Osteoradionecrosis of the jaws: definition, epidemiology, staging and clinical and radiological findings. A concise review

Introduction

Osteoradionecrosis (ORN) is a pernicious complication of radiotherapy in head and neck cancer. It was first described by Regaud¹ in 1922 and is still a clinical challenge. According to the most recent literature, ORN of the jaws is defined as exposed irradiated bone that fails to heal over a period of 3 months without any evidence of persisting or recurrent tumour2., 3., 4.. The mechanism of pathogenesis is still under investigation. However, the most frequently reported reason is radiation arteritis. Radiation arteritis leads to the development of a hypocellular, hypovascular and hypoxic environment, which results in a pathological outcome⁵.

The purpose of this paper is to explore the current theories on the definition and staging systems of ORN of the jaws. In addition, the epidemiology and the clinical and radiological findings of ORN of the jaws, are critically reviewed based on the available literature. Finally, critical issues and considerations are discussed and pathways for future research are proposed.

Methods

An electronic search of three databases (PubMed, Scopus and the Cochrane Library) was performed using the keywords ‘osteoradionecrosis’, ‘radiotherapy’, ‘osteonecrosis’, ‘osteoradionecrosis’ and ‘mandible’, ‘osteoradionecrosis’ and ‘jaw’, ‘osteoradionecrosis’ and ‘radiotherapy’, ‘osteoradionecrosis’ and ‘staging’, ‘osteoradionecrosis’ and ‘clinical findings’, ‘osteoradionecrosis’ and ‘radiological findings’, to identify literature published in English and German between January 1922 and April 2014 (Table 1). Abstracts were obtained for all the titles identified during the electronic search. Two reviewers (A.C. and T.Z.) independently screened titles and abstracts to eliminate articles that were completely off-topic. No other exclusion criteria were utilised. Following exclusion of the articles irrelevant to the topic and the removal of duplicating papers, the remaining articles were carefully reviewed and their findings were critically analysed.

Table 1.

Example of electronic search strategy in PubMed

Search strategy	Results
Osteoradionecrosis	1,980
Radiotherapy	286,652
Osteonecrosis	15,087
Osteoradionecrosis and mandible	636
Osteoradionecrosis and jaw	738
Osteoradionecrosis and radiotherapy	1,332
Osteoradionecrosis and staging	76
Osteoradionecrosis and clinical findings	177
Osteoradionecrosis and radiological findings	30

Results

Definition

Several attempts have been made in order to define ORN. Ewing⁶ was the first to use the term ‘radiation osteitis’ to describe changes in bone after radiotherapy. In the following years, several terms were used to name these changes in bone, such as radiation osteitis, ORN and avascular bone necrosis⁷. In 1974, Guttenberg⁸ proposed the term ‘septic ORN of the mandible’ to describe the stage of necrosis when irradiated bone becomes superficially infected, ending up with a high risk of involvement of deeper structures.

In 1983, Marx⁵ defined ORN as ‘an area >1 cm of exposed bone in a field of irradiation that failed to show any evidence of healing for at least 6 months’. Marx also reported that superficial contamination and no interstitial infection was present. In 1987, Marx and Johnson³ suggested the definition of ORN as: ‘The exposure of nonviable bone which fails to heal without intervention’. Epstein et al.⁹ defined ORN as ‘an ulceration or necrosis of the mucous membrane, with exposure of necrotic bone for more than 3 months’.

Widmark et al.¹⁰ described ORN as ‘a non-healing mucosal or cutaneous ulcer with denuded bone, lasting for more than 3 months’. Both Widmark and Marx excluded from their definition conditions with necrotic bone for which the mucosa and skin were intact. Store and Boysen¹¹ reported that exposed bone is sporadically present; radiological findings are evident in all cases of ORN. In 1997, Wong et al.¹² defined ORN as ‘a slow-healing radiation-induced ischemic necrosis of variable extent occurring in the absence of local primary tumor necrosis, recurrence or metastatic disease’.

According to the literature published in the last 15 years, ORN of the jaws is defined as exposed irradiated bone that fails to heal over a period of 3 months without evidence of persisting or recurrent tumour2., 3., 4., 13., 14., 15.. At the time of diagnosis, the necrosis might involve the bone superficially or deeply. It might be a process that progresses slowly or an active progressive state that can lead to a pathological fracture16., 17., 18..

Epidemiology

The prevalence of ORN varies widely in the literature (Table 2)16., 17., 19., 20., 21., 22., 23., 24., 25., 26., 27., 28., 29., 30., 31., 32., 33., 34., 35., 36., 37., 38., 39., 40., 41., 42., 43., 44., 45., 46., 47., 48. and ranges from 0.4% to 56%19., 20., 22., 24., 30., 31., 32., 33., 36., 49.. ORN usually affects patients over 55 years of age14., 36., 50., 51.. However, the most frequently reported prevalence rate is 5–15%7., 36., 52.. In up to 20% of patients with persistent ORN which does not respond to aggressive treatment, bone damage is reported to be caused by recurrent disease or a second primary tumour16., 53..

Table 2.

Incidence of osteoradionecrosis (ORN) in different studies

Study	Year	Period	Total number of patients	Prevalence of ORN (%)	Mean irradiation dose
Watson and Scarborough20	1938	1930–1937	1,819	12.9
Martin and Sugarbaker21	1940		103	25
MacComb22	1962	1952–1959	251	37.1
MacDougall et al.23	1963		364	5
Grant and Fletcher24	1966	1954–1962	176	37.5	10,000*
Rahn and Drane25	1967	1960–1962	120	44.2
Rankow and Weissman26	1971		176	6.3	5,000*
Wang27	1972		262	5.8
Cheng and Wang28	1974		76	17.1	2,500*
Marciani and Plezia29	1974		220	10.5
Bedwinek et al.30	1976	1966–1971	381	14.2	7,000*
Murray et al.31	1980	1966–1975	653	21.1	5,000*
Morrish et al.19	1981	1971–1977	100	22	7,500*
Epstein et al.32	1987	1977–1984	1,000	2.7	6,000*
Withers et al.33	1995	1976–1985	676	4.7	65†
Turner et al.34	1996	1980–1987	333	5.9	55†
Thorn et al.16	2000	1992–1998	80	74	60†
Storey et al.35	2001	1965–1995	83	6	60†
Reuther et al.36	2003	1969–1999	830	8.2	60†
Oh et al.37	2004	1989–2004	81	4.9	52.6†
Studer et al.38	2004	1980–1998	268	12.5	60†
Ben-David et al.39	2007	1996–2005	176	0	65†
Jham et al.40	2008	2003–2005	207	5.5	58.9†
Katsura et al.41	2008	1996–2003	39	15	60.7†
Gomez et al.42	2009	2000–2006	35	5	60†
Lee et al.43	2009	1990–2000	198	6.6	60†
Stenson et al.44	2010	1994–2008	27	18.4	62†
Gomez et al.45	2011	2000–2007	168	1.2	67.9†
Monnier et al.17	2011	2000–2007	73	40	66†
Crombie et al.46	2012	2000–2007	54	36	69.7†
Niewald et al.47	2013	1993–2001	99	12	72
Tsai et al.48	2013	2000–2008	402	7.5	70†

^*Radiation dose given in rads.

^†Radiation dose given in Grays (Gy).

Staging

Several staging or scoring systems of ORN have been proposed (Table 3)2., 11., 32., 54., 55., 56., 57., 58., 59.. These systems are based on response to hyperbaric oxygen (HBO) therapy, degree of bone damage, clinical–radiological findings, duration of bone exposure and treatment required.

Table 3.

Classification systems of osteoradionecrosis (ORN)

Study	Stages	Description of stages		Basis of stage
Coffin55	2	Minor	A series of small sequestra which separate spontaneously after varying periods of weeks or months. These areas cannot be demonstrated radiologically	Clinical and radiological findings
		Major	Necrosis occurring to an extent that involves the entire thickness of the jaw, and a pathological fracture is inevitable. This form can be obviously seen radiologically
Marx2	3	Stage I	Exposed alveolar bone without pathologic fracture, which responds to HBO therapy	Response to HBO therapy
		Stage II	Disease does not respond to HBO therapy, and requires sequestrectomy and saucerisation
		Stage III	Full-thickness bone damage or pathological fracture, usually requires complete resection and reconstruction with free tissue
Morton and Simpson56	3	Minor	Ulceration with exposed bone and a history of bony spicules that healed spontaneously over a period of months	Clinical findings and response to treatment
		Moderate	Exposed bone and small sequestra limited in nature and healing spontaneously with conservative treatment within 6–12 months
		Major	Large areas of exposed bone, with formation of large sequestra, possible fracture and sinus formation. These cases often require radical treatment
Epstein et al.32	3	Stage I Ia Ib	Resolved, healed No pathological fracture Pathological fracture	Disease progression
		Stage II IIa IIb	Chronic, persistent non-progressive No pathological fracture Pathological fracture
		Stage III IIIa IIIb	Active, progressive No pathological fracture Pathological fracture
Glanzmann and Gratz57	5	Stage 1	Bone exposure without signs of infection and persisting for at least 3 months	Duration of bone exposure and treatment necessity
		Stage 2	Bone exposure with signs of infection or sequester and without the signs of Grades 3–5
		Stage 3	Bone necrosis treated with mandibular resection with a satisfactory result
		Stage 4	Bone necrosis with persisting problems despite mandibular resection
		Stage 5	Death from ORN
Clayman54	2	Type I	Bone lysis occurs under intact gingiva or mucosa	Clinical findings
		Type II	A more aggressive type in which soft tissues break down, exposing the bone to saliva, and causing secondary contamination
Store and Boysen11	4	Stage 0	Mucosal defects only	Combination of radiological and clinical parameters
		Stage 1	Radiological evidence of necrotic bone with intact mucosa
		Stage 2	Positive radiological findings with denuded bone intra-orally
		Stage 3	Clinically exposed radionecrotic bone, verified by imaging techniques, along with skin fistulae and infection
Schwartz and Kagan58	3	Stage I	Minimal soft-tissue ulceration and limited exposed cortical bone. Patients are treated with conservative management	Imaging and clinical findings
		Stage II IIa IIb	Localised involvement of the mandibular cortex and underlying medullary bone Minimal soft tissue ulceration Presence of an oro-cutaneous fistula and mild soft-tissue necrosis
		Stage III	Full-thickness involvement of the bone, including the inferior border. Pathological fractures may also be present
Notani et al.59	3	Stage I	ORN confined to alveolar bone	Clinical findings
		Stage II	ORN limited to the alveolar bone and/or mandible above the level of the inferior alveolar canal
		Stage III	ORN involving the mandible below the level of the inferior alveolar canal and/or skin fistula and/or pathological fracture

HBO, hyperbaric oxygen.

Coffin⁵⁵ divided cases of ORN into two groups: minor and major. Morton and Simpson⁵⁶ subdivided ORN into three groups – ‘minor’, ‘moderate’ and ‘major’. In 1983, Marx² proposed a three-stage system for ORN. According to his protocol, patients are categorised as Stage I if they exhibit exposed bone in a field of radiation that has failed to heal for at least 6 months and do not have a pathological fracture, cutaneous fistula or osteolysis to the inferior border. In Stage I, all patients receive 30 sessions of HBO at 2.4 atmospheres absolute for 90 minutes at depth. Patients who respond to HBO alone (Stage I responder) demonstrate a softening of the radiated tissues and spontaneous sequestration of exposed bone with formation of granulation tissue. Each Stage I responder undergoes an additional 10 HBO sessions and then the tissues are allowed to heal completely. Stage II patients are those who do not respond to the 30 sessions of HBO. This group is characterised by a large amount of non-viable bone that makes resorption and sequestration from HBO-induced angiogenesis alone impossible. Consequently, careful surgical debridement is required. Stage II patients undergo transoral resection with limited soft-tissue reflection. In particular, surgical treatment includes extraction of involved dentition and a non-continuity bone resection to clinically bleeding bone. Wound flaps are closed primarily and the patient is given 10 postsurgical sessions of HBO. Tissues that heal without complication are treated with a prosthesis when required, similarly to Stage I responders. Stage III patients are characterised by having a large quantity of non-viable bone and/or soft tissue unable to be managed by HBO-induced angiogenesis alone or HBO combined with local sequestrectomy. In addition to 30 presurgical HBO treatments, each Stage III patient requires a continuity resection, stabilisation and 10 postsurgical sessions of HBO, and are scheduled for later (usually 3 months) reconstruction (Stage III-R). Stage III patients are therefore those who fail to respond to Stage I and Stage II treatment and those who initially present with a pathological fracture, cutaneous fistula or osteolysis to the inferior border2., 60..

Epstein et al.³² suggested a new staging system for ORN with three tiers based on clinical findings. Glanzmann and Gratz⁵⁷ proposed a system based on the duration of bone exposure and necessity of treatment. Clayman⁵⁴ introduced a classification of ORN related to the integrity of the overlying mucosa. According to this classification, type I includes cases of ORN in which bone lysis occurs under intact gingiva or mucosa. Type II includes more aggressive cases of ORN in which soft tissues break down and the bone is exposed to saliva, causing secondary contamination. This is defined as radiation osteomyelitis. Type I cases heal with conservative therapy; type II cases do not. In 2000, Store and Boysen¹⁶ introduced a new classification of ORN that is based on the presence or absence of clinical and radiological signs.

The most recently designed staging systems are these of Schwartz and Kagan⁵⁸ and Notani et al.⁵⁹ The system of Schwartz and Kagan⁵⁸ is based on clinical and radiological findings. Notani et al.⁵⁹ divided the cases into three grades based on the extent of the ORN lesion. Grade I is defined as ORN confined to the alveolar bone. Grade II is defined as ORN limited to the alveolar bone and/or the mandible above the level of the mandibular alveolar canal. In Grade III the ORN extends to the mandible under the level of the mandibular alveolar canal and a skin fistula and/or a pathological fracture is present.

Clinical findings

Clinical signs and symptoms of ORN include ulceration or necrosis of the mucosa with exposure of necrotic bone for longer than 3 months, pain, trismus and suppuration in the area (Figure 1a)32., 61., 62., 63., 64.. Neurological symptoms, such as pain, dysaesthesia or anaesthesia, as well as fetor oris, dysgeusia and food impaction in the area, are also usually present. Exposure of rough and irregular bone can result in physical irritation of adjacent tissues. Progression of ORN may lead to pathological fractures, intra-oral or extra-oral fistulae and local or systemic infection. Difficulties in mouth opening, mastication and speech frequently arise9., 65., 66., 67.. In patients treated with external beam radiation therapy (EBRT), osseous alterations usually appear in the body of the mandible (premolar and molar regions), whereas in those managed with brachytherapy, the lingual or buccal surfaces are affected⁶⁸.

Figure 1.

A 69-year-old man presented with pain and a non-healing wound in the left lower jaw. The patient was an active smoker, suffered from pharynx cancer and had received radiotherapy (external beam radiotherapy with standard field sizes, conventional fractionation and mean dose 64 Gy) and chemotherapy. Clinically, exposed necrotic bone in the left lower jaw, inflammation, swelling and inferior alveolar nerve hypesthesia was present (a). The orthopantomogram revealed pathologic fracture of the left lower jaw (b). The patient was diagnosed with osteoradionecrosis of the lower jaw and was scheduled to be treated surgically.

The diagnosis of septic ORN appears to be easier. Marked pain is the primary symptom. A thorough clinical examination will reveal intra- or extra-oral draining fistulae, ulcerations of the mucous membrane, exposed devitalised bone, haemorrhage, cellulitis or pathological fractures. A biopsy is mandatory for final diagnosis in order to exclude metastatic cancer⁸.

Radiological findings

Radiographs, computed tomography (CT) scans, magnetic resonance imaging (MRI), Doppler ultrasound, nuclear medicine and near-infrared spectroscopy are often indicated in order to detect ORN¹⁸.

ORN is not usually detectable radiographically in early stages⁶⁹. Imaging features are not correlated to the severity of ORN8., 23., 70., 71.. The described radiographic features range from normal appearance, to localised osteolytic areas, extensive osteolytic areas, sequestra and fracture (Figure 1b). Radioluncies indicating post extraction sockets will often remain visible for longer than 12 months (Figure 2a). The most definitive radiographic alterations in early disease are increased radiodensity, as well as a mixed radio-opaque/radiolucent lesion in which radiolucent areas represent bone destruction (Figure 2b)⁸.

Figure 2.

Orhtopantomograms of patients who have received radiotherapy (external beam radiotherapy with standard field sizes, conventional fractionation and mean dose 64 Gy) for oropharyngeal cancers. The patients underwent extractions of teeth of the lower jaw after completion of radiotherapy. The orthopantomogram revealed radiolucencies indicating post extraction sockets, which remained visible for longer than 12 months (a), as well as increased radiodensity with mixed radio-opaque/radiolucent lesions, where the radiolucent areas represent bone destruction (b).

Orthopantomogram (OPT) is the most frequently used imaging method for the diagnosis of ORN and is usually supplemented with other extra-oral or intra-oral radiographs. In an OPT, ORN is depicted as an undefined radiolucency, without sclerotic demarcation, which surrounds necrotic zone. Radiopaque areas can be identified when bone sequestra are formed. In order to be visible in an OPT, a substantial alteration in mineral content and extensive involvement of bone is required and this only occurs in later stages of ORN⁹. Ardran⁷² noted that a 30% loss of bone mineral content is necessary before any radiographic change can be seen.

CT⁷³ shows osseous abnormalities, such as focal lytic areas, cortical interruptions and loss of the spongiosa trabeculation on the symptomatic side, frequently accompanied by soft-tissue thickening. Such a picture may cause difficulties in differential diagnosis between ORN and recurrent tumour⁶⁸. In MRI with gadolinium administration, an abnormal marrow signal, cortical destruction and slight-to-mild irregular enhancement is demonstrated35., 74., 75., 76.. MRI has the advantage of excellent tissue contrast and high spatial resolution⁷⁷.

Bone scintigraphy permits estimation of the extension and location of the lesion. It shows high sensitivity (up to 100%) but low specificity (about 60%) for the diagnosis of ORN⁷⁷. Scintigraphy using 99mTc-marked diphosphonates (99mTc-MDP) allows highly sensitive depiction of mandibular lesions as a result of their altered phosphate metabolism. Pathophysiological changes in bone can be identified sooner when using 99mTc-MDP than when using conventional radiography because scan changes reflect osteoblastic activity and good blood flow⁷⁸. Low spatial resolution and over-projection by soft tissues are the major disadvantages of the method, which can be overcome with the use of single photon emission computed tomography (SPECT)⁷⁷. Finally, positron emission tomography (PET), which is a promising method for pretherapeutic assessment of the spread of squamous cell carcinomas, has been advocated as being efficient for differentiating between ORN and tumour recurrence⁷⁹.

Discussion and Conclusion

Irradiation of bone is the main prerequisite for development of ORN. A differential diagnosis is required to exclude recurrence of tumour and bisphosphonate-related osteonecrosis of the jaws. In contrast to the findings of Store and Boysen¹¹ mucosal breakdown or failure of healing is necessary for the diagnosis of ORN3., 9., 10..

The duration of bone exposure is still a matter of controversy. Some authors did not state the period of time that bone was exposed⁷¹. Other authors recommended a 2-month period of exposed bone before diagnosis73., 80., 81., or even 319., 82., 83. to 62., 5. months. There are also cases in which a late diagnosis is present. Berger and Symington⁸⁴ reported two late presentations: one 45 years after radium implant therapy; and the other 38 years after external beam treatment.

A very short waiting period can lead to over-diagnosis as mucosal radionecrosis can occur without ORN. Moreover, any surgery and/or extraction usually takes up to 1 month to heal. On the other hand, monitoring a trauma for a longer period (such as 6 months) is contraindicated; intervention prior to this time is certainly needed. For the aforementioned reasons, in order to diagnose ORN, recent literature2., 3., 4., 13., 14., 15. indicates that bone exposure should be at least 3 months in duration.

The variability in prevalence of ORN can probably be attributed to differences in the study populations observation periods and existence of pretreatment dental assessment and dental management of cohorts. The literature reports numerous factors which are associated with the risk of ORN development and also affect its prevalence. Total radiation dose, brachytherapy, fractionation, poor oral hygiene, alcohol, tobacco use, dental extractions, tumour size and location, staging and chemotherapy have been highlighted7., 19., 36., 44., 47., 53., 85., 86., 87., 88., 89., 90..

The late onset of ORN can be attributed to the late occurrence of oropharyngeal cancers (OPCs) and their complications (mean age of patients over 60 years). This can be explained by the fact that oral tissues tend to undergo prolonged exposure to potential carcinogens with advancing age. In addition, aging cells may be more susceptible to DNA damage. ORN is predominant in the mandible (the ratio between ORN in the mandible and ORN in the maxilla is 24:1) ⁹¹. The reason for this could be that the mandible has a restricted localised blood supply, which is often completely within the radiation field, whereas the maxilla has many anastomoses located outside the area of irradiation. Furthermore, mandibular bone density is different from maxillary bone density, and the mandible absorbs a higher amount of radiation during radiotherapy.

The prevalence of ORN has decreased since the 1990s⁹². Recent studies have shown a decrease in the prevalence of ORN to levels lower than 5%. According to Clayman⁵⁴, the application of megavoltage therapy resulted in a significant reduction of the overall prevalence of ORN from 11.8% before 1968 to 5.4% after this time. Wahl⁹³ described similar results and noted a prevalence of ORN of 3% during the period 1997 to 2006. Lee et al.⁴³ found that the frequency of ORN was 6.6% among 198 patients with either oral cavity or OPCs treated with radiation between 1990 and 2000. The overall reduction of ORN can be attributed to the advent of megavoltage radiotherapy, improved dental-preventive care and improved radiation techniques, including three-dimensional conformal radiotherapy (3D-CRT) and intensity-modulated radiotherapy (IMRT)7., 36., 38..

Marx’s staging system² is based on the use of, and response to, HBO. The advantages of this protocol include selection of patients who are able to respond to less aggressive treatments, use of minimal levels of HBO, resolution of the disease and preparation of patients’ tissues for reconstruction without further HBO2., 60..

Even though Marx’s staging system is the one still used by surgeons, it has been used less frequently from the mid-1990s to date. The reasons for this are that HBO is nowadays recommended for use as adjuvant therapy, its effectiveness has been questioned and most cases of ORN can be managed successfully without HBO according to the most recent literature58., 94., 95., 96.. The staging system of Epstein et al.³² is an improvement, but it is focussed on the presence or absence of a pathological fracture⁵⁸.

The classification system of Notani et al.⁵⁹ seems to be more accurate as it is based on: (i) the presence or absence of clinical and radiological signs, in contrast to other systems that are non-specific and are partly based on patients’ subjective interpretation; and (ii) pretreatment evaluation, not on treatment response or refractoriness. It is also simple and easily recalled⁹⁷ (For review see Chronopoulos⁹⁸

During the past 80 years, a number of theories about the origin of ORN have been proposed. Despite the controversy, the majority of authors agree that prerequisites for the diagnosis of ORN are: (i) previous irradiation of the affected bone; (ii) absence of recurrent tumour; (iii) presence of mucosal breakdown or failure of healing, resulting in bone exposure; and (iv) ‘necrosis’ of the overlying bone. The presence of pathological fracture, fistula formation or cellulitis is not necessary for the diagnosis. Mandibular ORN is predominant over maxillary ORN. ORN can lead to pain, fracture and sequestration of the bone, as well as to fistulae. Although the prevalence of ORN has decreased to levels lower than 5%, it still remains a pernicious complication of radiotherapy and a challenge to the clinician because of difficulty in its management. Owing to the severity of ORN and the difficulties in its early and accurate diagnosis and treatment, increased awareness of this complex entity among clinicians is desirable, with the aim to prevent its appearance or the development of severe complications. Further research is required in order to clarify its complex etiology, and guide new treatment strategies.

Acknowledgements

This review is included in the doctoral thesis of the first author, entitled ‘Clinical presentation and risk factors of osteoradionecrosis’.

Patient data were anonymised and de-identified before analysis. The clinical photograph and three radiographs were derived from the archive of the Department of Oral and Maxillofacial Surgery in Munich (LMU).

Conflicts of interest

We wish to confirm that there are no known conflicts of interest associated with this publication and there has been no significant financial support for this work that could have influenced its outcome. There was no funding received for this study.