Changes in the use of end points in clinical trials for elderly cancer patients over time

Transcription

1 Annals of Oncology 28: , 2017 doi: /annonc/mdx354 Published online 3 August 2017 ORIGINAL ARTICLE Changes in the use of end points in clinical trials for elderly cancer patients over time O. Le Saux 1,2 *, C. Falandry 3,4, H. K. Gan 5,6,7, B. You 1,2, G. Freyer 1,2 &J.Péron 1,8,9 1 Medical Oncology Department, Institut de Cancérologie des Hospices Civils de Lyon (IC-HCL), Pierre-Bénite; 2 Faculté demédecine Lyon-Sud, Lyon 1 University, EMR 3738, Oullins; 3 Geriatric Oncology Department, Centre Hospitalier Lyon Sud, Pierre-Bénite; 4 CarMen Biomedical Research Laboratory (Cardiovascular Diseases, Metabolism, Diabetology and Nutrition) INSERM UMR 1060, Université de Lyon, Oullins, France; 5 Olivia Newton-John Cancer Research Institute, Heidelberg; 6 School of Cancer Medicine, La Trobe University, Heidelberg; 7 Department of Medicine, Melbourne University, Melbourne, Australia; 8 Statistics Unit, Hospices Civils de Lyon, Pierre-Bénite; 9 CNRS, UMR 5558 Biometry and Evolutionary Biology Laboratory, Université Lyon 1, Villeurbanne, France *Correspondence to: Dr Olivia Le Saux, Medical Oncology Department, Centre Hospitalier Lyon Sud, 165, Chemin du Grand Revoyet, Institut de Cancérologie des Hospices Civils de Lyon, Pierre-Bénite, France. Tel: þ ; Fax: þ ; olivia.lesaux@gmail.com Background: Physicians need well-addressed clinical trials assessing benefits and harm of treatments to avoid under-treatment or over-treatment of elderly patients. The main objectives of this report were to present an overview of end points used in clinical trials dedicated to elderly patients; and to assess the evolution in chosen end points before and after the creation of the International Society of Geriatric Oncology in the early 2000s. Patients and methods: All phases I, II and III trials dedicated to the treatment of cancer among elderly patients published between 2001 and 2004 and between 2011 and 2014 were reviewed. All phase III clinical trials assessing cancer treatments among adults in the same periods were also reviewed to identify subgroup analyses of elderly patients among these trials. Results: Among phase III trials dedicated to elderly patients, overall survival was a common primary end point. Interestingly, tumor centered end points were very common in the first time period and very uncommon in the second time period, whereas composite end points were very uncommon in the first time period but very common in the second time period. Concerningly, disease-specific survival was very infrequently reported in dedicated clinical trials of elderly patients despite their importance in evaluating competing risk of death from non-oncology causes. The use of patient-reported outcomes (PROs) as a primary end point remained very uncommon but the reporting of PROs as a secondary end point tended to increase in the second time period, from 19% to 33% (P ¼ 0.10). Functional was infrequently reported. Conclusion: During the past decade, the use of clinically meaningful end points such as PROs and functional in elderly patients remained moderate. Yet, the use of PROs as a secondary end point tended to increase between the two time periods. Key words: end points, geriatric oncology, elderly, neoplasms Introduction With the ageing of the population, the importance of detailed oncology data from older populations has become increasingly relevant [1]. Given this demographic shift, medical societies such as the European Organization for Research and Treatment of Cancer (EORTC), the American Society of Clinical Oncology (ASCO) and the International Society for Geriatric Oncology (SIOG) established guidelines on clinical trial methodology in older cancer patients [2 5]. The need for specific trial designs dedicated to vulnerable or frail older patients was emphasized, as well as the need to increase the number of elderly patients included in clinical trials. Specifically, the need to perform some form of geriatric assessment (GA) and the need to choose appropriate outcome measures were stressed. A recent position paper of the SIOG [5] stated that overall survival (OS) is a crucial end point, but disease-specific survival (DSS) should also be assessed in trials of older cancer patients because of the risk of nononcological deaths resulting from other causes (e.g. other diseases, treatment toxicity) occur much more frequently in the older population. These guidelines also stated that health related quality of life (HRQoL) and preservation of functional capacity and independence were important for understanding more VC The Author Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For Permissions, please journals.permissions@oup.com.

2 Annals of Oncology comprehensively the effects of disease and treatment from the patient s perspective [6]. The SIOG creation and the publication of the first guidelines on geriatric oncology research methods were published in the early 2000s [7 9]. This date can be seen as the beginning of geriatric oncology as a scientific field. We undertook a review of changes on the use of end points in trials dedicated to elderly patients between the early 2000s and 10 years after as, it takes 10 years from the conception of a study until the publication of its final results. Methods Trial selection In January 2015, two researchers (OLS and JP) identified all reports of clinical trials (phase I, phase II, and phase III trials) assessing therapies for hematological or solid tumors and dedicated to the elderly (at least using a chronological landmark to define the elderly), and all phase III clinical trials among older adults in order to identify subgroup analyses of elderly patients. Included reports were published in English between 1 January 2001 and 31 December 2004 (first time period pre SIOG guidelines), or between 1 January 2011 and 31 December 2014 (second time period after SIOG guidelines) to assess the impact of the SIOG creation in 2000, and the impact of published guidelines on inclusion of elderly patients in oncology trials. A minimum of 10 years is necessary to observe the impact of such guidelines on clinical trials manuscripts frequently published, as several years will pass between writing a protocol and publishing results. The methodology of this systematic review has been published previously [10]. The full Search strategy is reported in supplementary material S1, available at Annals of Oncology online. Data extraction Data extraction was performed by the same authors who carried out the initial article selection (OLS and JP). The data were cross-verified by the two data extractors. The collected variables included study design, year of publication, tumor site, source of trial funding, journal impact factor, regions in which trials were conducted, type of investigational therapy, cancer stage, number of enrolled patients, primary age threshold defining elderly patients, proportion of patients in the elderly subgroup, maximum age, median age and end points. Definition of trial end points End points were classified as primary and secondary end points. When no clear primary end point was stated by authors, all reported end points were considered as secondary outcomes. When two end points were reported as primary end points or coprimary end points, both end points were considered as primary end points in this review. End points were further classified as OS, composite end point, tumor-centered end points, toxicity end points, patient-reported outcomes (PROs), DSS, impact on functional end points, feasibility end points, pharmacokinetics or pharmacodynamics end points, and economical end points. Composite end points were defined as time-to-event end points based on a combination of individual events including death and at least one other event related to the tumor evolution (supplementary material S2, available at Annals of Oncology online). Tumor-centered end points included all measure of tumor evolution (including radiological assessment, biological, or histological markers evolution) (supplementary material S2, available at Annals of Oncology online). PROs were defined as any report of the of a patient s health condition that comes directly from the patient, without interpretation of the patient s response by a clinician or anyone else [11]. PROs included mainly symptoms and health related QoL assessments. Maintenance of functional was defined as an end point assessing functioning and (in)dependence through validated instruments [5]. Cognitive function was defined as an end point based on validated cognitive measures [5]. DSS was defined as the time between randomization and the death from cancer. DSS is censored at the time of death from other cause. Clinically meaningful end points were defined as end points showing clinically meaningful benefit for the patient such as assessed by a longer or better life, and ideally both [12]. Analysis The nature of end points used in all clinical trials included in this review were described according to their time period of publication. A separate analysis of trial end points was performed for phase II trials and phase III trials, respectively. The evolution of end points used between the two time periods was investigated. For phase II trials, multivariate logistic regression was used to assess the trial characteristics associated with the choice of primary and secondary end points. Odds ratio and their 95% confidence intervals were calculated. All factors of interest and potential confounding factors were included in the multivariate models if they were associated with the choice of end point in univariate analysis (P < 0.10). The added value of each covariate was evaluated using a likelihood ratio test. Most of the analyses performed were descriptive. Qualitative data were described by percentage and statistical comparisons were performed when appropriate using chi-squared or Fisher exact test. Quantitative data were described by median values and interquartile ranges, and statistical comparisons were performed when appropriate using the Wilcoxon test or the Kruskall Wallis test. All analyses were performed using SPSS version 17.0 (SPSS Inc., Chicago, IL) or the R software version ( Results Original article Characteristics of the selected clinical trials A total of 1084 clinical trials (supplementary material S3, available at Annals of Oncology online) were included in the analysis: 366 from the first time period ( ) and 718 from the second time period ( ). We identified 263 clinical trials including only elderly patients (or elderly patients along with unfit patients impaired functional or comorbidities), over the two time periods: 27 phase I clinical trials, 193 phase II trials and 43 phase III clinical trials. We also identified 280 phase III clinical trials reporting at least one subgroup analysis of elderly patients, and 541 phase III clinical trials reporting no subgroup analysis of elderly patients. Study characteristics of included trials are presented in supplementary Tables S1 S3, available at Annals of Oncology online. Median age of patients included in dedicated phase I trials was 72 years (IQR: 68 76). Median age of patients included in dedicated phase II trials was 72 years (IQR: 72 76), and median age of patients included in dedicated phase III trials was 71 years (IQR: 68 74). All primary and overall end points in phase I, phase II, and phase III trials are reported in Table 1. In phase I trials, toxicity was the most frequently reported primary end point (100% and 93% of the cases in the first and second periods, respectively). In phase II trials, the most frequent primary end points were tumorcentered in the first period (46%), and also in the second period (68%). The use of a tumor-centered end point as a primary end point was statistically more frequent in the second time period Volume 28 Issue doi: /annonc/mdx

3 Original article Annals of Oncology Table 1. Evolution of end points in clinical trials OS DSS Composite Tumorcentered end point Toxicity Patientreported outcome Functional Feasibility PK-PD Economical Cognitive satus Other Phase (n¼3) 1 (33%) 0 (0%) 1 (33%) 2 (67%) 3 (100%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 1 (33%) (n¼24) 10 (42%) 0 (0%) 5 (21%) 16 (67%) 22 (92%) 1 (4%) 0 (0%) 4 (17%) 8 (33%) 0 (0%) 0 (0%) 13 (54%) (n¼3) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 3 (100%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) (n¼24) 0 (0%) 0 (0%) 0 (0%) 1 (4%) 20 (93%) 0 (0%) 0 (0%) 1 (4%) 2 (8%) 0 (0%) 0 (0%) 0 (0%) Phase 2 a (n¼68) 55 (81%) 2 (3%) 20 (29%) 64 (94%) 63 (93%) 9 (13%) 4 (6%) 12 (18%) 4 (6%) 0 (0%) 1 (1%) 18 (26%) (n¼125) 103 (82%) 1 (1%) 82 (66%) 115 (92%) 107 (86%) 22 (18%) 3 (2%) 12 (10%) 8 (6%) 3 (2%) 1 (1%) 44 (35%) (n¼68) 3 (4%) 0 (0%) 1 (1%) 31 (46%) 11 (16%) 0 (0%) 0 (0%) 5 (7%) 0 (0%) 0 (0%) 0 (0%) 1 (1%) (n¼125) 9 (7%) 0 (0%) 13 (10%) 85 (68%) 19 (15%) 1 (1%) 0 (0%) 4 (3%) 1 (1%) 0 (0%) 0 (0%) 1 (1%) Dedicated Phase 3 b (n¼15) 14 (93%) 0 (0%) 8 (53%) 15 (100%) 11 (73%) 3 (20%) 0 (0%) 1 (7%) 0 (0%) 0 (0%) 0 (0%) 4 (3%) (n¼28) 26 (93%) 0 (0%) 24 (86%) 21 (75%) 24 (86%) 12 (43%) 0 (0%) 2 (7%) 0 (0%) 0 (0%) 0 (0%) 7 (25%) (n¼15) 5 (33%) 0 (0%) 1 (7%) 6 (40%) 0 (0%) 1 (7%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) (n¼28) 14 (50%) 0 (0%) 13 (46%) 2 (7%) 0 (0%) 1 (4%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) Phase 3 with subgroup analysis of elderly c (n¼42) 40 (95%) 1 (1%) 20 (48%) 35 (83%) 37 (88%) 8 (19%) 0 (0%) 6 (14%) 1 (2%) 0 (0%) 2 (5%) 4 (10%) (n¼238) 226 (95%) 3 (1%) 207 (87%) 192 (81%) 202 (85%) 78 (33%) 5 (2%) 13 (5%) 18 (8%) 3 (1%) 0 (0%) 81 (34%) (n¼42) 22 (52%) 0 (0%) 6 (14%) 9 (21%) 1 (2%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) (n¼238) 84 (35%) 0 (0%) 126 (53%) 23 (10%) 6 (3%) 2 (1%) 0 (0%) 1 (0%) 0 (0%) 0 (0%) 0 (0%) 2 (1%) Phase 3 with no subgroup analysis of elderly d (n¼238) 200 (84%) 4 (2%) 111 (47%) 195 (82%) 210 (88%) 83 (35%) 7 (3%) 11 (5%) 4 (2%) 9 (4%) 1 (<1%) 58 (24%) (n¼303) 275 (91%) 9 (3%) 228 (75%) 225 (74%) 249 (82%) 82 (27%) 5 (2%) 7 (2%) 14 (5%) 3 (1%) 4 (1%) 78 (26%) (n¼238) 103 (43%) 0 (0%) 43 (18%) 73 (31%) 3 (1%) 8 (3%) 0 (0%) 0 (0%) 2 (1%) 0 (0%) 0 (0%) 1 (<1%) (n¼303) 108 (36%) 2 (1%) 117 (39%) 66 (22%) 5 (2%) 8 (3%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 1 (<1%) 5 (2%) Missing values are primary end points not clearly defined in the methodology of the clinical trial. a 24 missing values in and 8 missing values in b 2 missing values in and 0 missing value in c 7 missing values in and 8 missing values in d 24 missing values in and 10 missing values in OS, overall survival; DSS, disease-specific survival; PK-PD, pharmacokinetic pharmacodynamic end point Le Saux et al. Volume 28 Issue

4 Annals of Oncology (68%, P value ¼ ). This increased use of tumor-centered end point was independent of possible confounders in the multivariate regression model (supplementary Table S4, available at Annals of Oncology online). Most phase II trials included at least one tumor-centered end point, one toxicity end point, and OS. The use of a composite end point was more frequent in the second time period (66% versus 29%, P value < ). The use of PROs as secondary end points did not increase over time, and was essentially seen in trials in palliative non-hematologic settings (supplementary Table S5, available at Annals of Oncology online). All end points (both primary and secondary) used in dedicated phase III trials are reported in Figure 1. Primary end points used in dedicated phase III trials are reported in Figure 2. Overall, OS remained the main primary end point (33% and 50%, respectively, in the first and second period, P value ¼ 0.35). Between the first and second period, composite end points became more Original article prevalent as a primary end point (from 7% to 46%, P value ¼ 0.015) while the use of tumor-centered end points decreased (from 40% to 7%, P value ¼ 0.014). The use of PROs as secondary end points increased from 20% in the first time period to 43% in the second time period. Yet, this difference was not statistically significant (P value ¼ 0.19). The use of PROs as primary end point remained low (from 7% to 4%, P value ¼ 1). In these uncommon cases, PROs were never used as a co-primary end point. The evaluation of toxicity also increased slightly from 73% to 86% (P value ¼ 0.42). Feasibility as an end point was uncommon in both time periods (Figure 1). Cognitive function was never measured in dedicated phase III trials. In phase III trials with subgroup analyses of elderly patients (Table 1), the use of composite end points as a primary end point increased from 14% in the first period to 53% in the second period (P value < ), while the use of OS decreased from 52% to 35% (P value ¼ 0.040). The use of PROs as a primary end Frequency (%) OS DSS Composite Tumor-centred Toxicity PRO Functional Feasibility PK-PD Economical Cognitive Other Figure 1. All end points used in phase III dedicated clinical trials. OS, overall survival; DSS, disease-specific survival; PK-PD, pharmacokinetic pharmacodynamic end point. Use of each category of end points among 43 phase III clinical trials dedicated to elderly patients. The proportion of trials using each end point category is reported for the first time period (black) and the second time period (grey) Frequency (%) OS DSS Composite Tumor-centred Toxicity PRO Functional Feasibility PK-PD Economical Cognitive Other Figure 2. Primary end points used in phase III dedicated clinical trials. OS, overall survival; DSS, disease-specific survival, PK-PD, pharmacokinetic pharmacodynamic end point. Use of each category of primary end points among 43 phase III clinical trials dedicated to elderly patients. The proportion of trials using each end point category is reported for the first time period (black) and the second time period (grey). Volume 28 Issue doi: /annonc/mdx

5 Original article point remained very uncommon but the reporting of PROs as a secondary end point again showed a trend towards being more frequent in the second time period, from 19% to 33% (P value ¼ 0.10). On the contrary, among trials with no subgroup analysis dedicated to elderly patients, the reporting of PROs may even have decreased, between the two time periods (35% and 27%, P value ¼ 0.074). Functional was infrequently evaluated. This end point was used in 4 phase II trials in the first period (6%), 3 phase II trials in the second period (2%), and 5 phase III trials with a subgroup analysis on elderly patients in the second period (2%). It was never used in dedicated phase III trials. Functional was never used as a primary end point. Similarly, DSS was used in only 3 out of 263 clinical trials dedicated to elderly patients. Discussion We undertook the first comprehensive review of end points used to assess outcomes among elderly in oncology trials. A smaller study limited to breast cancer patients has previously been reported and showed that on nine clinical trials dedicated to elderly patients, only five included any patient-related end point (56%, P¼ 0.02) [13]. Another study focusing on hematological malignancies has shown that patient-centered outcomes such as QoL, health care utilization, and functional capacity were only incorporated in a small number of trials (<10%). Even in trials dedicated to older patients, the primary focus lied on standard tumor-centered outcomes, while patient-centered outcomes were included in less than one-fifth of studies [14]. Our study found that tumor-centered outcomes and OS were the first and second most common primary end points used in dedicated phase III trials during the first time period. Interestingly, during the second time period, OS was the most common primary end point used, whereas composite end points became the second most common primary end point, while tumor-centered end points were hardly used at all. This mightbeexplainedbythefactthatcompositeendpointsare considered as surrogate end points for OS in many cancer types and became more frequently used as a primary end point as it can accelerate the achievement of significant results and therefore new drug approvals [15, 16]. Perhaps more importantly, our study identified some glaring deficits in the end points used in these studies. DSS is considered to be an important data to help interpret OS data in this population with high competing risks of death from non-oncology comorbidities [17 19] even though DSS is limited by the fact that determining cause of death is extremely challenging in older patients. In DSS, deaths not related to the malignant disease are censored. As a consequence, this end point reflects directly the treatment antitumoral effect on survival, and does not take into account toxic death or death related to comorbidities. Yet, DSS was not used as a primary end point in any of the included studies. In contrast, PROs were reported for many (43%)trials.Theseresultsareinaccordancewiththeresultsreported by de Glas et al. [13] who showed that 56% of breast older cancer patients trials incorporated at least one patientrelated end point. PROs like HRQoL are a major concern for cancer patients, and it can be affected by symptoms caused by Annals of Oncology cancer, as well as by treatment-induced toxicity [20]. PROs may then be especially important for elderly patients. For example, elderly patients are less willing to compromise their HRQoL for a potential survival gain [21]. In a similar way, functional should be one of the major principles of cancer management in the elderly since a negative impact on patient s functional capacity will have a negative impact on survival as well [22]. The prolongation of active life expectancy seems much more important than the prolongation of life expectancy as such. The use of PROs might have been limited by some methodological concerns such as definition of functional dependence, optimal cut-off, and the fact that other handicaps can interfere with functionality while it might not have real impact on QoL [3]. Our study has some limitations. It is a retrospective systematic review of published trials. Therefore, we did not take into account publication bias [23], which can be significant, especially in elderly patients with comorbidities or impaired functional. Indeed, due to increased toxicity and decreased tolerance [24], authors and sponsors (such as industry [25]) may be reluctant to publish trial reports with negative results and editors might be less likely to accept it [26]. A second limitation is that we did not look at physiological age, such as assessed with a comprehensive GA, and limited our analyses to chronological age. A third and last limitation is that it takes many years from protocol writing to publication of results. As a consequence, the full effect of SIOG, EORTC and ASCO s guidelines may not have been reached in our second time period. However, the sample size, the robustness and the importance of the results are enough not to ignore the conclusions drawn. Whendesigning a clinical trial, the choice of primary and secondary end point is often challenging, as all end points have limits. This challenge is true among young and fit patients, but is even more important among elderly and frail patients. First we recommend that authors include systematically functional, and PROs when designing a trial among elderly patients. For phase III confirmatory trial, DSS should be considered as the disease-specific cause of death is believed to be recognized with enough accuracy. Moreover, a universal GA data set should be required for clinical trials. This GA should at least include a validated evaluation of functional [27] which add information compared to Eastern Cooperative Oncology Group performance in elderly cancer patients [28], an objective evaluation of comorbidities, cognitive function, depression, and nutrition in order to stratify analyses on different subgroups of older patients (vulnerable and frail). This is mandatory in order to extrapolate data in clinical practice. Elderly patients should not be excluded from non-dedicated clinical trials in order to have a minimum cohort of elderly representative of demographic data. We also recommend new trials focusing on geriatric interventions. To conclude, we showed that OS remains the dominant end point in trials including elderly oncology patients, as it is in oncology trials more generally. Yet, the number of clinical trials dedicated to elderly patients remained low. Given the specificities of elderly patients high competing risk of non-oncological death; importance of QOL; and importance of independence we believe that other end points deserve special attention. In particular, a focus on DSS, PROs and functional would be relevant Le Saux et al. Volume 28 Issue

6 Annals of Oncology Funding None declared. Disclosure The authors have no conflicts of interest to declare. References 1. Abazov VM, Abbott B, Abolins M et al. Measurement of dijet angular distributions at square root(s) ¼ 1.96 TeV and searches for quark compositeness and extra spatial dimensions. Phys Rev Lett 2009; 103: Lichtman SM. Geriatric oncology and clinical trials. Am Soc Clin Oncol Educ Book 2015; e127 e Pallis AG, Fortpied C, Wedding U et al. EORTC elderly task force position paper: approach to the older cancer patient. Eur J Cancer 2010; 46: Pallis AG, Ring A, Fortpied C et al. EORTC workshop on clinical trial methodology in older individuals with a diagnosis of solid tumors. Ann Oncol 2011; 22: Wildiers H, Mauer M, Pallis A et al. End points and trial design in geriatric oncology research: a joint European organisation for research and treatment of cancer Alliance for Clinical Trials in Oncology International Society Of Geriatric Oncology position article. J Clin Oncol 2013; 31: Husain LS, Collins K, Reed M, Wyld L. Choices in cancer treatment: a qualitative study of the older women s (>70 years) perspective. Psychooncology 2008; 17: Audisio RA, Bozzetti F, Gennari R et al. The surgical management of elderly cancer patients; recommendations of the SIOG surgical task force. Eur J Cancer 2004; 40: Lichtman SM, Balducci L, Aapro M. Geriatric oncology: a field coming of age. J Clin Oncol 2007; 25: Repetto L, Biganzoli L, Koehne CH et al. EORTC Cancer in the Elderly Task Force guidelines for the use of colony-stimulating factors in elderly patients with cancer. Eur J Cancer 2003; 39: Le Saux O, Falandry C, Gan HK et al. Inclusion of elderly patients in oncology clinical trials. Ann Oncol 2016; 27: Psachoulia K, Jamen F, Young KM, Richardson WD. Cell cycle dynamics of NG2 cells in the postnatal and ageing brain. Neuron Glia Biol 2009; 5: Wilson MK, Collyar D, Chingos DT et al. Outcomes and endpoints in cancer trials: bridging the divide. Lancet Oncol 2015; 16: e43 e de Glas NA, Hamaker ME, Kiderlen M et al. Choosing relevant endpoints for older breast cancer patients in clinical trials: an overview of all current Original article clinical trials on breast cancer treatment. Breast Cancer Res Treat 2014; 146: Hamaker ME, Stauder R, van Munster BC. On-going clinical trials for elderly patients with a hematological malignancy: are we addressing the right end points? Ann Oncol 2014; 25: Kundu MG, Acharyya S. Surrogacy of progression free survival for overall survival in metastatic breast cancer studies: meta-analyses of published studies. Contemp Clin Trials 2016; 53: Sidhu R, Rong A, Dahlberg S. Evaluation of progression-free survival as a surrogate endpoint for survival in chemotherapy and targeted agent metastatic colorectal cancer trials. Clin Cancer Res 2013; 19: Piccirillo JF, Feinstein AR. Clinical symptoms and comorbidity: significance for the prognostic classification of cancer. Cancer 1996; 77: de Vries SH, Postma R, Raaijmakers R et al. Overall and disease-specific survival of patients with screen-detected prostate cancer in the European randomized study of screening for prostate cancer, section Rotterdam. Eur Urol 2007; 51: Discussion Satariano WA, Ragland DR. The effect of comorbidity on 3-year survival of women with primary breast cancer. Ann Intern Med 1994; 120: Bottomley A, Vanvoorden V, Flechtner H et al. The challenges and achievements involved in implementing Quality of Life research in cancer clinical trials. Eur J Cancer 2003; 39: Yellen SB, Cella DF, Leslie WT. Age and clinical decision making in oncology patients. J Natl Cancer Inst 1994; 86: Reuben DB, Rubenstein LV, Hirsch SH, Hays RD. Value of functional as a predictor of mortality: results of a prospective study. Am J Med 1992; 93: Easterbrook PJ, Berlin JA, Gopalan R, Matthews DR. Publication bias in clinical research. Lancet 1991; 337: Wedding U, Honecker F, Bokemeyer C et al. Tolerance to chemotherapy in elderly patients with cancer. Cancer Control 2007; 14: Als-Nielsen B, Chen W, Gluud C, Kjaergard LL. Association of funding and conclusions in randomized drug trials: a reflection of treatment effect or adverse events? Jama 2003; 290: van Lent M, Overbeke J, Out HJ. Role of editorial and peer review processes in publication bias: analysis of drug trials submitted to eight medical journals. PLoS One 2014; 9: e Wildiers H, Heeren P, Puts M et al. International Society of Geriatric Oncology consensus on geriatric assessment in older patients with cancer. J Clin Oncol 2014; 32: Repetto L, Fratino L, Audisio RA et al. Comprehensive geriatric assessment adds information to Eastern Cooperative Oncology Group performance in elderly cancer patients: an Italian Group for Geriatric Oncology Study. J Clin Oncol 2002; 20: Volume 28 Issue doi: /annonc/mdx