Originally published as JCO Early Release 10.1200/JCO.2009.21.8438 on May 18 2009

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Google Scholar Articles by Browman, G. P. Articles by Smith, S. PubMed PubMed Citation Articles by Browman, G. P. Articles by Smith, S. Social Bookmarking                            What's this?

Prognostic Tools for Cancer Survival: A Secondary Role for Quality-of-Life Measurement

Department of Medical Oncology, School of Population and Public Health, University of British Columbia; and Vancouver Island Centre, British Columbia Cancer Agency, Vancouver, British Columbia, Canada

Tanya Berrang

Sally Smith

Vancouver Island Centre, British Columbia Cancer Agency, Vancouver, British Columbia, Canada

Data from cancer patients, including those with head and neck cancer, suggest that those treated for cure or long-term survival will accept significant morbidity to achieve small survival gains.^1,2 Even so, almost 25% of patients with head and neck cancer ranked cure as only the second or third most important outcome, with items related to symptoms and functional outcomes, including appearance, ranking among the top three most important.² Experienced oncologists understand that for a significant proportion of patients with cancer, maintaining or improving health-related quality of life (HRQOL) is an important objective when considering treatment choices.

However, HRQOL is a more abstract notion than is "life or death," and it is more difficult to measure. Oncologists have tended, particularly in the clinical research domain, to focus on survival as paramount, often at the expense of more relevant clinical outcomes. For example, in a meta-analysis of randomized, controlled trials on the use of palliative chemotherapy for incurable (recurrent and distant metastatic) head and neck cancer, it was found that reports of all studies included survival and tumor response but did not include outcomes that would measure palliative benefits.³

Several strategies have been explored to address this imbalance and to better align study end points with outcomes that matter to patients. Pain and emotional distress have been promoted as the fifth and sixth vital signs, respectively.^4,5 Canada has chosen the rebalance focus (now called the cancer journey) as one of the few priority areas for action in its national cancer control strategy.⁶ Motivated by the adage "you can't manage what you don't measure," a growing body of literature is promoting the measurement of pain, emotional distress, and HRQOL to focus attention on better managing these relatively neglected areas. In the early 1990s, the National Cancer Institute of Canada Clinical Trials Group instituted HRQOL measurement as a default requirement for all its comparative clinical trials.⁷

Clinical teams involved with management of head and neck cancer have benefited from a relatively active focus by researchers on quality-of-life measurement,^8,9 with several validated instruments now available. This reflects the dramatic effects of the disease itself and the impact of treatment across several HRQOL domains, including physical, social, and emotional.

With so much energy spent to refocus the clinical lens on the softer outcomes that matter to patients, it is ironic that the interest and curiosity of clinical oncologists with regard to the value of HRQOL measurement have peaked because of the link between HRQOL and survival, as if the field of HRQOL has gained more credibility because of its association with this outcome. This phenomenon was most obviously illustrated when Spiegel et al¹⁰ reported their observation that a psychosocial intervention designed to address emotional well-being in a randomized trial actually had an impact on survival. Despite several attempts, to our knowledge, this observation so far has not been reproduced.

In this issue of Journal of Clinical Oncology, Meyer et al¹¹ report on quality-of-life measurement as an independent prognostic indicator for survival in patients with early-stage (I and II) squamous cell head and neck cancer. This is the fourth in a series of reports^12–14 emanating from a well-designed and well-executed placebo-controlled, randomized trial. The primary objective of this trial was to evaluate the role of antioxidant medications (ie, beta carotene and vitamin E) in the occurrence of second primary cancers, with overall survival as one of several secondary outcomes. In the three previous reports of this trial, the investigators established that the use of antioxidant medications did not reduce the frequency of second primary cancers,¹² was associated with reduced overall survival,¹³ and was not associated with improvement in HRQOL,¹⁴ although there were mixed effects on acute adverse events depending on the subgroup analysis used.

In the previous reports in this series,^12–14 all comparisons were directed at the randomized groups, whereas the report by Meyer et al¹¹ in this issue of JCO presents results from a pooled analysis of the randomized groups. Multivariate statistics including a Cox proportional hazards model were used to determine the independent contributions of different factors to prognosis, including baseline HRQOL and the change scores at 6 months after completion of radiotherapy. Meyer et al report that for every 10-point difference in baseline HRQOL score, there was a corresponding 13% difference in risk of mortality in the expected direction (ie, favoring better HRQOL). The physical functioning domain was the only change score at 6 months that independently predicted survival, with a mortality hazard ratio of 0.75 at 1 year for every 10 points of change. However, over time, its predictive performance diminished.

As appropriately cited by Meyer et al,¹¹ there have been numerous similar observations of the predictive performance of HRQOL measures on survival. The strengths of the current study are its rigorous experimental design to address the primary question with methodologic strategies to minimize bias, documented high levels of compliance and data capture for the HRQOL instruments, and sophisticated statistical analytic approach. The study involved patients with early-stage disease, thus largely avoiding the thorny statistical and interpretive issues associated with missing data over time that plague most longitudinal studies of changes in HRQOL. The strength of the study might have been improved with the use of test and validation samples to ensure that the performance of the predictive model derived from the test population could be confirmed in a separate but comparable group. However, taking all methodologic issues into account, it is reasonable to conclude that the inferences drawn from the analyses are valid. Baseline HRQOL and the 6-month change score for the physical functioning domain are independent predictors of overall survival in patients with stage I and II head and neck cancer treated with radiotherapy.

Having established the validity of the findings, the question becomes: what is the clinical utility of HRQOL as a predictive variable in this clinical scenario? There is no universal definition of clinical utility that includes domains such as effectiveness (or performance) of the intervention (or test), adverse effects, accessibility, ease of administration and interpretation, acceptance by the health care provider and patient, and affordability.¹⁵ A useful construct is that clinical utility of an intervention represents the value conferred on the population of individuals who are exposed to the intervention, including those who benefit and those who do not. A classic illustration of the usefulness of this definition is the debate over prostate-specific antigen screening for prostate cancer. In this case, utility is considered in the context of not only those whose cancers have been detected or whose survival can be influenced but also the consequences for individuals and the health system related to misclassification (false positives and false negatives), which could lead to unnecessarily aggressive follow-up procedures that may do more harm than good for the population exposed and drive up costs and resource utilization.¹⁶

In a recent editorial in JCO, Bruera and Hui¹⁷ reviewed key issues associated with the development and evaluation of prognostic tools and models for clinical use in patients with advanced cancer (which readers are encouraged to consult). We will not reiterate their points here, except to comment on their conclusion that "future models will need to be sophisticated enough to provide highly accurate prognostic information, yet simple enough to allow application by busy clinicians in everyday practice." They explained the usefulness of models that cater to the usual decisional heuristics, such as using prognostic categories as opposed to complex formulas.¹⁸ In the clinical domain (and even among the seemingly healthy), there exists the reality of uncertainty related to time of death; we need to ask how finely honed prognostic tools need to be and whether there is a threshold beyond which greater precision adds little value.

In routine clinical care, prognostic factors may be useful in communicating a prognosis to a patient or evaluating alternative treatment choices. In early-stage head and neck cancer, the difference in HRQOL associated with alternative curative modalities (eg, surgery versus radiotherapy with or without chemotherapy) may in fact be starker than are differences in the magnitude of survival produced. HRQOL measures may contribute additional information statistically, but how should this information be factored into a conversation with the patient when there are so many other factors at play, many of which have more intuitive links with survival? In the population studied by Meyer et al,¹¹ overall survival, according to a previous report,¹² was more than 75% at 5 years. The current state of management and prognosis of early-stage head and neck cancer does not seem to render the findings of Meyer et al clinically useful at this time. The possible exception is related to the baseline and 6-month change scores for physical performance, which might allow clinicians to better select patients for more intensive and frequent follow-up when salvage might be effective for recurrent disease.

The more useful role for the observations of Meyer et al¹¹ is in connection with improving the design, analysis, or interpretation of future randomized trials of interventions in early-stage head and neck cancer. In the design phase, prognostic models that perform better than do traditional single factors (eg, T stage, N stage, overall stage, disease site, and performance status) can incorporate HRQOL data if prospective stratification is considered. The distribution of HRQOL scores between randomized groups at baseline could be used to check for comparability of groups (ie, how well the randomization worked). The 6-month change score in physical function might be used in the same way. The final analyses could incorporate HRQOL scores as variables in Cox modeling for time-to-event analyses. The conversion of HRQOL scores to actual health utility scores might also be useful in economic evaluation studies that are incorporated into comparative clinical trials to estimate variables such as quality-adjusted life years. However, the utility of HRQOL for clinical trials would need to take into account as key concerns the ease of use, acceptance by researchers and patients, and additional cost of measurement of such domains. Are the data compelling enough at this stage to warrant the additional complexity of incorporating such a measure routinely into trials of early-stage head and neck cancer?

The answer to this question is left for readers and researchers to contemplate. But let us not regress back to our old ways and attach significance to HRQOL only in relation to our attachment to survival as a clinical outcome. Measuring HRQOL should have value in its own right. As the field evolves, it should acquire greater clinical importance and expand the lessons we take away from clinical trials.

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The author(s) indicated no potential conflicts of interest.

AUTHOR CONTRIBUTIONS

Conception and design: George P. Browman

Manuscript writing: George P. Browman, Tanya Berrang, Sally Smith

Final approval of manuscript: George P. Browman, Tanya Berrang, Sally Smith

REFERENCES

1. Ravdin PM, Siminoff IA, Harvey JA: Survey of breast cancer patients concerning their knowledge and expectations of adjuvant therapy. J Clin Oncol 16:515–521, 1998.[Abstract]

2. List MA, Stracks J, Colangelo L, et al: How do head and neck cancer patients prioritize treatment outcomes before initiating treatment? J Clin Oncol 18:877–884, 2000.[Abstract/Free Full Text]

3. Browman GP, Cronin L: Standard chemotherapy in squamous cell head and neck cancer: What have we learned from randomized trials? Semin Oncol 21:311–319, 1994.[Medline]

4. Walid SW, Donahue SN, Darmohray DM, et al: The fifth vital sign: What does it mean? Pain Pract 8:417–422, 2008.[CrossRef][Medline]

5. Bultz BD, Carlson LE: Emotional distress: The sixth vital sign in cancer care. J Clin Oncol 23:6440–6441, 2005.[Free Full Text]

6. Canadian Partnership Against Cancer. Cancer journey. http://www.partnershipagainstcancer.ca/journey.

7. Osoba D: The Quality of Life Committee of the Clinical Trials Group of the National Cancer Institute of Canada: Organization and functions. Qual Life Res 1:211–218, 1992.[CrossRef][Medline]

8. Gotay CC, Moore TD: Assessing quality of life in head and neck cancer. Qual Life Res 1:5–17, 1992.[CrossRef][Medline]

9. Rogers SN, Ahad SA, Murphy AP: A structured review and theme analysis of papers published on "quality of life" in head and neck cancer: 2000-2005. Oral Oncol 43:843–868, 2007.[CrossRef][Medline]

10. Spiegel D, Kraemer HC, Bloom JR, et al: Effect of psychosocial treatment on survival of patients with metastatic breast cancer. Lancet 2:888–891, 1989.[CrossRef][Medline]

11. Meyer F, Fortin A, Gélinas M, et al: Health-related quality of life as a survival predictor for patients with localized head and neck cancer treated with radiation therapy. J Clin Oncol 27:2970–2976, 2009.[Abstract/Free Full Text]

12. Bairati I, Meyer F, Gélinas M, et al: A randomized trial of antioxidant vitamins to prevent second primary cancers in head and neck cancer patients. J Natl Cancer Inst 97:481–488, 2005.[Abstract/Free Full Text]

13. Bairati I, Meyer F, Jobin E, et al: Antioxidant vitamins supplementation and mortality: A randomized trial in head and neck cancer patients. Int J Cancer 119:2221–2234, 2006.[CrossRef][Medline]

14. Bairati I, Meyer F, Gélinas M, et al: Randomized trial of antioxidant vitamins to prevent acute adverse effects of radiation therapy in head and neck cancer patients. J Clin Oncol 23:5805–5813, 2005.[Abstract/Free Full Text]

15. Smart A: A multi-dimensional model of clinical utility. Int J Qual Health Care 18:377–382, 2006.[Abstract/Free Full Text]

16. Twombly R: Preventive Services Task Force recommends against PSA screening after age 75. J Natl Cancer Inst 100:1571–1573, 2008.[Free Full Text]

17. Bruera E, Hui D: Practical model for prognostication in advanced cancer patients: Is less more? J Clin Oncol 26:5843–5844, 2008.[Free Full Text]

18. Sloan JA, Loprinzi CL, Laurine JA, et al: A simple stratification factor prognostic for survival in advanced cancer: The good/bad/uncertain index. J Clin Oncol 19:3539–3546, 2001.[Abstract/Free Full Text]

CiteULike    Complore    Connotea    Del.icio.us    Digg    Facebook    Reddit    Technorati    Twitter    What's this?

This Article Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Save to my personal folders Download to citation manager Rights & Permissions Google Scholar Articles by Browman, G. P. Articles by Smith, S. PubMed PubMed Citation Articles by Browman, G. P. Articles by Smith, S. Social Bookmarking                            What's this?