Introduction

Ototoxicity, which can manifest as hearing impairment, tinnitus, and/or vertigo, is a recognized adverse effect associated with a group of antitumor therapies, including platinum chemotherapy, radiotherapy, or surgery involving the ear and auditory nerves [1]. Hearing impairment can lead to communication and social difficulties, ultimately reducing the quality of life. In children, hearing impairment can severely impair cognitive development as well as language and social skills. Although the structure of the human ear is formed at birth, the maturation of neural pathways and auditory structures continues during infancy and early childhood, making young children particularly vulnerable to radiotherapy-induced ototoxicity [2].

Radiation therapy (RT), which can be utilized as a single treatment modality or as an adjuvant treatment before and after surgery, is commonly used to treat patients with a variety of cancers. Patients with locally advanced and inoperable head and neck tumors are usually treated with cisplatin-based chemoradiotherapy. Cisplatin or carboplatin-based chemotherapy drugs are known to cause ototoxicity [3]. Winther et al. discovered that inner ear radiation in guinea pigs resulted in extensive degeneration of hair cells outside Corti organs. Concurrently, radiation therapy to the temporal bone led to Corti organ damage and auditory vestibular nerve atrophy [4]. In radiation therapy for head, neck, or brain malignancies, the middle ear, inner ear, and brainstem may be exposed to high doses of ionizing radiation [5]. The underlying physiological processes leading to hearing impairment may vary depending on the location of radiation-induced lesions. If hearing impairment arises from damage to middle ear components, such as eustachian tubes or ossicles, it is classified as conductive. In contrast, sensorineural hearing loss (SNHL) results from lesions in the cochlea or the auditory system’s posterior section [6].

Despite the prevalence and severity of ototoxicity following radiotherapy, it has seldom been reported in radiation oncology literature. The correlation between cochlear radiation dose and subsequent morbidity has rarely been documented. The objective of this study is to assess the incidence of various factors that may contribute to radiotherapy-induced ototoxicity.

Methods

This meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) [7].

Data sources and searches

Databases, including the Cochrane Library, PubMed, Embase, and Web of Science, were searched from inception until January 2023. Medical Subject Headings (MeSH) and text word combinations were employed to create three subsets of references: the first subset encompassed radiotherapy (such as intensity-modulated radiotherapy, proton radiotherapy, carbon ion radiotherapy, photon radiation, gamma knife, stereotactic radiotherapy, etc.); the second subset involved complications related to hearing (namely, ototoxicity, hearing impairment, and hearing loss); and the third subset pertained to cancer. After an initial screening of titles or abstracts, two independent reviewers (YH, HZ) assessed the full text of relevant publications and the reference lists for final inclusion. Additionally, references considered potentially relevant were searched and thoroughly evaluated.

Study selection

Studies were included based on the following criteria: (1) studies that reported hearing impairment in cancer patients due to RT as a first-line treatment; (2) hearing outcomes obtained from pure tone audiograms (either air and bone conduction or bone conduction alone) conducted before and after treatment; (3) studies providing the number of individuals evaluable for toxicity following radiotherapy and the number of individuals with hearing impairment; (4) studies that clearly defined hearing impairment and offered sufficient irradiation information to quantify the effect; and (5) studies that were randomized controlled trials, excluding one-arm trials. All criteria needed to be met for study inclusion. Exclusion criteria encompassed postoperative studies, single-arm studies, case reports, reviews, meeting minutes or abstracts, articles not published in English, and studies with cisplatin as monotherapy.

Data extraction

Two evaluators independently employed standardized forms to extract and summarize the following data: first author, year of publication, study ID, country, cancer type, radiotherapy design, radiotherapy mode, cochlear radiation dose, total number of patients, number of patients for safety analysis, standard version of general terms for adverse events, rate of hearing impairment, and frequency of tinnitus and vertigo symptoms. The standard for general terms of adverse events served as the most commonly used tool to evaluate the type and severity of adverse events in clinical practice, featuring a grading scale and clear definitions.

Quality assessment

Two reviewers (AZ, JW) evaluated the risk of bias based on the original studies, utilizing the Cochrane Collaboration’s tool. Five aspects of adequacy were assessed: random sequence generation, allocation concealment, blinding, outcome assessment, and outcome reporting [8]. Each item was assigned an assessment indicator related to risk of bias, classified as yes, no, or unclear. Any disagreements regarding study selection, data extraction, and quality assessment were resolved through discussions with statistical experts [9].

Data synthesis and statistical analysis

Meta-analysis was performed using R(4.2.1) statistical software (metafor and meta package). Fixed-effect or random-effects models were employed to estimate event rates and their corresponding 95% confidence intervals. Forest plots were constructed to summarize data and incidence for each analysis group. The Cochran Q statistic and the I2 statistic were utilized to assess statistical heterogeneity [10]. When I2 exceeded 25%, 50%, or 75%, it indicated low, medium, or high heterogeneity, respectively. If significant heterogeneity was present, a random-effects model was used. A simple analysis of funnel plots offered a useful test for possible bias in meta-analyses [11]. Otherwise, a fixed-effects model was applied. Meta-Analyst was used to generate pooled rates of different ototoxic events for treatment. Subgroup analysis was conducted according to median cochlear radiation dose. Finally, sensitivity analysis was performed to evaluate the stability of the results.

Results

Systematic review and characteristics

After 5192 duplicates were deleted and filtered by title and abstract, 286 of the 3134 records initially searched were reviewed in full. Due to insufficient data or lack of full text in meta-analysis, we excluded 20 studies. Finally, 28 eligible studies were included, including 25 prospective randomized controlled trial and 3 retrospective randomized controlled trials [12,13,14]. The radiotherapy modes included IMRT (intensity-modulated radiotherapy), SRT (Stereotactic radiotherapy), 3D-CRT (three-dimensional conformal RT), Conventional-RT, Proton-RT, Radiosurgery, HFRT (hyperfractionated RT), ART (accelerated RT). The median age ranged from 3 to 87. Patients under 18 years old participated in 3 studies, and the median age of the adult groups was greater than 18 years old. Twelve countries were included in the study: China (n = 9), U.S.A (n = 7), UK (n = 2), Sweden (n = 2), Canada (n = 1), Germany (n = 1), Japan (n = 1), Thailand (n = 1), Norway (n = 1), Singapore (n = 1), Spain (n = 1), Netherlands (n = 1). CTCAE (Common Terminology Criteria for Adverse Events) grading system was used to define ototoxic effects in 15 studies (i.e., Grade 1: Threshold shift of 15–25 dB averaged at two contiguous frequencies; Grade 2: Threshold shift of > 25 dB averaged at two contiguous frequencies; Grade 3: Threshold shift of > 25 dB averaged at three contiguous frequencies; Grade 4: > 80 dB at 2 kHz and above). Two study used Brock criteria to evaluate ototoxicity (i.e., Grade 0 to 1: < 40 dB on all frequencies or ≥ 40 dB at 8 kHz; Grade 2: ≥ 40 dB at 4 kHz; Grade 3 to 4: ≥ 40 dB at 2-1 kHz). Two studies used Gardner Robertson scale to judge ototoxicity. One study used the Pediatric Oncology Group (POG) to define the effects of ototoxicity. More information about the included study population and programs has been listed in Table 1.The detailed process of retrieval was shown in Fig. 1.

Table 1 Characteristics of included documents
Full size table
Fig. 1
figure 1

Flow chart of document screening

Full size image

Risk of bias assessment

The risk of bias was assessed for each included study (Additional file 1: Fig. S1). The overall risk of bias was low. Two studies did not mention the randomization process [18, 26, 34], while the third study did not conceal selective reporting bias [34]. Another study did not report measurements or determinations of whether results differed between experimental groups [30]. Despite these inclusions, some concerns regarding the risk of bias remained.

Ototoxicity

Hearing impairment

The 28 included studies compared the risk of all levels of hearing impairment effects in cancer patients receiving first-line therapy as radiation therapy (Fig. 2). The ratio between the experimental group and the control group under the random-effect model was 0.85 (95% CI, 0.71–1.00; P < 0.01; I2 = 75%). All RCTs were combined to compare the ratio of the experimental arm to the control arm, and the heterogeneity was high. Since the experimental design of each included RCT varied, subgroup analyses were performed based on population characteristics, original tumor, radiotherapy modality, and mean cochlear dose to explore potential sources of heterogeneity.

Fig. 2
figure 2

Summary total hearing impairment for all included studies

Full size image
Subgroup analysis of the association of RT with hearing impairment by irradiation design mode

Four trials involved the combination of Cetuximab and RT compared to the combination of Cisplatin and RT (Fig. 3A). Three trials were designed with Cetuximab + IMRT versus Cisplatin + IMRT, and one trial with Cetuximab + Conventional-RT versus Cisplatin + Conventional-RT. The combined OR (Odds Ratio) value was 0.42 (95% CI, 0.29–0.6; P = 0.65; I2 = 0%). The result suggested that Cetuximab combined with radiotherapy resulted in lower hearing impairment than Cisplatin combined with RT.

Fig. 3
figure 3

Subgroup Analysis of the Association of RT with Hearing impairment by Irradiation design mode

Full size image

Two trials compared ART with conventional-RT for ototoxicity (Fig. 3B). Pooled results suggest that hearing impairment after ART irradiation may be more severe than conventional radiotherapy (OR, 1.30, 95% CI, 0.72–2.35; P = 0.36; I2 = 0%). Two studies demonstrated the combination of ART and chemotherapy compared to conventional radiotherapy and chemotherapy (Fig. 3C). The results of the forest plot also showed that ART combined with chemotherapy is more ototoxic than conventional radiotherapy (OR, 1.08, 95% CI, 0.80–1.45; P = 0.94; I2 = 0%).

Two studies reported differences in ototoxicity between HFRT and conventional radiotherapy (Fig. 3D). Forest plot results showed that hearing impairment with HFRT was greater than loss with conventional radiotherapy (OR, 1.61, 95% CI, 0.9–2.88; P = 0.99; I2 = 0%). The advantages of IMRT can be seen in comparison with the hearing impaired population of Conventional-RT (OR, 0.53, 95% CI, 0.47–0.60; P = 0.73; I2 = 0%) (Fig. 3E).

Compared with SRT, RS irradiation caused more severe hearing damage (OR, 1.44, 95% CI, 1.00–2.07; P = 0.69; I2 = 0%) (Fig. 3F). Three trials covered chemotherapy combined with radiotherapy and radiotherapy alone. Two items were Chemotherapy + 3D-CRT compared with 3D-CRT, and one item was Chemotherapy + Conventional-RT compared with Conventional-RT. A summary analysis showed that chemotherapy combined with radiotherapy had a higher risk of ototoxicity than radiotherapy alone (OR, 1.06, 95% CI, 0.89–1.25; P = 0.54; I2 = 0%) (Fig. 3G).

In the comparison of surgery with radiotherapy, one trial involved transoral robotic surgery versus IMRT, while the other examined nasal endoscopic surgery versus IMRT. Surgical removal demonstrated higher hearing preservation than IMRT (OR, 0.59, 95% CI, 0.38–0.93; P = 0.28; I2 = 15%) (Fig. 3H).

Subgroup analysis of the association of RT with hearing impairment by age

In the summary analysis of hearing impairment by comparing age groups, three studies included children under 18 years of age, and the combined effect value of hearing impairment was 0.95 (95% CI, 0.75–1.19; P = 0.44; I2 = 0%) (Fig. 4). Twenty-five studies reported hearing loss in adults, with the combined value being 0.83 (95% CI, 0.69–1.00; P < 0.01; I2 = 76%). The probability of hearing toxicity in children is higher than in adults.

Fig. 4
figure 4

Subgroup Analysis of the Association of RT with hearing impairment by age

Full size image
Subgroup analysis of the association of RT with hearing impairment by tumor type

Oropharyngeal carcinoma appeared to have the lowest hearing impairment associated with radiotherapy (OR, 0.41, 95% CI, 0.26–0.64; P = 0.76; I2 = 0%). The hearing impairment of glioblastoma patients after irradiation was also relatively evident (OR, 1.08, 95% CI, 0.80–1.45; P = 0.94; I2 = 0%). Radiotherapy-associated hearing impairment in vestibular adenomas was high among all primary tumors included in the literature (OR, 1.14, 95% CI, 0.76–1.70; P = 0.05; I2 = 66%) (Fig. 5).

Fig. 5
figure 5

Subgroup Analysis of the Association of RT with hearing impairment by Tumor type

Full size image
Subgroup analysis of the association of RT with hearing impairment by mean cochlear radiation dose

One randomized controlled study covered experimental versus control arms and involved different average cochlear doses. Therefore, when calculating cochlear dose-related radiotherapy hearing impairment, it was divided into each arm and its corresponding dose. A total of 887 individuals reported mean cochlear radiation dose and hearing loss in 10 arms. The mean cochlear radiation dose of the 3 arms was in the range of 30-40 Gy, with the probability of total hearing impairment being 27% (95% CI, 0.19–0.35; P = 0.42; I2 = 0%). When the cochlear radiation dose was 40-50 Gy, the combined value of total hearing impairment was 28% (95% CI, 0.19–0.39; P < 0.01; I2 = 85%). When the cochlear radiation dose increased to 50-60 Gy, the probability of total hearing impairment was the highest, at 35% (95% CI, 0.26–0.44) (Additional file 1: Fig. S2).

Publication bias and sensitivity analysis

The funnel plots of hearing impairment included in the study were roughly symmetrical (Additional file 1: Fig. S3). The Egger test was also conducted to assess whether there was publication bias in this study. No significantly different results emerged, with p = 0.126 for Egger’s test. The combined effect value of the sensitivity analysis was 0.85 (95% CI, 0.72–1.00), indicating that the results were stable (Additional file 1: Fig. S4).

Discussion

The study encompassed 28 randomized controlled trials, involving 6,252 patients, to evaluate ototoxic effects in cancer patients after radiotherapy. Factors such as mean cochlear irradiation dose, primary tumor, radiotherapy modality (or technique), and patient age may influence the risk of hearing impairment. IMRT radiotherapy-associated ototoxicity was less common than conventional radiotherapy. Stereotactic radiotherapy appeared to be a better option for hearing protection than radiosurgery. Children are at a higher risk of hearing impairment than adults. Over half of patients with vestibular neuroadenoma experience hearing impairment after radiation therapy. The average cochlear radiation dose is strongly associated with hearing impairment, and the radiation dose to the cochlea must be precisely controlled. To the best of our knowledge, this is the first comprehensive meta-analysis to analyze ototoxic injury caused by radiotherapy.

Previous literature has discussed the relationship between ototoxicity and radiotherapy. Theunissen and others conducted a study on sensorineural hearing loss (SNHL) caused by radiotherapy of head and neck tumors, suggesting that factors influencing the risk of SNHL included cochlear radiation dose, population age, and follow-up time [40]. Radiation-related ototoxicity involving auditory structures is multifactorial in nature. Radiation affecting the external auditory canal may lead to increased soft tissue susceptibility to infection and may necessitate regular removal of the cochlea to keep it dry [41]. Sensorineural hearing loss typically occurs at doses greater than 30 Gy [42, 43]. The risk of ototoxicity increases in patients receiving combined treatments, such as radiotherapy and platinum chemotherapy [44, 45].

Discussions about the optimal treatment strategy for techniques, prescription dosing, and segmentation are based on the need to prioritize curing the tumor while maintaining an acceptable risk of complications. Pediatric brain and head and neck malignancies requiring dose escalation, as well as adult skull base malignancies, are internationally recognized indications for proton therapy, exhibiting good local control, survival, and acceptable toxicity rates [46, 47]. Due to the lack of robust prediction models of photons and protons for this toxicity, it is impossible to predict the risk of hearing loss based on the patient's disease and treatment characteristics. The Normal Tissue Complication Probability (NTCP) model has been employed in previous studies to guide clinical judgment of proton beam therapy (PBT) [48]. In S. Gaito et al.’s study, the risk reduction of secondary tumors with PBT was estimated to be considerable compared with conventional photon radiotherapy through modeling studies. The clinical benefit of PBT primarily depends on the location of the tumor relative to the organ at risk and on the prescribed dose [48]. A multicenter study evaluating proton therapy and volumetric modulated arc therapy (VMAT) by establishing an NTCP model demonstrated that the reduction of NTCP in the population had a significant impact on auditory toxicity (VMAT: 8.0%; Proton: 3.3%). A significant reduction in the median population was observed in the proton-radiotherapy program, which provided auditory complications as well as a reduced risk of secondary brain cancer [49]. Previous research has also shown that proton therapy can effectively lower the dose of normal tissue surrounding patients with low-grade glioma (LGG). Compared with proton therapy, IMRT poses a two-fold higher risk of secondary intracranial tumors [50]. Fortin et al. conducted photon intensity modulation and proton radiotherapy in 50 children. Using proton and photon RT dose distributions, the intelligence quotient (IQ) and hearing loss probability of each ear were estimated by a Monte Carlo model. They concluded that compared with photon RT, proton RT is expected to reduce the adverse effects of RT on IQ and hearing [51].

Early radiation-induced ototoxicity is associated with mucosal edema, inflammation, and scaling of the outer, middle, and inner ear tissues [52]. The outer ear, middle ear, and inner ear may be affected, but otitis media is more common. It is related to middle ear effusion and can cause hearing loss, earache, and otorrhea, which usually subside after a few weeks [53]. The most prevalent toxicity is sensorineural hearing impairment. Arterial microvascular fibrosis and obliterative endarteritis often occur in the blood vessels of the inner ear's spirochetes, leading to degeneration and atrophy of the smooth muscle of the inner ear and the outer hair cells of the cochlea [54]. These hair cells are located at the base of the cochlea and are more sensitive to ionizing radiation than internal hair cells. They are responsible for hearing high frequencies, so high frequencies are more susceptible than low frequencies [55].

In this meta-analysis, the influence of age, radiotherapy mode, primary tumor, and mean cochlear dose on ototoxic hearing impairment is discussed. It can be concluded that the dose delivered to the inner ear (more precisely, the cochlea), radiotherapy technique (three-dimensional conformation), tumor type, and the age of the radiotherapy population are closely related to radiation-induced ototoxicity. The primary sites of radiation-induced ototoxicity are the paranasal sinuses, nasal cavity, nasopharynx, and parotid glands [56]. In the case of large tumors, the risk is undoubtedly greater. The ototoxicity of radiotherapy alone is related to the total dose received by the cochlea. Most authors in this study selected a threshold dose between 30 and 60 Gy. The probability of total hearing loss between 30 and 40 Gy was 27% (95% CI: 0.19–0.35), while the probability of total ototoxicity between 50 and 60 Gy was 35% (95% CI: 0.26–0.44). Charlotte et al. performed pure-tone audiometry at 0.250–16 kHz before and after treatment in 101 patients with head and neck cancer treated with modulated radiotherapy. They indicated that high frequencies might be affected early [57]. Keilty et al. assessed hearing in 340 children receiving radiotherapy. Mean cochlear dose, time after radiotherapy, cisplatin dose, and carboplatin dose were associated with increased assessment grades of hearing loss. If the mean cochlear dose is > 4 Gy, the cumulative incidence of high-frequency hearing impairment (> 5 kHz) at 50 years after radiotherapy is greater than 30%. Children who are treated with RT, especially those also receiving chemotherapy, are at a higher risk of hearing impairment and should have lower cochlear restraint [58]. They recommend an average cochlear dose of ≤ 30 Gy as a target for reducing Hodgkin's lymphoma risk [58]. To understand ototoxicity in radiotherapy patients, it is important to collect results reported by the clinician at baseline and during follow-up, in addition to tests such as audiograms, as this may partially overcome the inadequacy of hearing testing [59].

In this study, the subgroup analysis of hearing impairment with different radiotherapy techniques was analyzed. From the results, it is evident that the impact of carbon ions on hearing loss is minimal, making it a more reliable choice. Before the advent of IMRT, traditional two-dimensional radiation therapy was used. The response to radiation therapy depends on the sensitivity of tumor cells to radiation. In theory, higher radiation sensitivity leads to better therapeutic effects. However, the damage caused by treatment to the surrounding normal tissues also increases. Heavy particles, such as carbon ions, release energy to centrally explode tumor cells, maximizing the effect of radiotherapy and reducing damage to surrounding healthy tissues. Huang et al. divided 26 pediatric patients receiving medulloblastoma treatment into two groups, receiving conventional radiotherapy or IMRT, respectively. The pure tone audiogram was detected, and hearing function was graded from 0 to 4 according to the toxicity standard of the pediatric oncology group. They concluded that, compared to the traditional RT group, the IMRT group had a lower average decibel hearing threshold at each frequency. The overall incidence of ototoxicity in the IMRT group was low. Thirteen percent of the IMRT group had grade 3 or 4 hearing loss, compared with 64% of the conventional RT group [60]. In the study by Erner et al., 54 patients with low-grade and middle-grade chondrosarcoma of the skull base received carbon ion radiotherapy. After a median follow-up period of 33 months, only one tumor patient had sensory hearing loss in the inner ear. The patient retained useful hearing and did not use hearing aids [61].

The first-line treatment for radiation ototoxicity is drugs, which can be injected into the ear with a vasoconstrictor [52]. If this treatment is not effective, tympanoplasty may be required to treat middle ear effusion. Bone conduction hearing aids can be provided if symptoms associated with damage to the outer or middle ear persist, such as when mucosa is irreversibly damaged. They allow bypassing of the outer and middle ear, directly stimulating the sensory cells of the cochlea. Hyperbaric oxygen therapy can also be used as a treatment option for most subjects. Xu et al. discussed the treatment of 96 nasopharyngeal carcinoma patients with effusion otitis media after radiotherapy. They were divided into 3 groups: simple auricular point plus aspiration, tympanic membrane fenestration plus cauterization, tympanic membrane fenestration plus tympanic membrane tube insertion. Finally, they concluded that intensive local care after eardrum insertion can effectively reduce the incidence of ear complications after radiotherapy [62].

Limitations

This study had several limitations. First, there are few RCTs providing average doses of cochlea, making it difficult to analyze the link between doses and ototoxicity on a large scale. Second, potential bias may exist due to differences in tumor stage and underlying disease in the included population. Third, the number of RCTs involved in each radiotherapy technique is small.

Conclusions

In this study, randomized controlled trials were analyzed to compare the association of various factors with radiotherapy-related ototoxicity. Radiation design patterns and doses, population characteristics, and tumor characteristics are all intricately linked to ototoxicity. Children treated with RT, particularly those receiving chemotherapy, have a higher risk of hearing impairment; therefore, their cochlear restraint should be lower. To gain a better understanding of hearing toxicity, it is crucial to collect clinician-reported results at baseline and during follow-up, in addition to tests such as audiograms.