Introduction

The optimal management of patients with head and neck squamous cell carcinoma requires a multidisciplinary approach. Cure for tumors is possible, even for locally advanced tumors, by surgery or radiation therapy (RT) alone, or combined treatments, including surgery, RT, and chemotherapy [1,2,3]. High survival rates can be achieved for local tumors using RT, with 5-year survival rates at > 80% for stages 1 and 2 and 60–70% for stages 3 and 4 tumors [4]. However, the occurrence of significant long-term treatment sequelae of RT may impact patients’ quality of life (QoL) [5, 6]. In particular, radiation-induced xerostomia (dry mouth), a common side effect of RT, may affect speech, swallowing, and overall oral problems [7].

Currently, the radiation methods used to treat head and neck squamous cell carcinoma include conventional RT and intensity-modulated radiation therapy (IMRT) [2]. IMRT is an advanced type of radiation therapy that uses computed tomography-based planning to deliver precise radiation doses to a malignant tumor or specific areas within the tumor. IMRT has superior advantages compared to conventional RT treatments. It allows the radiation dose to conform more precisely to the three-dimensional (3D) shape of the tumor by computer controlling the intensity of the radiation beam in multiple small volumes, which minimizes the radiation doses to organs at risk (OARs) [8]. This function is important particularly for OARs in the head and neck region, including the spinal cord, brain stem, optic pathway, parotid glands, and inner ear [8]. In addition, the high radiation beam dose conformed to the shape of the tumor and spared the dose to surrounding normal tissue, which allowed escalating the radiation dose for tumors and the possible reduction of side effects [9]. Small randomized controlled studies (RCTs) and a meta-analysis had previously provided evidence to support the potential benefits of IMRT over conventional radiation therapy [2, 9].

Among long-term outcomes, it is especially important to maximize and preserve QoL in patients with head and neck cancer. The instruments used to evaluate QoL vary between studies and may include tumor size, side effects of RT, global health status, emotional status, social function, cognitive function, and severity of xerostomia. Some studies have reported associations between QoL and survival before, during, and after treatment [8]. IMRT allows the sparing of OARs and other non-tumor tissues and has the potential to induce a less negative impact on QoL compared with conventional RT. Several previous RCTs reported significant reduction in moderate to severe xerostomia with IMRT compared with either two-dimensional RT (2D-RT) or 3D conformal radiation therapy (3D-CRT); however, the impact upon tumor and survival has been inconsistent due to small sample size and the associated low statistical power of individual studies.

In the last decade, several RCTs and meta-analyses have directly compared IMRT with either 2D-RT or 3D-RT in head and neck cancer. While nearly all studies reported the data of severe xerostomia, tumor and survival with IMRT and conventional RT, the impact upon different QoL domains were not included or meta-analysis was not performed. Therefore, we have performed the first meta-analysis including the most recently published data. Our hypothesis was that patients receiving IMRT may have better QoL than those receiving conventional RT in head and neck cancers. The present study aimed to compare the effects of IMRT with those of conventional radiotherapy, including 2D-RT and 3D-CRT, on QoL and specific QoL domains, including cognition, emotional function, and the severity of xerostomia in patients with head and neck cancer.

Methods

Search strategy

This study was performed in accordance with the PRISMA guidelines. PubMed, Cochrane, and EMBASE databases were searched to July 1, 2019, using the following search terms: (head and neck cancer) AND (radiotherapy OR radiation therapy OR 2D-CRT OR 3D-CRT OR IMRT OR Conformal proton beam radiation therapy OR stereotactic radiosurgery OR volumetric modulated arc therapy OR VMAT) AND (neuropsychological OR cognition OR quality of life) AND (nasopharyngeal carcinoma OR oral cancer OR buccal caner OR head and neck cancer) AND (squamous cell carcinoma) AND (radiotherapy OR radiation therapy OR radiosurgery) AND (neuropsychological OR cognition OR xerostomia). PubMed search filters: Abstract available, English, Clinical Trial; Embase search filters: Abstract, English, Human, Full text, Clinical Article.

RCTs, two-arm prospective and retrospective studies were included. Included studies were required to have evaluated patient-reported QoL and xerostomia between patients with head and neck cancer who received IMRT and patients with head and neck cancer who received any other type of radiotherapy. Studies were also required to have reported quantitative outcomes of interest. One-arm studies, cohort studies, review articles, letters, comments, editorials, case reports, proceedings, and personal communications were also excluded.

Study selection and data extraction

Studies identified by the search strategy were screened using a two-step process. First, the title and abstract of each article were examined and citations not meeting the inclusion criteria were discarded. Second, full-text review of the remaining studies was performed by two independent reviewers. If any uncertainties existed regarding eligibility, a third reviewer was consulted. The reference lists of the relevant studies were also hand searched to identify other studies that met the inclusion criteria. Data from the included studies were extracted by two independent reviewers and a third consulted when necessary to resolve any disagreements. The following data were extracted: name of first author, study design, interventions, patient number, age, gender, and site and stage of cancer. In addition, information regarding radiation and chemotherapy, and the percentage of patients with neck dissection or surgery was extracted.

Quality assessment

The quality of the included studies was performed using the Quality In Prognosis Studies (QUIPS) tool [10]. The QUIPS tool evaluates six sources of bias, including study participation, study attrition, prognostic factor measurement, confounding and account measurement, outcome measurement, and analysis. The quality of included studies was independently appraised by two reviewers, and any disagreements were resolved by a third reviewer.

Outcomes

The primary outcomes were QoL measured by global health status, emotional function, social function, and cognitive function. The secondary outcome was the severity of xerostomia.

Data analysis

Standardized mean difference (SMD) with corresponding 95% confidence intervals (CIs: lower and upper limits) was calculated for each individual study and for all studies combined, because measurements were determined by various instruments. A χ2-based test of homogeneity was performed and the inconsistency index (I2) and Q statistics were determined. If the I2 statistic was > 50%, a random-effects model (DerSimonian and Laird) was used. Otherwise, a fixed-effects model (inverse-variance method) was employed. Pooled effects were calculated and a 2-sided P value < 0.05 indicated statistical significance. Sensitivity analysis was carried out using the leave-one-out approach. In addition, publication bias was only assessed if there were > 10 studies, because > 10 studies are necessary to detect funnel plot asymmetry [11]. All analyses were performed using Comprehensive Meta-Analysis statistical software, version 2.0 (Biostat, Englewood, NJ, USA).

Results

Search results

A total of 2017 studies were identified in the initial research (Fig. 1). Of those, 1939 were excluded for not being relevant by reviewing titles and abstracts. In total, 78 studies underwent full-text review for eligibility and 71 were excluded for not reporting QoL, that is, being single-arm studies, duplicates, review articles, not comparing IMRT with another type of radiation, or not evaluating head and neck cancer and the data cannot be pooled. Finally, 7 studies were included for analysis in this study [3, 12,13,14,15,16,17].

Fig. 1
figure 1

Flow diagram of study selection

Full size image

Study characteristics

The main characteristics of these 7 included studies are summarized in Table 1, with a total of 761 patients, including 369 patients who received IMRT and 392 patients treated with conventional RT [3, 12,13,14,15,16,17]. Three studies were RCTs (Nutting [2011], Kam [2007], Pow [2006]), and the others were prospective or retrospective studies. The median age among the studies ranged from 18 to 65 years. The majority of participants were male, ranging from 59 to 81%. The site and stage of cancer varied across the studies.

Table 1 Summary of basic characteristics of selected studies
Full size table

Treatment protocols and QoL assessment

Table 2 summarizes the treatment protocols of the included 7 studies. Diverse radiotherapy protocols were used in patients with head and neck squamous cell carcinoma. The number of patients receiving either IMRT or RT was equally distributed. Chemotherapy was also administered to 35 to 63% of patients as a combination treatment with RT. Where reported, 9 to 62% of patients had received either neck dissection or neck surgery. The instruments used to evaluate QoL varied across studies (Supplemental Table S1). The studies included in our analysis used European Organization for Research and Treatment of Cancer Quality of Life questionnaire C30 (EORTC QLQ-C30), EORTC QLQ Module for Head and Neck Cancer (EORTC QLQ H&N35), Head and Neck Quality of Life instrument (HNQOL), University of Washington Quality of Life Questionnaire (UWQOL), Medical Outcomes Study Short-Form (36-item) Health Survey (SF36), and xerostomia questionnaire (XQ).

Table 2 Summary of treatment protocols of selected studies
Full size table

Meta-analysis

Global health status was firstly performed analyses to assess QoL. A total of five studies (Chen [2012], Nutting [2011], Vergeer [2009], Pow [2006], and Jabbari [2005]) provided information on global health status before and after treatment. Since a statistically significant heterogeneity was found when data from the 5 studies were pooled (Q statistic = 32.673, I2 = 87.76%, P < 0.01), a random effects model of analysis was used (Fig. 2a). The analysis revealed that patients who received IMRT had significantly better global health status compared with those who received conventional RT (pooled SMD = 0.80, 95% CI 0.26 to 1.35, P = 0.004).

Fig. 2
figure 2

Meta-analysis for global health status, emotional function, social function, cognitive function, and dry mouth

Full size image

We examined the emotional function of QoL and included three studies (Chen [2015], Vergeer [2009], and Pow [2006]). As shown in Fig. 2b, a fixed-effect model of analysis (Q statistic = 2.528, I2 = 20.90%, P = 0.282) revealed no significant differences between pooled effects in the emotional function between patients treated with IMRT and those who received conventional RT (pooled SMD = 0.06, 95% CI − 0.13 to 0.26, P = 0.53).

Three studies (Chen [2015], Vergeer [2009], and Pow [2006]) provided complete data for changes in social function with treatment (Fig. 2c). A significant observed heterogeneity was found in emotional function among these studies (Q statistic = 22.231, I2 = 91.00%, P < 0.001). However, this result in pooled effects indicated that social function was comparable between the two treatment groups using a random effects model (pooled SMD = 0.35, 95% CI − 0.38 to 1.08, P = 0.348).

We then examined the effects of treatment on cognitive function (Fig. 2d), and only two studies [Vergeer (2009), Pow (2006)] were included. No significant observed heterogeneity was found between the two treatment groups (Q statistic = 1.114, I2 = 10.24%, P = 0.291) so a fixed-effect model was then used. The overall analysis revealed that patients in the IMRT group had significantly better cognitive function compared with those in the conventional RT group (pooled SMD = 0.30, 95% CI 0.06 to 0.54, P = 0.013).

Six studies (Chen [2015], Nutting [2011], Vergeer [2009], Kam [2007], Pow [2006], and Jabbari [2005]) were used to evaluate the severity of patient-reported xerostomia (dry month) following treatments (Fig. 2e). A significant observed heterogeneity was found among the studies (Q statistic = 19.721, I2 = 74.65%; P = 0.001); therefore, a random effects model was used. The analysis revealed that patients who received IMRT had significantly lower scores for xerostomia than those who received conventional RT (pooled SMD = − 0.60, 95% CI − 0.97 to − 0.24, P = 0.001).

Sensitivity analysis

Sensitivity analysis using the leave-one-out approach was performed for outcomes of global health status, emotional status, social function, and the severity of xerostomia (Table 3). However, this analysis was not performed for cognitive function because of the small number of studies reporting cognitive function (n = 2). In general, sensitivity analysis of each outcome revealed that the magnitude of combined estimates did not vary markedly with the removal of the studies, indicating good reliability and that the data were not overly influenced by each study. However, the analysis of social function revealed that removal of the study of Chen et al. [3] resulted in the pooled IMRT/conventional RT becoming significant (pooled SMD = 0.64, 95% CI 0.03 to 1.25, P = 0.04), indicating that the pooled estimates might be influenced by this individual study. In addition, sensitivity analysis of the severity of xerostomia using the leave-one-out approach revealed that the magnitude of combined estimates did not vary markedly with the removal of the studies, indicating good reliability and that the data were not overly influenced by each study.

Table 3 Sensitivity analysis
Full size table

Quality assessment

Quality assessment indicated that higher percentages of biases of all included studies were at low risk for influencing the study population, measurement of prognostic factors, evaluation of outcomes, and analysis (Fig. 3a). Figure 3 b shows risk-of-bias summary, which was the quality assessment result of the individual study. Evident risk of bias for study attrition was found due to Jabbari et al. (2005) and confounding measurement and account due to Kam et al. (2007) and Chen et al. (2015). Overall, the included studies were of moderate to high quality.

Fig. 3
figure 3

Quality assessment a percentages of biases of all included studies and b risk-of-bias summary for individual studies

Full size image

Discussion

In patients with head and neck cancer, the use of IMRT or conventional RT is frequent in clinical practice, but differences in their impact on QoL and long-term sequelae may be debatable. The present meta-analysis included the most recently published data to July 1, 2019. Our focus was on the QoL of patients after treatment of head and neck cancer with either IMRT or conventional RT, so that clinicians can be more assertive in their decisions to use IMRT or conventional RT. Our findings revealed that patients in the IMRT group had significantly better global health status and cognitive function compared with patients in the conventional RT group (P ≤ 0.013). No differences were observed in emotional and social function between the two groups. This study also revealed a significant reduction in severity of patient-reported xerostomia in patients receiving IMRT than in those receiving conventional RT (P = 0.006), which may also translate into an improvement in xerostomia-specific QoL in patients with IMRT.

The use of chemoradiotherapy for head and neck cancer may expose patients to potential additional neurotoxicity due to the use of cytotoxic drugs. In addition, this patient population may be at a higher risk of reduction in cognitive function as several of the factors shown to be associated with the development of head and neck cancer are also known to negatively impact cognitive function, such as smoking, excess alcohol consumption, and poor diet [18]. Results of the present study indicated that patients with IMRT had better cognitive function than those receiving conventional RT; however, the findings must be interpreted with caution since only two studies were included in this analysis.

A systematic review by Tribius et al. (2011) [8] assessed whether IMRT was associated with better QoL compared with 2D-RT and 3D-CRT. The review included 14 studies, only one of which was an RCT. Those authors found that IMRT was associated with significant improvements in some QoL domains compared with 2D-RT. This was particularly evident in domains associated with salivary function and domains related to xerostomia, such as speech, swallowing, problems with teeth, and sticky saliva. However, evidence of the benefits of IMRT compared with those of 3D-CRT were less apparent.

Three prior meta-analyses evaluated the efficacy of IMRT compared with conventional RT in treating patients with head and neck cancer [1, 19, 20]. The meta-analysis of Marta et al. (2014) [1] included five phase III RCTs with a total of 871 patients. Consistent with our findings, a significant overall benefit of IMRT was found (hazard ratio [HR], 0.76; P < 0.0001) for xerostomia grades 2 to 4, and this benefit was seen up to 5 years following therapy. However, no differences were found between IMRT and conventional RT in loco-regional control (P = 0.35) and overall survival (P = 0.11). Zhang et al (2015) [19] performed a meta-analysis that compared clinical outcomes and late toxicities of IMRT and conventional RT in nasopharyngeal cancer, including eight studies with 3570 patients. Those authors found that IMRT was associated with better 5-year overall survival (odds ratio [OR], 1.51; P = 0.0001) and tumor local control (OR, 1.94; P < 0.0001) than conventional RT. In addition, the incidence of late xerostomia was lower in patients who received IMRT than in those who received conventional RT (OR, 0.18; P = 0.0004). The frequency of radiation-induced chronic toxicities of trismus and temporal lobe neuropathy were also significantly lower with IMRT compared with conventional RT (OR, 0.18; P = 0.03 for trismus; and OR, 0.44; P = 0.0003 for temporal lobe neuropathy). In the review and meta-analysis of Gupta et al. (2019) [20], which focused on xerostomia, the use of IMRT was associated with a 36% relative risk reduction in acute xerostomia and late xerostomia compared to all non-IMRT radiotherapies at all time points. Taken together, the findings of the present study and those of Tribius et al. [8], and the three meta-analyses of Marta et al. [1], Zhang et al. [19], and Gupta et al. [20], suggest that IMRT may have greater benefits than conventional RT with regard to specific QoL domains and xerostomia. However, future studies specifically designed and powered to test the benefits of IMRT over 2D/3D-RT in patients with head and neck cancer in regard to QoL and xerostomia would undoubtedly provide more conclusive evidence.

The present study included the most recently published data and was the first meta-analysis conducted on the QoL of patients following treatment of head and neck cancer with IMRT or conventional RT. The design of our study was very different from the previously published meta-analyses and/or systematic reviews. For example, in 2011, Tribius assessed whether IMRT was associated with QoL benefits versus 2D-RT and 3D-CRT but did not perform meta-analysis [8]. However, both Marta (2014) [1] and Zhang (2015) [19] performed meta-analyses regarding overall survival, loco-regional control, and incidence of radiation-induced late toxicities but did not evaluate QoL outcomes. Meanwhile, the present meta-analysis indicated that IMRT conveyed benefits in global health status and cognitive function, and reduced the severity of xerostomia compared with conventional RT. Results for specific QoL domains such as global health status and cognitive function have not been reported previously. Therefore, results of the present study may provide additional useful information for head and neck cancer management.

We examined heterogeneity among our included studies, mainly in terms of QoL and xerostomia outcomes. Heterogeneity in health-related QoL instruments was found between Chen (2012) and Jabbari (2005), who used the UWQOL questionnaire and the HNQOL instrument, respectively; and Nutting (2011), Vergeer (2009), and Pow (2006), who all used the EORTC scale. Tumor sites included in Chen (2012), Vergeer (2009), and Jabbari (2005) were more heterogeneous than Nutting (2011) and Pow (2006). It is well known that cancers at different sites in the head and neck produce different effects on health-related QoL and such site-specific assessment may yield more meaningful and useful QoL outcomes [21, 22]. Moreover, Chen (2012), Vergeer (2009), and Jabbari (2005) were nonrandomized studies. There may be case selection bias. Sensitivity analysis indicated that Chen et al. (2012) may have overly influenced the findings for social function, possibly by using theSF-36 while other studies used the EORTC QLQ-C30. For dry mouth, heterogeneity existed between Chen (2015), who used the SF-36, and Nutting (2011), Vergeer (2009), and Pow (2006), who all used the EORTC QLQ-H&N35 subscale scores. The RCT of Kam et al. (2007) [17] used a self-reported xerostomia questionnaire combined with a 100-mm visual analog scale (VAS) as previously described [23]. The instrument recorded responses to six items: overall mouth and tongue dryness, feeling of mouth/tongue during daytime, difficulty sleeping at night, difficulty speaking without first drinking liquids, difficulty chewing and swallowing food, and difficulty wearing dentures. Jabbari (2005) used a xerostomia questionnaire with eight items, four related to patient-reported dryness while eating or chewing and four related to dryness while not eating or chewing, also described previously [24]. Clearly, if we bring together a body of studies for meta-analysis, we will find variability in the data. However, investigating these differences ultimately gives us greater understanding of the effects of a specific intervention and the influencing factors.

Our study has several limitations that should be considered when interpreting the findings. The number of included studies was small, which prevented the analysis of several other QoL domains. Heterogeneity existed across the studies for type of instrument used to evaluate QoL, type of RT, types of cancer, radiation doses, and the time frame in which endpoints were evaluated. Because of the relatively small sample sizes, we could not perform subgroup analysis of different cancer types or the specific conventional RT used. However, although some heterogeneity was noted across included studies, study design, methodology, analyses, and reporting were quite similar. QoL data were also pooled at later time-points (6–24 months) to ascertain whether the significant benefit of QoL with IMRT persisted over time. The quality of included studies was judged to be from moderately low to having an unclear risk of bias. Nevertheless, no significant publication bias was detected for any of the outcome measures in this analysis. Herein, we have reported a meta-analysis with a smaller sample size, and with so-called small-study effects and reporting bias; however, the findings in smaller studies are more likely to be selected for publication based on statistical significance. It is important to show the observed heterogeneity in effects across multiple independent trials, even some that are much smaller, as heterogeneity occurs normally in clinical practice. While it may be difficult to conclude whether our findings were relevant in terms of evidence-based learning, this study has provided the most up-to-date insight regarding the outcomes of QoL in patients with IMRT or conventional RT in head and neck cancer. Our findings regarding xerostomia were limited because studies that only reported rates of xerostomia were not included in favor of patient-reported xerostomia and its inherent contribution to reduced quality of life, our primary outcome; excluding these studies could represent bias. Also, the severity of xerostomia was self-reported by patients and did not rely upon clinical evaluation. These limitations emphasize the need for prospective, randomized studies that include homogeneous patient populations, xerostomia rates, appropriate QoL instruments, and relevant clinical end points, as well as to measure the relative impact of different radiation doses on QoL, in order to determine the value of IMRT in head and neck cancer.

In conclusion, IMRT exhibits benefits in global health status and cognitive function, and reduces the severity of xerostomia compared with conventional RT. Results of this study provide clinicians with guidelines for making decisions on the use of IMRT versus conventional RT.