Introduction

Around 700 new cases of head and neck cancer are diagnosed annually in Finland with a population of 5.3 M [1]. Surgery and (chemo) radiotherapy [(C)RT] remain the treatment options for these tumors. CRT is the standard treatment for locally advanced (Stage III−IV) tumors [2–4] in most tumor locations in this patient population. Further, small and early-stage tumors can often be treated with RT alone as an alternative to surgery.

Definitive CRT for head and neck cancer usually results in good locoregional disease control [2–5], but treatment-related acute and long-term morbidity can be substantial. Radiation damage to the healthy tissues, such as mucosa and salivary glands, can lead to difficult local symptoms, which in addition to the distress caused by cancer diagnosis itself, can have a noticeable negative impact on patients’ health-related quality of life (HRQoL) [6]. After adopting intensity-modulated radiotherapy (IMRT) into clinical practice, it has become possible to reduce the radiation dosage to the healthy tissues compared with the former 2- or 3-dimensional techniques. However, this development has not entirely eliminated the treatment-related toxicity [7–9]. Therefore, in addition to the mere disease control rates, HRQoL remains an important outcome when managing this patient group.

To assess HRQoL of patients with head and neck cancer, a large variety of different instruments, both disease-specific and generic, have been developed [10–12]. However, there is still no consensus on which instrument would be preferable to use. During the last years, self-administered questionnaires have become the mainstay of HRQoL evaluation in head and neck cancer patients but the contents of the questionnaires vary widely [10]. Thus, the instruments should be selected based on the specific focus of the inquiry [13, 14].

In this study, HRQoL was assessed using the self-administered generic 15D instrument. The 15D is a widely used HRQoL instrument and it has been shown to be feasible in evaluating HRQoL of patients with, e.g., lung, prostate, breast, and colorectal cancer [15–19], and some non-cancerous head and neck conditions [20–22]. Its usefulness in head and neck cancer patients, however, has not yet been established.

The aim of this prospective cohort study was to evaluate the HRQoL of head and neck cancer patients scheduled for curative treatment with definitive RT, with or without concomitant chemotherapy, as well as to assess the feasibility of the 15D instrument in the evaluation of this patient group.

Patients and methods

Between November 2007 and July 2012, a total of 84 patients who presented with newly diagnosed carcinoma of pharynx, larynx or nasal cavity, and who were scheduled to receive definitive RT with or without concomitant chemotherapy at our Department of Oncology filled in the baseline 15D questionnaire.

Follow-up 15D questionnaires were then sent to these patients at 3, 6 and 12 months after answering the baseline questionnaire. Those patients who had answered at least to the last (12 months after the diagnosis) questionnaire, in addition to the baseline (prior to the treatments) questionnaire, were included in the analyses.

An institutional research approval was granted for the study (538/EO/2002).

Patients

Seven patients died during the 1-year follow-up and were excluded from the study. Further, 12 out of the 84 patients were excluded because of missing data at 12 months and one patient because of having two simultaneous malignancies.

A total of 64 patients with complete 15D data were included. Sixteen (25.0 %) of these patients were female. The mean age was 61.6 years (range, 40−81 years). The tumor site was larynx in 18 (28.1 %), pharynx in 44 (68.8 %) and nasal cavity in two (3.1 %) cases. Fifteen (23.4 %) of the patients had Stage I−II and 49 (76.6 %) had Stage III−IV tumors. The main patient and tumor characteristics are presented in Table 1.

Table 1 Patient and tumor characteristics
Full size table

Prior to treatment, the patients went through standard pretreatment evaluation including clinical head and neck examination, imaging (CT/MRI), and endoscopy. All diagnoses were histopathologically confirmed. The tumors were staged according to the seventh UICC TNM classification. Treatment-related toxicity was scored according to CTCAE v.3.0.

Treatment

Treatment characteristics are presented in Table 2.

Table 2 Treatment characteristics
Full size table

All patients were treated with IMRT. In all but one patient the treatment was given with 2 Gy daily fractions, and the treatment dose varied from 62.5−70 Gy (mean 69.3 Gy). One patient treated with the simultaneous integrated boost (SIB) technique received 50/2 Gy to the elective lymph node areas and 62.5/2.5 Gy to the primary tumor area.

Fifty-five (84.4 %) patients received combined CRT and ten (15.6 %) received RT alone. The used chemotherapeutic agent was weekly cisplatin alone in 49 cases (corresponding to 90.7 % of the patients receiving combined therapy), cetuximab in two (3.1 %) cases, cisplatin+panitumumab in one (1.6 %) case, and paclitaxel in another one (1.6 %). In addition, in one patient, the weekly cisplatin was switched to paclitaxel after four courses due to cisplatin-induced kidney failure. One patient with nasopharyngeal cancer scheduled for cisplatin-based CRT received additional three cycles of cisplatin+5-fluorouracil after having completed the CRT.

All patients received RT up to the planned dose (range 62.5−70 Gy). Treatment duration ranged from 37−66 days (mean, 50.4 days).

None of the patients underwent primary surgery. Salvage surgery was reserved for cases with residual disease after 3 months of completing the primary treatment.

15D

HRQoL was measured with the 15D questionnaire, which is a standardized, self-administered generic HRQoL instrument [23]. The questionnaire consists of 15 questions comprising mobility, vision, hearing, breathing, sleeping, eating, speech, excretion, usual activities, mental function, discomfort and symptoms, depression, distress, vitality, and sexual activity. For each of these dimensions, there are five different answer options scored from 1 to 5 where 1 represents the best option and 5 the worst. The 15D instrument yields both a profile of the 15 separate HRQoL dimensions as well as a single index quality of life score. The single index score (15D score), representing the overall HRQoL on a 0−1 scale (1 = full health, 0 = being dead) and the dimension level values, reflecting the goodness of the levels relative to no problems on the dimension (=1) and to being dead (=0) and used for drawing up a profile, are calculated from the health state descriptive system using a set of population-based preference or utility weights [23]. The minimal clinically important change in the 15D score has been estimated at 0.018 for improvement and −0.045 for deterioration [24].

All patients were asked to fill in the 15D questionnaire prior to and 3, 6 and 12 months after the treatment onset. The 15D results at different time points were compared with the baseline results of the patients themselves as well as with the scores of the age-standardized general population. The 15D data for the general population were obtained from the National Health 2000 Health Examination Survey [25] and the sample of the general population was weighted to reflect the age distribution of the present patients.

Statistical analysis

Data were analyzed using the SPSS for Windows statistical software version 19.0 (SPSS Inc., Chicago, IL, USA). The results are given as means (standard deviations, SD), or as percentages. Differences in the means between the groups and those over time within the groups were tested by independent samples t test and paired samples t test, respectively. As the distributions of all 15D variables were not normally distributed, also corresponding non-parametric tests (Mann–Whitney and Wilcoxon signed-rank, respectively) were applied. Apart from minor differences in the level of statistical significance, the results of parametric and non-parametric tests were quite similar. Therefore, only results of parametric tests are reported. P values < 0.05 were considered statistically significant.

Results

15D

Completed questionnaires were available for all of the included patients (n = 64) at baseline and 12 months, and 51 (79.7 %) patients filled in all four questionnaires. The remaining 13 patients (20.3 %) filled in three of the questionnaires. The number of patients who had answered to the 3- and 6-month questionnaires was 54 (84.4 %) and 61 (95.3 %), respectively.

The mean 15D scores of all patients (n = 64) at each follow-up time point (baseline and 3, 6 and 12 months), were 0.886 (0.10), 0.829 (0.12), 0.860 (0.12) and 0.862 (0.14), respectively, showing a clear clinically important and statistically significant (p < 0.001) decline at 3 months after the baseline but then significantly improving towards the end of the 12-month follow-up. Compared with the baseline, the p values for the difference in the means of the 15D scores at 6 and 12 months were 0.012 and 0.077, respectively (Fig. 1).

Fig. 1
figure 1

The mean 15D profiles and scores of head and neck cancer patients (N = 64) at baseline and 3, 6, and 12 months after the onset of treatment (mean age 61.6 years, 75 % men)

Full size image

The mean 15D scores at baseline or at 12 months did not differ in a statistically significant or clinically important manner from those of the standardized general population (Fig. 2).

Fig. 2
figure 2

The mean 15D profiles and scores of patients (n = 64) at baseline and at 12 months after the onset of treatment compared with those of age-standardized general population (n = 4,479)

Full size image

As for single dimensions compared with the baseline results of the study group, significantly lower scores were reported on the dimensions of “eating” (p < 0.01), “usual activities” (p < 0.001), “discomfort and symptoms” (p < 0.05), “vitality” (p < 0.05), and “sexual activity” (p < 0.01) at 3 months. At 6 and 12 months, these dimensions showed significant gradual improvement (Fig. 1).

During the 12-month follow-up, a clear trend was seen in the dimensions concerning mental well-being. Clear changes were seen on the dimensions of “depression” and “mental function”: compared with the baseline scores, these dimensions showed marked deterioration at 3 months after the treatment onset but improved notably during the follow-up, exceeding the baseline scores at the end of the follow-up. Moreover, the dimension of “distress” was at its lowest at baseline, and showed subsequent improvement towards the end of the follow-up. This trend was seen in all tumor subgroups (Fig. 3). At the end of the follow-up, the score for “depression” had significantly (p < 0.05) improved compared with the baseline. However, on none of these dimensions, the patients reached the levels of the general population during the 12-month follow-up (Fig. 2).

Fig. 3
figure 3

The mean 15D profiles of head and neck cancer patients at baseline (N = 64) and according to tumor location at 12 months after the onset of treatment

Full size image

Twelve-month HRQoL outcomes for the two major tumor site subgroups pharyngeal (n = 44) and laryngeal (n = 18) cancer are presented in Fig. 3. The mean 15D score seemed to be lower in the pharyngeal cancer subgroup compared both with the laryngeal cancer patients and the baseline score of all patients. At 12 months, the patients with pharyngeal cancer were markedly worse off on most dimensions (moving, breathing, sleeping, eating, usual activities, mental function, discomfort and symptoms and sexual activity) than the laryngeal cancer patients. However, the dimension of “speech” was markedly better in patients with pharyngeal cancer than in those with laryngeal cancer (0.912 vs. 0.877).

At baseline, no statistically significant or clinically important differences in the mean 15D scores were detected between the age-standardized general population and the study group. On the dimensions of “breathing”, “excretion” as well as “discomfort and symptoms”, the patients had significantly higher mean scores than the general population. However, the patient group presented with significantly lower mean scores for “speech” (p < 0.05), “depression” (p < 0.001), “distress” (p < 0.001), and “sexual activity” (p < 0.01) than the general population. Even though a clear improvement on these dimensions was seen during the 12-month follow-up, our patients did not reach the mean scores of the general population (Fig. 2).

The mean 15D scores tended to be lower at all measurement points in those patients who presented with a persistent disease at the end of the treatment (data not shown). This difference was not explained by primary disease stage or treatment-related acute toxicity.

Treatment-related acute toxicity

Treatment-related acute toxicity was mostly of mild to moderate grade (Grade 1−2). Occurrence of acute Grade 3−4 toxicity is depicted in Table 3. Most common Grade 3−4 side effects were radiation mucositis, cisplatin-induced leucopenia, infection and nausea. No treatment-related deaths were reported.

Table 3 Treatment-related acute toxicity Gr 3−4
Full size table

Treatment results

During the 1-year follow-up, locoregional failure was observed in five patients (7.8 %) and distant failure in one patient (1.6 %).

Three patients (4.7 %) (Stage III laryngeal carcinoma, Stage IVA hypopharyngeal carcinoma and Stage IVA oropharyngeal carcinoma, all primarily treated with CRT) had residual disease at 3 months after the treatment. They all went through salvage surgery. Modified radical neck dissection (MRND) was performed to all three patients with residual disease. In addition to the MRND, one patient underwent oropharyngeal resection and one patient underwent total laryngectomy.

Disease control during 1-year follow-up after salvage surgery was achieved in one patient with laryngeal cancer. The remaining two patients relapsed soon after the salvage surgery.

During the 1-year follow-up, additional two locoregional recurrencies were detected. In addition, one patient developed distant metastases but not local relapse. At 12 months, five (7.8 %) patients were alive with an active disease.

Discussion

We used the generic 15D instrument to assess the HRQoL of 64 patients receiving definitive CRT for head and neck cancer. We have previously studied salivary gland preservation after IMRT and treatment outcome in this patient population [26–29]. To our knowledge, the present results represent the first HRQoL results in the Finnish head and neck cancer patient population treated with IMRT. In the present series, the mean HRQoL score significantly declined during the first 3 months after treatment, as expected. Gradual recovery in both overall HRQoL as well as in the initially affected separate HRQoL dimensions was seen towards the end of the follow-up. These findings highlight the importance of multidisciplinary management efforts and supportive measures in this patient population.

Further, our aim was to add to the existing validation of the generic 15D instrument in various areas of head and neck conditions. The 15D is a widely used HRQoL questionnaire, but it has not previously been used in head and neck cancer patients. It has, however, been shown to be feasible for evaluation of HRQoL in patients with many other cancer types including lung cancer [15, 16], breast cancer [17], colorectal cancer [17, 18], and prostate cancer [17, 19]. It has also been used among patients with some non-cancerous head and neck conditions, such as juvenile-onset recurrent respiratory papillomatosis (JORRP) [20], and patients undergoing septoplasty [21] and tonsillectomy [22].

The advantage of generic HRQoL measures, as opposed to the disease-specific ones, is that they enable comparison between different patient populations and different disease entities [23]. This is crucial in e.g., treatment effectiveness analyses. As a widely used HRQoL instrument in Finland, the 15D allows comparison between (chemo)irradiated head and neck cancer patients and other patient populations, as well as general population.

For comparison, other malignant conditions that have been evaluated with the 15D are presented in Table 4. Most of these studies have been cross-sectional and thus not directly comparable with our results. Longitudinal HRQoL analyses with the 15D have been made among patients with various benign conditions [30–35], these are listed in Table 5.

Table 4 Mean (SD) 15D scores in different malignant conditions
Full size table
Table 5 Mean (SD) 15D scores in different patient groups before and 6 months after treatment
Full size table

In the present study, the mean 15D score declined significantly during the first 3 months after the treatment, as expected, but improved gradually towards the end of the follow-up. At 12 months, the mean 15D score did not differ in a statistically significant or clinically important manner from the baseline scores. Of the 15 HRQoL dimensions, the most affected ones were “eating”, “usual activities”, “discomfort and symptoms”, “vitality”, and “sexual activity”, all of which were markedly deteriorated at 3 months. A significant improvement over the 6- and 12-month follow-up was seen on all these dimensions. However, at the end of the follow-up, the mean scores of the dimensions of “eating” and “usual activities” remained statistically significantly lower compared with the baseline scores.

This kind of trend in HRQoL over the 12-month follow-up has also been seen in earlier studies. For example Hammerlid et al. [36] showed in a prospective longitudinal study of 135 patients with head and neck cancer that the QoL measured with SF-36 (Short Form Health Survey) and EORTC-QLQ-C30 (European Organisation for Research and Treatment of Cancer Quality of Life Questionnaire) was at its lowest at 3 months after the onset of the treatment and that it improved markedly towards 1 year even though some head and neck specific morbidity domains (dry mouth and senses) did not fully recover. Also Bjordal et al. [37] showed in a prospective longitudinal study of 218 patients that HRQoL (measured with EORTC-QLQ-C30) fully recovered at 1 year even though some site-specific domains (senses, sex, dry mouth, EORTC-QLQ-H&N35) remained impaired. Consequently, the present study confirms these earlier results. Further, in a recent study of 50 long-term survivors of head and neck cancer treated with IMRT, Chen et al. [38] showed that the patient-reported HRQoL remained good at 3 and 5 years after the diagnosis.

A crucial finding in our study was that the mean scores on the dimensions reflecting mental well-being, such as “distress”, “depression”, “vitality”, and “sexual activity” remained significantly lower throughout the 12-month follow-up compared with the general population. This might reflect the general features of this patient group. However, despite the significant deterioration at 3 months after the treatment onset, also these dimensions showed encouraging gradual improvement over the follow-up. More importantly, the mean scores for “depression” and “distress” were markedly higher at 12 months compared with the baseline. This is a strong signal for the multidisciplinary professional team managing this patient population to continue with every effort in supporting the patient in various challenges included in the treatment and rehabilitation process.

In our study, the pharyngeal cancer patients had consistently lower mean HRQoL scores compared with those with laryngeal cancer. A similar trend has also been reported previously [36, 37]. Furthermore, at 12 months, the mean scores for the dimension of “eating” showed a trend of being lower in patients with pharyngeal cancer (0.859 vs. 0.961). On the contrary, “speaking” seemed to be impaired in patients with laryngeal cancer compared with those with pharyngeal cancer (0.877 vs. 0.912). These differences are in accordance with the tumor location, and indicate that the 15D instrument is sensitive in detecting differences between different tumor subgroups. The patient material in this study was too small for further assessing how the oncological treatments should be customized or developed regarding the HRQoL aspect in different tumor site subgroups.

Those patients who had persistent disease at the end of follow-up had lower 15D scores compared with those who remained disease free. This difference was seen already at the baseline and it remained throughout the follow-up. A similar correlation between patient self-reported quality of life and the treatment outcome has been shown in previous studies. Fang et al. [39] showed in a series of 347 nasopharyngeal carcinoma patients treated with conformal RT that pretreatment quality of life, particularly physical functioning, correlates with disease free and overall survival. Also, Meyer et al. [40] showed in a series of 540 patients with an early-stage (Stages I−II) head and neck cancer that the patient self-reported physical functioning before and at 6 months after the treatment was a stronger prognostic factor on overall survival than some previously recognized patient and tumor related risk factors (disease stage, tumor site, age, smoking, alcohol intake, BMI), whereas de Graeff et al. [41] in their study of 208 surgically operated head and neck cancer patients concluded that cognitive functioning before the treatment onset was a strong predictor of disease control and survival. Although the patient material in our study was too small to further estimate the correlations between quality of life and survival, the present results further confirm the previously reported impact of HRQoL on overall survival and disease-free survival: patient-reported poor HRQoL can correlate negatively with overall outcome.

The present series of patients was small which is a limitation of our study. Seven of the 84 patients died during the 1-year follow-up, and one was excluded because of two simultaneous malignancies, resulting in a 12-month compliance of 84 % (64 respondents out of the 76 patients alive). Since the HRQoL data were analyzed only for those patients who had answered to at least the 12-month questionnaire in addition to the baseline, and the 12-month answers were often missing from those patients who had relapsed or deceased during the 1-year follow-up time, the present results may be biased in an optimistic direction. It is likely that the patients who dropped out or died during the follow-up had lower HRQoL than the survivors [39].

Treatment was generally well tolerated with no unexpected side effects. No differences in the treatment-related toxicity or treatment duration between these two groups were detected. One-year local control was 92 % with five locoregional failures, and 1-year overall relapse free survival was 91 % with one additional distant failure. Five patients presented with an active malignancy at the end of the follow-up.

Conclusions

The 15D HRQoL scores of the patients receiving definitive CRT for head and neck cancer declined on average significantly during the first 3 months after the treatment, but improved gradually towards the end of the follow-up. These changes highlight the importance of the multidisciplinary management and supportive measures in this patient population. In addition, the 15D instrument seems to detect differences between different tumor subgroups, which may be used in customizing the treatment schedules individually. The self-administered generic 15D instrument is a valuable addition to the armamentarium of HRQoL instruments available for head and neck cancer patients undergoing definitive CRT.