Discussion

With the increasing survival of patients following treatment including radiotherapy for head and neck malignancies,^{Reference Kadakia, Badhey, Inman, Mourad and Ducic3,Reference Funk, Karnell and Christensen4} the incidence of osteoradionecrosis of the temporal bone is set to increase. The hypoxic, hypovascular and hypocellular environment induced by radiotherapy leads to impaired collagen synthesis and cell production, with resultant tissue breakdown and increased prevalence of chronic infection in the ischaemic bone.^{Reference Pathak and Bryce5,Reference Ewing7,Reference Hao, Chen, Wei, Chen, Yeh and Su18}

Whilst some risk factors, such as diabetes mellitus, may be modifiable,^{Reference Herr, Vincent, Skotnicki, Ducic and Manolidis14} many, such as patient age, are not. Ramsden et al.^{Reference Ramsden, Bulman and Lorigan10} noted that the development of osteoradionecrosis was more frequently seen when the temporal bone was in close proximity to the focus of radiation.^{Reference Ramsden, Bulman and Lorigan10} The superficial location of the temporal bone and thin overlying soft tissue make this region particularly susceptible,^{Reference Ramsden, Bulman and Lorigan10} and unfortunately when treating malignancies in this region, especially of the ear, this focus cannot be changed.^{Reference Rudge1} Rudge^{Reference Rudge1} found that the extent of necrosis was proportional to the dose of radiation administered.^{Reference Rudge1} However, this was not supported by Pathek and Bryce^{Reference Pathak and Bryce5} or Sharon et al.^{Reference Sharon, Khwaja, Drescher, Gay and Chole11} In our study, patients were found on average to have received 45–60 Gy.

It has been recognised that there is a latency period between radiotherapy and the development of osteoradionecrosis.^{Reference Rudge1,Reference Pathak and Bryce5,Reference Ramsden, Bulman and Lorigan10} Pathek and Bryce^{Reference Pathak and Bryce5} reported an average latency of 8 years (range, 6–11 years) from radiation to the development of diffuse osteoradionecrosis,^{Reference Pathak and Bryce5} whilst Lovin et al.^{Reference Lovin, Hernandez, Elms, Choi, Lindquist and Moreno16} reported a mean time of 10 years.^{Reference Lovin, Hernandez, Elms, Choi, Lindquist and Moreno16} Our mean time to development of osteoradionecrosis post-radiation was shorter, at 2.8 years, but ranged from 1.2 to 15 years, again showing a wide variation. Currently there does not seem to be a relationship between the latency period and the severity of osteoradionecrosis.^{Reference Ramsden, Bulman and Lorigan10}

Consistent with our results, Yuhan et al.^{Reference Yuhan, Nguyen, Svider, Raza, Hotaling and Chan15} showed that otalgia and otorrhoea are the most common initial symptoms.^{Reference Yuhan, Nguyen, Svider, Raza, Hotaling and Chan15} The most commonly used classification for osteoradionecrosis is by Ramsden et al.^{Reference Ramsden, Bulman and Lorigan10} and divides cases into localised, where bone erosion is limited to the EAC, and diffuse, where it affects more of the ear and mastoid.^{Reference Ramsden, Bulman and Lorigan10,Reference Kammeijer, Van Spronsen, Mirck and Dreschler19}

As reflected in our protocol (Figure 3), it is widely established that limited disease associated with minimal symptoms should be managed conservatively with regular aural toilet and antibiotic therapy.^{Reference Ramsden, Bulman and Lorigan10,Reference Herr, Vincent, Skotnicki, Ducic and Manolidis14} This is particularly as the goal of treatment for localised disease is directed at symptom control^{Reference Lovin, Hernandez, Elms, Choi, Lindquist and Moreno16} rather than complete removal or resolution of the necrotic bone.^{Reference Yuhan, Nguyen, Svider, Raza, Hotaling and Chan15}

The algorithm proposed by Sharon et al.^{Reference Sharon, Khwaja, Drescher, Gay and Chole11} also recommended initial conservative management, with culture-directed topical antibiotics, topical antiseptics, periodic clinic debridement and pain management,^{Reference Sharon, Khwaja, Drescher, Gay and Chole11} as did Kammeijer et al.^{Reference Kammeijer, Van Spronsen, Mirck and Dreschler19} Yuhan et al.^{Reference Yuhan, Nguyen, Svider, Raza, Hotaling and Chan15} performed a systematic review on the management of osteoradionecrosis of the temporal bone and found that 89 per cent of all cases treated conservatively had adequately resolved presenting symptoms at last follow up.^{Reference Yuhan, Nguyen, Svider, Raza, Hotaling and Chan15}

Pathak and Bryce^{Reference Pathak and Bryce5} found that being over 60 years of age was an indicator of those who failed under conservative management within 2 years^{Reference Pathak and Bryce5} and these cases may therefore need closer monitoring and aural toileting. These patients are more likely to progress to stage 2 management (Figure 3), which we propose necessitates local debridement in the form of canalplasty, meatoplasty or mastoidectomy and consideration of hyperbaric oxygen therapy.

The systemic review performed by Yuhan et al.^{Reference Yuhan, Nguyen, Svider, Raza, Hotaling and Chan15} found that 21.5 per cent of cases received conservative management and 60.9 per cent underwent surgical management.^{Reference Yuhan, Nguyen, Svider, Raza, Hotaling and Chan15} Hyperbaric oxygen was used in conjunction with surgical treatment in 11.3 per cent of cases.^{Reference Yuhan, Nguyen, Svider, Raza, Hotaling and Chan15} Kammerijer et al.^{Reference Kammeijer, Van Spronsen, Mirck and Dreschler19} outlined their guidelines to reflect the localised and diffuse osteoradionecrosis described by Ramsden et al.^{Reference Ramsden, Bulman and Lorigan10} They suggest that when those with localised osteoradionecrosis are not symptomatically controlled, they should be managed as per diffuse A osteoradionecrosis, defined as computed tomography evidence supporting diffuse disease associated with little pain and infection, and intact functional hearing.^{Reference Kammeijer, Van Spronsen, Mirck and Dreschler19} They also suggest sequestrectomy in these cases.^{Reference Kammeijer, Van Spronsen, Mirck and Dreschler19}

Hyperbaric oxygen therapy aims to increase tissue oxygenation, which in turn promotes neovascularisation and wound healing.^{Reference Hao, Chen, Wei, Chen, Yeh and Su18} Sharon et al.^{Reference Sharon, Khwaja, Drescher, Gay and Chole11} recommended other therapies in the form of intravenous (IV) antibiotics and hyperbaric oxygen in patients who developed increasing pain, progressive infection and cranial neuropathies despite conservative management.^{Reference Sharon, Khwaja, Drescher, Gay and Chole11} Our algorithm includes IV antibiotics as an option in stage 1, dependent on clinical findings. There is some evidence from small randomised controlled trials to support hyperbaric oxygen therapy in late radiation-induced tissue injury.^{Reference Bennett, Kaylie, Warren and Jackson20} Reported clinical outcomes have been variable, with Sharon et al.^{Reference Sharon, Khwaja, Drescher, Gay and Chole11} finding that of their six patients who received hyperbaric oxygen therapy, no patients achieved resolution of the necrotic bone. It was, however, felt to aid post-operative healing in one patient.^{Reference Sharon, Khwaja, Drescher, Gay and Chole11}

It should be noted that although hyperbaric oxygen is often incorporated into treatment algorithms for osteoradionecrosis, evaluation of its benefit is currently limited by small patient cohorts, concurrent surgical management and varied treatment protocols.^{Reference Herr, Vincent, Skotnicki, Ducic and Manolidis14} The use of hyperbaric oxygen is affected by geographical limitations with only eight chambers across England.²¹ It is also a significant undertaking for patients both physically and mentally, with the most common protocol involving 40 treatments,^{Reference Forner, Dieleman, Shaw, Kanatas, Butterworth and Kjeller22} each typically lasting 90 minutes.²³ We propose a more manageable 10 treatments. The risks of treatment include temporary visual problems, Eustachian tube dysfunction and seizures.^{Reference Shaw, Butterworth, Silcocks, Tesfaye, Bickerstaff and Jackson24}

Whilst conservative management has been shown to be appropriate for patients with limited disease, it may not be sufficient, and refractory disease may arise. Sharon et al.^{Reference Sharon, Khwaja, Drescher, Gay and Chole11} found that 18 out of 33 patients in their study went on to require surgical management because of intractable pain, persistent infection or the development of cholesteatoma.^{Reference Sharon, Khwaja, Drescher, Gay and Chole11} Our study showed a mean time between medical therapy and surgery with curative intent of 21 months.

Kammerijer et al.^{Reference Kammeijer, Van Spronsen, Mirck and Dreschler19} reserved subtotal petrosectomy for those with diffuse B disease, who they define as those in whom there is severe pain and infection and/or no functional hearing because of the risk–benefit balance of the more extensive surgery.^{Reference Kammeijer, Van Spronsen, Mirck and Dreschler19} Yuhan et al.^{Reference Yuhan, Nguyen, Svider, Raza, Hotaling and Chan15} found that less than 60 per cent of mastoidectomies led to complete resolution, but over 90 per cent of lateral temporal bone resections resolved.^{Reference Yuhan, Nguyen, Svider, Raza, Hotaling and Chan15} Sharon et al.^{Reference Sharon, Khwaja, Drescher, Gay and Chole11} specified that the aim of surgery was to gain symptom control rather than complete removal of necrotis bone.^{Reference Pathak and Bryce5,Reference Sharon, Khwaja, Drescher, Gay and Chole11} This view was justified in that it has been shown that there are increased risks of operating in radiated temporal bone, including higher than expected rates of facial nerve dehiscence, oval window and lateral canal fistulae, dural exposure, CSF leak and lateral canal procedures.^{Reference Bennett, Kaylie, Warren and Jackson20} These risks are thought to occur because of the poor blood supply to the EAC, which is then further decreased by surgery.^{Reference Pathak and Bryce5} However, it should be noted that our algorithm differs in its consideration and recommendation of reconstruction in appropriate cases.

As in our advanced cases, the symptom severity and potentially life-threatening nature of diffuse disease makes more aggressive management necessary.^{Reference Kadakia, Badhey, Inman, Mourad and Ducic3,Reference Pathak and Bryce5,Reference Ramsden, Bulman and Lorigan10} Flexibility from local flaps or free flaps is needed to reconstruct defects of the lateral temporal bone. With advancing techniques in reconstruction, cosmesis is now of increasing focus rather than simply covering the resultant defect,^{Reference Kadakia, Badhey, Inman, Mourad and Ducic3} and this is reflected in stage 3 of our algorithm. However, significant complications continue to be reported in those undergoing such surgery, including persistently discharging fistulae.^{Reference Sharon, Khwaja, Drescher, Gay and Chole11}

Free flaps can provide much better functional and aesthetic results compared with local and regional flaps.^{Reference Lóderer, Vereb, Paczona, Janovszky and Piffkó25} However, there are a number of considerations when selecting an appropriate flap. These include the size and location of the defect, the blood supply of both the flap and the recipient site, and, importantly, the aim of the reconstruction in terms of aesthetic and functional outcomes.^{Reference Lóderer, Vereb, Paczona, Janovszky and Piffkó25}

The anterolateral thigh chimeric flap, which relies on the descending branch of the lateral circumflex femoral artery, has been described for use in the reconstruction of large soft tissue defects in this region and to correct facial palsy.^{Reference Lóderer, Vereb, Paczona, Janovszky and Piffkó25} This flap is easily obtained and offers sufficient muscle bulk to fill large defects. There is also access to redundant motor nerves such as a branch of the motor nerve to the vastus lateralis which is suitable for grafting.^{Reference Revenaugh, Knott, Scharpf and Fritz26} It is also possible to alter the thickness of the subcutaneous fat in the anterolateral region to achieve appropriate flap thickness at the reconstruction site.^{Reference Ren, Wu, Wang, Zhang, Tan and Gong27} This mitigates the complications of fat liquefaction and seroma associated with fat transfer and dermal grafts.^{Reference Lóderer, Vereb, Paczona, Janovszky and Piffkó25} Lóderer et al.^{Reference Lóderer, Vereb, Paczona, Janovszky and Piffkó25} reported that there was minimal morbidity at the donor site.^{Reference Lóderer, Vereb, Paczona, Janovszky and Piffkó25} This single-stage approach is generally associated with fewer complications and better neural regeneration than a multistep approach.^{Reference Lóderer, Vereb, Paczona, Janovszky and Piffkó25} Direct nerve anastomosis or cable grafting is the preferred reconstruction technique, yielding the best functional outcomes in patients undergoing facial nerve sacrifice.^{Reference Lóderer, Vereb, Paczona, Janovszky and Piffkó25}

Our surgically treated patients in this report obtained good symptomatic relief and long-term control of osteoradionecrosis. However, our findings are based on a small cohort with heterogenicity in presentation, making it difficult to draw overall conclusions. To the best of our knowledge, no multicentre study has yet been performed, and given the rarity of this presentation it is unlikely that this will be undertaken in the near future. In addition, consideration of the interpretation of outcomes is needed. Consensus is required on what constitutes successful treatment, be it complete eradication of exposed bone, resolution of symptoms or reduction of symptoms to acceptable levels.^{Reference Sharon, Khwaja, Drescher, Gay and Chole11}

Osteoradionecrosis is a rare and difficult clinical entity requiring co-ordinated MDT input. The key guiding principle is to restore vascularity to the dying bone and facial nerve. Outcomes are promising, with 50 per cent of cases now disease-free. We advocate the protocol described here as part of a multidisciplinary approach to this complex condition. Larger multicentre studies may provide a better understanding of the optimum treatment strategy in any individual patient.