Prognostic factors for overall survival in elderly patients with advanced ovarian cancer treated with chemotherapy: Results of a pooled analysis of three GINECO phase II trials

Gynecologic Oncology 143 (2016) 22–26

Contents lists available at ScienceDirect

Gynecologic Oncology journal homepage: www.elsevier.com/locate/ygyno

Prognostic factors for overall survival in elderly patients with advanced ovarian cancer treated with chemotherapy: Results of a pooled analysis of three GINECO phase II trials Fabien Tinquaut a, Gilles Freyer b, Franck Chauvin a, Nicolas Gane c, Eric Pujade-Lauraine d, Claire Falandry e,⁎ a

Hygee Center, Regional center for Cancer Prevention, Saint Priest en Jarez, France Medical Oncology unit, Centre Hospitalier Lyon Sud, Université Lyon 1, Pierre-Bénite, France c ARCAGY (Association de Recherche sur le CAncers Gynécologiques), Paris, France d Unité Cancer de la Femme et Recherche Clinique, Hôpitaux Universitaires Paris Centre, site Hôtel-Dieu, Université Paris Descartes, Paris, France e Geriatrics and Oncology Unit, Centre Hospitalier Lyon Sud, Université Lyon 1, Pierre-Benite, France b

H I G H L I G H T S • Treatment completion rates were 73% in this elderly specific population. • Toxicities were moderate and manageable. • Being “depressed”, hypoalbuminemia b 35 g/L, and FIGO stage IV impaired OS.

a r t i c l e

i n f o

Article history: Received 9 September 2015 Received in revised form 10 March 2016 Accepted 13 March 2016 Available online 25 August 2016 Keywords: Elderly Geriatrics Ovarian cancer Prognostic factors

a b s t r a c t Background. The GINECO led three multicentric prospective phase II studies, Elderly Woman Ovarian Trials 1 (EWOT1), EWOT2, and EWOT3, to evaluate the impact of geriatric covariates on the outcome of elderly patients treated with six courses of first-line chemotherapy for FIGO stage III\\IV ovarian cancer. This pooled analysis was designed to evaluate the validity of the geriatric vulnerability parameters identified in EWOT3 (Falandry et al., 2013). Patients and methods. From 1997 to 2011, 266 patients were recruited: 83 in EWOT1, 72 in EWOT2, and 111 in EWOT3, which evaluated respectively a 4-weekly carboplatin-cyclophosphamide regimen, a 3-weekly standard carboplatin-paclitaxel doublet and a carboplatin monotherapy. All patients were analyzed in this pooled analysis for treatment completion, toxicity, and overall survival. Results. The global treatment completion rate was 73% and ranged from 68% in EWOT2 to 74% in EWOT3. Toxicities were generally manageable: neutropenia was more frequent in EWOT2 and thrombopenia in EWOT1 and EWOT3. In multivariate analysis, covariates associated with decreased survival were: being “depressed” according to the investigators' assessment, hypoalbuminemia b 35 g/L, and FIGO stage IV. In addition, a Hospital Anxiety and Depression Scale (HADS) score N 14 and Instrumental Activities of Daily Living (IADL) score b 25 confirmed a deleterious impact in the EWOT2 + EWOT3 population subanalysis. Conclusions. Despite moderate heterogeneity among the studies, this pooled analysis confirmed the deleterious effects on overall survival of emotional disorders (“depressed”, as assessed by investigators or the HADS score), and decreased functionality (IADL score), in addition to FIGO stage. © 2016 Elsevier Inc. All rights reserved.

1. Introduction Epithelial ovarian cancer is the leading cause of gynecologic cancer– related deaths in Western countries [1]. There are two main reasons for the high morbidity associated with ovarian cancer. First, approximately ⁎ Corresponding author at: Geriatrics and Oncology Unit, Centre Hospitalier Lyon Sud, 165 Chemin du Grand Revoyet, 69495, Pierre-Benite, France. E-mail address: [email protected] (C. Falandry).

http://dx.doi.org/10.1016/j.ygyno.2016.03.018 0090-8258/© 2016 Elsevier Inc. All rights reserved.

75% of ovarian cancer cases are diagnosed at advanced stages (International Federation of Gynecology and Obstetrics [FIGO] stages III\\IV) [1]. Second, survival is related to the age at diagnosis [2,3] and the highest incidence of and mortality from ovarian cancer are reported among women 75–79 years old [4]. However, although elderly patients represent the greatest proportion of those with ovarian cancer, this population remains undertreated. Indeed, several studies show that increasing age is associated with decreased use of chemotherapy in patients with advanced ovarian cancer [5,6]. The undertreatment of

F. Tinquaut et al. / Gynecologic Oncology 143 (2016) 22–26

elderly ovarian cancer patients may be attributed to the fact that treatment-related decisions are primarily based on their chronological age and not on their overall physical and mental health. Therefore, a multidisciplinary evaluation, the comprehensive geriatric assessment (CGA), was developed to identify clinical predictors of morbidity and mortality in geriatric medicine settings [7]. In the oncology setting, variables of CGA, such as functional status, may predict survival and chemotherapy toxicity [7–11]. In this context, the French National Group of Investigators for the study of Ovarian and Breast Cancer (GINECO) used the CGA to successively assess the feasibility of three different chemotherapy schedules in elderly patients with ovarian advanced cancer. Since 1997, three phase II trials incorporating the CGA have been conducted: Elderly Woman Ovarian Trials (EWOT) 1 [12], EWOT2 [13], and EWOT3 [14]. A geriatric vulnerability score (GVS) was developed on the basis of the survival score of EWOT3; the GVS integrates the following items: albuminemia b35 g/L, ADL score b 6, IADL score b 25, lymphopenia b1 × 103/mm3, and HADS N14 [14]. Patients having 3 or more of these vulnerability parameters are considered vulnerable. To test the validity of these geriatric prognostic factors, we performed a pooled analysis of the EWOT trials. 2. Methods 2.1. Study design EWOT1–3 were open-label, multicenter, prospective phase II trials. Chemo-naïve elderly patients with advanced ovarian cancer were enrolled and received carboplatin and cyclophosphamide in EWOT1 (1998–2000), carboplatin and paclitaxel in EWOT2 (2001–2004), and carboplatin only in EWOT3 (2007–2011). Each trial was approved by the Independent Ethics Committee of Lyon University Hospital. EWOT2 and EWOT3 were centrally registered according 2005 guidelines (EWOT2: NCT00231075; EWOT3: INCA-RECF0456/EUDRACT 2006-005504-13). All patients provided written informed consent before participation. 2.2. Patients Key eligibility criteria in the three trials were similar. They included age ≥ 70 years, FIGO stage III\\IV ovarian epithelial carcinoma, and life expectancy of at least 3 mo (see Supplementary Text S1 for detailed inclusion and exclusion criteria). 2.3. Treatment Patient treatments were described previously for each of the three trials [12–14]. Briefly, EWOT1 patients received carboplatin AUC (area under the curve) 5 mg min/mL and cyclophosphamide 600 mg/m2 every 4 wk. EWOT2 patients received carboplatin AUC 5 mg min/mL and paclitaxel 175 mg/m2 every 3 wk. EWOT3 treatment consisted of carboplatin AUC 5 mg min/mL every 3 wk. Patients received six cycles in the absence of disease progression or unacceptable toxicity. In accordance with the protocol design, treatment could be pursued upon the investigators' decision. 2.4. Comprehensive geriatric assessment In all three trials, a CGA was performed at the inclusion visit [12–14]. CGA domains included functional status, comorbidity, emotional status, nutrition, and medication as described in Supplementary Text S2. Since our previous results demonstrated a statistical correlation between emotional disorders and lymphopenia in EWOT1 and EWOT2 (unpublished data), an association that has been previously suggested [15–17], blood lymphocyte count was also included in our explanatory model.

23

2.5. Toxicity Safety data were analyzed according to the National Cancer Institute's Common Toxicity Criteria version 2.0. Severe toxicity was defined as any grade ≥ 3 toxicity.

2.6. Statistical methods The primary endpoint of this pooled analysis was overall survival (OS), from inclusion to patient death. We used Cox proportional hazards models to determine the relationship between each covariate (i.e., patient characteristics and CGA parameters) and OS for each study. The overall pooled hazard ratio (HR) and their 95% confidence intervals (CI) were then calculated. Chi-square heterogeneity tests were carried out. I2 statistics, expressing the proportion of variability in the results attributable to heterogeneity versus sampling error, were calculated, with I2 statistic b25% indicating low heterogeneity, 25%–50% moderate heterogeneity, and N50% high heterogeneity. When moderate heterogeneity was observed, a random effects model was used to pool HR. In our final model, we included explanatory variables with a univariate P value of b0.10, including albuminemia (b 35 vs ≥35 g/L), HADS score (N 14 vs ≤14), and FIGO stage (IV vs III) as categorical variables. Reduced model selection was carried out using a backward step-down by applying the stopping rule of the Akaike information criterion. All analyses were carried out using meta-analysis packages for the R statistical software program (R Foundation for Statistical Computing, Vienna, Austria).

3. Results 3.1. Patient characteristics Pooled analysis included a total of 266 patients (EWOT1, n = 83; EWOT2, n = 72; EWOT3, n = 111). Patient and study characteristics are summarized in Table 1. Overall, mean age was 77.1 ± 4.9 years; performance status was 0–1 in 61%; 28% had FIGO stage IV disease; and 25% had optimal primary surgical treatment. Despite similar inclusion criteria, there was some heterogeneity among the three patient groups. The population in EWOT3 tended to be more vulnerable than those in the two other studies, with a higher median age and a higher rate of altered performance status (≥2). Furthermore, the proportion of patients with optimal primary cytoreductive surgery was 16% in the EWOT3 trial versus 21% and 40%, respectively, in EWOT1 and EWOT2. Table 1 Patient characteristics.

Treatment years Treatment Age, years Median [range] Mean ± SD Performance status 0–1 (%) 2–3 (%) FIGO initial stage III (%) IV (%) Histological subtype Serous papillary (%)

EWOT1 N = 83

EWOT2 N = 72

EWOT3 N = 111

Total N = 266

1998–2000 Carboplatin + Cyclophosphamide

2001–2004 Carboplatin + Paclitaxel

2007–2010 Carboplatin

1998–2012 –

76 [70–90]

75 [70–89]

78 [70–93]

76 [70–93]

76.6 ± 1.1 N = 62

76.0 ± 4.4 N = 72

78.1 ± 5.1 N = 111

77.1 ± 4.9 N = 245

37 (60) 25 (40) N = 82 62 (76) 20 (24)

53 (74) 19 (26) N = 72 56 (78) 16 (22)

59 (53) 52 (47) N = 110 71 (65) 39 (35)

149 (61) 96 (39) N = 264 189 (72) 75 (28)

61 (73)

52 (71)

65 (59)

178 (67)

24

F. Tinquaut et al. / Gynecologic Oncology 143 (2016) 22–26

Table 2 Rates of missing geriatric data in EWOT trials.

Albuminemia Lymphocytes HADS OA depression IADL ADL BMI PS FIGO stage

EWOT1(N = 83)

EWOT2(N = 72)

EWOT3(N = 111)

28 (33.7%) 9 (10.8%)

18 (25%) 3 (4.2%) 30 (41.7%) 15 (20.8%) 15 (20.8%)

6 (5.4%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)

19 (22.9%)

0 (0%) 21 (25.3%) 0 (0%)

0 (0%) 0 (0%) 0 (0%)

3.2. Comprehensive geriatric assessment The distribution of CGA parameters in the three trials is presented in Supplementary Table S1. Noteworthy, the quality of geriatric data collection increased over time, illustrated by a decrease in missing data (Table 2). Overall, 130 patients (57%) reported taking four or more drugs daily. The most common comorbid conditions were hypertension (40%) and heart disease (20%). Approximately one-quarter of patients (26%) had a low BMI (b21), and half had hypoalbuminemia (b 35 g/L). In EWOT1, 61 patients (73%) reported complete autonomy at home, whereas in EWOT2 and EWOT3, the mean IADL score (n = 168/183 patients) was 22.0 ± 4.3, with IADL b 25 in 55% of patients. In EWOT2 and EWOT3, HADS mean score (n = 153/183 patients) was 13.5 ± 6.5, with 21% of patients identified as suffering emotional distress. Concurrently, the proportion of patients considered depressed according to oncologist assessment (OA) varied from 9% in EWOT2 to 22% in EWOT1. Lymphopenia b 1 × 103/mm3 was reported in more than one-quarter of patients. 3.3. Treatment completion and safety A total of 193 patients (73%) completed planned chemotherapy: 72% in EWOT1 (i.e., carboplatin plus cyclophosphamide), 68% in EWOT2 (i.e., carboplatin plus paclitaxel), and 74% in EWOT3 (i.e., carboplatin). The major reasons for treatment discontinuation in the three trials were severe chemotherapy toxicity and disease progression. Hematological and nonhematological toxicities are shown in Supplementary Table S2. Neutropenia tended to occur more often in patients receiving carboplatin plus paclitaxel than in those receiving either single-agent carboplatin or carboplatin plus cyclophosphamide. Conversely, thrombocytopenia was more frequently reported in these patients. The incidence of nonhematological toxicities was low in all three trials, irrespective of the chemotherapy regimen. 3.4. Prognostic factors for overall survival Median OS was 21.6 mo (95% CI 13.4–29.8 mo), 25.9 mo (95% CI 17.4–34.5 mo), and 17.4 mo (95% CI 13.3–21.4 mo) in EWOT1, 2, and 3, respectively. The results of univariate analysis for each trial, as well as pooled HR, are plotted in Fig. 1. Overall, heterogeneity among the studies was moderate. Results of the pooled analysis showed that the risk for overall mortality was higher in women aged ≥ 76 years (HR 1.68, 95% CI 1.11–2.53), with hypoalbuminemia (HR 2.21, 95% CI 1.27–3.85), with stage IV disease (HR 2.28, 95% CI 1.62–3.22), with emotional distress (HADS total score N 14; HR 1.79, 95% CI 0.98–3.27), or with lymphopenia (HR 1.58, 95% CI 1.10–2.29) (Table 3). Among these significant parameters, after adjustment on trial heterogeneity (random), multivariate analysis identified the following prognostic factors as affecting OS independently: patient considered depressed, hypoalbuminemia b35 g/L, and FIGO stage IV (Table 4, model a). Because paclitaxel use was previously proposed as a potential risk factor for poorer OS [13], possibly explaining some of this heterogeneity, the “paclitaxel”

Fig. 1. Univariate analyses of clinical and comprehensive geriatric assessment covariates (binary) in EWOT 1, 2 and 3.

covariate was tested but was not linked with OS in univariate analysis (HR 0.93, 95% CI, 0.64–1.36). In the subgroup of patients in EWOT2 and EWOT3 only (n = 183), the following additional covariates were associated in univariate analysis with a higher risk of mortality: HADS N14 and IADL score b 25 (Fig. 1, Supplementary Fig. 1). In this subpopulation, the best-fitting model

F. Tinquaut et al. / Gynecologic Oncology 143 (2016) 22–26 Table 3 Univariate analyses of clinical and comprehensive geriatric assessment covariates (binary) in all EWOT trials population (n = 266).

Age (≥76 vs b76 years) Albuminemia (b35 vs ≥35 g/L) FIGO stage (IV vs III) PS (≤1 vs N1) HADS (N14 vs ≤14) Rate of lymphocyte (b1 vs ≥ 1G/L) Depression (Yes vs No) BMI (N25 vs b25) Hemoglobine PINI IADL (b25 g/L vs N25) PNN

HR

95%-CI

I-squared

P-value

1.67 2.08 2.28 1.9 1.67 1.57 2.32 1.54 1.06 1.01 1.85 1.22

[1.18; 2.35] [1.40; 3.10] [1.62; 3.22] [1.3; 2.7] [1.10; 2.55] [1.08; 2.28] [1.51; 3.57] [1.10; 2.16] [0.95; 1.20] [1; 1.01] [1.20; 2.85] [1.14; 1.30]

28.10% 36.40% 0.01% 43.% 37.70% 26.30% 57.60% 0.01% 0.01% 0.00% 0.00% 83.10%

0.003 0.0003 b 0.0001 0.0004 0.02 0.02 0.0001 0.01 0.3 b 0.0001 0.006 b 0.0001

included the following covariates: HADS N14, hypoalbuminemia b35 g/L, and FIGO stage IV (Table 3, model f). 4. Discussion This pooled analysis of EWOT1–3 was planned after analysis of EWOT3, in order to externally validate the geriatric vulnerability parameters identified in the most recent EWOT study. A geriatric vulnerability score (GVS) was developed on the basis of the survival score; the GVS integrates the following items: albuminemia b 35 g/L, ADL score b 6, IADL score b 25, lymphopenia b1 × 103/mm3, and HADS N 14. A similar survival model had been developed that including the item “depressed”, Table 4 Cox regression models for overall survival: in the EWOT1 + EWOT2 + EWOT3 population (models a, b) and in the EWOT2 + EWOT3 population only (models c–f). Hazard ratio

95%CI

P value

Model a: EWOT1 + EWOT2 + EWOT3 population “Depressed” 2.3 Albuminemia (b35 g/L) 2.0 FIGO (stage IV) 2.8 BMI (N25) 1.5 Trial value (random)

1.5–3.7 1.3–3.0 1.9–4.3 0.9–2.2

0.0003 0.001 b0.0001 0.051 0.016

Model b: EWOT1 + EWOT2 + EWOT3 population “Depressed” 2.3 Albuminemia (b35 g/L) 2.0 FIGO (stage IV) 2.8 Trial value (random)

1.5–3.7 1.3–3.0 1.8–4.3

0.0004 0.001 1.7 × 10−6 0.006

Model c: EWOT2 + EWOT3 population only “Depressed” 1.7 Albuminemia (b35 g/L) 2.0 FIGO (stage IV) 2.2 BMI (N25) 1.8 HADS (N14) 1.4 IADL (score b 25) 1.5 Trial value (random)

0.9–2.9 1.3–3.3 1.3–3.5 1.1–2.7 0.9–2.3 0.9–2.5

0.06 0.004 0.001 0.01 0.17 0.13 0.94

Model d: EWOT2 + EWOT3 population only “Depressed” 1.9 Albuminemia (b35 g/L) 2.1 IADL score b 25 4.5 FIGO (stage IV) 2.2 Trial value (random)

1.1–3.3 1.3–3.3 0.9–2.5 1.4–3.5

0.01 0.001 0.09 0.001 0.9

Model e: EWOT2 + EWOT3 population only HADS (N14) 1.6 Albuminemia (b35 g/L) 1.9 IADL score b 25 1.3 FIGO (stage IV) 2.3 Trial value (random)

1.0–2.5 1.2–3.1 0.8–2.2 1.4–3.6

0.05 0.006 2.33 0.0004 0.22

Model f: EWOT2 + EWOT3 population only HADS (N14) 1.6 Albuminemia (b35 g/L) 2.0 FIGO (stage IV) 2.4 Trial value

1.0–2.5 1.22–3.2 1.5–3.8

0.04 0.005 0.0001 0.53

25

rather than HADS N 14, but the latter was preferred for its better subjective reliability among investigators. In EWOT3, with cut-off ≥3, the GVS discriminated between two groups with significantly different OS, treatment completion, severe adverse events, and unplanned hospital admissions rates [14]. Since GVS will be included in the design of future studies by the Gynecologic Cancer Intergroup (GCIG), particularly the EWOC-1 study of GINECO [18] for the identification of vulnerable patients, an analysis of the consistency of GVS items in all previous EWOT studies was needed. Of the five vulnerability parameters measured by GVS, four were included in this pooled analysis; two were valuable in the EWOT1 + EWOT2 + EWOT3 analysis (albuminemia and lymphopenia), and two in the EWOT2 + EWOT3 analysis only (IADL and HADS). Given the similarity between the survival models containing either HADS N14 or the “depressed” item, the latter was considered as a surrogate marker for the former. According to univariate analyses, “depressed”, age, hypoalbuminemia, low BMI, FIGO stage IV, and lymphopenia were associated with a worse prognosis in the total pooled population. HADS N14 and IADL score b 25 were associated with an increased risk of death in the EWOT2 + EWOT3-only population. According to multivariate Cox models including all patients included in EWOT1 + EWOT2 + EWOT3, two domains of the CGA, namely hypoalbuminenia and “depressed,” as well as FIGO stage IV, were independent prognostic factors of overall survival in elderly patients with advanced ovarian cancer. In the EWOT2 + EWOT3 population, the best-fitting models included either depression, hypoalbuminemia, FIGO stage IV, and IADL score b 25 or HADS, hypoalbuminemia, and FIGO stage IV as independent explanatory covariates of poor survival (Table 2, models c–e), validating the consistency of the GVS. The “depressed” item was previously described as a negative prognostic factor in a pooled analysis of EWOT1 and EWOT2 [13]. Moreover, a meta-analysis of the association between depression and mortality in cancer patients found that this association is stronger in older patients than in younger patients [19]. Nevertheless, in a previously published study of EWOT3, the prognostic impact of the item “depressed” on overall survival was lower and we hypothesized a possible heterogeneity between trials [14]. According this pooled analysis, the heterogeneity test I2 was 0.573 for this item, which is consistent with moderate-to-high heterogeneity, but this effect was not significant (P = 0.096). Considering the signification of this item, an analysis of the agreement between assessment tools of emotional status was performed with EWOT3 database. It revealed that “depressed” and HADS N14 should be considered more as indicators of emotional disorders rather than depression itself, since the agreement with DSMIV criteria for depression was low [20]. Nutritional status, another parameter of the CGA, should also be considered as a central element in the evaluation of elderly patients with advanced ovarian cancer. Indeed, an association between poor prognosis and lower levels of serum albumin has been reported in several studies of patients with ovarian cancer [21], although it was not statistically significant in previous analyses of EWOT1 and EWOT2 [12,13]. Moreover, the findings of a retrospective study indicated that improving nutritional status might positively affect survival in ovarian cancer patients [22]. More generally, the results of this pooled analysis are evidence of the continuing effort of GINECO to better understand heterogeneity in elderly populations and to collect data of better quality (see Table 2). Both the median age of the study population (76 years) and the high degree of functional losses, as assessed by IADL and ADL, are consistent with the wide inclusion criteria of the three trials. Nevertheless and regardless of chemotherapy regimens, the completion rate was 73%, varying from 68% in EWOT2 to 74% in EWOT3. Our results are in line with those of previous clinical trials on chemotherapy feasibility in elderly ovarian cancer patients [23,24], with completion rates ranging from 65% to 77%. Chemotoxicities remained manageable, suggesting that chemotherapy should be considered

26

F. Tinquaut et al. / Gynecologic Oncology 143 (2016) 22–26

even for vulnerable patients with advanced ovarian cancer, provided they are offered adequate attention and care. There are limitations to this pooled analysis. First, the EWOT data were covered over a long period of time, introducing potential biases due to changes in treatment practices. Second, geriatric parameters such as emotional status and functional status were evaluated with different—and increasing numbers of—assessment tools in the successive trials. Third, heterogeneity tests showed weak-to-high heterogeneity between trials among the covariates analyzed, suggesting possible confounding roles for the study population and/or treatment regimens. Conversely, the external validity of our study is strengthened by the wide inclusion criteria and the long duration of data collection, which ensured a study population that reflects the variability found among older adults and at different time periods. In its ongoing effort to provide close-to-everyday-life clinical data about elderly patients, GINECO is currently opening EWOC-1, an international trial that will evaluate three chemotherapy protocols of carboplatin ± paclitaxel in vulnerable patients who were identified as having at least 3 of the 5 vulnerability parameters of the GVS [18]. This trial, conducted under the auspices of the GCIG, will investigate and compare the current standard of carboplatin and paclitaxel, as evaluated in EWOT2; a carboplatin monotherapy, as proposed in EWOT3; and an adaptive protocol of weekly carboplatin (AUC 2 mg mL/min) and paclitaxel (60 mg/m2; day 1, day 8, and day 15 every 4 wk) that was developed by the Multicentre Italian Trial in Ovarian Cancer (MITO) cooperative group in Italy [24]. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ygyno.2016.03.018. Funding EWOT3 was supported by research grants from the French Ministry of Health (27-39) (Program Hospitalier de Recherche Clinique 2006, 27–39) and La Fondation de France (Grant N°2006010589). Disclosure The authors have declared no conflicts of interest. Acknowledgements The authors would like to acknowledge Douglas Micheau-Bonnier and Benedicte Votan from the GINECO study office and the Delegation a la Recherche Clinique et a l'Innovation of the Hospices Civils de Lyon. We also thank the following investigators who participated in the trial: Pr. H. Cure (Institut Jean Godinot, Reims), Dr. J. Salvat (Hopitaux du Leman, Thonon-les-Bains), Dr. M. Combe (Centre Hospitalier du Mans, Le Mans), Dr. M.C. Kaminsky (Centre Alexis Vautrin-Brabois, Vandoeuvre-les-Nancy), Dr. I. Ray-Coquard (Centre Leon Berard, Lyon), Dr. G. Yazbek, (Institut Jean Godinot, Reims), Dr. J. Meunier (Centre Hospitalier Regional d'Orleans, Orleans), Dr. J. Cretin (Clinique Bonnefon, Ales), Dr. L. Chauvenet, (Hopital Hotel-Dieu, Paris), Dr. J. Provencal (Centre Hospitalier de la region d'Annecy, Pringy), Dr. M. Fabbro (Hopital Arnaud de Villeneuve, Montpellier), Dr. M.N. Certain (Centre Hospitalier d'Auxerre, Auxerre), Dr. J.P. Guastalla (Centre Leon Berard, Lyon), Dr. Selar kalla (Groupe Hospitalier Saint-Joseph, Paris), Pr. J. Alexandre (Hopital Hotel-Dieu, Paris), Dr. E. Legouffe (Clinique de Valdegour, Nimes), Dr. F. Savinelli (Groupe Hospitalier Saint-Joseph, Paris), Dr. J.M. Tigaud (Hopital Hotel-Dieu, Paris), Dr. R. Largillier (Centre Azureen de Cancerologie, Mougins), Pr. O. Gisserot (Hopital d'Instruction des Armees Sainte-Anne, Toulon), Dr. C. Ligeza-Poisson (Clinique Mutualiste de l'Estuaire, Saint-Nazaire), Dr. G. Deplanque (Groupe Hospitalier Saint-Joseph, Paris), Dr. P. Deguiral (Clinique Mutualiste de l'Estuaire, Saint-Nazaire), Dr. E. Luporsi (Centre Alexis Vautrin-Brabois, Vandoeuvre-les-Nancy), Dr. F. Priou

(Centre Hospitalier Departemental Les Oudairies, La Roche-sur-Yon), Dr. F. Rousseau (Institut Paoli Calmettes, Marseille), and Dr. A. Le Rol (Hopital Perpetuel Secours, Levallois-Perret). We thank Laetitia Finzi (Clinsearch), Anne-Catherine Vignon, Richard Dupont (Hospira) for writing assistance and Paul Kretchmer (San Francisco Edit) for language editing.

References [1] A. Jemal, R. Siegel, J. Xu, E. Ward, Cancer statistics, 2010, CA Cancer J. Clin. 60 (5) (2010) 277–300. [2] D. Pectasides, G. Fountzilas, G. Aravantinos, A. Bamias, H.P. Kalofonos, D. Skarlos, et al., Epithelial ovarian carcinoma in younger vs older women: is age an independent prognostic factor? The Hellenic Oncology Cooperative Group experience, Int. J. Gynecol. Cancer 17 (5) (2007) 1003–1010. [3] S. Tingulstad, F.E. Skjeldestad, T.B. Halvorsen, B. Hagen, Survival and prognostic factors in patients with ovarian cancer, Obstet. Gynecol. 101 (5 Pt 1) (2003) 885–891. [4] L.A. Ries, Ovarian cancer. Survival and treatment differences by age, Cancer 71 (2 Suppl.) (Jan 15 1993) 524–529. [5] V. Sundararajan, D. Hershman, V.R. Grann, J.S. Jacobson, A.I. Neugut, Variations in the use of chemotherapy for elderly patients with advanced ovarian cancer: a population-based study, J. Clin. Oncol. 20 (1) (Jan 1 2002) 173–178. [6] T.L. Jorgensen, S. Teiblum, M. Paludan, L.O. Poulsen, A.Y. Jorgensen, K.H. Bruun, et al., Significance of age and comorbidity on treatment modality, treatment adherence, and prognosis in elderly ovarian cancer patients, Gynecol. Oncol. 127 (2) (2012) 367–374. [7] M. Extermann, A. Hurria, Comprehensive geriatric assessment for older patients with cancer, J. Clin. Oncol. 25 (14) (2007) 1824–1831. [8] M. Extermann, I. Boler, R.R. Reich, G.H. Lyman, R.H. Brown, J. Defelice, et al., Predicting the risk of chemotherapy toxicity in older patients: the chemotherapy risk assessment scale for high-age patients (CRASH) score, Cancer 9 (2011) 26646. [9] A. Hurria, K. Togawa, S.G. Mohile, C. Owusu, H.D. Klepin, C.P. Gross, et al., Predicting chemotherapy toxicity in older adults with cancer: a prospective multicenter study, J. Clin. Oncol. 29 (25) (2011) 3457–3465. [10] M. Extermann, I. Boler, R. Reich, G.H. Lyman, R.H. Brown, J. DeFelice, et al., The chemotherapy risk assessment scale for high-age patients (CRASH) score: design and validation, J. Clin. Oncol. 28 (15 s) (2010) (abstr 9000). [11] P. Maione, F. Perrone, C. Gallo, L. Manzione, F. Piantedosi, S. Barbera, et al., Pretreatment quality of life and functional status assessment significantly predict survival of elderly patients with advanced non-small-cell lung cancer receiving chemotherapy: a prognostic analysis of the multicenter Italian lung cancer in the elderly study, J. Clin. Oncol. 23 (28) (2005) 6865–6872. [12] G. Freyer, J.F. Geay, S. Touzet, J. Provencal, B. Weber, J.P. Jacquin, et al., Comprehensive geriatric assessment predicts tolerance to chemotherapy and survival in elderly patients with advanced ovarian carcinoma: a GINECO study, Ann. Oncol. 16 (11) (2005 Nov) 1795–1800. [13] O. Trédan, J.F. Geay, S. Touzet, R. Delva, B. Weber, J. Cretin, et al., Carboplatin cyclophosphamide or carboplatin pacitaxel in elderly with advanced ovarian cancer? Analysis of two consecutive trials from the GINECO, Ann. Oncol. 18 (2) (2007) 256–262. [14] C. Falandry, B. Weber, A.M. Savoye, F. Tinquaut, O. Tredan, E. Sevin, et al., Development of a geriatric vulnerability score in elderly patients with advanced ovarian cancer treated with first-line carboplatin: a GINECO prospective trial, Ann. Oncol. 2013 (2013) 22. [15] S.D. Targum, L.E. Marshall, P. Fischman, D. Martin, Lymphocyte subpopulations in depressed elderly women, Biol. Psychiatry 26 (6) (1989 Oct) 581–589. [16] D.F. Darko, J.C. Gillin, S.G. Risch, S. Golshan, K. Bulloch, S.M. Baird, Peripheral white blood cells and HPA axis neurohormones in major depression, Int. J. Neurosci. 45 (1–2) (Mar 1989) 153–159. [17] C. Garcia-Rizo, E. Fernandez-Egea, B.J. Miller, C. Oliveira, A. Justicia, J.K. Griffith, et al., Abnormal glucose tolerance, white blood cell count, and telomere length in newly diagnosed, antidepressant-naive patients with depression, Brain Behav. Immun. 28 (2013) 49–53. [18] http://www.clinicaltrials.gov/ct2/show/NCT02001272. [19] M. Pinquart, P.R. Duberstein, Depression and cancer mortality: a meta-analysis, Psychol. Med. 40 (11) (2010) 1797–1810. [20] M. Filbet, W. Rhondali, G. Freyer, V. Adam, M. Dezerelle, G. Abgrall-Barbry, et al., Correlations between depression according DSM-IV-TR (DSM) criteria, three validated scales, oncologist assessment, and clinical psychiatric interview in elderly advanced ovarian cancer (AOC) patients (pts): a GINECO study, J. Clin. Oncol. 30 (2012) ((Suppl):Abstr 9080). [21] D. Gupta, C.G. Lis, Pretreatment serum albumin as a predictor of cancer survival: a systematic review of the epidemiological literature, Nutr. J. 9 (69) (2010) 1475–2891. [22] D. Gupta, C.G. Lis, P.G. Vashi, C.A. Lammersfeld, Impact of improved nutritional status on survival in ovarian cancer, Support Care Cancer 18 (3) (2010) 373–381. [23] F. Hilpert, A. du Bois, E.R. Greimel, J. Hedderich, G. Krause, L. Venhoff, et al., Feasibility, toxicity and quality of life of first-line chemotherapy with platinum/paclitaxel in elderly patients aged ≥70 years with advanced ovarian cancer—a study by the AGO OVAR Germany, Ann. Oncol. 18 (2) (2006) 282–287. [24] S. Pignata, E. Breda, G. Scambia, C. Pisano, V. Zagonel, D. Lorusso, et al., A phase II study of weekly carboplatin and paclitaxel as first-line treatment of elderly patients with advanced ovarian cancer. A multicentre Italian trial in ovarian cancer (MITO-5) study, Crit. Rev. Oncol. Hematol. 66 (3) (2008) 229–236.