- Email: [email protected]
Development of a geriatric vulnerability score in elderly patients with advanced ovarian cancer treated with first-line carboplatin: a GINECO prospective trial
original articles
Annals of Oncology Annals of Oncology 24: 2808–2813, 2013 doi:10.1093/annonc/mdt360 Published online 22 September 2013
Development of a geriatric vulnerability score in elderly patients with advanced ovarian cancer treated with first-line carboplatin: a GINECO prospective trial C. Falandry1*, B. Weber2, A.-M. Savoye3, F. Tinquaut4, O. Tredan5, E. Sevin6, L. Stefani7, F. Savinelli8, M. Atlassi9, J. Salvat10, E. Pujade-Lauraine11 & G. Freyer12
Received 12 February 2013; revised 16 July 2013; accepted 18 July 2013
: Two previous GINECO elderly specific studies in advanced ovarian cancer (AOC) patients highlighted the prognostic value of geriatric covariates for overall survival (OS). Patients and methods: This open-label prospective trial was designed to identify the impact of geriatric covariates on OS in AOC patients ≥70 years treated with first-line carboplatin. Results: Geriatric covariates of the 111 patients included median age 79 years (≥80 years: 41%); performance status (PS) ≥2: 47%; ≥3 major comorbidities: 24%; ≥4 comedications: 68%; activities of daily living (ADL) score <6: 55%; instrumental activities of daily living (IADL) score <25: 69%; Hospital Anxiety and Depression Scale (HADS) >14: 37%. The median OS was 17.4 months. Overall, 74% of patients completed the six planned chemotherapy cycles. Grade 3–4 haematological toxic effects were frequent (50%) but manageable. Grade 3–4 non-haematological toxicities included fatigue (15%), anorexia (12%), infections (9%) and thrombosis (2%). A survival score = exp(0.327*GVS) was developed, where the geriatric vulnerability score (GVS) is the sum of the following (each assigned a value of one): albuminaemia <35 g/l; ADL score <6; IADL score <25; lymphopaenia <1 G/l; and HADS >14. With a cut-off ≥3, GVS discriminated two groups with significantly different OS, treatment completion, severe adverse events and unplanned hospital admissions rates. Conclusions: The GVS is a valuable tool for identifying vulnerable patients when treating an elderly AOC population. Key words: elderly, geriatrics, ovarian cancer, prognostic factors
introduction Ovarian cancer is the leading cause of death from gynaecological cancer in Europe [1] and the Western world [2]. Incidence and mortality increase with age, incidence peaking between 75 and 79 years and mortality between 80 and 84 years. About 50% of cases are diagnosed in women over 65 [1]. Most cases, particularly among elderly patients, are advanced stages (i.e. FIGO stages III-IV) [3]. Age has long been recognized as an independent poor prognostic factor [4], due in part to suboptimal management and excessive toxicity leading to dose limitations and treatment termination.
*Correspondence to: Dr Claire Falandry, Geriatrics and Oncology Unit, Centre Hospitalier Lyon Sud, 165 Chemin du Grand Revoyet, 69495, Pierre-Benite, France. Tel: +33-4-78-86-15 80; Fax: +33-4-78-86-57-27; E-mail: [email protected]
In 1997, the French National Group of Investigators for the Study of Ovarian and Breast Cancer (GINECO) initiated a programme evaluating the role of geriatric assessment in patients over 70 years treated with carboplatin– cyclophosphamide (CC) and standard carboplatin AUC5– paclitaxel (CP) [5, 6]. In a retrospective review of these trials, multivariate analysis of overall survival (OS) revealed a significant negative impact of age and emotional disorders according to the Hospital Anxiety and Depression Scale (HADS) [7]. Stage IV disease and lymphopaenia at study entry were also significantly associated with poor survival. Although patient characteristics favoured the CP group, survival curves were strictly comparable. Paclitaxel (Taxol) use was also an independent poor prognostic factor for survival [hazard ratio (HR) = 2.42, P = 0.001] [6]. While some clinicians consider standard treatment of advanced ovarian cancer (AOC; debulking surgery with six CP cycles) appropriate in elderly
© The Author 2013. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please email: [email protected].
Downloaded from http://annonc.oxfordjournals.org/ at Mount Allison University on June 26, 2015
1 Geriatrics and Oncology Unit, Centre Hospitalier Lyon Sud, Université Lyon 1, Pierre-Benite; 2Medical Oncology Department, Centre Alexis-Vautrin, Vandoeuvre-lesNancy; 3Medical Oncology Department, Institut Jean Godinot, Reims; 4Hygee Center, Regional center for Cancer Prevention, Saint Priest en Jarez; 5Medical Oncology Department, Centre Léon Bérard, Lyon; 6Medical Oncology Department, Centre François Baclesse, Caen; 7Medical Oncology Department, Centre Hospitalier de la Region d’Annecy, Pringy; 8Medical Oncology Department, Groupe Hospitalier Saint Joseph, Paris; 9Medical Oncology Department, Centre Hospitalier du Mans, Le Mans; 10 Medical Oncology Department, Hôpitaux du Léman, Thonon-les-Bains; 11Medical Oncology Unit, Hôpital Hôtel-Dieu and Université Paris Descartes, Paris; 12 Medical Oncology Unit, Centre Hospitalier Lyon Sud, Université Lyon 1, Pierre-Bénite, France
original articles
Annals of Oncology
patients [8] following proof of feasibility in a selected group [9], these data nonetheless question the value of combining paclitaxel with platinum in elderly AOC patients. The current trial extends the GINECO programme, aiming to delineate geriatric covariates to identify vulnerable elderly AOC patients treated with six carboplatin cycles.
patients and methods study design
patient population Eligible patients were ≥70 years old, with histologically or cytologically proven epithelial FIGO stage III and IV ovarian cancer. Cytology consistent with ovarian cancer was sufficient if associated with both a CA125 rise and a radiological pelvic mass. Patients were ineligible if they had prior chemo- or radiotherapy, planned interval debulking surgery, inadequate hepatic function [transaminases > 3 × upper limit of normal (ULN), total bilirubin > 2 × ULN] , haematological function (neutrophils <1.5 × 109/l, platelets <100 × 109/l) or severe renal insufficiency (creatinine clearance <30 ml/min).
assessments A pre-inclusion geriatric assessment reviewed the patient’s medical charts, nutrition and functionality and an extensive psychocognitive assessment (Supplementary Text S2, available at Annals of Oncology online). Toxicity was assessed throughout using the National Cancer Institute Common Toxicity Criteria, version 3.0.
statistical analyses Sample size was calculated based on the hypothesis that emotional disorders increase the risk of death at 2 years by threefold (bilateral test) [5, 6]. With an expected risk of death at 24 months of 85%, 105 patients were required with an α error = 5% and β = 10%. Assuming a 5% rate of non-assessable patients, 110 patients were planned. Survival curves were estimated using the Kaplan–Meier method in all enrolled patients (intent-to-treat). Univariate analyses were carried out using log-rank tests. Geriatric variables reaching P < 0.2 and considered clinically relevant were included in a Cox model to identify the optimal combined set of geriatric risk factors, termed geriatric vulnerability parameters. These were used to predict survival by calculating a geriatric vulnerability score (GVS). The GVS was calculated as follows: survival score = exp(β*N[criteria]), where β is the parameter obtained from the Cox model and N[criteria] is the
Volume 24 | No. 11 | November 2013
results patient characteristics and geriatric assessment One hundred and eleven patients were recruited between August 2007 and January 2010 in 21 centres. The baseline characteristics are shown in Table 1. The median follow-up was 16.4 months (range 0.2–49.6).
feasibility and safety Treatment exposure and safety are summarised in Table 2. Of the 109 treated patients, 82 received six cycles without severe toxicity or progression [treatment completion rate of 73.9%, 95% confidence interval (95% CI) 65.7%–82.1%]. Of the 27 patients who discontinued early, eight died, three following progression and five due to possibly treatment-related toxicity (septicaemia, two patients; pulmonary embolism, three patients).
tumour response and OS A total of 98 (88%) patients were assessable for response; 20 patients (18%) had a complete response, 27 (24%) had a partial response, 31 (28%) had stable disease and 20 (18%) had progression. At last follow-up, 75 patients (68%) had died and 13 (12%) were progression-free. The median PFS and OS were 9.2 months (95% CI 7.7–10.6 months) and 17.4 months (95% CI 13.3–21.4 months), respectively.
prognostic factors for OS Geriatric risk factors associated with decreased survival in univariate analysis and deemed clinically significant were identified (Table 3). When several covariates were available for the same geriatric dimension, one was selected as best fitting for implementation in the survival model. For nutritional assessment, albuminaemia <35 g/l was chosen over body mass index (BMI) and prognostic inflammatory and nutritional index (PINI) for its good reproducibility and segregation power. For psychological assessment, the HADS score >14 was chosen over the investigator’s estimation of emotional disorders as it is more reproducible and has previously been validated [6]. We used two assessment methods for functionality, given that both basal and instrumental activities of daily living (activities of daily living, ADL, and instrumental activities of daily living, IADL, respectively) had a significant impact on survival and are known to be different dimensions of functionality. An equal value was assigned to each geriatric vulnerability criteria: albuminaemia
doi:10.1093/annonc/mdt360 |
Downloaded from http://annonc.oxfordjournals.org/ at Mount Allison University on June 26, 2015
This open-label phase II multicentric trial was approved by a national ethics committee and conducted in accordance with the declaration of Helsinki and Good Clinical Practice guidelines. All patients provided written informed consent. Data were externally audited by the sponsor. Patients received carboplatin dosed at area under the curve (AUC) 5 mg/ml/min for 30 min every 3 weeks for up to six cycles (see Supplementary Text S1, available at Annals of Oncology online for dose reductions and comedications). The primary end point was to confirm the impact of psychogeriatric covariates on OS. Secondary objectives included assessment of chemotherapy feasibility, chemotherapy-related toxicity, progression-free survival (PFS), tumour response, and the impact of psychogeriatric covariates on treatment tolerance. Chemotherapy feasibility was defined as the completion of six cycles without disease progression, death due to any cause or investigator or patient decision to stop treatment due to chemotherapy-related toxicity.
total number of geriatric vulnerability parameters. Patients were assigned to one of the two subgroups (high-risk and low-risk) using the survival density function. A threshold was determined [receiver operation characteristic (ROC) curves] to discriminate the two subpopulations using the GVS (Supplementary Figures S1 and S2, available at Annals of Oncology online). A univariate survival analysis was carried out using each vulnerability group after adjustment for FIGO stage (IV versus III). The model was internally validated using the integrated Brier score [10] on the whole population and on 100 bootstrap samples. All tests were two-sided. Analyses were carried out using R statistical package (R Foundation for Statistical Computing, Austria) and Splus, version 6.2 (Insightful Corp., WA).
original articles Table 1. Patient and disease characteristics and geriatric assessment
Annals of Oncology Table 2. Treatment exposure and adverse events
Percentage of patients (n = 111) 79 (71–93) 41 47
75 6 11 8 1 64 35 75 16 50 8 24 32 41 27 55 69 17 61 22 25 29 37 36
ADL, activities of daily living; IADL, instrumental activities of daily living; BMI, body mass index; MMS, Mini-Mental Scale; HADS, Hospital Anxiety and Depression Scale; GDS, Geriatric Depression Scale.
<35 g/l, ADL score <6, IADL score <25, lymphopaenia at inclusion <1 × 109/l and HADS score >14.
development and validation of a survival model using the GVS The selected criteria were used to calculate the survival score = exp(β*N[criteria]) with β = 0.327, i.e. an HR of 1.387 (95% CI 1.082–1.777) for each criterion (see Supplementary Text S3 and Supplementary Figures S1–3, available at Annals of Oncology online). Given that N[criteria] is a discrete numeric function from 0 to 5, a Cox model was carried out using the likelihood maximum analysis (Table 4) along with a log-rank analysis of survival, the latter confirming good segregation of patients with a GVS of 0–2 versus those with a GVS of 3–5 (Figure 1). Among patients with a GVS of ≥3, the HR for | Falandry et al.
Percentage of patients (n = 111) Completion of six or more treatment cycles Early treatment discontinuation No treatment Toxicitya Disease progression Unplanned hospital admission Death during treatment Possibly treatment-related Disease progression Haematological toxicity, grade 3–4 (n = 109) Neutropaenia Thrombocytopaenia Anaemia Leukopaenia Non-haematological toxicity, grade 3–4 b(n = 108) Fatigue Anorexia Neurotoxicity Infection Dyspnoea Pain Nausea/vomiting Pulmonary embolism Thrombosis Alopecia
74 26 2 10 14 41 7 5 2 30 25 17 11 15 12 11 9 6 5 5 5 2 2
Includes five (5%) fatal events. Except alopecia (grade 2 only) and neurotoxicity (grade ≥ 2).
a
b
premature death was 2.94 (95% CI 1.79–4.84, P < 0.0001) in univariate analysis (median survival 21.7 versus 11.5 months) and 2.89 (95% CI 1.74–4.78, P < 0.0001) in multivariate analysis after adjustment for FIGO stage (HR 2.19 for stage IV, 95% CI 1.34–3.58, P = 0.002).
impact of GVS on treatment completion and toxicity GVS significantly segregated two groups with different treatment completion rates: 82.1% (95% CI 71.8% to 92.5%) for patients with a GVS <3 versus 65.5% (95% CI 52.5% to 78.4%) for patients with a GVS ≥3, with an odds ratio (OR) for the GVS ≥3 group of 0.41 (95% CI 0.17–0.99, P = 0.044). Patients with a GVS ≥3 experienced more grade ≥3 non-haematological toxicities (54.6 versus 21.4%, OR = 4.40; 95% CI 1.92–10.08, P = 0.0002), more serious adverse events (52.7 versus 28.6%, OR = 2.79; 95% CI 1.27–6.11, P = 0.009) and more unplanned hospital admissions (52.8 versus 30.4%, OR = 2.57; 95% CI 1.17–5.63, P = 0.017).
discussion We report here the third study in a series of elderly specific trials in ovarian cancer, reflecting GINECO’s focus on optimising treatment outcome by exploiting geriatric covariates. A total of 111 patients over the age of 70 were prospectively enrolled. Although eligibility criteria were similar to those of the two
Volume 24 | No. 11 | November 2013
Downloaded from http://annonc.oxfordjournals.org/ at Mount Allison University on June 26, 2015
Median age in years (range) ≥80 years Performance status (PS, ECOG) ≥2 Tumor assessment Histology Serous Endometrioid/clear cell Undifferentiated Other/unspecified FIGO stage IIb III IV Primary cytoreduction Complete Incomplete Unknown Geriatric assessment ≥3 comorbidities No. of comedications 1–3 4–6 ≥7 Functional assessment ADL score <6 IADL score <25 Home care Nutritional assessment Albuminaemia <35 g/l BMI <21 kg/m2 Lymphocyte count <1G/l Psychocognitive assessment MMS score <25 HADS score >14 GDS score >10
original articles
Annals of Oncology Table 3. Association between patient characteristics and overall survival (OS) Univariate (log-rank) Hazard ratio 2.39 1.86 0.32
0.002 0.02 0.026
1.62
0.07
2.16 2.00 1.79
0.006 0.003 0.041
2.36 3.04 0.90 2.12
0.003 <0.001 0.72 0.004
1.67 1.44 1.08
0.12 0.13 0.79
ADL, activities of daily living; IADL, instrumental activities of daily living; BMI, body mass index; HADS, Hospital Anxiety and Depression Scale; MMS, Mini Mental Scale; PINI, prognostic inflammatory and nutritional index;
Table 4. Impact of additional GVS criteria (maximum likelihood analysis)
Geriatric vulnerability score (GVS) criteria ADL score <6 IADL score <25 Albuminaemia <35 g/l Lymphocyte count <1 G/l HADS score >14 N criteria 0 1 2 3 4 5 ≥3 GVS criteria
Risk ratio
P value
1.39 – – – – –
0.01 – – – – –
1 1.31 1.3 3.37 3.8 4.1 2.94
– 0.57 0.55 0.008 0.005 0.01 0.0006
ADL, activities of daily living; IADL, instrumental activities of daily living; HADS, Hospital Anxiety and Depression Scale.
previous studies, the population in this study had a poorer prognostic factor profile on the basis of geriatric and oncologic parameters (Supplementary Table S1 and Supplementary Text S4, available at Annals of Oncology online). These characteristics are of note given that they conflict with the biases typically observed in elderly patients in clinical trials, which favour fitter patients with fewer comorbid conditions [11]. This study may have been biased towards including more vulnerable patients, partly attributed to investigators’ concerns over recruiting fit
Volume 24 | No. 11 | November 2013
Figure 1. Kaplan–Meier estimates of OS according to the number of GVS criteria.
patients to a trial using carboplatin monotherapy, as well as to the competition of large randomised trials without age limits. At 74% (95% CI 66% to 82%), the treatment completion rate was similar to one of the earlier studies, 76% (95% CI 66% to 85%) and slightly higher than the other at 68% (95% CI 57% to 79%) [5, 6]. Incidence of adverse events, severe adverse events and unplanned hospital admissions were, however, substantial, including 41% of patients with at least one unplanned hospital admission, lasting a median of 7 days. Of the 11 patients who stopped treatment prematurely due to toxicity, 5 died from their complication. The high number of treatment-related deaths with carboplatin monotherapy highlights just how lifethreatening chemotherapy-induced complications can be in a vulnerable elderly population. With a median of 17.4 months, OS in this trial was lower than in the two earlier trials (21.6 and 25.9 months) [6]. However OS, which we consider a better outcome measure than the response rate or PFS in this population, must be interpreted in light of the poor geriatric risk factor profile. This study confirmed the effect of psychogeriatric covariates identified in the two earlier trials on survival. In Trial 1, independent prognostic factors identified for OS were depression (P = 0.003), FIGO stage IV (P = 0.007) and more than six comedications daily (P = 0.043). The impact of emotional disorders was confirmed in trial 2, patients with a HADS score of >14 having a significantly increased risk of death (P = 0.015). In a grouped analysis of the two trials (158 patients), independent prognostic factors for poorer OS were increasing age (P = 0.013), baseline symptoms of depression (P < 0.001), FIGO stage IV (P = 0.001) and paclitaxel (Taxol) use (P = 0.025) [12]. In the current study, the effect of emotional disorders was less significant in univariate analysis, but was integrated in the scoring model for its clinical significance and on the basis of a meta-analysis of the three studies which confirms the effect of depression on OS (unpublished data). Based on the published data, very fit elderly patients, corresponding to the population included in large, randomised trials, can receive current treatment standards; however, little is
doi:10.1093/annonc/mdt360 |
Downloaded from http://annonc.oxfordjournals.org/ at Mount Allison University on June 26, 2015
Oncologic covariates FIGO stage IV Performance status (PS) ≥2 Optimal cytoreduction Geriatric covariates Age ≥80 Functional assessment ADL score <6 IADL score <25 Three or more comorbidities Nutritional assessment Albuminaemia <35 g/l PINI ≥10 BMI <21 kg/m² Lymphocyte count <1 G/l Psychocognitive assessment Emotional disorders (investigator’s assessment) HADS score >14 MMS score >24
P value
original articles
acknowledgements The authors acknowledge Douglas Micheau-Bonnier, Nicolas Gane and Bénédicte Votan from the GINECO study office, the Délégation à la Recherche Clinique et à l’Innovation of the Hospices Civils de Lyon and Raymonde Maraval-Gaget. We also thank the following investigators who participated in the trial: Pr. H. Curé (Institut Jean Godinot, Reims), Dr J. Salvat (Hôpitaux du Léman, Thonon-les-Bains), Dr M. Combe (Centre Hospitalier du Mans, Le Mans), Dr M.C. Kaminsky (Centre Alexis Vautrin-Brabois, Vandœuvre-lès-Nancy), Dr I. Ray-Coquard (Centre Léon Bérard, Lyon), Dr G. Yazbek, (Institut Jean Godinot, Reims), Dr J. Meunier (Centre
| Falandry et al.
Hospitalier Régional d’Orléans, Orléans), Dr J. Cretin (Clinique Bonnefon, Alès), Dr L. Chauvenet, (Hôpital Hôtel-Dieu, Paris), Dr J. Provençal (Centre Hospitalier de la région d’Annecy, Pringy), Dr M. Fabbro (Hôpital Arnaud de Villeneuve, Montpellier), Dr M.N. Certain (Centre Hospitalier d’Auxerre, Auxerre), Dr J.P. Guastalla (Centre Léon Bérard, Lyon), Dr S. Kalla (Groupe Hospitalier Saint-Joseph, Paris), Pr. J. Alexandre (Hôpital Hôtel-Dieu, Paris), Dr E. Legouffe (Clinique de Valdegour, Nîmes), Dr F. Savinelli (Groupe Hospitalier Saint-Joseph, Paris), Dr J.M. Tigaud (Hôpital HôtelDieu, Paris), Dr R. Largillier (Centre Azuréen de Cancérologie, Mougins), Pr. O. Gisserot (Hôpital d’Instruction des Armées 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, Vandœuvre-lès-Nancy), Dr F. Priou (Centre Hospitalier Départemental Les Oudairies, La Roche-sur-Yon), Dr F. Rousseau (Institut Paoli Calmettes, Marseille), Dr A. Le Rol (Hôpital Perpétuel Secours, LevalloisPerret). We thank Sarah MacKenzie for language editing (funded by ARCAGY-GINECO).
funding This work was supported by research grants from the French Ministry of Health (Programme Hospitalier de Recherche Clinique 2006, 27–39) and La Fondation de France (Grant N° 2006010589).
disclosure The authors have declared no conflicts of interest.
references 1. Oberaigner W, Minicozzi P, Bielska-Lasota M et al. Survival for ovarian cancer in Europe: the across-country variation did not shrink in the past decade. Acta Oncol 2012; 51: 441–453. 2. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2013. CA Cancer J Clin 2013; 63: 11–30. 3. Surveillance Epidemiology and End Results (SEER). 2013 http://seer.cancer.gov/ statfacts/html/ovary.html. (4 February 2013, date last accessed). 4. Thigpen T, Brady MF, Omura GA et al. Age as a prognostic factor in ovarian carcinoma. The Gynecologic Oncology Group experience. Cancer. 1993; 71: 606–614. 5. Freyer G, Geay JF, Touzet S et al. Comprehensive geriatric assessment predicts tolerance to chemotherapy and survival in elderly patients with advanced ovarian carcinoma: a GINECO study. Ann Oncol 2005; 16: 1795–1800. 6. Trédan O, Geay JF, Touzet S et al. Carboplatin cyclophosphamide or carboplatin pacitaxel in elderly with advanced ovarian cancer? Analysis of two consecutive trials from the GINECO. Ann Oncol 2007; 18: 256–262. 7. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983; 67: 361–370. 8. Zola P, Ferrero A. Is carboplatin–paclitaxel combination the standard treatment of elderly ovarian cancer patients? Ann Oncol 2007; 18: 213–214. 9. Hilpert F, du Bois A, Greimel ER et al. Feasibility, toxicity and quality of life of firstline chemotherapy with platinum/paclitaxel in elderly patients aged ≥70 years with advanced ovarian cancer—a study by the AGO OVAR Germany. Ann Oncol 2006; 18: 282–287.
Volume 24 | No. 11 | November 2013
Downloaded from http://annonc.oxfordjournals.org/ at Mount Allison University on June 26, 2015
known about more vulnerable patients. Our study shows that in addition to oncologic risk factors, different treatment strategies should be applied to low-risk versus high-risk elderly patients on the basis of a geriatric assessment, as has been suggested [12]. Future elderly specific prospective trials should focus on vulnerable patients, aided by tools such as the GVS. Among the five GVS risk factors, age over 80 was excluded because it was dependent on other parameters of the geriatric assessment in multivariate analysis. The best-fitting model included two dependent scores evaluating functionality (ADL and IADL), two dependent biological values (albuminaemia and lymphopaenia) and assessment of emotional disorders using the HADS. Other elements of the geriatric assessment were excluded because they did not add information to the fiveitem model in multivariate analysis. Functionality has been considered a core element in geriatric assessment since being shown to correlate with both OS and survival in healthy condition in the elderly [13, 14]. In a female population, such as in this study, IADL evaluation is rarely compromised by social habits which contribute to heterogeneity between men and women. Furthermore, there is mounting evidence suggesting that IADL correlates strongly with incident dementia and provides insight into cognitive assessment [15, 16]. Nutrition also plays a key role in geriatric assessment. In the context of cancer, where albuminaemia may be negatively influenced by inflammation, the PINI was also calculated and showed better discriminating power than albuminaemia. Nonetheless, albuminaemia <35 g/l was preferred because of better availability and good reproducibility. In contrast, the BMI had poor discriminating power, likely due to the artefactual impact of ascites. Lymphopaenia is integral to both the geriatric and the oncologic fields, correlating with morbi-mortality in the former [17] and with the risk of chemo-induced febrile neutropaenia [18] and poorer survival [19, 20] in the latter. In the two earlier GINECO trials, lymphopaenia was associated with decreased survival in a multivariate model (unpublished data). In summary, with a cut-off score of 3, the GVS can identify two groups with significantly different OS outcomes, treatment completion rates, grade 3–4 non-haematological toxic effects, serious adverse events and unplanned hospital admissions. It provides a valuable tool to identify vulnerable patients in future elderly AOC trials, and will be used in the Elderly Woman Ovarian Cancer GINECO/GCIG trial currently being initiated.
Annals of Oncology
original articles
Annals of Oncology 10. Gerds TA, Schumacher M. Consistent estimation of the expected Brier score in general survival models with right-censored event times. Biomed J 2006; 48: 1029–1040. 11. Townsley CA, Selby R, Siu LL. Systematic review of barriers to the recruitment of older patients with cancer onto clinical trials. J Clin Oncol 2005; 23: 3112–3124. 12. Balducci L, Yates J. General guidelines for the management of older patients with cancer. Oncology 2000; 14: 221–227. 13. Reuben DB, Rubenstein LV, Hirsch SH et al. Value of functional status as a predictor of mortality: results of a prospective study. Am J Med 1992; 93: 663–669. 14. Wang L, van Belle G, Kukull WB et al. Predictors of functional change: a longitudinal study of nondemented people aged 65 and older. J Am Geriatr Soc 2002; 50: 1525–1534. 15. Luck T, Luppa M, Wiese B et al. Prediction of incident dementia: impact of impairment in instrumental activities of daily living and mild cognitive impairment—results
16. 17. 18. 19.
20.
from the German Study on Ageing, Cognition, and Dementia in Primary Care Patients. Am J Geriatr Psychiatry 2012; 20: 943–954. Regal P, Heatherington E. Baseline instrumental activities of daily living and incident dementia. J Am Geriatr Soc 2012; 60: 1189–1190. Cretel E, Veen I, Pierres A et al. Immune profile of elderly patients admitted in a geriatric short care unit. Rev Med Interne 2011; 32: 275–282. Blay JY, Chauvin F, Le Cesne A et al. Early lymphopenia after cytotoxic chemotherapy as a risk factor for febrile neutropenia. J Clin Oncol 1996; x14: 636–643. Ray-Coquard I, Cropet C, Van Glabbeke M et al. Lymphopenia as a prognostic factor for overall survival in advanced carcinomas, sarcomas and lymphomas. Cancer Res 2009; 69: 5383–5391. Tredan O, Ray-Coquard I, Chvetzoff G et al. Validation of prognostic scores for survival in cancer patients beyond first-line therapy. BMC Cancer 2011; 11: 1471–2407.
Association of somatic DNA methylation variability with progression-free survival and toxicity in ovarian cancer patients J. M. Flanagan1, C. S. Wilhelm-Benartzi1, M. Metcalf1, S. B. Kaye2 & R. Brown1,3* 1
Epigenetics Unit, Department of Surgery and Cancer, Ovarian Cancer Action Research Centre, Imperial College London, London; 2Section of Medicine; 3Section of Molecular Pathology, Institute for Cancer Research, Sutton, UK
Received 4 February 2013; revised 23 July 2013; accepted 24 July 2013
Background: We have addressed whether inter-individual methylation variation in somatic (white blood cells, WBCs) DNA of ovarian cancer patients provides potential for prognostic and/or pharmacoepigenetic stratification. Patients and methods: WBC DNA methylation was analysed by bisulphite pyrosequencing at ataxia telangiectasia mutated (ATM), estrogen receptor 1 (ESR1), progesterone receptor (PGR), mutL homologue 1 (MLH1), breast cancer susceptibility gene (BRCA1), secreted frizzled-related protein 1 (SFRP1), stratifin (SFN), retinoic acid receptor beta (RARB) loci and the repetitive element LINE1 in 880 SCOTROC1 trial patients [ paclitaxel (Taxol)–carboplatin versus docetaxel (Taxotere)–carboplatin as primary chemotherapy for stage Ic–IV epithelial ovarian cancer]. Results: We observed no significant associations (P < 0.005, after correction for multiple testing) for progression-free survival (PFS) using test and validation sets. However, we did identify mean SFN methylation associated with PFS (hazard ratio, HR = 1.01 per 1% increase in methylation, q = 0.028); particularly in the paclitaxel (HR = 1.01, q = 0.006), but not in the docetaxel arm in stratified analyses. Furthermore, higher methylation within the ESR1 gene was associated with CA125 response (odds ratio, OR = 1.06, q = 0.04) and with neuropathy (HR = 0.95, q = 0.002), but only in the paclitaxel arm of the trial. Conclusions: This is the first study linking DNA methylation variability in WBC to clinical outcomes for any tumour type; the data generated on novel prognostic and pharmacoepigenetic DNA methylation biomarkers in the circulation now need independent further evaluation. Key words: clinical trial, DNA methylation, ovarian cancer, pharmacoepigenetics, prognosis, toxicity
*Correspondence to: Prof Robert Brown, Epigenetics Unit, Division of Cancer, Department of Surgery and Cancer, Faculty of Medicine, Imperial College London. 4th floor IRDB, Hammersmith Campus, Du Cane Road, London W12 0NN, UK. Tel: +44-20-75941804; Fax: +44-20-75942129; E-mail: [email protected]
© The Author 2013. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please email: [email protected].
Downloaded from http://annonc.oxfordjournals.org/ at Mount Allison University on June 26, 2015
Annals of Oncology 24: 2813–2818, 2013 doi:10.1093/annonc/mdt370 Published online 10 October 2013
Annals of Oncology Annals of Oncology 24: 2808–2813, 2013 doi:10.1093/annonc/mdt360 Published online 22 September 2013
Development of a geriatric vulnerability score in elderly patients with advanced ovarian cancer treated with first-line carboplatin: a GINECO prospective trial C. Falandry1*, B. Weber2, A.-M. Savoye3, F. Tinquaut4, O. Tredan5, E. Sevin6, L. Stefani7, F. Savinelli8, M. Atlassi9, J. Salvat10, E. Pujade-Lauraine11 & G. Freyer12
Received 12 February 2013; revised 16 July 2013; accepted 18 July 2013
: Two previous GINECO elderly specific studies in advanced ovarian cancer (AOC) patients highlighted the prognostic value of geriatric covariates for overall survival (OS). Patients and methods: This open-label prospective trial was designed to identify the impact of geriatric covariates on OS in AOC patients ≥70 years treated with first-line carboplatin. Results: Geriatric covariates of the 111 patients included median age 79 years (≥80 years: 41%); performance status (PS) ≥2: 47%; ≥3 major comorbidities: 24%; ≥4 comedications: 68%; activities of daily living (ADL) score <6: 55%; instrumental activities of daily living (IADL) score <25: 69%; Hospital Anxiety and Depression Scale (HADS) >14: 37%. The median OS was 17.4 months. Overall, 74% of patients completed the six planned chemotherapy cycles. Grade 3–4 haematological toxic effects were frequent (50%) but manageable. Grade 3–4 non-haematological toxicities included fatigue (15%), anorexia (12%), infections (9%) and thrombosis (2%). A survival score = exp(0.327*GVS) was developed, where the geriatric vulnerability score (GVS) is the sum of the following (each assigned a value of one): albuminaemia <35 g/l; ADL score <6; IADL score <25; lymphopaenia <1 G/l; and HADS >14. With a cut-off ≥3, GVS discriminated two groups with significantly different OS, treatment completion, severe adverse events and unplanned hospital admissions rates. Conclusions: The GVS is a valuable tool for identifying vulnerable patients when treating an elderly AOC population. Key words: elderly, geriatrics, ovarian cancer, prognostic factors
introduction Ovarian cancer is the leading cause of death from gynaecological cancer in Europe [1] and the Western world [2]. Incidence and mortality increase with age, incidence peaking between 75 and 79 years and mortality between 80 and 84 years. About 50% of cases are diagnosed in women over 65 [1]. Most cases, particularly among elderly patients, are advanced stages (i.e. FIGO stages III-IV) [3]. Age has long been recognized as an independent poor prognostic factor [4], due in part to suboptimal management and excessive toxicity leading to dose limitations and treatment termination.
*Correspondence to: Dr Claire Falandry, Geriatrics and Oncology Unit, Centre Hospitalier Lyon Sud, 165 Chemin du Grand Revoyet, 69495, Pierre-Benite, France. Tel: +33-4-78-86-15 80; Fax: +33-4-78-86-57-27; E-mail: [email protected]
In 1997, the French National Group of Investigators for the Study of Ovarian and Breast Cancer (GINECO) initiated a programme evaluating the role of geriatric assessment in patients over 70 years treated with carboplatin– cyclophosphamide (CC) and standard carboplatin AUC5– paclitaxel (CP) [5, 6]. In a retrospective review of these trials, multivariate analysis of overall survival (OS) revealed a significant negative impact of age and emotional disorders according to the Hospital Anxiety and Depression Scale (HADS) [7]. Stage IV disease and lymphopaenia at study entry were also significantly associated with poor survival. Although patient characteristics favoured the CP group, survival curves were strictly comparable. Paclitaxel (Taxol) use was also an independent poor prognostic factor for survival [hazard ratio (HR) = 2.42, P = 0.001] [6]. While some clinicians consider standard treatment of advanced ovarian cancer (AOC; debulking surgery with six CP cycles) appropriate in elderly
© The Author 2013. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please email: [email protected].
Downloaded from http://annonc.oxfordjournals.org/ at Mount Allison University on June 26, 2015
1 Geriatrics and Oncology Unit, Centre Hospitalier Lyon Sud, Université Lyon 1, Pierre-Benite; 2Medical Oncology Department, Centre Alexis-Vautrin, Vandoeuvre-lesNancy; 3Medical Oncology Department, Institut Jean Godinot, Reims; 4Hygee Center, Regional center for Cancer Prevention, Saint Priest en Jarez; 5Medical Oncology Department, Centre Léon Bérard, Lyon; 6Medical Oncology Department, Centre François Baclesse, Caen; 7Medical Oncology Department, Centre Hospitalier de la Region d’Annecy, Pringy; 8Medical Oncology Department, Groupe Hospitalier Saint Joseph, Paris; 9Medical Oncology Department, Centre Hospitalier du Mans, Le Mans; 10 Medical Oncology Department, Hôpitaux du Léman, Thonon-les-Bains; 11Medical Oncology Unit, Hôpital Hôtel-Dieu and Université Paris Descartes, Paris; 12 Medical Oncology Unit, Centre Hospitalier Lyon Sud, Université Lyon 1, Pierre-Bénite, France
original articles
Annals of Oncology
patients [8] following proof of feasibility in a selected group [9], these data nonetheless question the value of combining paclitaxel with platinum in elderly AOC patients. The current trial extends the GINECO programme, aiming to delineate geriatric covariates to identify vulnerable elderly AOC patients treated with six carboplatin cycles.
patients and methods study design
patient population Eligible patients were ≥70 years old, with histologically or cytologically proven epithelial FIGO stage III and IV ovarian cancer. Cytology consistent with ovarian cancer was sufficient if associated with both a CA125 rise and a radiological pelvic mass. Patients were ineligible if they had prior chemo- or radiotherapy, planned interval debulking surgery, inadequate hepatic function [transaminases > 3 × upper limit of normal (ULN), total bilirubin > 2 × ULN] , haematological function (neutrophils <1.5 × 109/l, platelets <100 × 109/l) or severe renal insufficiency (creatinine clearance <30 ml/min).
assessments A pre-inclusion geriatric assessment reviewed the patient’s medical charts, nutrition and functionality and an extensive psychocognitive assessment (Supplementary Text S2, available at Annals of Oncology online). Toxicity was assessed throughout using the National Cancer Institute Common Toxicity Criteria, version 3.0.
statistical analyses Sample size was calculated based on the hypothesis that emotional disorders increase the risk of death at 2 years by threefold (bilateral test) [5, 6]. With an expected risk of death at 24 months of 85%, 105 patients were required with an α error = 5% and β = 10%. Assuming a 5% rate of non-assessable patients, 110 patients were planned. Survival curves were estimated using the Kaplan–Meier method in all enrolled patients (intent-to-treat). Univariate analyses were carried out using log-rank tests. Geriatric variables reaching P < 0.2 and considered clinically relevant were included in a Cox model to identify the optimal combined set of geriatric risk factors, termed geriatric vulnerability parameters. These were used to predict survival by calculating a geriatric vulnerability score (GVS). The GVS was calculated as follows: survival score = exp(β*N[criteria]), where β is the parameter obtained from the Cox model and N[criteria] is the
Volume 24 | No. 11 | November 2013
results patient characteristics and geriatric assessment One hundred and eleven patients were recruited between August 2007 and January 2010 in 21 centres. The baseline characteristics are shown in Table 1. The median follow-up was 16.4 months (range 0.2–49.6).
feasibility and safety Treatment exposure and safety are summarised in Table 2. Of the 109 treated patients, 82 received six cycles without severe toxicity or progression [treatment completion rate of 73.9%, 95% confidence interval (95% CI) 65.7%–82.1%]. Of the 27 patients who discontinued early, eight died, three following progression and five due to possibly treatment-related toxicity (septicaemia, two patients; pulmonary embolism, three patients).
tumour response and OS A total of 98 (88%) patients were assessable for response; 20 patients (18%) had a complete response, 27 (24%) had a partial response, 31 (28%) had stable disease and 20 (18%) had progression. At last follow-up, 75 patients (68%) had died and 13 (12%) were progression-free. The median PFS and OS were 9.2 months (95% CI 7.7–10.6 months) and 17.4 months (95% CI 13.3–21.4 months), respectively.
prognostic factors for OS Geriatric risk factors associated with decreased survival in univariate analysis and deemed clinically significant were identified (Table 3). When several covariates were available for the same geriatric dimension, one was selected as best fitting for implementation in the survival model. For nutritional assessment, albuminaemia <35 g/l was chosen over body mass index (BMI) and prognostic inflammatory and nutritional index (PINI) for its good reproducibility and segregation power. For psychological assessment, the HADS score >14 was chosen over the investigator’s estimation of emotional disorders as it is more reproducible and has previously been validated [6]. We used two assessment methods for functionality, given that both basal and instrumental activities of daily living (activities of daily living, ADL, and instrumental activities of daily living, IADL, respectively) had a significant impact on survival and are known to be different dimensions of functionality. An equal value was assigned to each geriatric vulnerability criteria: albuminaemia
doi:10.1093/annonc/mdt360 |
Downloaded from http://annonc.oxfordjournals.org/ at Mount Allison University on June 26, 2015
This open-label phase II multicentric trial was approved by a national ethics committee and conducted in accordance with the declaration of Helsinki and Good Clinical Practice guidelines. All patients provided written informed consent. Data were externally audited by the sponsor. Patients received carboplatin dosed at area under the curve (AUC) 5 mg/ml/min for 30 min every 3 weeks for up to six cycles (see Supplementary Text S1, available at Annals of Oncology online for dose reductions and comedications). The primary end point was to confirm the impact of psychogeriatric covariates on OS. Secondary objectives included assessment of chemotherapy feasibility, chemotherapy-related toxicity, progression-free survival (PFS), tumour response, and the impact of psychogeriatric covariates on treatment tolerance. Chemotherapy feasibility was defined as the completion of six cycles without disease progression, death due to any cause or investigator or patient decision to stop treatment due to chemotherapy-related toxicity.
total number of geriatric vulnerability parameters. Patients were assigned to one of the two subgroups (high-risk and low-risk) using the survival density function. A threshold was determined [receiver operation characteristic (ROC) curves] to discriminate the two subpopulations using the GVS (Supplementary Figures S1 and S2, available at Annals of Oncology online). A univariate survival analysis was carried out using each vulnerability group after adjustment for FIGO stage (IV versus III). The model was internally validated using the integrated Brier score [10] on the whole population and on 100 bootstrap samples. All tests were two-sided. Analyses were carried out using R statistical package (R Foundation for Statistical Computing, Austria) and Splus, version 6.2 (Insightful Corp., WA).
original articles Table 1. Patient and disease characteristics and geriatric assessment
Annals of Oncology Table 2. Treatment exposure and adverse events
Percentage of patients (n = 111) 79 (71–93) 41 47
75 6 11 8 1 64 35 75 16 50 8 24 32 41 27 55 69 17 61 22 25 29 37 36
ADL, activities of daily living; IADL, instrumental activities of daily living; BMI, body mass index; MMS, Mini-Mental Scale; HADS, Hospital Anxiety and Depression Scale; GDS, Geriatric Depression Scale.
<35 g/l, ADL score <6, IADL score <25, lymphopaenia at inclusion <1 × 109/l and HADS score >14.
development and validation of a survival model using the GVS The selected criteria were used to calculate the survival score = exp(β*N[criteria]) with β = 0.327, i.e. an HR of 1.387 (95% CI 1.082–1.777) for each criterion (see Supplementary Text S3 and Supplementary Figures S1–3, available at Annals of Oncology online). Given that N[criteria] is a discrete numeric function from 0 to 5, a Cox model was carried out using the likelihood maximum analysis (Table 4) along with a log-rank analysis of survival, the latter confirming good segregation of patients with a GVS of 0–2 versus those with a GVS of 3–5 (Figure 1). Among patients with a GVS of ≥3, the HR for | Falandry et al.
Percentage of patients (n = 111) Completion of six or more treatment cycles Early treatment discontinuation No treatment Toxicitya Disease progression Unplanned hospital admission Death during treatment Possibly treatment-related Disease progression Haematological toxicity, grade 3–4 (n = 109) Neutropaenia Thrombocytopaenia Anaemia Leukopaenia Non-haematological toxicity, grade 3–4 b(n = 108) Fatigue Anorexia Neurotoxicity Infection Dyspnoea Pain Nausea/vomiting Pulmonary embolism Thrombosis Alopecia
74 26 2 10 14 41 7 5 2 30 25 17 11 15 12 11 9 6 5 5 5 2 2
Includes five (5%) fatal events. Except alopecia (grade 2 only) and neurotoxicity (grade ≥ 2).
a
b
premature death was 2.94 (95% CI 1.79–4.84, P < 0.0001) in univariate analysis (median survival 21.7 versus 11.5 months) and 2.89 (95% CI 1.74–4.78, P < 0.0001) in multivariate analysis after adjustment for FIGO stage (HR 2.19 for stage IV, 95% CI 1.34–3.58, P = 0.002).
impact of GVS on treatment completion and toxicity GVS significantly segregated two groups with different treatment completion rates: 82.1% (95% CI 71.8% to 92.5%) for patients with a GVS <3 versus 65.5% (95% CI 52.5% to 78.4%) for patients with a GVS ≥3, with an odds ratio (OR) for the GVS ≥3 group of 0.41 (95% CI 0.17–0.99, P = 0.044). Patients with a GVS ≥3 experienced more grade ≥3 non-haematological toxicities (54.6 versus 21.4%, OR = 4.40; 95% CI 1.92–10.08, P = 0.0002), more serious adverse events (52.7 versus 28.6%, OR = 2.79; 95% CI 1.27–6.11, P = 0.009) and more unplanned hospital admissions (52.8 versus 30.4%, OR = 2.57; 95% CI 1.17–5.63, P = 0.017).
discussion We report here the third study in a series of elderly specific trials in ovarian cancer, reflecting GINECO’s focus on optimising treatment outcome by exploiting geriatric covariates. A total of 111 patients over the age of 70 were prospectively enrolled. Although eligibility criteria were similar to those of the two
Volume 24 | No. 11 | November 2013
Downloaded from http://annonc.oxfordjournals.org/ at Mount Allison University on June 26, 2015
Median age in years (range) ≥80 years Performance status (PS, ECOG) ≥2 Tumor assessment Histology Serous Endometrioid/clear cell Undifferentiated Other/unspecified FIGO stage IIb III IV Primary cytoreduction Complete Incomplete Unknown Geriatric assessment ≥3 comorbidities No. of comedications 1–3 4–6 ≥7 Functional assessment ADL score <6 IADL score <25 Home care Nutritional assessment Albuminaemia <35 g/l BMI <21 kg/m2 Lymphocyte count <1G/l Psychocognitive assessment MMS score <25 HADS score >14 GDS score >10
original articles
Annals of Oncology Table 3. Association between patient characteristics and overall survival (OS) Univariate (log-rank) Hazard ratio 2.39 1.86 0.32
0.002 0.02 0.026
1.62
0.07
2.16 2.00 1.79
0.006 0.003 0.041
2.36 3.04 0.90 2.12
0.003 <0.001 0.72 0.004
1.67 1.44 1.08
0.12 0.13 0.79
ADL, activities of daily living; IADL, instrumental activities of daily living; BMI, body mass index; HADS, Hospital Anxiety and Depression Scale; MMS, Mini Mental Scale; PINI, prognostic inflammatory and nutritional index;
Table 4. Impact of additional GVS criteria (maximum likelihood analysis)
Geriatric vulnerability score (GVS) criteria ADL score <6 IADL score <25 Albuminaemia <35 g/l Lymphocyte count <1 G/l HADS score >14 N criteria 0 1 2 3 4 5 ≥3 GVS criteria
Risk ratio
P value
1.39 – – – – –
0.01 – – – – –
1 1.31 1.3 3.37 3.8 4.1 2.94
– 0.57 0.55 0.008 0.005 0.01 0.0006
ADL, activities of daily living; IADL, instrumental activities of daily living; HADS, Hospital Anxiety and Depression Scale.
previous studies, the population in this study had a poorer prognostic factor profile on the basis of geriatric and oncologic parameters (Supplementary Table S1 and Supplementary Text S4, available at Annals of Oncology online). These characteristics are of note given that they conflict with the biases typically observed in elderly patients in clinical trials, which favour fitter patients with fewer comorbid conditions [11]. This study may have been biased towards including more vulnerable patients, partly attributed to investigators’ concerns over recruiting fit
Volume 24 | No. 11 | November 2013
Figure 1. Kaplan–Meier estimates of OS according to the number of GVS criteria.
patients to a trial using carboplatin monotherapy, as well as to the competition of large randomised trials without age limits. At 74% (95% CI 66% to 82%), the treatment completion rate was similar to one of the earlier studies, 76% (95% CI 66% to 85%) and slightly higher than the other at 68% (95% CI 57% to 79%) [5, 6]. Incidence of adverse events, severe adverse events and unplanned hospital admissions were, however, substantial, including 41% of patients with at least one unplanned hospital admission, lasting a median of 7 days. Of the 11 patients who stopped treatment prematurely due to toxicity, 5 died from their complication. The high number of treatment-related deaths with carboplatin monotherapy highlights just how lifethreatening chemotherapy-induced complications can be in a vulnerable elderly population. With a median of 17.4 months, OS in this trial was lower than in the two earlier trials (21.6 and 25.9 months) [6]. However OS, which we consider a better outcome measure than the response rate or PFS in this population, must be interpreted in light of the poor geriatric risk factor profile. This study confirmed the effect of psychogeriatric covariates identified in the two earlier trials on survival. In Trial 1, independent prognostic factors identified for OS were depression (P = 0.003), FIGO stage IV (P = 0.007) and more than six comedications daily (P = 0.043). The impact of emotional disorders was confirmed in trial 2, patients with a HADS score of >14 having a significantly increased risk of death (P = 0.015). In a grouped analysis of the two trials (158 patients), independent prognostic factors for poorer OS were increasing age (P = 0.013), baseline symptoms of depression (P < 0.001), FIGO stage IV (P = 0.001) and paclitaxel (Taxol) use (P = 0.025) [12]. In the current study, the effect of emotional disorders was less significant in univariate analysis, but was integrated in the scoring model for its clinical significance and on the basis of a meta-analysis of the three studies which confirms the effect of depression on OS (unpublished data). Based on the published data, very fit elderly patients, corresponding to the population included in large, randomised trials, can receive current treatment standards; however, little is
doi:10.1093/annonc/mdt360 |
Downloaded from http://annonc.oxfordjournals.org/ at Mount Allison University on June 26, 2015
Oncologic covariates FIGO stage IV Performance status (PS) ≥2 Optimal cytoreduction Geriatric covariates Age ≥80 Functional assessment ADL score <6 IADL score <25 Three or more comorbidities Nutritional assessment Albuminaemia <35 g/l PINI ≥10 BMI <21 kg/m² Lymphocyte count <1 G/l Psychocognitive assessment Emotional disorders (investigator’s assessment) HADS score >14 MMS score >24
P value
original articles
acknowledgements The authors acknowledge Douglas Micheau-Bonnier, Nicolas Gane and Bénédicte Votan from the GINECO study office, the Délégation à la Recherche Clinique et à l’Innovation of the Hospices Civils de Lyon and Raymonde Maraval-Gaget. We also thank the following investigators who participated in the trial: Pr. H. Curé (Institut Jean Godinot, Reims), Dr J. Salvat (Hôpitaux du Léman, Thonon-les-Bains), Dr M. Combe (Centre Hospitalier du Mans, Le Mans), Dr M.C. Kaminsky (Centre Alexis Vautrin-Brabois, Vandœuvre-lès-Nancy), Dr I. Ray-Coquard (Centre Léon Bérard, Lyon), Dr G. Yazbek, (Institut Jean Godinot, Reims), Dr J. Meunier (Centre
| Falandry et al.
Hospitalier Régional d’Orléans, Orléans), Dr J. Cretin (Clinique Bonnefon, Alès), Dr L. Chauvenet, (Hôpital Hôtel-Dieu, Paris), Dr J. Provençal (Centre Hospitalier de la région d’Annecy, Pringy), Dr M. Fabbro (Hôpital Arnaud de Villeneuve, Montpellier), Dr M.N. Certain (Centre Hospitalier d’Auxerre, Auxerre), Dr J.P. Guastalla (Centre Léon Bérard, Lyon), Dr S. Kalla (Groupe Hospitalier Saint-Joseph, Paris), Pr. J. Alexandre (Hôpital Hôtel-Dieu, Paris), Dr E. Legouffe (Clinique de Valdegour, Nîmes), Dr F. Savinelli (Groupe Hospitalier Saint-Joseph, Paris), Dr J.M. Tigaud (Hôpital HôtelDieu, Paris), Dr R. Largillier (Centre Azuréen de Cancérologie, Mougins), Pr. O. Gisserot (Hôpital d’Instruction des Armées 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, Vandœuvre-lès-Nancy), Dr F. Priou (Centre Hospitalier Départemental Les Oudairies, La Roche-sur-Yon), Dr F. Rousseau (Institut Paoli Calmettes, Marseille), Dr A. Le Rol (Hôpital Perpétuel Secours, LevalloisPerret). We thank Sarah MacKenzie for language editing (funded by ARCAGY-GINECO).
funding This work was supported by research grants from the French Ministry of Health (Programme Hospitalier de Recherche Clinique 2006, 27–39) and La Fondation de France (Grant N° 2006010589).
disclosure The authors have declared no conflicts of interest.
references 1. Oberaigner W, Minicozzi P, Bielska-Lasota M et al. Survival for ovarian cancer in Europe: the across-country variation did not shrink in the past decade. Acta Oncol 2012; 51: 441–453. 2. Siegel R, Naishadham D, Jemal A. Cancer statistics, 2013. CA Cancer J Clin 2013; 63: 11–30. 3. Surveillance Epidemiology and End Results (SEER). 2013 http://seer.cancer.gov/ statfacts/html/ovary.html. (4 February 2013, date last accessed). 4. Thigpen T, Brady MF, Omura GA et al. Age as a prognostic factor in ovarian carcinoma. The Gynecologic Oncology Group experience. Cancer. 1993; 71: 606–614. 5. Freyer G, Geay JF, Touzet S et al. Comprehensive geriatric assessment predicts tolerance to chemotherapy and survival in elderly patients with advanced ovarian carcinoma: a GINECO study. Ann Oncol 2005; 16: 1795–1800. 6. Trédan O, Geay JF, Touzet S et al. Carboplatin cyclophosphamide or carboplatin pacitaxel in elderly with advanced ovarian cancer? Analysis of two consecutive trials from the GINECO. Ann Oncol 2007; 18: 256–262. 7. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scand 1983; 67: 361–370. 8. Zola P, Ferrero A. Is carboplatin–paclitaxel combination the standard treatment of elderly ovarian cancer patients? Ann Oncol 2007; 18: 213–214. 9. Hilpert F, du Bois A, Greimel ER et al. Feasibility, toxicity and quality of life of firstline chemotherapy with platinum/paclitaxel in elderly patients aged ≥70 years with advanced ovarian cancer—a study by the AGO OVAR Germany. Ann Oncol 2006; 18: 282–287.
Volume 24 | No. 11 | November 2013
Downloaded from http://annonc.oxfordjournals.org/ at Mount Allison University on June 26, 2015
known about more vulnerable patients. Our study shows that in addition to oncologic risk factors, different treatment strategies should be applied to low-risk versus high-risk elderly patients on the basis of a geriatric assessment, as has been suggested [12]. Future elderly specific prospective trials should focus on vulnerable patients, aided by tools such as the GVS. Among the five GVS risk factors, age over 80 was excluded because it was dependent on other parameters of the geriatric assessment in multivariate analysis. The best-fitting model included two dependent scores evaluating functionality (ADL and IADL), two dependent biological values (albuminaemia and lymphopaenia) and assessment of emotional disorders using the HADS. Other elements of the geriatric assessment were excluded because they did not add information to the fiveitem model in multivariate analysis. Functionality has been considered a core element in geriatric assessment since being shown to correlate with both OS and survival in healthy condition in the elderly [13, 14]. In a female population, such as in this study, IADL evaluation is rarely compromised by social habits which contribute to heterogeneity between men and women. Furthermore, there is mounting evidence suggesting that IADL correlates strongly with incident dementia and provides insight into cognitive assessment [15, 16]. Nutrition also plays a key role in geriatric assessment. In the context of cancer, where albuminaemia may be negatively influenced by inflammation, the PINI was also calculated and showed better discriminating power than albuminaemia. Nonetheless, albuminaemia <35 g/l was preferred because of better availability and good reproducibility. In contrast, the BMI had poor discriminating power, likely due to the artefactual impact of ascites. Lymphopaenia is integral to both the geriatric and the oncologic fields, correlating with morbi-mortality in the former [17] and with the risk of chemo-induced febrile neutropaenia [18] and poorer survival [19, 20] in the latter. In the two earlier GINECO trials, lymphopaenia was associated with decreased survival in a multivariate model (unpublished data). In summary, with a cut-off score of 3, the GVS can identify two groups with significantly different OS outcomes, treatment completion rates, grade 3–4 non-haematological toxic effects, serious adverse events and unplanned hospital admissions. It provides a valuable tool to identify vulnerable patients in future elderly AOC trials, and will be used in the Elderly Woman Ovarian Cancer GINECO/GCIG trial currently being initiated.
Annals of Oncology
original articles
Annals of Oncology 10. Gerds TA, Schumacher M. Consistent estimation of the expected Brier score in general survival models with right-censored event times. Biomed J 2006; 48: 1029–1040. 11. Townsley CA, Selby R, Siu LL. Systematic review of barriers to the recruitment of older patients with cancer onto clinical trials. J Clin Oncol 2005; 23: 3112–3124. 12. Balducci L, Yates J. General guidelines for the management of older patients with cancer. Oncology 2000; 14: 221–227. 13. Reuben DB, Rubenstein LV, Hirsch SH et al. Value of functional status as a predictor of mortality: results of a prospective study. Am J Med 1992; 93: 663–669. 14. Wang L, van Belle G, Kukull WB et al. Predictors of functional change: a longitudinal study of nondemented people aged 65 and older. J Am Geriatr Soc 2002; 50: 1525–1534. 15. Luck T, Luppa M, Wiese B et al. Prediction of incident dementia: impact of impairment in instrumental activities of daily living and mild cognitive impairment—results
16. 17. 18. 19.
20.
from the German Study on Ageing, Cognition, and Dementia in Primary Care Patients. Am J Geriatr Psychiatry 2012; 20: 943–954. Regal P, Heatherington E. Baseline instrumental activities of daily living and incident dementia. J Am Geriatr Soc 2012; 60: 1189–1190. Cretel E, Veen I, Pierres A et al. Immune profile of elderly patients admitted in a geriatric short care unit. Rev Med Interne 2011; 32: 275–282. Blay JY, Chauvin F, Le Cesne A et al. Early lymphopenia after cytotoxic chemotherapy as a risk factor for febrile neutropenia. J Clin Oncol 1996; x14: 636–643. Ray-Coquard I, Cropet C, Van Glabbeke M et al. Lymphopenia as a prognostic factor for overall survival in advanced carcinomas, sarcomas and lymphomas. Cancer Res 2009; 69: 5383–5391. Tredan O, Ray-Coquard I, Chvetzoff G et al. Validation of prognostic scores for survival in cancer patients beyond first-line therapy. BMC Cancer 2011; 11: 1471–2407.
Association of somatic DNA methylation variability with progression-free survival and toxicity in ovarian cancer patients J. M. Flanagan1, C. S. Wilhelm-Benartzi1, M. Metcalf1, S. B. Kaye2 & R. Brown1,3* 1
Epigenetics Unit, Department of Surgery and Cancer, Ovarian Cancer Action Research Centre, Imperial College London, London; 2Section of Medicine; 3Section of Molecular Pathology, Institute for Cancer Research, Sutton, UK
Received 4 February 2013; revised 23 July 2013; accepted 24 July 2013
Background: We have addressed whether inter-individual methylation variation in somatic (white blood cells, WBCs) DNA of ovarian cancer patients provides potential for prognostic and/or pharmacoepigenetic stratification. Patients and methods: WBC DNA methylation was analysed by bisulphite pyrosequencing at ataxia telangiectasia mutated (ATM), estrogen receptor 1 (ESR1), progesterone receptor (PGR), mutL homologue 1 (MLH1), breast cancer susceptibility gene (BRCA1), secreted frizzled-related protein 1 (SFRP1), stratifin (SFN), retinoic acid receptor beta (RARB) loci and the repetitive element LINE1 in 880 SCOTROC1 trial patients [ paclitaxel (Taxol)–carboplatin versus docetaxel (Taxotere)–carboplatin as primary chemotherapy for stage Ic–IV epithelial ovarian cancer]. Results: We observed no significant associations (P < 0.005, after correction for multiple testing) for progression-free survival (PFS) using test and validation sets. However, we did identify mean SFN methylation associated with PFS (hazard ratio, HR = 1.01 per 1% increase in methylation, q = 0.028); particularly in the paclitaxel (HR = 1.01, q = 0.006), but not in the docetaxel arm in stratified analyses. Furthermore, higher methylation within the ESR1 gene was associated with CA125 response (odds ratio, OR = 1.06, q = 0.04) and with neuropathy (HR = 0.95, q = 0.002), but only in the paclitaxel arm of the trial. Conclusions: This is the first study linking DNA methylation variability in WBC to clinical outcomes for any tumour type; the data generated on novel prognostic and pharmacoepigenetic DNA methylation biomarkers in the circulation now need independent further evaluation. Key words: clinical trial, DNA methylation, ovarian cancer, pharmacoepigenetics, prognosis, toxicity
*Correspondence to: Prof Robert Brown, Epigenetics Unit, Division of Cancer, Department of Surgery and Cancer, Faculty of Medicine, Imperial College London. 4th floor IRDB, Hammersmith Campus, Du Cane Road, London W12 0NN, UK. Tel: +44-20-75941804; Fax: +44-20-75942129; E-mail: [email protected]
© The Author 2013. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For permissions, please email: [email protected].
Downloaded from http://annonc.oxfordjournals.org/ at Mount Allison University on June 26, 2015
Annals of Oncology 24: 2813–2818, 2013 doi:10.1093/annonc/mdt370 Published online 10 October 2013