This work was performed at the Sydney Head and Neck Cancer Institute, Royal Prince Alfred Hospital, Camperdown, Australia.
Head and Neck
The effect of cancer stage and treatment modality on quality of life in oropharyngeal cancer
Article first published online: 5 AUG 2013
© 2013 The American Laryngological, Rhinological and Otological Society, Inc.
Volume 124, Issue 1, pages 151–158, January 2014
How to Cite
Oates, J., Davies, S., Roydhouse, J. K., Fethney, J. and White, K. (2014), The effect of cancer stage and treatment modality on quality of life in oropharyngeal cancer. The Laryngoscope, 124: 151–158. doi: 10.1002/lary.24136
Presented at the Cancer Nurses Society of Australia 14th Winter Congress, Sydney, New South Wales, Australia, July 21–23, 2011; the Multinational Association of Supportive Care in Cancer International Symposium on Supportive Care, New York, New York, U.S.A., June 28–30, 2012; and the 17th International Conference on Cancer Nursing (Biennial Meeting of the International Society of Nurses in Cancer Care), Prague, Czech Republic, September 9–13, 2012.
The authors have no funding, financial relationships, or conflicts of interest to disclose.
- Issue published online: 20 DEC 2013
- Article first published online: 5 AUG 2013
- Manuscript Revised: 11 MAR 2013
- Manuscript Accepted: 11 MAR 2013
- Manuscript Received: 16 SEP 2012
- Head and neck cancer;
- oropharyngeal cancer;
- quality of life;
- surgery and postoperative radiotherapy;
To examine changes in health-related quality of life among oropharyngeal cancer patients by stages and across treatment types among advanced cancer patients.
Individual prospective cohort study.
All newly diagnosed patients with oropharyngeal cancer treated with curative intent were routinely assessed. The European Organization for Research and Treatment of Cancer (EORTC) both the Main Module quality-of-life questionnaire (QLQ-C30) and the Head and Neck Cancer (HNC) Module (QLQ-H&N35) were administered at diagnosis and 3, 6, and 12 months thereafter. Complete case analysis was used following assessment of missing data. The proportion of patients with clinically significant deterioration (changes of ≥10 points) from baseline were calculated for each follow-up time point and compared by stage (I/II vs. III/IV) and then treatment type (chemotherapy and radiotherapy [CRT] vs. surgery and postoperative radiotherapy [S&PORT]).
Deterioration in most domains was most frequent for stage III/IV patients at 3 months (both modules), whereas stage I/II patients experienced this at 6 months (QLQ-C30) and 12 months (H&N35). Among stage III/IV patients, this happened at all time points for S&PORT patients (QLQ-C30) versus 12 months for CRT patients (H&N35). The number of patients reporting deterioration was lower for most domains at 12 months compared to earlier periods, although dry mouth remained a problem for most patients (60%–85% across treatment/stage groups).
Our preliminary findings suggest that general and disease-specific deterioration is of most concern for stage I/II patients at 6 and 12 months and at 3 months for advanced cancer patients. For stage III/IV patients receiving S&PORT, general deterioration remains a problem after diagnosis, whereas for CRT patients, disease-specific deterioration is of most concern at 12 months.
Level of Evidence
4. Laryngoscope, 124:151–158, 2014
Head and neck cancer (HNC) and its treatment affects both the physical and psychosocial aspects of health-related quality of life (HRQOL). Most HNC patients experience a worsening in quality of life (QOL) during treatment, but then return to pretreatment (baseline) levels by 12 months after treatment,[2-5] although some disease-specific items such as dry mouth, senses, or saliva remain below baseline levels.[2, 4]
Stage of disease, cancer site, and treatment type are predictors of post-treatment HRQOL, particularly disease-specific symptoms. Findings of statistically significant differences in HRQOL scores favoring patients receiving a single therapy compared to combination therapies[5-7] are not consistent across studies. Comparisons of the effect of chemotherapy and radiotherapy (CRT) to surgery and postoperative radiotherapy(S&PORT) on HRQOL have likewise been inconclusive regarding which therapy is associated with more disease-specific symptoms.[8, 9] Further research is required to ascertain the impact of treatment modality on HRQOL, particularly for patients with oropharyngeal cancers.7
Since 2001, prospective longitudinal assessment of HRQOL pre- and post-treatment has been incorporated into standard clinical practice in a tertiary head and neck oncology service at a major metropolitan cancer center in Sydney, Australia. The aim of this article was to examine clinically significant HRQOL changes over time by cancer stage and by treatment modality in oropharyngeal cancer patients.
MATERIALS AND METHODS
Study Design and Sample
The study design is a prospective cohort study. The patient population was drawn from patients being treated at the Sydney Cancer Centre Head and Neck Unit, Royal Prince Alfred Hospital for HNC. Eligibility criteria were treatment with curative intent, no recurrence, English-language ability, no cognitive impairment, and a first diagnosis of HNC. This study presents data collected from 2002 to 2010; data collection, as part of routine multidisciplinary care at the Sydney Cancer Centre Head and Neck Unit, is ongoing. To date, there have been no refusals to participate.
Data Collection Measures
HRQOL was assessed using the European Organization for Research and Treatment of Cancer (EORTC) quality-of-life questionnaire (QLQ), both the Main Module (QLQ-C30) and the Head and Neck Cancer (HNC) Module (QLQ-H&N35). These questionnaires are valid and reliable[11-13] and have been tested widely. The QLQ-C30 includes general function and symptom items and the QLQ-H&N35 assesses HNC-specific symptoms. As per EORTC scoring guidelines, scores were linearly transformed to a 0- to 100-point scale. For function scales (QLQ-C30), 100 represents maximum function and 0 represents minimum function, therefore higher scores indicate better function. For symptom scales (QLQ-C30, H&N35) 100 represents maximum symptoms and 0 represents minimum symptoms, therefore higher scores indicate worse symptoms. This article reports scale measures only.
Information on patient survival status was obtained from the death registry of New South Wales and the database at the Sydney Cancer Centre Head and Neck Unit, which contains patient information including notifications of patient death. Information about patient diagnosis, stage of disease (staged as per American Joint Committee on Cancer recommendations), and treatment type and dose was obtained from hospital records.
Handling of missing data
It was apparent there were missing data due to death and individuals lost to follow-up, and therefore a number of checks were required to determine the best approach for analyzing the data (i.e., complete case or imputation of missing data). Logistic regression analysis could not identify any significant association between clinical characteristics (cancer stage, treatment type, global health status from the previous period) and missing data (P > 0.10 in all analyses).
Nonparametric tests were used to explore whether there were differences in change over time between the different datasets created to explore missing values: 1) complete case; 2) complete case and deceased patients, with worst possible scores (0 for function scores, 100 for symptom scores)[17, 18] assumed for deceased patients; 3) imputation of all missing data; and 4) worst possible scores assumed for deceased patients, other missing data imputed. Imputation was by single imputation using the expectation maximization algorithm (n = 100 iterations) for all HRQOL items with clinical characteristics included as auxiliary variables. Little's missing completely at random (MCAR) test was employed for datasets 2 and 4, and in both cases the null hypothesis of MCAR was not rejected. The median HRQOL scores and pattern of changes in the scores over time were compared for all datasets and were similar if not identical across datasets, supporting a complete case analysis approach.
All analyses were conducted using IBM SPSS version 19 (IBM, Armonk, NY). Clinical significance is of relevance when employing HRQOL assessments as part of clinical practice, therefore clinically significant changes in HRQOL outcomes, defined to be a change of 10 points or more,[20, 21] were assessed. Clinically significant change has been described as a difference in score that is large enough to have an implication for the patient's treatment or care. Changes in each domain for each time point for individuals were computed by taking the difference for each HRQOL domain from the baseline (e.g., 3 months–baseline). As per recommendation, the proportion of patients with clinically significant deterioration compared to baseline (decrease of ≥10 points for function scales, increase of ≥10 points for symptom scales), clinically significant improvement (increase of ≥10 points for function scales, decrease of ≥10 points for symptom scales), or clinically stable HRQOL was estimated. Recently, there have been suggestions to consider other approaches to clinical significance, and changes in scores that vary across the domains of the main module have been suggested. The differences in scores compared to baseline were also assessed using the recent guidelines of Cocks et al., which are available for the main (QLQ-C30) module, specifically the proportion of patients with trivial, small, medium, and large deterioration by the Cocks et al. guidelines.
Clinically significant changes in HRQOL scores are reported in preference to P values, as statistically significant changes in HRQOL measures do not necessarily imply clinical relevance. For readability only, results relating to clinically significant deterioration are presented.
The proportion of patients with clinically significant deterioration could not be compared across treatment groups directly, because the treatments were not evenly balanced across stages. CRT was used predominantly in stage III/IV patients (95%), as was S&PORT (72%), whereas radiotherapy alone was predominantly used in stage I/II patients (57%). Proportions were therefore compared by cancer stage (early: stage I/II, advanced: stage III/IV) and then treatment type within the advanced cancer group, as most patients were in that group. Due to the low number of patients with stage III/IV cancer receiving radiotherapy alone (n = 3), comparisons within this group were restricted to patients receiving combination therapy.
The study was approved by the Human Research Ethics Committee of the Sydney Southwest Area Health Service (Royal Prince Alfred Hospital Zone). Because the study is ongoing, the application is resubmitted for review and approval on an annual basis. Informed consent was obtained from all patients prior to data collection.
Demographic and Clinical Characteristics
Ninety-six participants have participated since 2002, and 60 patients completed questionnaires at all time points. Table 1 presents the demographic and clinical characteristics of the full sample. When survival status was assessed in April 2011, just over half of the patients (59%) were still alive. The most common cancer site was the tonsil (56%), and most patients were stage IV at diagnosis (62%).
|Survival status as of April 2011|
|Soft palate||5 (5)|
|Tongue base||37 (39)|
|Surgery only||8 (8)|
|Radiotherapy only||11 (11)|
|Surgery and postoperative radiotherapy||21 (22)|
|Chemotherapy and radiotherapy||56 (58)|
Fifty-six patients received CRT. Among these patients, 27 (48%) received cisplatin third weekly, whereas seven (13%) received cetuximab weekly, nine (16%) received cisplatin weekly, and two (4%) received carboplatin weekly. A further six (11%) were enrolled in a trial of tirapazamine, and regimen information was unknown for the five (9%) treated at another hospital.
Among CRT patients, 44 had percutaneous endoscopic gastrostomy (PEG) tubes inserted pretreatment. Of these, six (11%) had PEG tubes until 2 months following treatment, 15 (27%) at 3 months, and 11 (20%) at 6 months. Although 11 (20%) were PEG dependent, three (5%) refused PEG and four (7%) had a comorbidity that excluded PEG use. Information on PEG tubes was not available for the five (9%) treated at another hospital.
For those 88 patients who received radiotherapy, 43 (49%) had a dose of <70 Gy, and 21 (24%) had a dose of ≥70 Gy. Intensity-modulated radiotherapy (IMRT) was given to 11 (13%) patients; as the hospital introduced IMRT in 2008, all patients receiving IMRT have been treated since that time. Information on radiotherapy dose is not available for the 13 patients (15%) who were treated at another hospital.
Table 2 compares the proportion of patients with clinically significant worsening for the main HRQOL module. The highest proportion of stage III/IV patients with worsened scores compared to baseline occurred at 3 months, whereas for stage I//II patients, the highest proportion occurred at 6 months. Between the two stages, at each time point a higher proportion of patients with stage III/IV cancer experienced clinically significant deterioration across more domains compared to patients with stage I/II cancer. The disparity between the two groups was greatest at 3 months, where stage III/IV patients reported a higher proportion of clinically worse scores (13 of the 15 domains) compared to stage I/II (one of the 15 domains). By 6 and 12 months, the differences between the groups were less pronounced, and the number of patients experiencing deterioration was lower. However, one-third or more of patients still experienced clinically significant deterioration at 12 months: global health status (31% of stage III/IV), role function (33% of stage III/IV), fatigue (38% of stage III/IV), and insomnia (36% of stage I/II and 33% of stage III/IV).
|3 Months||6 Months||12 Months||When Highest Proportion Within Each Group Occurs|
|Global health status|
|Stage I/II||18%||36%||9%||6 months|
|Stage III/IV||58%||35%||31%||3 months|
|Stage I/II||18%||27%||27%||6 and 12 months|
|Stage III/IV||33%||25%||21%||3 months|
|Stage I/II||9%||27%||9%||6 months|
|Stage III/IV||50%||40%||33%||3 months|
|Stage III/IV||29%||27%||25%||3 months|
|Stage I/II||9%||27%||9%||6 months|
|Stage III/IV||25%||23%||15%||3 months|
|Stage I/II||36%||36%||27%||3 and 6 months|
|Stage III/IV||29%||29%||21%||3 and 6 months|
|Stage I/II||27%||46%||18%||6 months|
|Stage III/IV||58%||50%||38%||3 months|
|S&PORT||69%||69%||54%||3 and 6 months|
|Nausea and vomiting|
|Stage I/II||9%||0%||0%||3 months|
|Stage III/IV||10%||2%||0%||3 months|
|Stage I/II||18%||18%||9%||3 and 6 months|
|Stage III/IV||48%||27%||17%||3 months|
|Stage I/II||0%||18%||18%||6 and 12 months|
|Stage III/IV||19%||8%||8%||3 months|
|Stage I/II||27%||36%||36%||6 and 12 months|
|Stage III/IV||46%||38%||33%||3 months|
|S&PORT||39%||39%||31%||3 and 6 months|
|Stage I/II||9%||18%||18%||3 and 6 months|
|Stage III/IV||42%||33%||19%||3 months|
|Stage I/II||0%||18%||0%||6 months|
|Stage III/IV||10%||13%||10%||6 months|
|S&PORT||0%||15%||15%||6 and 12 months|
|Stage III/IV||6%||0%||0%||3 months|
|Stage I/II||9%||18%||0%||6 months|
|Stage III/IV||25%||25%||15%||3 and 6 months|
|Stage I/II worse||1/15||4/15||4/15|
|Stage III/IV worse||13/15||6/15||8/15|
For the comparison across treatment types, the greatest deterioration in the main module was reported at 3 months, as these were all stage III/IV patients. Between the two treatment types, a higher proportion of patients receiving S&PORT experienced clinically significant deterioration in more domains compared to patients receiving CRT, although differences between the two groups were smaller by 12 months. S&PORT was consistently worse than CRT across all time periods for cognitive function, fatigue, and pain, whereas CRT was consistently worse for physical function.
Clinically significant deterioration in fatigue was reported by just over half to two-thirds of both treatment groups at 3 months. Although a smaller proportion of the CRT group reported worsening at 12 months, over 50% of the S&PORT group were still reporting clinically significant deterioration at 12 months. Nearly one-third of S&PORT patients continued to report deterioration in the domains of global health status, role function, and insomnia at 12 months.
By the Cocks et al. guidelines, for the symptom scales, when clinically significant deterioration was present, such deterioration was entirely medium and/or large. For the function scales, at least 50% of patients had medium and/or large clinically significant deterioration when such deterioration was present. The highest proportion of patients with small and/or trivial deterioration for the function scales at any time point was 33.4% (physical function at 6 months compared to baseline).
Table 3 compares the proportion of patients with clinically significant worsening for the disease-specific HRQOL module. For the comparison across disease stages, results were similar for the head and neck module. Within each of the two stages, the highest proportion with clinically significant worsening occurred at 3 months for stage III/IV and at 12 months for stage I/II. Between the two stages a higher proportion of stage III/IV patients reported worsening than stage I/II at 3 months (11 of 13 domains compared to two of 13) and 6 months (eight of 13 compared to two of 13).
|3 Months||6 Months||12 Months||When Highest Proportion Within Each Group Occurs|
|Stage I/II||9%||18%||0%||6 months|
|Stage III/IV||29%||17%||17%||3 months|
|Stage I/II||9%||0%||9%||3 and 12 months|
|Stage III/IV||42%||27%||19%||3 months|
|Stage I/II||18%||18%||27%||12 months|
|Stage III/IV||48%||31%||23%||3 months|
|Stage III/IV||23%||19%||10%||3 months|
|CRT||19%||19%||13%||3 and 6 months|
|Stage I/II||36%||18%||18%||3 months|
|Stage III/IV||58%||42%||17%||3 months|
|Stage I/II||18%||9%||9%||3 months|
|Stage III/IV||23%||4%||4%||3 months|
|Stage I/II||27%||36%||36%||6 and 12 months|
|Stage III/IV||44%||52%||40%||6 months|
|S&PORT||46%||46%||23%||3 and 6 months|
|Stage I/II||9%||18%||27%||12 months|
|Stage III/IV||25%||31%||25%||6 months|
|Stage I/II||36%||27%||0%||3 months|
|Stage III/IV||29%||33%||25%||6 months|
|Stage I/II||82%||73%||73%||3 months|
|Stage III/IV||73%||67%||60%||3 months|
|Stage I/II||18%||27%||27%||6 and 12 months|
|Stage III/IV||40%||29%||23%||3 months|
|Stage I/II||0%||9%||18%||12 months|
|Stage III/IV||21%||10%||6%||3 months|
|Stage I/II||9%||9%||0%||3 and 6 months|
|Stage III/IV||29%||19%||8%||3 months|
|Stage I/II worse||2/13||2/13||5/13|
|Stage III/IV worse||11/13||8/13||5/13|
By 12 months, both groups reported a similar number of deteriorated symptoms. Despite the overall trend toward improvement, over one-third of patients from both groups continued to report deterioration at 12 months in the sexuality domain (stage I/II: 36%, stage III/IV: 40%), and a majority of patients from both groups reported deterioration for dry mouth (stage I/II: 73%, stage III/IV: 60%). Furthermore, there was greater deterioration over time for five domains for stage I/II: senses (18% to 18% and 27%), sexuality (27% to 36% and 36%), teeth (9% to 18% to 27%), sticky saliva (18% to 27% to 27%), and coughing (0% to 9% to 18%).
For comparison across treatments, the greatest deterioration in the head and neck module was reported at 3 months. Between the two treatment types, a higher proportion of patients receiving S&PORT experienced clinically significant deterioration in more domains compared to CRT patients at 3 months, although by 12 months this had reversed, when higher proportions of CRT patients reported clinically significant deterioration. Deterioration in dry mouth symptoms was reported by a majority of both treatment groups at 3 months (66% for CRT, 85% for S&PORT) and was still reported by more than 50% of both groups at 12 months. For S&PORT patients, worsening of pain and sticky saliva remained relatively high at 12 months (both 31%), whereas 44% of CRT patients reported worsening sexuality at 12 months.
The prevalence of clinically significant deterioration among patients with oropharyngeal cancer varied by cancer stage and by treatment type among patients with advanced cancer. Furthermore, differences were apparent across the HRQOL modules, with proportionally more advanced cancer and S&PORT patients experiencing deterioration in domains of the main module but proportionally more stage I/II and CRT patients experiencing deterioration in domains of the head and neck-specific module.
These findings, although preliminary due to the small number of patients in the study, are broadly consistent with other findings of deterioration during treatment and then subsequent improvement within the next 12 months.[2, 24] Differences in terms of length of follow-up and type of treatment may help explain why our findings for dry mouth and sexuality show much greater deterioration than reported in a number of other studies.[2, 26, 27] Furthermore, other studies examining HRQOL in oropharyngeal cancer have used different instruments[28, 29]or have not been prospective,[7-9, 30] which may also explain the divergence in findings. Few studies have also reported clinical significance as part of their analysis,[24, 31] presenting an additional challenge. In this study, a difference of at least 10 points was considered clinically significant, and an assessment of the differences in scores against newly released guidelines relating to interpretation of differences in the main module indicated that most clinically significant deterioration seen in this study was medium to large.
In contrast to Tschudi et al., we found differences among the treatment types in both the main and head and neck specific modules. In our study, S&PORT patients experienced deterioration in higher proportions in more domains of the main module, whereas the reverse was true for the head and neck-specific module. Tschudi et al. found differences to be most pronounced in the head and neck module only. The domains in which we found greater deterioration for S&PORT patients also differed from those reported in Tschudi et al.
However, like Boscolo-Rizzo and colleagues, we found that patients receiving CRT performed comparably better in the domains of the main module but worse in the domains of teeth, opening mouth, and dry mouth in the head and neck module. For S&PORT patients, we found less deterioration in the domains Boscolo-Rizzo et al. found to be more problematic for this patient group. We also found that differences remained after treatment, in contrast to the findings of Mowry et al. A possible explanation for these divergences is length of follow-up and study design; we used a prospective design and followed patients for 1 year after treatment, whereas all three studies looked at patients at least 2 years post-treatment.
Comparing our findings with regard to stage and HRQOL is difficult, as most studies looking across stage in HNC do not focus specifically on oropharyngeal cancer patients. Previous work has identified that unsurprisingly, patients with advanced-stage cancer have worse symptom and function scores and greater deterioration over time than patients with early-stage cancer. Our results supported this in that patients with advanced cancer had greater deterioration at all time periods compared to early-stage patients, and was particularly marked at 3 months, although the differences between the two groups were much less pronounced by 12 months. Only social function appeared to be a consistently worse problem for early-stage cancer patients.
In this study, xerostomia remained a significant complaint at 12 months; these results may improve since IMRT was performed at this institution 7 years after the introduction of prospective QOL assessment. IMRT has been shown to have fewer oral-related symptoms and better global QOL outcomes compared to two-dimensional radiotherapy and three-dimensional conformal radiotherapy.
A strength of this study is the thorough assessment of missing data, which lends support for the ultimate decision to use complete case analysis. Furthermore, although attrition did occur in this study, primarily due to death, missing data due to not completing aspects of the questionnaire were minimal. In particular, the problem of patients not completing the sexuality item[6, 14, 33] did not occur in this study. The integration of prospective HRQOL assessment into routine clinical practice and the high participation rate may make these findings particularly relevant for clinicians treating patients with oropharyngeal cancer with curative intent.
This study also had several limitations. Although missingness was rigorously assessed, definitive conclusions regarding the missingness mechanism are not possible, therefore we cannot rule out bias. Due to the small numbers of patients in the early-stage cancer group, the findings for this group cannot be considered definitive. Similarly, the small numbers in each treatment group among the advanced cancer patients mean that this study is an exploratory assessment, and the findings are useful in indicating further areas for research and assessment rather than providing clear guidelines for clinical practice. This was a prospective study conducted over a long time period, and during this time chemoradiation treatment changed; as noted above, IMRT was introduced several years after the prospective assessment commenced. A subset analysis to examine HRQOL variations across differing treatment groups was not possible due to the small number of patients.
This study is also one of few to carry out prospective assessment of HRQOL in oropharyngeal cancer. It is also one of very few to examine the impact of combination therapies on HRQOL in patients with oropharyngeal carcinomas,[8, 9] rather than assessing a combination therapy versus a primary therapy or some mixture of combination and/or primary therapies.[27, 35]
Among patients with oropharyngeal cancer, HRQOL varied across cancer stage and then across treatment modality for patients with advanced cancer. Although improvement was seen in most domains by 12 months after treatment, our findings indicate that closer attention may need to be paid to early-stage patients at 6 and 12 months after diagnosis, and to advanced-stage patients at 3 months after diagnosis. In addition, attention should be paid to general symptoms and function after diagnosis for advanced cancer patients receiving S&PORT, as well as a focus on head and neck-specific symptoms 3 months after diagnosis for this group. Advanced cancer patients receiving CRT may need additional attention at 12 months postdiagnosis for head and neck specific symptoms. Further research is required to confirm these findings.
The authors gratefully acknowledge their participants, without whom the study would not have been possible. The assistance of Claire Chang, Louise Acret, and Eleanor Whitford with data entry and literature searching is also acknowledged. The authors also thank the audiences at the Cancer Nurses Society of Australia Winter Congress (2011), Multinational Association for Supportive Care International Symposium (2012), and International Conference on Cancer Nursing (2012) for helpful feedback. Attendance at the Winter Congress was supported by the Cancer Council NSW.
- 15The EORTC QLQ-C30 Scoring Manual. 3rd ed. Brussels, Belgium: European Organisation for Research and Treatment of Cancer; 2001., , , et al.
- 16Edge SB, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A, eds. AJCC Cancer Staging Manual. 7th ed. New York, NY: Springer-Verlag; 2009.
- 34Applied Missing Data Analysis. New York, NY: The Guilford Press; 2010..