors associated with decreased bone densities.

Methods

Retrospective pilot study (2007–2017) of osteoradionecrosis patients after 66- to 70-gray radiation therapy. Cortical and cancellous mandibular bone densities ipsilateral/contralateral to tumour were measured on pre-radiation therapy and at three-month and one-year post-radiation therapy positron emission tomography/computed tomography. Two-way analysis of variance and linear regression were used as exploratory analyses, given the pilot nature of the study.

Results

Sixteen of 67 head and neck cancer patients had osteoradionecrosis (41–67 years of age; nine African Americans, seven Caucasians). Bone density did not significantly change from baseline within either racial group. African Americans had higher bone densities than Caucasians (pre-radiation therapy ipsilateral cortical: 1044 ± 229.8 vs. 965.6± 278.7 Hounsfield units). At one year, African Americans had higher contralateral cortical (+35.5 per cent; p = 0.008) and cancellous (+109.4 per cent; p = 0.028) bone densities. Multivariable models showed lower contralateral cortical (p = 0.039) and cancellous (p = 0.005) bone densities in Caucasians.

Conclusion

In this small pilot study, mandibular bone densities appeared higher in African Americans, with a greater post-radiation therapy reduction observed in Caucasians; these findings are exploratory and require validation in larger cohorts.

Keywords

bone densityhead and neck cancerosteonecrosisosteoradionecrosisradiotherapy

Information

Type
Main Article
Information
The Journal of Laryngology & Otology , First View , pp. 1 - 9
Copyright
© The Author(s), 2026. Published by Cambridge University Press on behalf of J.L.O. (1984) LIMITED.

Introduction

Approximately 50–60 per cent of head and neck cancer patients receive radiation therapy (RT). The current management of these malignancies relies on RT for advanced-stage disease and organ preservation strategies.Reference Rivero, Shamji and Kolokythas1 Osteoradionecrosis (ORN) of the mandible is one of the late and devastating side effects of head and neck RT.Reference Singh, Kitpanit, Neal, Yorke, White and Yom2, Reference Vasudevan, Candelo, Sharifi, Ma, Patel and Routman3 Although the underlying cancer is eradicated, ORN causes long-term pain and may require additional surgeries with associated morbidity.Reference Chronopoulos, Zarra, Ehrenfeld and Otto4 Medication-induced osteonecrosis of the jaw, specifically due to bisphosphonate use, has been primarily linked to the anti-angiogenic and bone-resorption–inhibiting properties of bisphosphonates.Reference Kün-Darbois and Fauvel5 Furthermore, the overall likelihood of developing osteonecrosis due to bisphosphonates rises with more frequent dosing, larger individual doses and prolonged treatment duration.Reference Otto, Pautke, Van den Wyngaert, Niepel and Schiødt6

Several other risk factors associated with the development of ORN include site and size of tumour, dentition status, the dosage and fractionation of RT, smoking and trauma.Reference Rivero, Shamji and Kolokythas1, Reference Chronopoulos, Zarra, Ehrenfeld and Otto4 Mechanistically, ORN is believed to be due to radiation-induced disruption of intra-osseous arteries.Reference Kün-Darbois and Fauvel5 Some studies have suggested that radiation may trigger a cascade of tissue changes marked by reduced cellularity, impaired vascularity and hypoxia that disrupt normal mucosal healing and ultimately lead to chronic non-healing wounds and osteonecrosis.Reference Marx7 Our institutional data over 10 years revealed that of 2,323 patients diagnosed with jaw osteonecrosis, 65 per cent were Caucasian, suggesting a higher burden of disease in Caucasian patients compared to African-American patients at our institution. Similarly, Quispe et al. reported higher osteonecrosis rates in Caucasian compared to African Americans.Reference Quispe, Shi and Burton8 Regarding overall bone density (BD) differences between races, the National Health and Nutrition Examination Survey (NHANES) data demonstrated that African Americans exhibit femoral neck and total-site BD 6–11 per cent higher than Caucasians.Reference Noel, Santos and Wright9, Reference HM, Horowitz, Morley, Fleming, Jensen and Caaccione10 Furthermore, the trajectory of age-related bone loss in African Americans appears markedly attenuated, with annual hip BD decline roughly half that observed in Caucasians.Reference Noel, Santos and Wright9 Therefore, we wanted to see if this racial difference in ORN prevalence could also be associated with differences in BD loss with radiation. Our study aimed to compare BDs of cancellous and cortical mandibular bone between African-American and Caucasian ORN patients in pre- and post-RT scenarios.

Materials and methods

Study design and population

This pilot study is an institutional review board (IRB) -approved retrospective chart review of ORN cases at our institution, limited to head and neck cancer patients treated with radiotherapy who developed ORN within five years and had sufficient positron emission/computed tomography (PET/CT) imaging available for analysis from 2007 to 2017.

Data collection

Data collected included patient demographics (age, gender and race/ethnicity), appropriate medical history (details about head and neck cancer subsite, treatment and follow-up) and relevant medical history (xerostomia, dental extractions and smoking).

Exposure and covariates

Imaging studies, including pre- and post-RT PET/CT scans, were performed. BDs were assessed by measuring Hounsfield units (HU) in predetermined areas of the mandible on the pre- and one-year post-RT PET/CT scans. Our method of using CT-derived HU as a validated surrogate measure of bone mineral density is well established in the literature, including for mandibular and craniofacial bone assessment in both implant and oncologic contexts.Reference Park, Lee, Jeong and Kwon11, Reference Shapurian, Damoulis, Reiser, Griffin and Rand12 The CT portion of each PET/CT scan was displayed in the bone window, and the BDs of the retromolar trigone and canine regions were measured bilaterally, both ipsilateral and contralateral to the tumour (Fig. 1). All PET/CT scans were acquired at a single institution using a consistent scanner platform and standardised acquisition and reconstruction protocols throughout the study period. BD measurements were performed by a single experienced radiologist to minimise inter-observer variability. The retromolar trigone and canine areas were chosen because these sites contain the densest cortical and cancellous bone, respectively.Reference Park, Lee, Jeong and Kwon11 BD of cortical and cancellous mandibular bone was compared between pre- and post-RT scans separately for the ipsilateral and contralateral sides of the tumour in both African-American and Caucasian patients. These cortical and cancellous BD measurements were performed at three different time points: (1) prior to initiation of RT, (2) three months following RT and (3) one year following RT. Other factors, including xerostomia, dental extractions, smoking and RT dosage for each patient, were also evaluated along with BD.

Figure 1.

Axial CT images showing the regions of interest used for mandibular bone density measurements. (A) The canine region (cancellous bone density measurement sites). (B) The retromolar trigone region (cortical bone density measurement sites), bilaterally ipsilateral and contralateral to the tumour. These standardised sites were selected as they represent the regions of maximal cancellous and cortical bone density in the mandible, respectively. CT = computed tomography.

Two axial CT images showing mandibular regions for bone density measurements.

Statistical analysis

Results were statistically analysed by utilising Graphpad PRISM software version 10.5.0. Descriptive statistics were calculated by computing the mean, median and range for continuous variables, while categorical variables were summarised using frequencies and percentages. Prior to two-way analysis of variance (ANOVA), normality was assessed using the Shapiro–Wilk test and homogeneity of variance was confirmed using Bartlett’s test. Two-way ANOVA was employed to analyse BD, with race (African American vs. Caucasian) and time (pre-RT, three months post-RT and one year post-RT) specified as factors, thereby enabling assessment of intraracial changes over time (pre-RT vs. three months vs. one year within each race) and inter-racial differences at corresponding time points (African American vs. Caucasian at pre-RT, three months post-RT and one year post-RT). Non-parametric correlations were used to test for binary correlations between variables. Univariate and multivariate linear regression analysis were used to analyse the factors associated with BD variations, such as age, race and others.

Results

Patient characteristics

Out of 67 identified ORN cases, 16 were ultimately analysed due to limited pre-RT and post-RT scan availability. The mean age of these patients was 57.75 (SD = 6.59) years, ranging from 41 to 67 years, and there were 15 males and one female. Of the 16 patients, 10 (62.5 per cent) had tonsillar cancer, five (31.3 per cent) had base of tongue cancer and one (6.2 per cent) had soft palate cancer. Nine patients were African Americans, and seven were Caucasians. All patients in both groups had Stage IV squamous cell carcinoma and received bilateral radiotherapy at a minimum dose of 66 Gy. Five of the African-American and five of the Caucasian patients underwent surgical management (Table 1). Twelve out of 16 were smokers, five suffered from xerostomia, one was diabetic and two underwent post-RT dental extraction. Mean age was similar between African-American and Caucasian patients (56.7 vs. 59.1 years; p = 0.67). Both groups were predominantly male (89 per cent vs. 100 per cent; p = 0.99). Smoking rates were comparable (78 per cent vs. 71 per cent; p = 0.99), as were rates of xerostomia (33 per cent vs. 29 per cent; p = 1.000), dental extraction (11 per cent vs. 14 per cent; p = 0.99) and diabetes (11 per cent vs. 0 per cent; p = 0.99) (Table 2).

Table 1.

Patient distribution and demographics

16 stage IV squamous cell carcinoma patients, mostly male smokers, had bilateral radiotherapy; osteoradionecrosis was managed mainly conservatively, surgically, or via debridement. See long description.
Table 1 Long description

Patient level data are listed for 16 patients, including age in years, gender, race, tumor site, TNM, stage, treatment, sides of radiation therapy, smoking, xero stomia, DM, dental extraction, and treatment of ORN. Ages range from 41 to 67 years; 15 patients are male and 1 is female. Race is African American for 10 patients and Caucasian for 6 patients. Sites include tonsil in 10 patients, base of tongue in 5 patients, and soft palate in 1 patient. TNM entries span T1 to T4b with nodal categories N1, N2a, N2b, and N2c, and all patients have M0; stage is listed as squamous cell carcinoma stage four for every patient. Treatment is chemoradiation for 6 patients, surgery plus radiation therapy for 5 patients, and surgery plus chemoradiation for 5 patients; sides of radiation therapy are bilateral for all patients. Smoking is yes for 12 patients and no for 4 patients; xero stomia is yes for 5 patients and no for 11 patients; DM is yes for 1 patient and no for 15 patients; dental extraction is yes for 2 patients and no for 14 patients. Treatment of ORN is conservative for 8 patients, surgery for 4 patients, and debridement for 4 patients.

AA = African American; BOT = base of tongue; C = Caucasian; CRT = chemoradiation; F = female; M = male; N = no; ORN = osteoradionecrosis; RT = radiation therapy; SCC = squamous cell carcinoma; Y = yes.

Table 2.

Baseline demographic and clinical characteristics by race

Both groups were entirely stage four and received bilateral radiotherapy with dose at least 66 gray; other baseline measures were similar between African American and Caucasian patients. See long description.
Table 2 Long description

Baseline demographic and clinical variables are listed for African American patients, n 9, and Caucasian patients, n 7, with a p value for each row. Mean age was 56.7 years with a range of 41 to 67 for African American patients and 59.1 years with a range of 55 to 65 for Caucasian patients, p value 0.670 with footnote marker a. Gender counts were 8 male and 1 female for African American patients and 7 male and 0 female for Caucasian patients, p value 0.99 with footnote marker b. Smoking was 7 patients, 78 percent, in African American patients and 5 patients, 71 percent, in Caucasian patients, p value 0.99 with footnote marker b; xerostomia was 3 patients, 33 percent, and 2 patients, 29 percent, p value 0.99 with footnote marker b. Diabetes was 1 patient, 11 percent, in African American patients and 0 patients, 0 percent, in Caucasian patients, p value 0.99 with footnote marker b; dental extraction was 1 patient, 11 percent, and 1 patient, 14 percent, p value 0.99 with footnote marker b. Stage four was 9 patients, 100 percent, and 7 patients, 100 percent, with p value listed as not applicable; bilateral radiotherapy was 9 patients, 100 percent, and 7 patients, 100 percent, with p value listed as not applicable; radiotherapy dose greater than or equal to 66 gray was 9 patients, 100 percent, and 7 patients, 100 percent, with p value listed as not applicable. Surgical osteoradionecrosis management was 5 patients, 56 percent, in African American patients and 5 patients, 71 percent, in Caucasian patients, p value 0.633 with footnote marker b.

a Mann–Whitney U-test.

b Fisher’s exact test. No statistically significant between-group differences were identified for any baseline variable (all p > 0.05).

F = female; M = male; NA = not applicable; ORN = osteoradionecrosis; RT = radiotherapy.

Pre- vs. post-RT mandibular BD analysis for Caucasians

In Caucasian patients, mandibular BD remained statistically stable from pre-RT through the one-year post-RT follow-up. On the ipsilateral side of the tumour, neither cortical nor cancellous BD changed significantly across any pairwise comparison, pre-RT versus three months, three months versus one year or pre-RT versus one year (all p > 0.05; Table 3). Ipsilateral cortical BD rose modestly three months before returning toward baseline by one year, while ipsilateral cancellous BD was essentially unchanged throughout. On the contralateral side, both cortical and cancellous BD showed a consistent downward trend across follow-up, declining by 14.6 per cent and 21.0 per cent, respectively, from baseline to one year. Despite this directional decrease, none of the contralateral changes reached statistical significance. Overall, mandibular BD did not change significantly before versus after RT for Caucasian patients on either side.

Table 3.

Mean mandibular bone density (HU) by time point within each racial group, with within-group pairwise comparisons

Across both races, mandibular cortical and cancellous bone density values in HU are similar from pre RT to 3 months and 1 year, with all within group p values above 0.05. See long description.
Table 3 Long description

Mean mandibular bone density is reported in Hounsfield units for two races, by side and bone type, at pre RT, 3 months, and 1 year, with p values for pre versus 3 months, 3 months versus 1 year, and pre versus 1 year. For Caucasian ipsilateral cortical bone, values are 965.6 pre RT, 1009.0 at 3 months, and 929.8 at 1 year, with p values 0.99, 0.93, and 0.98. For Caucasian ipsilateral cancellous bone, values are 268.8, 262.6, and 269.8, with p values 0.99, 0.99, and 0.99. For Caucasian contralateral cortical bone, values are 1043.0, 960.0, and 890.1, with p values 0.91, 0.94, and 0.53, and contralateral cancellous values are 247.3, 222.4, and 195.3, with p values 0.99, 0.99, and 0.94. For African American ipsilateral cortical bone, values are 1044.0, 1049.0, and 983.8, with p values 0.99, 0.92, and 0.93, and ipsilateral cancellous values are 320.8, 287.3, and 353.3, with p values 0.95, 0.75, and 0.95. For African American contralateral cortical bone, values are 1118.0, 1197.0, and 1206.0, with p values 0.85, 0.99, and 0.77, and contralateral cancellous values are 384.1, 374.6, and 409.0, with p values 0.99, 0.97, and 0.98. All listed within group comparisons are reported as non significant with p values greater than 0.05.

Note: All within-group comparisons were non-significant (p > 0.05).

HU = Hounsfield units; RT = radiotherapy.

Pre- vs. post-RT mandibular BD analysis for African Americans

A similar pattern of statistical stability was observed in African-American patients. On the ipsilateral side, neither cortical nor cancellous BD changed significantly across any time-point comparison (all p > 0.05; Table 3). Ipsilateral cancellous BD dipped at three months and then recovered to above baseline by one year, but these shifts were not significant. On the contralateral side, both cortical and cancellous BD remained stable or trended slightly upward over follow-up, in contrast to the downward contralateral trend seen in Caucasian patients. As in the Caucasian group, mandibular BD in African-American patients did not change significantly on either side from pre-RT to post-RT.

Pre-RT baseline mandibular BD inter-racial analysis

At baseline, mean mandibular BD was higher in African-American than in Caucasian patients at all four measurement sites (Table 4). The differences were modest for cortical bone (+8.1 per cent ipsilateral; +7.2 per cent contralateral) but more pronounced for cancellous bone, particularly on the contralateral side, where African-American patients had 55.3 per cent higher BD than Caucasian patients. Although African-American patients had consistently higher mean baseline BD on both sides, none of these inter-racial differences reached statistical significance (all p > 0.05).

Table 4.

Inter-racial comparison of mean mandibular bone density (HU); difference expressed as African American relative to Caucasian

At 1 year on the contralateral side, African American cancellous bone density was 409.0 HU versus 195.3 HU in Caucasian, with a 109.4% difference and p value 0.028. See long description.
Table 4 Long description

Mandibular bone density in Hounsfield units is listed for African American and Caucasian groups, with percent difference and p value, by time point, side, and bone type. Pre RT ipsilateral cortical was 1044.0 HU versus 965.6 HU, difference 8.1%, p value 0.51; ipsilateral cancellous was 320.0 versus 268.8, difference 19.3%, p value 0.50. Pre RT contralateral cortical was 1118.0 versus 1043.0, difference 7.2%, p value 0.51; contralateral cancellous was 384.0 versus 247.3, difference 55.3%, p value 0.17. At 3 months, ipsilateral cortical was 1049.0 versus 1009.0, difference 4.0%, p value 0.78; ipsilateral cancellous was 287.3 versus 262.6, difference 9.4%, p value 0.79. At 3 months contralateral cortical was 1197.0 versus 960.0, difference 24.6%, p value 0.08; contralateral cancellous was 374.6 versus 222.4, difference 68.4%, p value 0.20. At 1 year, ipsilateral cortical was 983.8 versus 929.8, difference 5.8%, p value 0.65; ipsilateral cancellous was 353.3 versus 269.8, difference 30.9%, p value 0.26. At 1 year contralateral cortical was 1206.0 versus 890.1, difference 35.5%, p value 0.008; contralateral cancellous was 409.0 versus 195.3, difference 109.4%, p value 0.028.

Note: Shaded rows and bold values indicate statistically significant differences (p < 0.05).

HU = Hounsfield units; RT = radiotherapy.

Post-RT mandibular BD inter-racial analysis

The higher mean BD observed in African-American patients persisted through follow-up. At three months, African-American patients had higher mean BD than Caucasian patients at every site, with the largest differences again in the contralateral mandible (+24.6 per cent cortical; +68.4 per cent cancellous); however, these differences remained non-significant (Table 4). At one year, the inter-racial difference became statistically significant in the contralateral mandible: African-American patients had markedly higher contralateral cortical BD (1206.0 vs. 890.1 HU; +35.5 per cent; p = 0.008) and contralateral cancellous BD (409.0 vs. 195.3 HU; +109.4 per cent; p = 0.028) than Caucasian patients. Ipsilateral differences, while still favouring African-American patients, did not reach significance at either time point. Figure 2 illustrates these inter-racial comparisons across follow-up.

Figure 2.

Intraracial and inter-racial comparisons of mandibular bone density by race and time in (A) cortical ipsilateral, (B) cancellous ipsilateral, (C) cortical contralateral and (D) cancellous contralateral bone. Within each racial group, bone density did not change significantly across any time point (all p > 0.05). Key finding: At one-year post-RT, African-American patients had significantly higher contralateral cortical (**p = 0.008) and cancellous (*p = 0.028) bone density compared to Caucasians, reflecting greater radiation-induced bone density loss in Caucasian patients. **p < 0.01; *p < 0.05. ns = not significant; AA = African American; RT = radiotherapy.

A grouped bar chart set showing mandibular bone density over time for Caucasian and African American groups across four sites. See long description.
Figure 2 Long description

A grouped bar chart titled, Cortical Ipsilateral. The horizontal axis labels are, Pre RT BD, 3 months Post RT BD, 1 year Post RT BD. The vertical axis label is, Bone Density (HU), with tick labels 0, 500, 1000, 1500. Legend entries are, Caucasian and African American. For each time point, the African American bar is higher than the Caucasian bar. The Caucasian bars are near 1000 at all three time points. The African American bars are near 1100 at all three time points. Brackets above comparisons are labeled, ns. The image B showing a grouped bar chart titled, Cancellous Ipsilateral. The horizontal axis labels are, Pre RT BD, 3 months Post RT BD, 1 year Post RT BD. The vertical axis label is, Bone Density (HU), with tick labels 0, 200, 400, 600. Legend entries are, Caucasian and African American. For each time point, the African American bar is higher than the Caucasian bar. The Caucasian bars are near 250 at all three time points. The African American bars are near 300 at Pre RT BD, near 280 at 3 months Post RT BD and near 350 at 1 year Post RT BD. Brackets above comparisons are labeled, ns. The image C showing a grouped bar chart titled, Cortical Contralateral. The horizontal axis labels are, Pre RT BD, 3 months Post RT BD, 1 year Post RT BD. The vertical axis label is, Bone Density (HU), with tick labels 0, 500, 1000, 1500. Legend entries are, Caucasian and African American. At Pre RT BD, the Caucasian bar is near 1000 and the African American bar is near 1100. At 3 months Post RT BD, the Caucasian bar is near 1000 and the African American bar is near 1200. At 1 year Post RT BD, the Caucasian bar is near 900 and the African American bar is near 1200. Brackets above comparisons are labeled, ns. A bracket at the 1 year Post RT BD pair is labeled, double asterisk. The image D showing a grouped bar chart titled, Cancellous Contralateral. The horizontal axis labels are, Pre RT BD, 3 months Post RT BD, 1 year Post RT BD. The vertical axis label is, Bone Density (HU), with tick labels 0, 200, 400, 600. Legend entries are, Caucasian and African American. At Pre RT BD, the Caucasian bar is near 250 and the African American bar is near 350. At 3 months Post RT BD, the Caucasian bar is near 200 and the African American bar is near 300. At 1 year Post RT BD, the Caucasian bar is near 200 and the African American bar is near 400. Brackets above comparisons are labeled, ns. A bracket at the 1 year Post RT BD pair is labeled, asterisk. Across the four bar charts, each chart compares Caucasian and African American bone density at three time points. The African American bars are higher than the Caucasian bars in all four charts. The only marked non ns group differences are at 1 year Post RT BD in Cortical Contralateral with double asterisk and in Cancellous Contralateral with asterisk.

We further evaluated factors associated with reduced BD. On univariate linear regression analysis, we found that there was a statistically significant reduction of contralateral cortical BD in Caucasians (β = -316.3; p = 0.03) (Table 5). Similarly, multivariate analysis of contralateral cortical BD showed a significant reduction of contralateral cortical BD in Caucasians (β = -310.5; p = 0.039) (Table 6). On univariate analysis of contralateral cancellous bone, we found there was a statistically significant reduction of contralateral cancellous BD with Caucasians (β = -213.7; p = 0.02) (Table 7). This association was also reported in multivariate linear regression (β = -284.6; p = 0.005) (Table 8).

Table 5.

Univariate linear regression analysis of factors associated with contralateral cortical Bone density change in the ORN patients

Race shows a p value of 0.03 for Caucasian versus African American, with an estimate of negative 316.3 and a 95% confidence interval from negative 597.8 to negative 34.75. See long description.
Table 5 Long description

Univariate linear regression results are listed for contralateral cortical bone density change in ORN patients, with columns for Variables, Estimate, 95% CI, and p Value. Age has an estimate of negative 6.956, with a 95% confidence interval from negative 31.02 to 17.11 and a p value of 0.54. For Sex, Female is the reference; Male has an estimate of negative 280.9, with a 95% confidence interval from negative 925.8 to 364.0 and a p value of 0.36. For Subsite, BOT is the reference; Soft palate has an estimate of 383.1, with a 95% confidence interval from negative 330.9 to 1097 and a p value of 0.26, and Tonsil has an estimate of 120.6, with a 95% confidence interval from negative 257.3 to 498.4 and a p value of 0.50. For Race, African American is the reference; Caucasian has an estimate of negative 316.3, with a 95% confidence interval from negative 597.8 to negative 34.75 and a p value of 0.03.

Note: Bold value indicates statistically significant differences (p < 0.05).

BOT = base of tongue; CI = confidence interval; ORN = osteoradionecrosis; ref = reference.

Table 6.

Multivariate linear regression analysis of factors associated with contralateral cortical bone density change in the ORN patients

Race showed a significant association: Caucasian had an estimate of minus 310.5 with a 95% confidence interval from minus 602.8 to minus 18.16 and p value 0.039. See long description.
Table 6 Long description

Multivariate linear regression results are listed for contralateral cortical bone density change, with columns for Variables, Estimate, 95% confidence interval, and p Value. Sex uses Female as the reference; Male has an estimate of minus 387.3, confidence interval from minus 995.8 to 221.2, and p Value 0.19. Subsite uses base of tongue as the reference; Soft palate has an estimate of 402.3, confidence interval from minus 206.2 to 1011, and p Value 0.17. Tonsil has an estimate of 295.0, confidence interval from minus 50.87 to 640.9, and p Value 0.08. Race uses African American as the reference; Caucasian has an estimate of minus 310.5, confidence interval from minus 602.8 to minus 18.16, and p Value 0.039.

Note: Bold value indicates statistically significant differences (p < 0.05).

BOT = base of tongue; CI = confidence interval; ORN = osteoradionecrosis; ref = reference.

Table 7.

Univariate linear regression analysis of factors associated with contralateral cancellous bone density change in the ORN patients

Race shows the only p value below 0.05, with Caucasian versus African American estimate minus 213.7 and 95% confidence interval minus 389.2 to minus 38.3. See long description.
Table 7 Long description

Univariate linear regression results are listed for contralateral cancellous bone density change, with columns for estimate, 95% confidence interval, and p value. Age has an estimate of 4.57, with a 95% confidence interval from minus 11.32 to 20.46 and p value 0.54. For sex, female is the reference group and male has an estimate of minus 172.1, with a 95% confidence interval from minus 599.4 to 255.1 and p value 0.40. For subsite, base of tongue is the reference group; soft palate has an estimate of 0.1400, with a 95% confidence interval from minus 483.7 to 484.0 and p value 0.99, and tonsil has an estimate of 37.36, with a 95% confidence interval from minus 204.6 to 279.3 and p value 0.74. For race, African American is the reference group and Caucasian has an estimate of minus 213.7, with a 95% confidence interval from minus 389.2 to minus 38.3 and p value 0.02.

Note: Bold value indicates statistically significant differences (p < 0.05).

BOT = base of tongue; CI = confidence interval; ORN = osteoradionecrosis; ref = reference.

Table 8.

Multivariate linear regression analysis of factors associated with contralateral cancellous bone density change in the ORN patients

Caucasian race has an estimate of minus 284.6 with a p value of 0.005, compared with African American as the reference group. See long description.
Table 8 Long description

Multivariate linear regression results are listed for contralateral cancellous bone density change in osteoradionecrosis patients, with columns for Variables, Estimate, 95% confidence interval, and p Value. Age has an estimate of 13.14, with a 95% confidence interval from minus 1.025 to 27.31 and a p value of 0.06. For Subsite, base of tongue is the reference; soft palate has an estimate of minus 184.7 with a 95% confidence interval from minus 565.7 to 196.3 and a p value of 0.31, and tonsil has an estimate of 111.8 with a 95% confidence interval from minus 77.97 to 301.6 and a p value of 0.22. For Race, African American is the reference; Caucasian has an estimate of minus 284.6 with a 95% confidence interval from minus 464.7 to minus 104.4 and a p value of 0.005.

Note: Bold value indicates statistically significant differences (p < 0.05).

BOT = base of tongue; CI = confidence interval; ORN = osteoradionecrosis; ref = reference

Discussion

Most patients with head and neck cancer receive RT with or without chemotherapy in a definitive or adjuvant setting. Patients receiving RT have short-term and long-term complications. ORN is one of the serious long-term complications of ionising radiation in the craniofacial region.Reference Sathasivam, Davies and Boyd13 The risk factors commonly associated with the development of ORN are age, sex, medical comorbidities, type of dentoalveolar surgery, trauma and dosage of RT.Reference Raguse, Hossamo, Tinhofer, Hoffmeister, Budach and Jamil14 ORN commonly occurs in the mandible due to its sparse blood supply, as opposed to other bones like the maxilla, which has an abundant blood supply.

Although ORN constitutes one of the common long-term complications, there has been reduction of ORN due to improved dental care, better radiation techniques like intensity-modulated radiation therapy (IMRT) and three-dimensional conformal radiation therapy (3D-CRT) and the advent of megavoltage radiation.Reference Chronopoulos, Zarra, Ehrenfeld and Otto4 Although there are known risk factors for osteonecrosis, we have institutional data that supports Caucasians are more prone to developing osteonecrosis. In line with this, the literature also reports trends in higher jaw osteonecrosis rates in Caucasians on bisphosphonate therapy.Reference Quispe, Shi and Burton8, Reference Villa, Castiglioni, Peretti, Omodei, Ferrieri and Abati15 Quispe et al. reported in their cohort that Caucasian patients receiving bisphosphonates had a significantly higher incidence of jaw osteonecrosis (15 per cent) compared to African-American patients (2 per cent).Reference Quispe, Shi and Burton8 Multiple studies have also consistently shown that African-American women have approximately half the risk of fractures compared to their Caucasian counterparts.Reference Griffin, Ray, Fought and Melton16–Reference Barrett-Connor, Siris, Wehren, Miller, Abbott and Berger18 All these findings add further focus to determine the jaw BD variation in ORN between Caucasians and the African-American population. In our study, both Caucasian and African-American ORN patients exhibited a comparable decrease in mean BD on the ipsilateral (direct radiation) side compared to their pre-RT baselines. However, on the contralateral side, African-American patients showed minimal to no reduction in mean BD, whereas Caucasian patients demonstrated a notable decrease relative to their pre-radiation measurements. Hence, the true difference on the contralateral side appears to be an increased susceptibility to radiation-induced BD loss in Caucasians. It is well known that radiation affects the bone primarily by reducing the number of structural units rather than the size. Local high-dose radiation has been associated with atrophy of trabecular bone, which can lead to spontaneous fractures. Studies have shown that the irradiation causes reduced blood flow to the mature bone and decreased bone mineral density.Reference Hopewell19 Radiation also increases vascular permeability, causing haemorrhage and fibrosis. Furthermore, it has been observed that osteoblast and osteoclast activity is directly impacted by radiation, leading to increased bone resorption.Reference Pacheco and Stock20

According to Park et al., the thickness and density of the mandible increase from anterior to posterior. When measurements were taken at the sites of maximal BD, cortical BD was consistently greatest in the retromolar region, whereas cancellous BD was greatest in the canine region.Reference Park, Lee, Jeong and Kwon11 BD measurements contralateral to the tumour may be more representative, as measurements made on the ipsilateral side could be skewed due to local tumour erosion. Our key finding is that mandibular BD did not change statistically within each racial group from pre-RT to post-RT. However, when comparing between races, BD loss at the one-year post-RT time point was statistically significantly greater in Caucasian patients than in African-American patients, especially in the contralateral mandible. On CT, ORN of the mandible typically demonstrates a heterogeneous mixture of destructive and reparative changes rather than a uniform loss of bone. Reported findings include cortical defects, osteolysis, alteration of the trabecular architecture, lingual and buccal cortical thinning or disruption, pathological fractures, sequestrum formation and areas of dense bone sclerosis, often with loss of normal cancellous trabeculae and bicortical involvement. Many of these components have opposing effects on CT-based BD measurements: cortical defects and osteolysis tend to lower HU values, whereas bone sclerosis, sequestra and healing fractures can increase local HU.

In the absence of ORN, radiation alone is expected to produce a more uniform decline in BD due to demineralisation and microstructural damage, and this pattern was more evident on the contralateral, non-ORN side in our cohort. In contrast, on the ipsilateral side, where ORN and tumour bed changes are more common, the BD change likely reflects an admixture of both destructive and sclerotic processes within the same region of interest. The presence of sclerotic foci and sequestra may partially offset the HU-lowering effects of osteolysis and cortical disruption, resulting in a smaller net reduction in measured BD on the ipsilateral side despite clinically significant radiation-related injury. This complex balance between bone loss and reactive sclerosis provides a plausible explanation for the relatively attenuated decrease in BD observed ipsilaterally compared with the contralateral mandible. This radiographic pattern is well supported in existing literature. Patil et al. reported that, at the time ORN was clinically suspected, CT commonly demonstrated mandibular resorption and cortical erosions in all cases, along with sclerosis in 86 per cent and fragmentation in 83 per cent.Reference Patil, Shah, Mahajan, Sable, Shukla and Bornak21 On subsequent imaging, sclerosis became even more prominent, occurring in 92 per cent of cases, while erosive changes continued to progress in 77 per cent. Miyamoto et al. further suggested that sclerosis and resorption may reflect successive bone responses to radiation-induced injury, rather than mutually exclusive findings.Reference Miyamoto, Tanaka, Kogi, Yamaya, Kawai and Ohashi22 In line with these observations, the ISOO-MASCC-ASCO guideline defines ORN radiographically as a lytic or mixed sclerotic bone lesion, thereby recognising that osteolytic destruction and sclerosis may coexist within the same disease process.Reference Peterson, Koyfman, Yarom, Lynggaard, Ismaila and Forner23 The difference in the contralateral BD before and after radiation was statistically significant between African Americans and Caucasians, showing a considerably higher decrease in BD after RT in Caucasians. It is important to acknowledge that race is a self-identified, socially constructed category rather than a direct biological variable,Reference Kaplan and Bennett24, Reference Bonham, Green and Pérez-Stable25 and the observed BD differences likely reflect a complex interplay of factors co-varying with race, including skeletal biology, vitamin D metabolism, socioeconomic status and healthcare access.Reference Williams and Sternthal26 African Americans are known to have higher baseline bone mineral density and lower osteoporosis rates compared to Caucasians, attributed to a combination of genetic, hormonal and mechanical factors rather than race.Reference Barrett-Connor, Siris, Wehren, Miller, Abbott and Berger18, Reference Ettinger, Sidney, Cummings, Libanati, Bikle and Tekawa27 While our findings are consistent with this literature, they should not be interpreted as evidence that race is a direct biological determinant of radiation response. Rather, race likely serves as a proxy for underlying factors that future studies should aim to measure and adjust for directly, and the observed differences should be regarded as exploratory associations pending further mechanistic investigation.Reference Bonham, Green and Pérez-Stable25 This was true for both cortical and cancellous bone and may indicate that Caucasians are more susceptible to ORN. While we did not observe a statistically significant reduction in contralateral cancellous BD in Caucasians (vs. African Americans) in the difference analysis, both univariate and multivariate linear regression models demonstrated a significant decrease. With a larger sample size in future studies, this difference is likely to become more pronounced.

Considering that African Americans have a lower prevalence of osteoporosis,Reference Aloia28 efforts were made to correlate BD and ORN. Although precautions are needed to avoid ORN in all cases, additional care concerning dentition, prevention of xerostomia and smoking cessation will help Caucasian patients. This pilot study is limited by its small sample size (n = 16), which restricts statistical power and generalisability. The multivariable regression findings carry a risk of overfitting and should be considered exploratory and hypothesis-generating. Larger sample size studies would be required to validate the outcomes of this multivariate analysis. Of 67 identified ORN cases over 10 years, only 16 had adequate serial PET/CT imaging, reflecting the rarity of ORN and the stringent imaging requirements. The exclusion of the remaining 51 out of 67 initial ORN cases was solely due to the absence of adequate pre- and post-RT PET/CT imaging, rather than any systematic clinical or demographic differences, minimising the likelihood of clinically driven selection bias. Key clinical variables, including smoking, xerostomia, dental extractions and diabetes, were collected but could not be simultaneously adjusted in the regression models due to the small sample size. Race remained a significant predictor of contralateral BD change after adjustment for sex and tumour subsite; however, residual confounding from these variables cannot be excluded and represents an important limitation. Future larger studies should prospectively adjust for these covariates to better isolate the independent effect of race on radiation-induced BD loss. Larger sample size studies are needed to enhance the validity of our pilot findings of racial differences in contralateral side BD and explore further potential differences on the ipsilateral radiation side, allowing for a more robust evaluation and the inclusion of additional covariates to assess independent risk factors for BD variation. Future prospective studies should also focus on comparing BD changes between patients who develop ORN and those who do not following radiotherapy, to better understand the potential association between lower BD and the increased prevalence of ORN in Caucasian patients.

  • • Mandibular osteoradionecrosis (ORN) is a major late complication after high-dose head and neck radiotherapy (RT)

  • • Bone quality/density is relevant to ORN risk, but longitudinal mandibular bone density tracking is not routinely reported

  • • Racial differences in baseline bone density are recognised in general populations, but mandibular patterns after radiotherapy in ORN patients are not well defined

  • • Uses positron emission tomography/computed tomography (PET/CT) -derived Hounsfield units to track cortical and cancellous mandibular bone density pre-RT, three months and one year post-RT in ORN patients

  • • Shows mandibular bone density remained broadly stable over time within each racial group

  • • Finds African-American patients had higher mandibular bone density than Caucasian patients at baseline and at one year, with greater relative reduction in Caucasian patients, suggesting race may be an important factor to consider in ORN-related bone density variation

Conclusion

Our pilot study shows that cortical and cancellous mandibular BDs are higher in African Americans compared to Caucasians, and the percent decrease in BD post-RT is greater in Caucasians. Caucasians have an increased association of BD loss in the contralateral mandible compared to African-American ORN patients. This may be associated with why Caucasians are at a higher risk for osteonecrosis and possibly ORN. Further study with a larger sample size and multi-institutional data will confirm these findings.

Author contributions

Cherie-Ann O. Nathan: project conceptualization, data analysis, manuscript preparation, manuscript review and editing. Srivatsa Surya Vasudevan: data extraction, data analysis, manuscript preparation, manuscript review and editing. Abhijit Gundale: data extraction, data analysis, manuscript preparation, manuscript review and editing. Eleni Mijalis: data extraction, data analysis, manuscript preparation, manuscript review and editing. Amol Takalkar: data analysis, manuscript review and editing. Kent Burton: data analysis, manuscript review and editing. Tara Moore-Medlin: data analysis, manuscript review and editing. Jerry McLarty: data analysis, manuscript review and editing.

Competing interests

No conflict of interest to disclose.

Funding/support

This study was department funding and approved by the Louisiana State University Health Sciences Center Insitutional Review Board (LSUHS IRB).

Footnotes

S.S.V. and A.G. contributed equally to the work

Cherie-Ann O. Nathan takes responsibility for the integrity of the content of the paper

References

Rivero, JA, Shamji, O, Kolokythas, A. Osteoradionecrosis: a review of pathophysiology, prevention and pharmacologic management using pentoxifylline, α-tocopherol, and clodronate. Oral Surg Oral Med Oral Pathol Oral Radiol 2017;124:464–7110.1016/j.oooo.2017.08.004CrossRefGoogle ScholarPubMed
Singh, A, Kitpanit, S, Neal, B, Yorke, E, White, C, Yom, SK, et al. Osteoradionecrosis of the jaw following proton radiation therapy for patients with head and neck cancer. JAMA Otolaryngol Head Neck Surg 2023;149:151–910.1001/jamaoto.2022.4165CrossRefGoogle ScholarPubMed
Vasudevan, SS, Candelo, E, Sharifi, A, Ma, DJ, Patel, SH, Routman, DM, et al. Survival, tumor control, and safety outcomes of proton therapy in sinonasal cancer population: a systematic review and meta-analysis. Head Neck 2025;47:1291–130510.1002/hed.28082CrossRefGoogle ScholarPubMed
Chronopoulos, A, Zarra, T, Ehrenfeld, M, Otto, S. Osteoradionecrosis of the jaws: definition, epidemiology, staging and clinical and radiological findings. A concise review. Int Dent J 2018;68:22–3010.1111/idj.12318CrossRefGoogle ScholarPubMed
Kün-Darbois, J-D, Fauvel, F. Medication-related osteonecrosis and osteoradionecrosis of the jaws: update and current management. Morphologie 2021;105:170–8710.1016/j.morpho.2020.11.008CrossRefGoogle ScholarPubMed
Otto, S, Pautke, C, Van den Wyngaert, T, Niepel, D, Schiødt, M. Medication-related osteonecrosis of the jaw: prevention, diagnosis and management in patients with cancer and bone metastases. Cancer Treat Rev 2018;69:177–8710.1016/j.ctrv.2018.06.007CrossRefGoogle ScholarPubMed
Marx, RE. Osteoradionecrosis: a new concept of its pathophysiology. J Oral Maxillofac Surg 1983;41:283–810.1016/0278-2391(83)90294-XCrossRefGoogle ScholarPubMed
Quispe, D, Shi, R, Burton, G. Osteonecrosis of the jaw in patients with metastatic breast cancer: ethnic and socio-economic aspects. Breast J 2011;17:510–1310.1111/j.1524-4741.2011.01119.xCrossRefGoogle ScholarPubMed
Noel, SE, Santos, MP, Wright, NC. Racial and ethnic disparities in bone health and outcomes in the United States. J Bone Miner Res 2021;36:1881–90510.1002/jbmr.4417CrossRefGoogle ScholarPubMed
HM, Perry III, Horowitz, M, Morley, JE, Fleming, S, Jensen, J, Caaccione, P, et al. Aging and bone metabolism in African American and Caucasian women. J Clin Endocrinol Metab 1996;81:1108–17Google Scholar
Park, H-S, Lee, Y-J, Jeong, S-H, Kwon, T-G. Density of the alveolar and basal bones of the maxilla and the mandible. Am J Orthod Dentofacial Orthop 2008;133:30–710.1016/j.ajodo.2006.01.044CrossRefGoogle ScholarPubMed
Shapurian, T, Damoulis, PD, Reiser, GM, Griffin, TJ, Rand, WM. Quantitative evaluation of bone density using the Hounsfield index. Int J Oral Maxillofac Implants 2006;21:290–7Google ScholarPubMed
Sathasivam, HP, Davies, GR, Boyd, NM. Predictive factors for osteoradionecrosis of the jaws: a retrospective study. Head Neck 2018;40:46–5410.1002/hed.24907CrossRefGoogle ScholarPubMed
Raguse, J-D, Hossamo, J, Tinhofer, I, Hoffmeister, B, Budach, V, Jamil, B, et al. Patient and treatment-related risk factors for osteoradionecrosis of the jaw in patients with head and neck cancer. Oral Surg Oral Med Oral Pathol Oral Radiol 2016;121:215–21.e110.1016/j.oooo.2015.10.006CrossRefGoogle ScholarPubMed
Villa, A, Castiglioni, S, Peretti, A, Omodei, M, Ferrieri, GB, Abati, S. Osteoporosis and bisphosphonate-related osteonecrosis of the jaw bone. ISRN Rheumatol 2011;2011:65402710.5402/2011/654027CrossRefGoogle ScholarPubMed
Griffin, MR, Ray, WA, Fought, RL, Melton, LJ III. Black-white differences in fracture rates. Am J Epidemiol 1992;136:1378–8510.1093/oxfordjournals.aje.a116450CrossRefGoogle ScholarPubMed
Baron, JA, Barrett, J, Malenka, D, Fisher, E, Kniffin, W, Bubolz, T, et al. Racial differences in fracture risk. Epidemiology 1994;5:42–710.1097/00001648-199401000-00008CrossRefGoogle ScholarPubMed
Barrett-Connor, E, Siris, ES, Wehren, LE, Miller, PD, Abbott, TA, Berger, ML, et al. Osteoporosis and fracture risk in women of different ethnic groups. J Bone Miner Res 2005;20:185–9410.1359/JBMR.041007CrossRefGoogle ScholarPubMed
Hopewell, JW. Radiation-therapy effects on bone density. Med Pediatr Oncol 2003;41:208–1110.1002/mpo.10338CrossRefGoogle ScholarPubMed
Pacheco, R, Stock, H. Effects of radiation on bone. Curr Osteoporos Rep 2013;11:299–30410.1007/s11914-013-0174-zCrossRefGoogle ScholarPubMed
Patil, V, Shah, P, Mahajan, A, Sable, N, Shukla, A, Bornak, G, et al. Decoding osteoradionecrosis of the jaw: radiological progression and a novel CT-based grading system. Cancers (Basel) 2026;18:18710.3390/cancers18020187CrossRefGoogle Scholar
Miyamoto, I, Tanaka, R, Kogi, S, Yamaya, G, Kawai, T, Ohashi, Y, et al. Clinical diagnostic imaging study of osteoradionecrosis of the jaw: a retrospective study. J Clin Med 2021;10:470410.3390/jcm10204704CrossRefGoogle ScholarPubMed
Peterson, DE, Koyfman, SA, Yarom, N, Lynggaard, CD, Ismaila, N, Forner, LE, et al. Prevention and management of osteoradionecrosis in patients with head and neck cancer treated with radiation therapy: ISOO-MASCC-ASCO Guideline. J Clin Oncol 2024;42:1975–9610.1200/JCO.23.02750CrossRefGoogle ScholarPubMed
Kaplan, JB, Bennett, T. Use of race and ethnicity in biomedical publication. JAMA 2003;289:2709–1610.1001/jama.289.20.2709CrossRefGoogle ScholarPubMed
Bonham, VL, Green, ED, Pérez-Stable, EJ. Examining how race, ethnicity, and ancestry data are used in biomedical research. JAMA 2018;320:1533–410.1001/jama.2018.13609CrossRefGoogle ScholarPubMed
Williams, DR, Sternthal, M. Understanding racial-ethnic disparities in health: sociological contributions. J Health Soc Behav 2010;51:S15–2710.1177/0022146510383838CrossRefGoogle ScholarPubMed
Ettinger, B, Sidney, S, Cummings, SR, Libanati, C, Bikle, DD, Tekawa, IS, et al. Racial differences in bone density between young adult black and white subjects persist after adjustment for anthropometric, lifestyle, and biochemical differences. J Clin Endocrinol Metab 1997;82:429–34Google ScholarPubMed
Aloia, JF. African Americans, 25-hydroxyvitamin D, and osteoporosis: a paradox. Am J Clin Nutr 2008;88:545S–50S10.1093/ajcn/88.2.545SCrossRefGoogle ScholarPubMed
Figure 0

Figure 1. Axial CT images showing the regions of interest used for mandibular bone density measurements. (A) The canine region (cancellous bone density measurement sites). (B) The retromolar trigone region (cortical bone density measurement sites), bilaterally ipsilateral and contralateral to the tumour. These standardised sites were selected as they represent the regions of maximal cancellous and cortical bone density in the mandible, respectively. CT = computed tomography.

Figure 1

Table 1. Patient distribution and demographicsTable 1 long description.

Figure 2

Table 2. Baseline demographic and clinical characteristics by raceTable 2 long description.

Figure 3

Table 3. Mean mandibular bone density (HU) by time point within each racial group, with within-group pairwise comparisonsTable 3 long description.

Figure 4

Table 4. Inter-racial comparison of mean mandibular bone density (HU); difference expressed as African American relative to CaucasianTable 4 long description.

Figure 5

Figure 2. Intraracial and inter-racial comparisons of mandibular bone density by race and time in (A) cortical ipsilateral, (B) cancellous ipsilateral, (C) cortical contralateral and (D) cancellous contralateral bone. Within each racial group, bone density did not change significantly across any time point (all p > 0.05). Key finding: At one-year post-RT, African-American patients had significantly higher contralateral cortical (**p = 0.008) and cancellous (*p = 0.028) bone density compared to Caucasians, reflecting greater radiation-induced bone density loss in Caucasian patients. **p < 0.01; *p < 0.05. ns = not significant; AA = African American; RT = radiotherapy.Figure 2 long description.

Figure 6

Table 5. Univariate linear regression analysis of factors associated with contralateral cortical Bone density change in the ORN patientsTable 5 long description.

Figure 7

Table 6. Multivariate linear regression analysis of factors associated with contralateral cortical bone density change in the ORN patientsTable 6 long description.

Figure 8

Table 7. Univariate linear regression analysis of factors associated with contralateral cancellous bone density change in the ORN patientsTable 7 long description.

Figure 9

Table 8. Multivariate linear regression analysis of factors associated with contralateral cancellous bone density change in the ORN patientsTable 8 long description.