Chemoradiation-induced hearing loss remains a major concern for head and neck cancer patients

Abstract

Objective: Review of the literature regarding hearing loss in patients with head and neck cancer treated with chemoradiation. Design: Studies in the literature are reviewed that pertain to hearing loss sustained in head and neck cancer patients receiving cisplatin-based chemoradiation. Personal observations noted while treating these patients are also detailed. Study sample: PubMed was searched for pertinent articles regarding hearing loss in head and neck cancer patients receiving cisplatin chemotherapy and/or radiation. Results: Studies on the incidence and severity of hearing loss in head and neck cancer patients are limited, but those studies suggest that the risk of hearing loss is greater with higher-dose regimens. Conclusions: Newer cisplatin chemotherapy regimens using lower, weekly doses may be associated with a lower incidence and severity of hearing loss; however, large prospective studies are needed. Such information will be paramount to effective pre-treatment counselling of head and neck cancer patients.

Key Words:

• Conditions/pathology/disorders
• hearing conservation/hearing loss prevention
• adult or general hearing screening
• demographics/epidemiology

Introduction

Cisplatin is a cytotoxic chemotherapy drug that is widely used for head and neck squamous cell carcinoma (HNSCC) and other malignancies. It has several adverse effects including nephrotoxicity, myelosuppression, nausea/vomiting, neuropathy, and ototoxicity. Cisplatin-induced ototoxicity results in high-frequency sensorineural hearing loss (SNHL), with damage to the outer hair cells and stria vascularis of the organ of Corti (Theunissen et al, 2015a). For HNSCC, cisplatin is given as concomitant therapy along with radiation, either as primary treatment or adjuvant treatment following surgical resection of the tumour. As seen in patients with other tumour types, the degree of cisplatin-induced SNHL in HNSCC patients correlates with the cumulative dose received (Zuur et al, 2007a). While much has been written on cisplatin-induced hearing loss in paediatric patients and patients with other tumour types, in this review we focus specifically on the hearing outcomes of head and neck cancer patients who undergo cisplatin-based chemoradiation. This group of patients is distinct from patients with other disease types who receive cisplatin, since the hearing loss induced by cisplatin may be exacerbated by concomitant radiation to the cochlea.

Existing studies on cisplatin ototoxicity in head and neck cancer patients

To summarise existing knowledge on hearing loss related to cisplatin-based chemotherapy and/or radiation in head and neck cancer patients, an extensive PubMed search was performed using the following search terms in several combinations: head and neck cancer, cisplatin, radiation, hearing loss, ototoxicity, and tinnitus. Articles that were not in English or included fewer than five patients were excluded. Given that the vast majority of head and neck cancer involves adults, no studies involving children were found.

Existing studies on cisplatin-induced hearing loss are limited in number, and most include small numbers of patients. Comparison is difficult due to variations in cancer therapy and methodology used to define ototoxicity (see Brewer and King in this edition). For example, the American Speech-Language-Hearing Association (ASHA) criteria define ototoxic hearing loss as a threshold shift >20 dB in one frequency or >10 dB in two or more frequencies up to 12 kHz (American Speech-Language-Hearing Association, 1994). The Common Terminology Criteria for Adverse Events (CTCAE) version 4.03 criteria define ototoxic hearing loss as a threshold shift >15 dB in two or more frequencies up to 8 kHz (National Cancer Institute, 2010). Thus, use of the ASHA criteria may enable detection of ototoxic hearing loss in a single high frequency with more sensitivity than CTCAE criteria. Despite these limitations and sources of variability, some information on hearing loss related to cisplatin chemoradiation in HNSCC patients is available, summarised here and in Table 1.

Table 1. Summary of pertinent studies on hearing loss and tinnitus in head and neck cancer patients treated with platinum chemotherapy and/or radiation.

Study	Number of patients	Study design	Hearing loss criteria	Tumour sites included	Treatments	Hearing loss reported	Tinnitus assessed?
Bhandare et al (2007)	325	Retrospective review	≥15 dB threshold shift	Sinonasal, nasopharynx	RT, CRT (chemotherapy unspecified)	RT: 18% incidence CRT: 30% incidence	No
Cheraghi et al (2015)	29	Prospective study	CTCAE	Sinonasal, nasopharynx, parotid salivary	RT, CRT (low-dose cisplatin)	RT: 51% incidence CRT: 77% incidence	No
Hitchcock et al (2009)	62	Prospective study using mathematical modelling	≥10 dB threshold shift	Sinonasal, nasopharynx, oral cavity, oropharynx, larynx, hypopharynx	RT, CRT (low-dose or high-dose cisplatin)	Low-dose cisplatin:<10 Gy RT: 9.5 dB at 8 kHz40 Gy RT: 18.9 dB at 8 kHz High-dose cisplatin:<10 Gy RT: 21.5 dB at 8 kHz40 Gy RT: 38.4 dB at 8 kHz	No
Hwang et al (2015)	92	Prospective study	Threshold shifts recorded over time	Nasopharynx	RT, CRT (cisplatin, regimen not specified)	RT: 19.9 dB mean threshold shift byAC, 10.6 dB threshold shift by BC CRT: 14.3 dB mean threshold shift byAC, 9.6 dB threshold shift by BC	No
Ishikawa et al (2015)	18	Prospective study	Threshold shift >20 dB at one frequency	Sinonasal, nasopharynx, oropharynx, larynx, hypopharynx	RT with cisplatin (intra-arterial + sodium thiosulphate or high-dose intravenous)	RT + intra-arterial cisplatin: hearingloss criteria noted for meanthreshold shifts at 8 and 10 kHz RT intravenous, high dose cisplatin:hearing loss criteria noted for meanthreshold shifts at 8, 10 and 12 kHz	Yes
Lee et al (2015)	211	Prospective study	LENT-SOMA criteria for tinnitus	Nasopharynx, larynx, hypopharynx, oropharynx, oral cavity	RT	11.6% developed tinnitus	Yes
Low et al (2006)	115	Randomised, single-blinded, controlled trial of RT vs. CRT	Change in median thresholds for treatment group	Nasopharynx	RT vs. CRT (high-dose cisplatin)	CRT group had worse median thresholds at 0.5, 1, 2, and 4 kHz	No
Mujica-Mota et al (2013)	N/A	Systematic review of 20 publications	Variable	Primarily nasopharynx, parotid salivary	RT (variable techniques)	Incidence ranged from 0 to85% from 0.5 to 4 kHz, 27–95% in frequencies >4 kHz	No
Niemensivu et al (2016)	22	Prospective study	>10 dB threshold shift at 4 and 8 kHz or in PTA	Sinonasal, nasopharynx, oral cavity, oropharynx, larynx, hypopharynx	CRT (low-dose cisplatin)	13.6% incidence	Yes
Pan et al (2005)	35	Prospective study	Mean threshold shifts	Sinonasal, salivary gland, nasopharynx, oral cavity, oropharynx, skin	RT, CRT (chemotherapy not specified)	Mean threshold shifts of 8 dB at 2–4 kHz, 17.5 dB at 8 kHz after 24 months	No
Pearson et al (2006)	37 (15 with pre-/post-treatment audiograms)	Retrospective study	10 dB or greater threshold shift in PTA or at 4 or 8 kHz	Sinonasal, nasopharynx, oral cavity, oropharynx, larynx, hypopharynx	CRT (high-dose cisplatin)	53% incidence of change in PTA 100% incidence of change at 4 kHz 87% incidence of change at 8 kHz	Yes
Rades et al (2008)	128	Retrospective study	CTCAE	Nasopharynx, oral cavity, oropharynx, larynx, hypopharynx	CRT (high-dose cisplatin or cisplatin/5-FU)	Two patients in high-dose cisplatin group with grade 2–3 ototoxicity	No
Theunissen et al (2014)	158 (67 with data at 4.5 years)	Prospective study	CTCAE	Oral cavity, oropharynx, hypopharynx	CRT: intra-arterial or intravenous cisplatin	Intra-arterial cisplatin: 68% incidence (4.5 years) Intravenous cisplatin: 81% incidence at 4.5 years	No
Theunissen et al (2015a)	N/A	Systematic review of 21 publications	Variable	Variable	RT, CRT (variable)	Incidence ranging from 17 to88% depending on study and criteria used	No
Theunissen et al (2015b)	156	Retrospective study to create predictive model of hearing loss	BC hearing thresholds	Not specified	CRT (high-dose cisplatin)	N/A: predictive model created to predict hearing loss based on baseline thresholds, cumulative cisplatin dose and radiation dose to the cochlea	No
Zuur et al (2007a)	146	Retrospective study	Threshold shift ≥5 dB in PTA	Oral cavity, oropharynx, larynx, hypopharynx	CRT (intra-arterial cisplatin + sodium thiosulphate)	20% incidence	No
Zuur et al (2007b)	158	Randomized phase III trial	CTCAE	Oral cavity, oropharynx, hypopharynx	CRT (intra-arterial cisplatin vs. intravenous cisplatin + sodium thiosulphate)	Intra-arterial cisplatin: 91% incidence Intravenous cisplatin: 88% incidence	No

RT: radiation therapy; CRT: chemoradiation therapy; AC: air conduction; BC: bone conduction; PTA: pure tone average.

High-dose cisplatin regimens (100 mg/m²)

Traditionally, patients have been treated with radiation and cisplatin at a dose of 100 mg/m² every three weeks for a total of three doses, based on large clinical trials showing that this cisplatin regimen improves patient survival (Bernier et al, 2004; Cooper et al, 2004). This cisplatin regimen has been found to induce hearing loss in a large proportion of patients. The reported incidence of hearing loss after two to three doses of cisplatin at 100 mg/m² ranges from 17 to 88%, depending on how the hearing loss is defined in different studies (Pearson et al, 2006; Zuur et al, 2007b; Theunissen et al, 2015a; Niemensivu et al, 2016). A systematic review showed significant heterogeneity among studies in tumour subsite and treatment regimens, making it difficult to determine the true incidence and severity of cisplatin ototoxicity in patients with HNSCC (Theunissen et al, 2015a). One study of 13 patients treated with this chemoradiation regimen found that threshold shifts at 8 kHz ranged from 21.5 to 38.4 dB, depending on the dose of radiation (Hitchcock et al, 2009). A study including 61 patients found that 10% developed grade 2–3 hearing loss, defined as >25 dB shift at 2–3 frequencies for monitored patients or subjective hearing loss in non-monitored patients (CTCAE v4). The proportion of these patients who actually had baseline and post-treatment audiograms was not reported (Rades et al, 2008). Another report showed that more than half of patients treated with high-dose cisplatin chemoradiation had a 30 dB or greater hearing loss (threshold shift) at 4 and 8 kHz (Pearson et al, 2006). In a small cohort of nine patients, five had disturbing tinnitus that interfered with daily activities (Niemensivu et al, 2016). Factors that may predict more severe hearing loss after cisplatin chemoradiation include age, pre-existing hearing loss, total dose of cisplatin, and the total radiation dose to the cochlea. The dose of radiation reaching the cochlea may be higher in patients with cancer of the nasopharynx, parotid salivary gland, and paranasal sinuses than with other subsites such as the oral cavity, oropharynx, or larynx (Hitchcock et al, 2009; Theunissen et al, 2015a, 2015b).

Low-dose cisplatin regimens (40 mg/m²)

Recently, many oncologists have shifted to a different chemoradiation paradigm for HNSCC, in which cisplatin is given at a lower dose of 40 mg/m² weekly for six or seven doses along with radiation. This paradigm shift has gained popularity due to anecdotal reports of lower toxicity, despite the absence of any large trials directly comparing toxicity or efficacy of low-dose, weekly cisplatin with the traditional regimen of three higher doses. However, recent retrospective studies suggest that the lower-dose regimen is equally effective (Melotek et al, 2016). Radiation regimens have also changed, with most centres now offering intensity-modulated radiation therapy (IMRT) to limit off-target adverse effects of radiation, such as hearing loss related to radiation of the cochlea. A recent study of patients receiving low-dose, weekly cisplatin with IMRT for HNSCC showed no hearing loss at low frequencies and only minor threshold shifts (0–20 dB, mean 5.2 dB) at 4–8 kHz. Less than half of patients experienced tinnitus, which was slight to moderate and did not interfere with daily activities (Niemensivu et al, 2016). However, only 22 patients were included in this study, and 10 patients (45%) were unable to complete all six cycles of low-dose cisplatin (Niemensivu et al, 2016). Another small study of nine patients treated with three-dimensional conformal radiation therapy and five or more weekly doses of cisplatin showed that seven patients (77%) had SNHL by CTCAE criteria, and the other two patients had detectable threshold shifts only at the high frequency of 12 kHz (Cheraghi et al, 2015). An additional study including 27 HNSCC patients treated with low-dose, weekly cisplatin and radiation showed threshold shifts of 9.5–18.9 dB at 8 kHz, depending on the radiation dose. However, this study gave minimal information on comparison to baseline audiograms (Hitchcock et al, 2009). All three of these studies included patients with carcinomas of the nasopharynx and paranasal sinuses, who may have increased hearing loss related to increased radiation to the cochlea and/or Eustachian tube dysfunction (Hitchcock et al, 2009; Niemensivu et al, 2016).

Other regimens

Though not commonly used in the United States, several other countries have adopted the use of intra-arterial cisplatin for HNSCC. This strategy consists of infusing cisplatin directly into the artery feeding the tumour, and sodium thiosulphate is often given intravenously at the same time to limit the systemic toxicities of cisplatin (Zuur et al, 2007a, 2007b; Ishikawa et al, 2015). One small recent study and a previous randomised, phase II clinical trial comparing this regimen to standard high-dose intravenous cisplatin and radiation in HNSCC found that the resulting SNHL was less severe in the patients receiving intra-arterial cisplatin and sodium thiosulphate, with lower threshold shifts overall and a tendency to affect only the highest frequencies (10–12 kHz; Zuur et al, 2007b; Ishikawa et al, 2015). This cisplatin regimen is technically more challenging, and large studies comparing the chemotherapeutic efficacy of this regimen to standard intravenous regimens are not available.

In patients who are unable to tolerate cisplatin due to kidney disease, severe pre-existing hearing loss or other factors, alternative agents such as carboplatin or cetuximab are given with radiation in HNSCC. Carboplatin is another platinum-based chemotherapeutic drug that has been found to cause ototoxicity in similar fashion to cisplatin in animal studies (Husain et al, 2005). Clinical studies on SNHL with carboplatin in HNSCC are limited, with most performed in the 1990s in conjunction with outdated radiation techniques; these studies suggest that ototoxicity does occur with carboplatin, but in less severe fashion than with cisplatin (Bauer et al, 1992; De Serdio et al, 1997). Carboplatin is anecdotally noted to be less toxic to the inner ear and kidney than cisplatin, but the efficacy of this drug in comparison to cisplatin is unknown. A recent meta-analysis of studies using cisplatin or carboplatin found no statistically significant difference in response rate between the two drugs, though the study may have been underpowered (Aguiar et al, 2016).

Cetuximab is a targeted monoclonal antibody against the epidermal growth factor receptor, which has been FDA approved for treatment of HNSCC and also used as an alternative to cisplatin. To our knowledge, hearing loss related to cetuximab has not been reported. Several studies [reviewed by Husain et al (2016)] suggest that the efficacy of cetuximab may be inferior to that of cisplatin, thus cetuximab is primarily used in patients with kidney disease or other conditions that prevent the use of traditional platinum chemotherapy.

Radiation to the cochlea as an exacerbating factor

Radiotherapy to the middle and inner ear structures, including the cochlea, Eustachian tube, and mastoid, may mechanistically result in a radiation-related decline in hearing. Radiotherapy-related ototoxicity may be worse in patients with tumours of the nasopharynx, skull base, and paranasal sinuses than in those with tumours involving the oral cavity, larynx or pharynx, due to high dose regions adjacent to the cochlea (Hitchcock et al, 2009; Walker et al, 2011; Niemensivu et al, 2016). One study of nasopharyngeal carcinoma (NPC) patients showed that radiation alone caused significant hearing loss, possibly related to Eustachian tube dysfunction and sequelae of otitis media, though SNHL was worse in patients with NPC who also received cisplatin (Low et al, 2006). However, with the advent of IMRT, doses to the cochlea and surrounding structures are now substantially lower. Another study illustrated that surrounding structures are also important. Patients with NPC treated with cisplatin chemoradiation tend to have a mixed conductive/sensorineural hearing loss, illustrating that both middle and inner ear dysfunction may result with use of radiation near the Eustachian tube and cochlea, respectively (Hwang et al, 2015). In a study of radiation for skull base tumours, a dose of >30 Gy was associated with chronic otomastoid opacification (Walker et al, 2011).

Dose constraints for the cochlea to avoid tinnitus have been largely under-reported. At least one normal tissue complication probability model utilising information from 211 patients suggests limiting the cochlear dose to <32 Gy for a 20% probability of clinically relevant tinnitus, which was defined to include intermittent, persistent or refractory tinnitus (Lee et al, 2015). Hitchcock et al (2009) noted in their study of 62 patients with HNSCC that a cumulative radiation dose to the cochlea of 40 Gy was associated with SNHL, whereas radiation doses less than 40 Gy did not seem to cause significant hearing loss. The hearing loss was more severe in patients also receiving cisplatin. In another study of 29 patients with HNSCC, three-dimensional conformal radiation therapy alone caused SNHL in half of the patients, and in 77% of patients who also received cisplatin (Cheraghi et al, 2015). A systematic review of patients with head and neck cancer treated with radiation alone found the incidence of SNHL to be highly variable, depending on the site of the tumour and the radiation technique utilised, but was more common with doses >45 Gy (Mujica-Moto et al, 2013). In summary, treatment of HNSCC patients with radiation can result in sensorineural and conductive (mixed) hearing loss, which may be more severe in patients with tumours near the ear or in patients who are also treated with cisplatin.

With more widespread use of IMRT, SNHL is expected to be less common. In a series of 36 patients treated with IMRT alone for head and neck cancer, many with anatomic sites near the ear, showed average threshold shifts of only 1.8, 2.9, and 7.3 dB at low, high and ultra-high frequencies, respectively (Theunissen et al, 2014). A larger study of 101 patients treated with IMRT alone showed minimal SNHL in most patients, but 18% of patients had a loss of 10 dB or greater at 8–12.5 kHz (Zuur et al, 2009). While IMRT is now the standard of care at most tertiary care medical centres, it is not yet available in all parts of the world. Thus, some patients may still receive significant radiation doses to the cochlea. More work is needed to accurately characterise dose-volume constraints for radiotherapy to the inner ear structures.

Clinical observations from our oncology practices

Most patients with HNSCC seen in our multidisciplinary practice receive the low-dose, weekly regimen of cisplatin (40 mg/m²) with concurrent IMRT. Administration of chemotherapy in this fashion allows for the ability to receive a majority of the cumulative dose of chemotherapy in case one is missed due to toxicity. It is not unusual for us to see patients who complain of tinnitus, often after the first dose. We also continue to see a few patients with clinically relevant, permanent hearing loss affecting their quality of life. Some patients also experience serous otitis media from radiotherapy, which may require intervention or drainage. We hypothesise that the newer cisplatin regimen and IMRT are largely responsible for a decreased incidence in SNHL in our patients, but this deserves further study with late-responding tissue reaction. With regard to radiotherapy-related dose volume constraints, for patients with base of skull or nasopharynx tumours we follow QUANTEC recommendations (Marks et al, 2010) for <30% cochlea volume getting less than 45 Gy. For tumours that do not require more than prophylactic dose close to the base of skull, we use a 30 Gy limit, or even a 10 Gy limit where possible, taking into account the patient’s age, baseline hearing function, need for cisplatin, and availability of hearing rehabilitation options (Pan et al, 2005; Bhandare et al, 2007; Hua et al, 2008). We do query our patients about pre-existing hearing loss and counsel them regarding the risks of hearing loss with cisplatin. Our ability to counsel patients would be improved if there were more data in the literature on the incidence and severity of hearing loss with these new treatment paradigms. We routinely arrange for pre-treatment and post-treatment audiograms in the majority of patients, so that patients with hearing loss can be offered follow up and hearing aid evaluation. To study this in more rigorous fashion, we have recently opened a prospective clinical to track hearing before and after low-dose cisplatin chemoradiation in head and neck cancer patients at Johns Hopkins University and the National Institutes of Health.

Conclusions

In summary, the paradigm for cisplatin-based chemoradiation for HNSCC is changing, in an attempt to limit toxicities and increase the proportion of patients who can successfully complete the recommended course of treatment. It has been suggested based on small studies and anecdotal evidence that this new cisplatin regimen, combined with modern radiotherapy techniques such as IMRT, may have a lower risk for causing sensorineural hearing loss. This finding needs to be explored in a larger, prospective study. Definitive information on the incidence and severity of SNHL in head and neck cancer patients treated with modern cisplatin chemoradiation regimens would enable us to improve our pre-treatment counselling of patients on their risk of sustaining treatment-related hearing loss and associated changes in quality of life. In the meantime, pre-treatment and post-treatment audiograms should be offered to all HNSCC patients treated with chemoradiation, and these patients should be counselled routinely about the possibility of treatment-related hearing loss.

Abbreviations
ASHA	=	American Speech-Language-Hearing Association
CTCAE	=	Common Terminology Criteria for Adverse Events
FDA	=	US Food and Drug Administration
HNSCC	=	head and neck squamous cell carcinoma
IMRT	=	intensity-modulated radiation therapy
NPC	=	nasopharyngeal carcinoma
QUANTEC	=	Quantitative Analysis of Normal Tissue Effects in         the Clinic
SNHL	=	sensorineural hearing loss

Declaration of interest

No potential conflict of interest was reported by the authors.

Funding

This work was supported by NIH, National Institute on Deafness and Other Communication Disorders intramural project number ZIA-DC-DC000090.