|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2003 American Society for Clinical Oncology Early Change in Patient-Reported Health During Lung Cancer Chemotherapy Predicts Clinical Outcomes Beyond Those Predicted by Baseline Report: Results From Eastern Cooperative Oncology Group Study 5592
From Evanston Northwestern Healthcare and Northwestern University, Evanston, and Rush-Presbyterian St. Luke’s Medical Center, Chicago, IL; University of Colorado Health Sciences Center, Denver, CO; and Vanderbilt University, Nashville, TN. Address reprint requests to David T. Eton, PhD, 1001 University Place, Suite 100, Evanston, IL 60201; email: d-eton{at}northwestern.edu.
Purpose: To determine the ability of longitudinal patient-reported health (PRH) scores to enhance prediction of clinical outcomes beyond baseline scores. Patients and Methods: In 573 advanced non–small-cell lung cancer patients enrolled in a phase III clinical trial, we used baseline and 6-week follow-up PRH scores to predict best response to treatment, disease progression, and survival. Using regression analyses, we tested the predictive ability of the five subscales of the Functional Assessment of Cancer Therapy–Lung (physical, functional, social/family, emotional well-being, and the lung cancer subscale) as well as the trial outcome index (TOI) aggregate score. Results: After clinical factors were controlled for, baseline physical well-being (PWB) and TOI scores predicted all three clinical outcomes. A higher baseline PWB score was associated with a better response to treatment (odds ratio, 1.09; P < .001) and lower risk of death (risk ratio, 0.95; P < .001). Higher baseline TOI score was associated with a lower risk of disease progression (risk ratio, 0.98; P < .001). These two baseline predictors (PWB and TOI) were then used along with 6-week change scores to classify patients into four groups: low baseline-declined, low baseline-improved, high baseline-declined, and high baseline-improved. Patients with low baseline-declined PWB scores showed the worst responses to treatment and survived the shortest duration. Patients with low baseline-declined TOI scores had the shortest time to progression. Conclusion: The physical aspects of baseline PRH and PRH change during chemotherapy are significant predictors of clinical outcomes in lung cancer. This has implications for patient stratification in clinical trials and may aid decision-making in clinical practice.
PATIENT-REPORTED health (PRH) data such as scores from health-related quality-of-life measures are becoming an integral part of outcomes assessment in cancer clinical trials. They have traditionally been used to supplement clinical data (ie, objective tumor response and time to disease progression) to determine treatment efficacy.1,2 PRH data also may be valuable to the clinician because they provide information on subjective patient outcome that can inform therapeutic course. Several investigators have evaluated PRH as a prognostic factor. PRH data have been found to be predictive of length of survival and tumor response, independent of other disease-related factors.3–24 This marks an important secondary usage of PRH data that could have profound implications for the conduct of clinical trials, while also providing important prognostic information. The importance of such data is evident in advanced lung cancer—a disease that profoundly affects patient-reported outcomes.25,26 In predicting objective outcomes, some researchers have used performance status (physician-rated status, patient-rated status, or both) as a prognostic factor.3–5 Recently, Sloan et al5 found that a brief index incorporating both physician and patient-rated performance status ratings predicted survival in advanced colorectal and lung cancer patients. Performance status ratings are useful, but they may not capture the full breadth of a patient’s subjective health status.27 Multidimensional measures of PRH and/or health-related quality of life (HRQL) provide health status information in a host of different domains. Numerous studies6–21,24 have used scores from such measures to predict objective clinical outcome, principally survival, in treated patients. Commonly used measures include the Functional Living Index–Cancer,18,19 the European Organization for the Research and Treatment of Cancer’s Core Quality-of-Life Questionnaire (EORTC QLQ-C30),6,7,21,24 and Linear Analog Self-Assessment Scales.10,12 PRH scores have predicted survival in a variety of patient samples. These include heterogeneous (mixed cancer) samples6–9 and site-specific samples, such as advanced breast,10,11 malignant melanoma,12,13,22 colorectal,14 head and neck,15 and esophageal cancers.16 Studies have also targeted advanced lung cancer patients.17–21,28 Though many different PRH domains have been used, scores of physical symptoms (ie, pain, fatigue), physical well-being, and overall HRQL have been the best predictors of survival. Throughout most of this work, only baseline (usually pretreatment) PRH scores have typically been evaluated. However, longitudinal PRH scores may add to the predictive value. For example, Coates et al10 found that on-treatment improvements in physical well-being, mood, and pain were associated with longer survival in advanced breast cancer patients. Blazeby et al16 have recently shown that 6-month changes in emotional functioning scores on the EORTC QLQ-C30 were associated with survival duration in esophageal patients. Improvements in emotional functioning (from pretreatment baseline to 6-month follow-up) were associated with longer survival. Although they are helpful, these findings do not lead to definitive conclusions because of the small sample sizes used. The purpose of this analysis was to evaluate whether longitudinal PRH data in lung cancer patients receiving first-line chemotherapy for advanced disease might have clinical relevance beyond that provided by baseline testing. We used the Functional Assessment of Cancer Therapy–Lung (FACT-L) as the basis for prediction.
In Eastern Cooperative Oncology Group (ECOG) study 5592, 599 advanced non–small-cell lung cancer (NSCLC) patients were randomly assigned to receive one of three chemotherapeutic regimens: cisplatin 75 mg/m2 intravenously [IV] over 1 hour plus etoposide 100 mg/m2 IV for 3 days, cisplatin 75 mg/m2 IV over 1 hour plus standard-dose paclitaxel 135 mg/m2 IV over 24 hours, or cisplatin 75 mg/m2 over 1 hour plus high-dose paclitaxel 250 mg/m2 IV over 24 hours with granulocyte colony-stimulating factor support (5 µg/kg/d). Newly diagnosed patients with stage IIIB or IV disease, an ECOG performance status rating of 0 (normal activity) or 1 (some symptoms, but ambulatory), and no evidence of brain metastases were eligible for the study. Twenty-six patients were excluded from the final analysis because they failed to meet inclusion criteria or withdrew from the protocol before any treatment; thus, the final sample size was 573. The primary objective of E5592 was to compare survival across treatments (complete results were reported elsewhere).29 A secondary objective was to compare PRH across treatments. Assessments were made at baseline (pretreatment) and at 6 weeks, 12 weeks, and 6 months. The 6-week assessment corresponded to the start of the third cycle of treatment; the 12-week assessment corresponded to the start of the fifth cycle of treatment. Using a joint mixed effects and survival model, Bonomi et al29 found that patients treated on a paclitaxel arm showed greater improvements in the physical aspects of HRQL than patients receiving standard regimen cisplatin plus etoposide. We determined the ability of PRH to predict clinical outcomes using baseline (pretreatment) and follow-up PRH data. A 6-week follow-up time point was selected to minimize the loss of data because of mortality or on-study progression. PRH was assessed with version 2 of the FACT-L.30 The FACT-L consists of four general and one lung cancer symptom-specific subscale. General subscales include physical well-being (PWB; seven items), social/family well-being (SWB; seven items), emotional well-being (EWB; five items), and functional well-being (FWB; seven items). The seven-item lung cancer subscale (LCS) assesses symptoms commonly reported by lung cancer patients (eg, shortness of breath, loss of weight, tightness in chest). A 21-item trial outcome index (TOI) is derived by adding PWB, FWB, and LCS scores. Among FACT-L scores, the TOI is thought to be the most relevant and precise indicator of PRH available for lung cancer patients in an oncology clinical trial.30 All FACT-L items are rated on five-point scales ranging from 0 for not at all to 4 for very much. Higher scores are representative of better PRH or fewer symptoms. Demographic and clinical data were abstracted from ECOG case report forms. Date of birth; sex; ethnicity; performance status; weight loss in the prior 6 months; primary, metastatic, and systemic disease symptoms; stage of disease; and associated chronic diseases were recorded at baseline. Physician-rated ECOG performance status was also obtained.31 Follow-up case report forms recorded treatment toxicities, best objective response to treatment, date of progression, and date last known to be alive or date of death. For baseline analyses, time to disease progression and duration of survival were determined from the start of treatment. For follow-up analyses, time to disease progression and duration of survival were determined from the date of the 6-week assessment.
Statistical Analysis Relations between change in PRH and clinical outcomes were determined by analyses that defined four groups of patients on the basis of their initial and 6-week PRH scores: high baseline-improved, low baseline-improved, high baseline-declined, and low baseline-declined. High and low baseline scores were determined from median splits of baseline PRH scores. Change status was determined from baseline to 6-week change scores. Improved patients recorded a higher 6-week than baseline PRH score (6-week PRH - baseline PRH > 0), whereas declined patients recorded a lower 6-week than baseline PRH score (6-week PRH - baseline PRH < 0). Note that these particular clinical groupings were developed for convenience and to maximize power in our statistical tests. Certainly, other more clinically defined category groupings are possible; that is, those using clinically meaningful change (CMC) scores. However, use of the CMC score leads to six groupings (high baseline-improved, high baseline-no change, high baseline-declined, low baseline-improved, low baseline-no change, and low baseline-declined), some of which have small sample sizes. The high baseline-improved group had an especially small sample using CMC criteria. Indeed, this is consistent with the clinical picture for advanced lung cancer patients, many of whom experience rapid declines in function. Hierarchical multivariate regression analyses were performed to determine whether change in PRH score over time added to the prediction of clinical outcome independent of other clinical factors. When treatment toxicity was added and the same stepwise approach was used as above, logistic regression was applied to treatment response data, and Cox proportional hazards regressions were applied to time to progression and survival duration. Cox survival curves were plotted to model relations between PRH change and the time to event outcomes. We chose Cox instead of Kaplan-Meier plots because the former allowed us to control for covariates and because PRH scores were not the basis for randomization in the original trial design.
Of 599 patients enrolled and randomly assigned to treatment in E5592, 26 were excluded from the final analyses. These exclusions included 12 patients withdrawn from the study (five patients withdrew consent, three patients died before starting treatment, two patients had brain metastases before starting therapy, one patient contracted infection, and one patient had cardiac arrhythmia requiring medication) and 14 patients who failed to meet inclusion criteria. Baseline demographic and clinical characteristics are shown in Table 1  . The majority of the patients were male, white, ambulatory with some symptoms, and had stage IV disease. Average age at study entry was 60.6 years. Most patients had more than one primary disease symptom but one or fewer metastatic and systemic symptoms. Almost half of the patients had at least one associated chronic condition (ie, cardiovascular or respiratory disease). About one third of the patients had lost 5% or more of their usual body weight in the 6 months before study entry.
Univariate Associations of Baseline Predictors With Clinical Outcomes Spearman correlation coefficients (rho) were used to screen for relations between demographic and clinical variables and objective clinical outcomes (best response to treatment, time to progression, and survival duration). Results appear in Table 2 . Better outcomes were associated with better performance status at baseline, lower stage of disease at baseline, treatment on a paclitaxel arm, and fewer primary, metastatic, and systemic symptoms at baseline. Greater toxicity was associated with better outcomes. This is likely indicative of the amount of chemotherapy received (ie, the number of completed cycles).
Table 3 shows univariate associations between baseline PRH scores and the outcomes. Better PWB and FWB and fewer lung cancer symptoms as measured by the LCS were associated with better clinical outcome. The TOI, an aggregate of the PWB, FWB, and LCS scores, was also strongly associated with outcome. (Note that correlations of the three component scores of the TOI [PWB, FWB, and LCS] ranged from 0.42 to 0.59.)
Multivariate Baseline PRH Prediction Model A series of stepwise multivariate regression analyses were performed to determine the ability of baseline PRH scores to predict clinical outcomes. Logistic regression models were used for best response to treatment (progression v complete response/partial response/stable disease) and Cox proportional hazards regressions were used for time to progression and survival duration. In this sample of patients, the 25th percentile for time to progression was 48 days and the 75th percentile for time to progression was 226 days. The 25th percentile for survival duration was 136 days and the 75th percentile for survival duration was 483 days.
For best response to treatment, the baseline PWB score of the FACT-L was the first variable to enter the model. Higher baseline PWB was associated with a better response to treatment (odds ratio [OR], 1.09; 95% confidence interval [CI], 1.05 to 1.14). Subsequently, stage of disease and paclitaxel treatment entered the model (-2 log likelihood, 562.01; overall
We also determined the added value of baseline PWB score in predicting best response to treatment beyond the clinical variables by repeating the regression analysis in a hierarchical fashion. Clinical variables (baseline stage of disease, paclitaxel treatment, metastatic symptoms, and systemic symptoms) were entered as a block in step 1, followed by the baseline PWB score in step 2. Baseline PWB score explained additional variation in best response to treatment (OR, 1.09; 95% CI, 1.04 to 1.13) beyond the clinical variables (   2 (df = 1) = 18.26, P < .001). In practical terms, this indicates that an increase of 1 point on the PWB subscale corresponds to a 9% relative increase in the probability that a patient will achieve at least stable disease.
Baseline PRH was also found to be a strong and independent predictor of time to progression and survival duration. Baseline TOI score was the strongest predictor of time to progression. Higher baseline TOI score was associated with a lower risk of disease progression (risk ratio [RR], 0.98; 95% CI, 0.97 to 0.99). Subsequently, stage of disease, number of metastatic symptoms, paclitaxel treatment, and baseline performance status entered the model (-2 log likelihood, 4,042.62; overall
To determine the added value of the TOI score in predicting time to progression beyond the clinical variables, we used a hierarchical regression. The clinical variables (baseline stage of disease, paclitaxel treatment, baseline performance status, metastatic symptoms, and systemic symptoms) were entered as a block in step 1, followed by the baseline TOI score in step 2. Baseline TOI score explained additional variation in time to progression (RR, 0.98; 95% CI, 0.97 to 0.99) beyond the clinical variables ( 2 (df = 1) = 19.44, P < .001). Practically, this indicates that a 1-point increase in the TOI score corresponds to a 2% relative reduction in the risk of disease progression at any given time.
Baseline PWB was found to be the strongest predictor of survival duration. Higher baseline PWB score was associated with a lower risk of death (RR, 0.95; 95% CI, 0.94 to 0.97). After entry of the PWB score, number of metastatic symptoms, baseline performance status, and paclitaxel treatment entered the regression (-2 log likelihood = 4,560.98; overall
Finally, we determined whether the baseline PWB score added value to the prediction of survival duration beyond clinical data. Clinical variables (baseline stage of disease, paclitaxel treatment, baseline performance status, prior 6-month weight loss, and primary, metastatic, and systemic symptoms) were entered as a block in step 1, followed by the baseline PWB score in step 2. Baseline PWB score explained additional variation in survival duration (RR, 0.95; 95% CI, 0.94 to 0.97) beyond the clinical variables ( 2 (df = 1) = 22.05, P < .001). Practically, this indicates that a 1-point increase in the PWB score corresponds to a 5% relative reduction in the risk of death at any given time.
Enhancing Prediction of Clinical Outcomes Using Change in PRH Scores
We used a hierarchical regression to determine whether change in PRH scores could predict best response to treatment independent of other factors. Only patients with a documented response occurring after the 6-week assessment were included in this analysis. Baseline clinical variables (stage of disease and paclitaxel treatment) and on-treatment toxicities (maximum hematologic toxicity, maximum neuro-cognitive toxicity, and maximum overall toxicity) were controlled in the first step of the model. When the PWB change pattern was entered in the second step, a significant increase resulted in
Next, we determined the associations between TOI change pattern and time to progression and between PWB change pattern and survival duration. We calculated time to progression and survival duration from the 6-week PRH assessment point to ensure that longitudinal scores were always predicting a future event. For time to progression, baseline clinical variables and on-treatment toxicities were entered as a block in step 1, followed by the TOI change pattern in step 2. When the TOI change pattern was entered in the second step, a significant decrease resulted in the -2 log likelihood statistic from the prior step (
For survival duration, baseline clinical variables and on-treatment toxicities were entered as a block in step 1, followed by the PWB change pattern in step 2. When the PWB change pattern was entered on the second step, a significant decrease resulted in the -2 log likelihood statistic from the prior step ( 2 (3) = 8.79, P < .05). Figure 2 shows the curves for survival duration based on PWB change, controlling for other clinical factors. Low baseline-declined patients survived the shortest time; high baseline-improved patients survived the longest time (Fig 2).
Similar to other studies that have used PRH or HRQL as a prognostic indicator,3–5,7–21,23,24,28 we found strong associations between PRH data and clinical outcomes of chemotherapy. When the PWB subscale and TOI aggregate index of the FACT-L were used, baseline and follow-up PRH scores were predictive of clinical outcomes in lung cancer patients participating in ECOG study 5592. Patients with higher baseline PWB scores showed better responses to chemotherapy and had a lower risk of death than patients with lower baseline PWB scores. Patients with higher baseline TOI scores had a lower risk of disease progression than patients with lower baseline TOI scores. For longitudinal assessments, patients with high baseline PWB scores who declined in PWB at 6 weeks showed better tumor responses to chemotherapy than did patients with low baseline PWB scores who declined at 6 weeks. Patients with high baseline PWB scores who improved at 6 weeks also showed better tumor responses to chemotherapy than did low baseline-declined patients, although the groups were not significantly different. This may be because of the lower number of high baseline-improved patients with tumor response data (n = 26). Patients with high baseline TOI scores who improved in TOI score at 6 weeks were less likely to progress rapidly while receiving treatment than were patients with low baseline TOI scores who declined at 6 weeks. Finally, patients with high baseline PWB scores who improved at 6 weeks were more likely to survive longer than were patients with low baseline PWB scores who declined at 6 weeks. Alhough it might be argued that both the PWB and TOI are simply surrogates for other clinical indicators, we found that all of these associations still held even when other clinical factors (ie, stage of disease, performance status, and number of metastatic symptoms) were forced into the model. Correlations between PRH and clinical indicators (stage of disease, performance status, and number of metastatic symptoms) were low to moderate (rho values ranged from -0.01 to -0.33). Why did the PWB and TOI scores predict different clinical outcomes? Although it is possible that there is some conceptual explanation for this difference, we believe that the reason is primarily statistical, given that these predictors are overlapping in content and therefore inherently correlated. In a subsequent analysis, we entered only the TOI (dropping the PWB, FWB, and LCS subscale scores) and found that it was predictive of all three clinical outcomes. Hence, the TOI probably does not enter the regression equations of response to treatment and survival duration merely because one of its component scores enters first. These findings may support a means of stratifying patients to treatment arms in future clinical trials. Stratification is crucial in a clinical trial because of patient heterogeneity. The greater the precision in which patients are stratified a priori, the better the chances that potential confounding influences will be controlled, thus allowing for the true effects of treatment to be uncovered. Although it is impossible to stratify on all potential confounding influences, it is important to account for those factors that are most predictive of clinical outcome. Traditionally, in advanced cancer patients, stage of disease, symptoms, and performance status have been used as stratification factors. A brief index that includes performance status (rated by both the physician and patient) and a physician estimate of survival has recently been proposed as a simple means of stratification.5 Such brief indices have the advantage of economy by producing less burden to patients and clinicians. However, what they achieve in brevity, they lack in completeness. Use of formal PRH assessment data may improve on the above-proposed stratification approach and certainly improves on the commonly used physician-rated performance status alone. We recommend that the TOI score be used for this purpose because it provides information on general PWB, FWB, and lung cancer symptoms; factors resonant with both patient and clinician. Although the data do not support complete replacement of traditional means of stratification, they do support use of the TOI as a supplementary means of stratification in lung cancer clinical trials. Our findings show that longitudinal PWB and TOI scores may also be helpful to the practicing clinician. Longitudinal PRH assessments are useful in assessing the outcome of treatment, but they may also have use in treatment planning.33,34 Declining PRH scores may signal to a clinician that a change in treatment protocol is needed, especially if baseline PRH was low. Observed changes in PRH scores may also facilitate patient-physician communication. Many patients express a desire to talk about health-related issues with their physician during the course of therapy, particularly aspects of physical functioning and symptomatology.35 Such information may help to initiate end-of-life discussions in those patients with poor and declining PRH; presumably, the patients with the shortest survival duration.36 Effective patient-physician communication has been associated with lower patient distress and improved physical outcome.37 Efforts to incorporate clinical assessment of symptoms and PRH during the course of clinical care of NSCLC patients have shown that the practice is feasible and appealing to both patients and clinicians. In a pilot study of a lung cancer disease management computer platform using PRH data clinically, more than 75% of patients reported that the program was easy to use.38 Furthermore, the majority of oncologists used the information learned from the assessment during the patient’s visit. PRH data can foster thoughtful discussion about treatment; however, these data would usually indicate (not necessarily determine) that a change in treatment protocol be considered. Ultimately, decisions about treatment are best made using all available clinical data. Another possibility is that patient perception of health may in some way influence the course of disease. One study has shown a survival advantage for patients who attended a year-long psychosocial intervention designed to reduce distress and improve HRQL.39 Although our data are consistent with this explanation, they do not imply any causal relation. The viability of this hypothesis is most appropriately tested in a controlled intervention trial. At the very least, we can conclude that both the PWB and TOI scores of the FACT-L provide additional information about patient status that is not currently provided in traditional clinical parameters of function—information that could enlighten the clinician about the future course of the disease. Our findings are limited in two ways. First, our estimates of time to disease progression and survival duration may be somewhat biased because of missing data at the 6-week follow-up time point. At 6 weeks, 35% of the total sample had missing PRH data. On further analysis, we found that those patients with missing PRH data at 6 weeks had poorer function at baseline than those patients with data at both baseline and 6 weeks (complete cases). Compared with complete cases, those patients with missing 6-week scores had worse baseline performance status and had significantly worse baseline PRH (lower PWB, FWB, LCS, TOI, and FACT-L scores). Moreover, patients with missing 6-week PRH data progressed and died sooner than those with complete data. Hence, our estimates of clinical outcomes are likely to be biased because only complete cases entered the correlation analyses. However, it is likely that our analyses are underestimating time to progression and overestimating survival duration in those patients with the worst PRH (ie, those patients who worsen over time), because most of those with missing follow-up data have the poorest baseline function and the worst clinical outcomes. Second, it is possible that different longitudinal PRH cutoff points would yield a different result. The choice of using a median split to create high and low baseline PRH scores and improved versus declined as a means of categorizing change scores does not reflect any formal clinical criteria. These cutoff points were selected mainly for their ease in interpretability and because they elicit sufficient sample sizes for the analysis. More formal clinically based criteria, such as the use of normative comparisons and CMC40 estimates, might provide for more conceptually distinct groupings. Similarly, it is important to note that a certain degree of error is inherent in the measurement of PRH. Hence, within-person variability in PRH scores may limit our findings because the true category grouping for some patients may have been different. In conclusion, both baseline and follow-up PRH scores predicted best response to treatment, time to disease progression, and survival duration independent of other clinical factors. Baseline FACT-L PWB scores were strongly associated with best response to treatment and survival duration; baseline FACT-L TOI scores were strongly associated with time to disease progression. Beyond this, changes in PWB scores were associated with response to treatment and survival duration; changes in TOI scores were associated with time to disease progression. In future advanced lung cancer clinical trials, baseline TOI data can be considered as a stratification variable. Longitudinal PWB and TOI data may have some clinical use in determining when treatment might be altered or when discussion of more palliative treatment options should be considered.
We gratefully acknowledge Mark Sorensen for statistical input.
Supported in part by Public Health Service grants from the National Cancer Institute, National Institutes of Health, and the Department of Health and Human Services (CA23318, CA66636, CA21115, CA17145, and CA49957). The contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Cancer Institute.
1. Moore MJ, Osoba D, Murphy K, et al: Use of palliative end points to evaluate the effects of mitoxantrone and low-dose prednisone in patients with hormonally resistant prostate cancer. J Clin Oncol 12:689–694, 1994[Abstract] 2. Thatcher N, Hopwood P, Anderson H: Improving quality of life in patients with non-small cell lung cancer: Research experience with gemcitabine. Eur J Cancer 33:S8–S13, 1997 (suppl 1) 3. Stanley KE: Prognostic factors for survival in patients with inoperable lung cancer. J Natl Cancer Inst 65:25–32, 1980[Medline] 4. Loprinzi CL, Laurie JA, Wieand HS, et al: Prospective evaluation of prognostic variables from patient-completed questionnaires. J Clin Oncol 12:601–607, 1994[Abstract]
5. 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 6. Coates A, Porzsolt F, Osoba D: Quality of life in oncology practice: Prognostic value of EORTC QLQ-C30 scores in patients with advanced malignancy. Eur J Cancer 33:1025–1030, 1997[CrossRef][Medline] 7. Dancey J, Zee B, Osoba D, et al: Quality of life scores: An independent prognostic variable in a general population of cancer patients receiving chemotherapy. Qual Life Res 6:151–158, 1997[Medline] 8. Chang VT, Thaler HT, Polyak TA, et al: Quality of life and survival: The role of multidimensional symptom assessment. Cancer 83:173–179, 1998[CrossRef][Medline] 9. Llobera J, Esteva M, Rifa J, et al: Terminal cancer: Duration and prediction of survival time. Eur J Cancer 36:2036–2043, 2000[CrossRef][Medline] 10. Coates A, Gebski V, Signorini D, et al: Prognostic value of quality-of-life scores during chemotherapy for advanced breast cancer. J Clin Oncol 10:1833–1838, 1992[Abstract] 11. Kramer JA, Curran D, Piccart M, et al: Identification and interpretation of clinical and quality of life prognostic factors for survival and response to treatment in first-line chemotherapy in advanced breast cancer. Eur J Cancer 36:1498–1506, 2000[CrossRef][Medline] 12. Coates A, Thomson D, McLeod GRM, et al: Prognostic value of quality of life scores in a trial of chemotherapy with or without interferon in patients with metastatic malignant melanoma. Eur J Cancer 29A:1731–1734, 1993[CrossRef][Medline]
13. Butow PN, Coates AS, Dunn SM: Psychosocial predictors of survival in metastatic melanoma. J Clin Oncol 17:2256–2263, 1999 14. Earlam S, Glover C, Fordy C, et al: Relation between tumor size, quality of life, and survival in patients with colorectal liver metastases. J Clin Oncol 14:171–175, 1996[Abstract] 15. de Graeff A, de Leeuw JRJ, Ros WJG, et al: Sociodemographic factors and quality of life as prognostic indicators in head and neck cancer. Eur J Cancer 37:332–339, 2001[CrossRef][Medline]
16. Blazeby JM, Brookes ST, Alderson D: The prognostic value of quality of life scores during treatment for oesophageal cancer. Gut 49:227–230, 2001 17. Kaasa S, Mastekaasa A, Lund E: Prognostic factors for patients with inoperable non-small cell lung cancer, limited disease. Radiother Oncol 15:235–242, 1989[CrossRef][Medline]
18. Ruckdeschel JC, Piantadosi S: Quality of life in lung cancer surgical adjuvant trials. Chest 106:324S–328S, 1994 (suppl 6) 19. Ganz PA, Lee JJ, Siau J: Quality of life assessment: An independent prognostic variable for survival in lung cancer. Cancer 67:3131–3135, 1991[CrossRef][Medline] 20. Herndon JE, Fleishman S, Kornblith AB, et al: Is quality of life predictive of the survival of patients with advanced nonsmall cell lung carcinoma? Cancer 85:333–340, 1999[CrossRef][Medline] 21. Montazeri A, Milroy R, Hole D, et al: Quality of life in lung cancer patients as an important prognostic factor. Lung Cancer 31:233–240, 2001[CrossRef][Medline] 22. Brown JE, Butow PN, Culjak G, et al: Psychosocial predictors of outcome: Time to relapse and survival in patients with early stage melanoma. Br J Cancer 83:1448–1453, 2000[CrossRef][Medline] 23. Buccheri GF, Ferrigno D, Tamburini M, et al: The patient’s perception of his own quality of life might have an adjunctive prognostic significance in lung cancer. Lung Cancer 12:45–58, 1995[CrossRef][Medline] 24. Jerkeman M, Kaasa S, Hjermstad M, et al: Health-related quality of life and its potential prognostic implications in patients with aggressive lymphoma: A Nordic Lymphoma Group trial. Med Oncol 18:85–94, 2001[CrossRef][Medline]
25. Montazeri A, Gillis CR, McEwen J: Quality of life in patients with lung cancer: A review of literature from 1970–1995. Chest 113:467–481, 1998 26. Sarna L: Lung cancer, in Holland JC (ed): Psycho-Oncolgy. New York, Oxford University Press, 1998, pp 340–348 27. Osoba D: Lessons learned from measuring health-related quality of life in oncology. J Clin Oncol 12:608–616, 1994[Abstract] 28. Moinpour C, Lyons B, Grevstad P, et al: Quality of life in advanced non-small-cell lung cancer: Results of a Southwest Oncology Group randomized trial. Qual Life Res 11:115–126, 2002[CrossRef][Medline]
29. Bonomi P, Kim KM, Fairclough D, et al: Comparison of survival and quality of life in advanced non-small-cell lung cancer patients treated with two dose levels of paclitaxel combined with cisplatin versus etoposide with cisplatin: Results of an Eastern Cooperative Oncology Group trial. J Clin Oncol 18:623–631, 2000 30. Cella D, Bonomi AE, Lloyd SR, et al: Reliability and validity of the Functional Assessment of Cancer Therapy-Lung (FACT-L) quality of life instrument. Lung Cancer 12:199–220, 1995[CrossRef][Medline] 31. Zubrod CG, Schneiderman M, Frei E, et al: Appraisal of methods for the study of chemotherapy of cancer in man: Comparative therapeutic trial of nitrogen mustard and triethylene thiophosphoramide. J Chronic Dis 11:7–33, 1960 32. Cox DR: Regression models and life tables. J R Stat Soc B 34:187–220, 1972 33. Jacobsen PB, Davis K, Cella D: Assessing quality of life in research and clinical practice. Oncology 16:133–139 (suppl) 34. Davis K, Cella D: Assessing quality of life in oncology clinical practice: A review of barriers and critical success factors. J Clin Outcomes Manage 9:327–332, 2002
35. Detmar S, Aaronson N, Wever L, et al: How are you feeling? Who wants to know? Patients’ and oncologists’ preferences for discussing health-related quality-of-life issues. J Clin Oncol 18:3295–3301, 2000
36. von Gunten C, Ferris F, Emanuel L: Ensuring competency in end-of-life care: Communication and relational skills. J Am Med Assoc 284:3051–3057, 2000 37. Stewart M: Effective physician-patient communication and health outcomes: A review. Can Med Assoc J 152:1423–1433, 1995[Abstract] 38. Chang CH, Cella D, Masters GA, et al: Real-time clinical application of quality-of-life assessment in advanced lung cancer. Clin Lung Cancer 4:104–109, 2002[Medline] 39. Spiegel D, Bloom J, Kraemer H, et al: Effect of psychosocial treatment on survival of patients with metastatic breast cancer. Lancet 2:888–891, 1989[CrossRef][Medline] 40. Cella D, Eton DT, Fairclough DL, et al: What is a clinically meaningful change on the Functional Assessment of Cancer Therapy-Lung (FACT-L) Questionnaire? J Clin Epidemiol 55:285–295, 2002[CrossRef][Medline] Submitted July 22, 2002; accepted January 2, 2003.
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2003 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
0.10 were retained for further analyses. Baseline variables associated with best response to treatment (progression versus complete response/partial response/stable disease) were entered into a logistic regression model; those associated with time to progression or survival duration were entered into Cox proportional hazards regression models.
) Between Demographic and Clinical Characteristics and Clinical Outcomes
