SIOG 2017 - Abstract Submission – Oral Presentations

J O U R NA L O F G E R I AT R I C O N C O L O G Y 8 S 1 ( 2 0 1 7 ) S 1 7 – S 3 7

A v a i l a b l e o n l i n e a t w w w. s c i e n c e d i r e c t . c o m

ScienceDirect

SIOG 2017 – Abstract Submission – Oral Presentations O01 OLDER CANCER PATIENTS’ PERCEPTION OF THEIR PREOPERATIVE ACTIVITY LEVEL COMPARED TO THE PEOPLE OF THEIR AGE STRONGLY CORRELATES WITH THEIR FRAILTY STATUS & OVERALL SURVIVAL A. Shahrokni1,*, A. Vickers2, S. Sarraf1, B. Korc-Grodzicki1 1 Medicine/Geriatrics and Oncology, 2Bio-statistics, Memorial Sloan Kettering Cancer Center, NYC, United States

Introduction: Surgery remains the mainstay for the cure of cancer among older cancer patients. Various frailty assessments have been developed to predict postoperative outcomes. Objectives: To assess the correlation between patients’ perception of their preoperative activity level compared to the people of their age with their frailty status and overall survival (OS). Methods: All older cancer patients who presented to Memorial Sloan Kettering Cancer Center completed geriatric assessment. They were asked to assess their overall activity level compared to people of their age. Those who mentioned they were much less limited, somewhat less limited, or same as others were categorized as “not limited” and those who mentioned they were somewhat more limited, much more limited were categorized as “limited”. We also measured frailty using the modified frailty index (m-FI). In order to exclude minor surgeries, we only included patients who required two days or longer hospital postsurgical stay. We used Cox regression analysis to assess the association between patients’ perception of their activity level and OS after adjusting for frailty status. We compared the OS of patients who were fit & not limited with patients who were either frail and/or limited. Results: In total, 879 patients (median age 80) were included. The median time of follow-up was 329 days (IQ range 155– 511). 20.8% and 14.8% of patients thought they were much less limited and somewhat less limited than others respectively. 40.4% thought they were as active as others. 15.6% and 8.4% thought they were somewhat more and much more limited respectively. As measured by m-FI, the prevalence of frailty was 34.8%, 57.5%, 52.8% among those who were much more active, somewhat more active and as active as others, respectively. 1879-4068/Published by Elsevier Ltd.

Fig. 1 (abstract O01) – Survival function for patterns 1–3. The prevalence of frailty was 74.4% and 85.9% among those who thought they were somewhat less and much less active than others, respectively. After adjusting for frailty, those who were limited had shorter OS when compared to those who were not limited (HR=1.85, P=0.005). Compared to patients who were fit and not limited, those who were either frail or limited had a poorer OS (HR=2.0, p<0.001). Those who were frail and limited had the poorest OS (HR=3.8, p<0.001). Conclusion: Older cancer patients’ perception of their preoperative activity level compared to people of their same age is accurate and has a strong correlation with their frailty status and OS. Future studies should assess whether this single question has high sensitivity/specificity to serve as a rapid frailty screening test before surgery. Disclosure of interest: None declared Keywords: Frailty, geriatric assessment, outcomes, surgery

S18

ABSTRACTS

O02 A NOVEL GERIATRIC ASSESSMENT TOOL THAT PREDICTS POSTOPERATIVE SURGICAL AND GERIATRIC COMPLICATIONS IN OLDER ADULTS WITH CANCER Y. Pollock1,*, C.-L. Chan2, K. Hall2, K. Diehl3, L. Min2 1 Div. of Hematology/Oncology, University of California San Francisco, San Francisco, 2Div. of Geriatric Medicine, 3Dept. of Surgery, University of Michigan, Ann Arbor, United States

Introduction: Comprehensive geriatric assessment (CGA) has been shown to predict surgical outcomes among oncology patients in a number of studies [1,2]. However completing a CGA can be time consuming, which is a huge barrier to implement it in most surgery clinics. Objectives: We aimed to assess the ability of a new abbreviated geriatric assessment tool, the Vulnerable Elderly Surgical Pathways and outcomes Assessment (VESPA), to predict post-operative surgical complications, geriatric complications, length of stay, and post-discharge functional or nursing needs among older adults undergoing oncologic surgeries. Methods: From 2008 to 2011, assessments were completed using the VESPA tool for patients age 70 seen in an university pre-operative clinic. The VESPA assessed functional status, mood, cognition, and mobility, and can be completed in <10 minutes. We selected the subset of patients who underwent a variety of oncologic surgery and evaluated association of the VESPA score with surgical complications, geriatric complication, post-discharge functional dependence or nursing needs, and length of stay. We used t-test, 2, or fisher’s exact test to compare the complication rates between high and low VESPA scores (9 vs. <9). Logistic regression was used to test whether VESPA score predicts all the post-operative complications measured in this study. Results: A total of 476 patients who underwent oncologic surgeries received geriatric assessment using VESPA. The study cohort comprised of a wide range of malignancies including skin, gastrointestinal, urologic, breast, head and

neck, and lung cancers. Compared to patients with low VESPA scores (<9), patients with high VESPA scores (9) had longer length of stay (mean 6.6 vs. 2.0 days; p<0.001), more geriatric complications (39.5% vs. 5.7%; p<0.001), surgical complications (29.5% vs. 11.8%; p<0.001), and post-discharge functional dependence or nursing needs (76.0% vs. 31.7%; p<0.001). Each additional point on the VESPA scale was associated with increased probability of geriatric complications (OR=1.3 [95% CI=1.2–1.4]; p<0.001), surgical complications (OR=1.2 [95% CI=1.1–1.2]; p<0.001), and post-discharge functional or nursing needs (OR=1.3 [95% CI=1.2–1.3]; p<0.001) (Figure 1). Conclusion: The VESPA tool adds to the existing body of research [3–4] by providing risk prediction for not only surgical complications, but also postoperative adverse outcomes that are unique to older adults including delirium, falls, malnutrition, and functional needs. It is also a practical tool that can feasibly fit into the current workflow of a surgical oncology clinic. References: [1] Feng MA, McMillan DT, Crowell K, et al. Geriatric assessment in surgical oncology: A systematic review. J Surg Res. 2015;193(1):265–272. [2] Huisman MG, Kok M, De Bock GH, et al. Delivering tailored surgery to older cancer patients: Preoperative geriatric assessment domains and screening tools - A systematic review of systematic reviews. Eur J Surg Oncol. 2017;43(1):1– 14. [3] Ghignone F, Van Leeuwen BL, Montroni I, et al. The assessment and management of older cancer patients: A SIOG surgical task force survey on surgeons’ attitudes. Eur J Surg Oncol. 2016;42(2):297–302. [4] Kenig J, Olszewska U, Zychiewicz B, et al. Cumulative deficit model of geriatric assessment to predict the postoperative outcomes of older patients with solid abdominal cancer. J Geriatr Oncol. 2015;6(5):370–379. Disclosure of interest: None declared Keywords: Functional status, geriatric assessment, postoperative complication

O03 THE EFFECT OF LIVE MUSIC ON POSTOPERATIVE PAIN IN ELDERLY PATIENTS H. van der Wal- Huisman1,*, E. Heineman1, B. L. van Leeuwen2 1 Surgery, 2Surgical oncology, University Medical Center Groningen, Groningen, Netherlands

Fig. 1 (abstract O02) – Adjusted prediction with 95 Cis for different complications. Prediction based on logistic regression for different complications. Mean age (78) was used to estimate the probabilities.

Introduction: Many patients experience pain after surgery. Pain has a negative influence on physical and mental recovery which may result in a variety of postoperative complications and functional impairment. Music is a universal and nonpharmaceutical intervention for pain and therefore seems suitable for the growing number of elderly patients needing (oncological) surgery as treatment. However, the effect of (live) music on postoperative pain among elderly patients is unknown.

ABSTRACTS

Objectives: In this pilot study, which is a unique collaboration between a university medical center and a conservatoire, we obtained experience in performing live music for elderly patients after surgery and examined the effects on postoperative pain. Methods: A pilot study, consisting of six standalone weeks, with a pre-post-test design with follow-up was conducted between September 2015 and May 2016 among postoperative patients, aged 60 years. The intervention contained: live music, person centered improvisation and existing repertoire, performed by professional musicians, from the conservatoire, for approximately 10–15 minutes, one session a day on three different surgical wards. Pain was measured with a Visual Analogue Scale (VAS; score 0–10) before the music session, 30 min and3 hours after the session. Higher VAS score indicated more pain. The control group, measured in separated weeks but on the same times and the same wards, consisted of patients who did not receive the intervention. Patients could participate with a maximum of seven days. Wilcoxon Signed Rank Test was performed to determine a difference within the groups and the Mann Whitney U test to compare VAS scores between both groups. Results: Eighty-eight music sessions among 38 patients, mean age 70.5 years (SD 6.5), were measured and 90 sessions were measured among 46 patients, mean age 70.3 years (SD 6.6) in the control group. The average participated music sessions was 2.32 (SD 1.78). VAS scores decreased within the music group, indicating less pain, between pretest and posttest, and, pretest and follow-up (p=0.000). No difference was found within the control group. There was no difference in pre-test scores between the music and control group (p=0.56); however, there was on the post-test scores (p=0.013) and follow-up (p=0.003). Table 1 (abstract O03) – VAS-scores Music group (mean)

Control group (mean)

Pre-test score

1.25 (SD 1.85)

1.49 (SD 2.14)

0.56

Post-test score

0.66 (SD 1.49)*1

1.41 (SD 2.12)

0.013*

Follow-up score 0.52 (SD 1.01)*1

1.42 (SD 2.12)

0.003*

p value

1

Significant difference within music group between pre-test vs posttest and pre-test vs follow-up

Conclusion: Music has a positive effect on experienced pain among elderly patients after surgery and is therefore suitable as a non-pharmaceutical intervention. Given the results of this pilot study, further research in a larger population is necessary to determine whether music is of influence on the dosage of pain medication. Disclosure of interest: None declared Keywords: Elderly patients, music, pain, postoperative

S19

O04 THE CARE PROGRAM: OCCUPATIONAL AND PHYSICAL THERAPY FOR OLDER ADULTS WITH CANCER M. Pergolotti1,*, A. M. Deal2, G. R. Williams3, L. McCarthy2, A. L. Bryant2, H. B. Muss2 1 Occupational Therapy, Colorado State University, Fort Collins, 2 University of North Carolina at Chapel Hill, Chapel Hill, 3 University of Alabama at Birmingham, Birmingham, United States

Introduction: Older adults with cancer experience physical and psychological deficits that can negatively impact their ability to participate in meaningful activity and decrease their overall quality of life (QOL). Occupational and physical therapy (OT/PT) address physical, functional and social needs to improve QOL however, these services remain underutilized and not tested in geriatric oncology. Objectives: The purpose of this randomized control trial was to examine the impact of a Cancer Rehabilitation (CARE) program, consisting of outpatient OT/PT, on the functional status, global QOL and the self-reported possibilities for activity in older adults with cancer. Methods: Participants were randomly assigned to either the CARE program or usual care (no rehabilitation). Inclusion criteria for participants was >65 years of age, cancer diagnosis or recurrence within past 5 years, and at least one functional limitation. OT and PT were structured by specific evaluation measures only, OT addressed functional activities and participation, and PT addressed strength/endurance needs. Outcome measures included functional status (Nottingham Extended Activities of Daily Living Scale [NEADL] (range 0–22)), global mental and physical health and participation in social roles (Patient-Reported Outcomes Measurement Information System® [PROMIS®] (range 0-100), and activity expectation and self-efficacy (Possibilities for Activities Scale [PActS]. Participants from both groups were administered each of the measures at baseline and again at three months. T-tests were used to compare groups. Results: Fifty-one adults were randomized: median age 73 years, 55% male, 92% White, 33% with Leukemia/lymphoma, 26% Breast, 22% Colorectal, 67% in active treatment, and 37% with Stage 3 or 4. After 3 months, both groups experienced a significant decline in functional status (p=0.046; p=0.005), but change in functional status (–1.5 UC, –1.1 CARE, p=0.637), physical health status (0.0 UC, 2.4 CARE, p=0.121) and participation in social roles (0.11 in UC, 3.71 CARE, p=.088) between group were not significant. However, change in mental health (–1.0 in UC, 3.0 CARE, p=0.032) and change in PActS scores between the treatment and control group (p=0.03), were significant, even when baseline scores were controlled for (p=0.02). Conclusion: This OT/PT intervention had a significant benefit on the mental health, activity expectation and selfefficacy in this sample of older adults with cancer. Future studies will have longer time between baseline and follow-up, and larger sample sizes. At a minimum, this study suggests cancer rehabilitation can improve participation in meaningful activity, and improve mental health in older adults with cancer. Reference:

S20

ABSTRACTS

[1] Pergolotti M, Cutchin M, Muss H. Predicting participation in meaningful activity for older adults with cancer. Quality of Life Research. 2015;24(5):1217–1222. doi:10.1007/s11136014-0849-7 Disclosure of interest: None declared

Table 1 (abstract O05) – Aggregate findings

Keywords: Activities, geriatric assessment, health related quality of life, participation, rehabilitation

O05 SURVIVAL OF AML PATIENTS AGED 70 AND OLDER. MODELING FROM DATA OBTAINED IN MORE THAN 13000 PATIENTS M. Extermann1,2,*, M. Sehovic2, T. Reljic3, J. Kim4, J. Lancet1,5, B. Djulbegovic5 1 Dpt of Oncology Sciences, University of South Florida, 2Senior Adult Oncology, Moffitt Cancer Center, 3University of South Florida, 4 Biostatistics core, 5Malignant hematology, Moffitt Cancer Center, Tampa, United States

Introduction: Little is known about the prognosis and response to treatment of patients aged 70+ with AML, as well as the impact of cytogenetics, performance status (PS), and comorbidity. Objectives: Obtain systematic data to support clinical decision modeling for older AML patients. Methods: We conducted a systematic review of clinical trials enrolling patients 70 and older with AML and extracted subgroup data, either from published data, or by queries to the investigators. Treatment was categorized into 4 categories: intensive chemotherapy, low-dose chemotherapy, hypomethylating agents (HMA), and best supportive care (BSC). Data were collected on cytogenetics (favorable, intermediate, poor), ECOG PS (0–1 vs 2+), and comorbidity (Charlson 0 vs 1+). Results: We obtained data from 13416 patients. The median age was 75 years. Time points for survival reporting varied from study to study, and aggregate findings are presented in Table 1. Patient treated with hypomethylating agents had the best prognosis. Patients treated with intensive chemotherapy and HMAs had a better survival than those treated with BSC or low-dose chemotherapy. Out of 65 studies, 29 reported data on cytogenetics, 31 on ECOG PS, and 4 on comorbidity. The impact of cytogenetics and ECOG PS on OS is significant, while the impact of comorbidity is more difficult to assess based on the small numbers assessed. For example, for patients treated with intensive chemotherapy, the 1-year survival of patients with good-intermediate cytogenetics is 38% vs 18% with poor cytogenetics; 38% with ECOG PS 0-1, vs 27% with ECOG PS 2 or more. Conclusion: This to our knowledge the largest data set on AML patients aged 70 and older. Active treatment with HMAs or intensive chemotherapy is associated with better survival in these patients. Cytogenetics and ECOG PS significantly influence the results. Data are being integrated in a decision model to allow personalized treatment discussions with patients.

Overall survival 1 year

2 years

3 years

5 years

Intensive chemo

0.34 (0.26–0.42)

0.19 (0.15–0.23)

0.12 (0.08–0.16)

0.11 (0.09–0.14)

Low-dose chemo

0.11 (0.06–0.18)

0.12 (0.05–0.21)

0.08 (0.05–0.12)

No data

HMA

0.45 (0.35–0.55)

0.17 (0.10–0.26)

0.09 (0.06–0.13)

0.05 (0.02–0.12)

BSC

0.17 (0.14–0.19)

0.06 (0.02–0.11)

0.04 (0.02–0.06)

0.02 (0.00–0.06)

Disclosure of interest: M. Extermann Grant / Research Support from: GTx, M. Sehovic: None declared, T. Reljic: None declared, J. Kim: None declared, J. Lancet Grant / Research Support from: Pfizer, Consultant for: Asterias Biotherapeutics; Baxalta; Bio-Path Holdings, Inc; Biosight; Boehringer Ingelheim; Celator; Celgene; ERYTECH Pharma; Janssen; Jazz Pharmaceuticals; Karyopharm Therapeutics; Novartis, B. Djulbegovic: None declared Keywords: AML, clinical decision making, prognosis, systematic review

O06 MAJOR IMPACT OF MILD COGNITIVE IMPAIRMENT (MCI) AND INFLAMMATORY STATUS IN OLDER PATIENTS RECEIVING CHEMOTHERAPY FOR HEMATOLOGICAL MALIGNANCIES S. Dubruille1,*, C. Kenis2, Y. Libert3, M. Delforge4, L. Dal Lago5, M. Roos5, C. Borghgraef5, A. Salaroli1, M. Maerevoet1, D. Razavi3, H. Wildiers2, D. Bron1 1 Department of Hematology, Institut Jules Bordet, Brussels, 2 Department of Oncology, University Hospitals Leuven, Leuven, 3 Clinic of Psycho-Oncology, Institut Jules Bordet, Brussels, 4 Department of Hematology, University Hospitals Leuven, Leuven, 5 Onco-geriatry Unit, Institut Jules Bordet, Brussels, Belgium

Introduction: Patients “clinically fit” to receive chemotherapy for a malignant hemopathy, are a heterogeneous population covering fit and vulnerable patients. Patients with geriatric syndromes and/or irreversible comorbidities are usually excluded from standard dose chemotherapy. Major progresses have been achieved to identify older patients who should be treated with standard dose of chemotherapy. However, a reliable “frailty score” remains urgently needed to better define the vulnerable population that does not benefit from chemotherapy. In the literature, two clinical (Mild Cognitive Impairment (MCI) and gastro-duodenal ulcer) and two biological (low hemoglobin level and IL-6 level) factors are frequently correlated with poor overall survival (OS) and/or chemotherapy-related toxicity. Objectives: To establish the reliability of a simple clinicobiological tool for the screening of vulnerable patients with malignant hemopathies presenting unacceptable

ABSTRACTS

chemotherapy-related toxicity or disappointing results defined as a poor OS (<6 months). Methods: This prospective multicentric study was conducted in the institute Jules Bordet (Brussels) and in the University Hospitals of Leuven (Leuven). A Comprehensive Geriatric Assessment (CGA) was performed to 269 consecutive patients (65–90 years) with malignant hemopathies admitted to receive chemotherapy. A screening tool composed of 0 to 4 of the prognostic factors (MCI [Mini Mental State Examination], presence of gastro-duodenal ulcer [Charlson Comorbidity Index], anemia [hemoglobin level] and IL-6 level [CRP]) was studied in our population. Univariate and multivariate Cox proportional hazards model were used to predict OS. Results: Two hundred and thirteen patients were evaluable for the clinico-biological tool (NHL, n=126; CLL, n=23; MM, n=29; AML, n=18; ALL, n=5; LMMC, n=8, MDS, n=4). Eighty-four percent had a more favorable prognosis (NHL, CLL or MM) and fifty-four percent had a first diagnosis of cancer. A “frailty” scoring system (range 0–4 items) was developed, based on our predictive factors for poor survival: Mild Cognitive Impairment (MCI) (MMSE <27, n=65), duodenal ulcer (n=24), anemia (HB <11 g/dl, n=99) and CRP 2 mg/l, (n=167). The population was stratified into 3 groups: fit (score=0–1, n=90), vulnerable (score= 2, n=82) and “frail” (score= 3 or 4, n=41). The one-year OS was 77% in fit, 62% in vulnerable (hazard ratio (HR)=1.83; 95% CI=1.05–3.19; p=0.033) and 39% in “frail” patients (HR=3.90; 95% CI=2.18–6.98; p<0.001) with a median survival of 3 months. Conclusion: In our selected population of patients referred to receive chemotherapy for malignant hemopathies, our “frailty score” helps clinician to predict a poor OS. This “frailty score” detects unsuspected “frail” patients who may benefit from palliative care. Further prospective analyses in a larger cohort of malignant hemopathies are ongoing to validate this score. Disclosure of interest: None declared Keywords: Hematological malignancies, inflammatory status, mild cognitive impairment, older, survival

O07 A LINK BETWEEN PSOAS MUSCLE AREA AND OUTCOMES IN ELDERLY PATIENTS WITH DIFFUSE LARGE B-CELL LYMPHOMA TREATED WITH R-CHOP A. Schlesinger1–3,*, I. Vaxman1–3, N. Goldberg1–3, S. Lando1–3, A. Gafter-Gvili1–3, P. Raanani1–3, M. Lahav1–3, L. Vidal1–3 1 Acute Geriatrics, Rabin Medical Center, 2Senior Adult Oncology, Davidoff Cancer Center, Petah Tikva, 3Internal Medicine, Tel Aviv University, Tel Aviv, Israel

Introduction: Sarcopenia, can be part of normal aging or result from cachexia. Multiple studies found correlation between reduced muscle mass and negative outcomes in chronic diseases and solid tumors. Little research exists on similar relationship and on longitudinal changes in muscle mass during treatment in elderly hematological patients.

S21

Objectives: To evaluate the muscle mass during the course of chemotherapy in patients 70 years or older with DLBCL and its effect on survival and toxicity as well as its changes during progression of the treatment. Methods: We performed a retrospective cohort study of patients diagnosed between the 2007 and 2014 and treated with RCHOP in our institution. We collected demographic, clinical and laboratory data as well as chemotherapy doses, timing and complications, treatment response, and survival. We evaluated muscle mass by adding bilateral psoas muscle cross-sectional areas at the level of the third lumbar vertebra on PET CT images and correcting it to patient’s height. The ratio was defined as muscle index and expressed in cm2/m2. The change in muscle index after vs. before treatment was estimated. The effect of pre-treatment muscle index and the above mentioned variables was estimated in a univariate Cox regression analysis. Variables potentially associated with mortality were entered into a Cox regression multivariate analysis. Results: Ninety-three patients (50% female) were included in the cohort. Median age was 78 years (range 70–90). Sixty percent had an IPI score of 3 or more. Mean pre- treatment muscle index was 4.66 cm2/m2, median 4.4 cm2/m2. End of treatment PET CT was available for 76 patients. Mean post treatment index was 4.2 cm2/m2, median 3.92 cm2/m2. A decrease in muscle index was observed in 76% of patients. A negative correlation was shown between pretreatment index and days of hospitalization in cycles 1–2 (p=0.007, r=–0.28). Pre-treatment muscle index was not associated with overall survival (p=0.43). In a sub-group analysis by sex, a higher muscle index was associated with a longer overall survival in men (HR 0.59, 95% CI 0.44 to 78, p<0.001), but not in women. In a multivariate model, the variables that were associated with overall survival were albumin, age, and gender (p<0.05). Conclusion: Based on our data including 93 elderly patients with DLBCL, a higher muscle mass before RCHOP was associated with longer overall survival among men but not among women. No association was found between muscle mass measured as total psoas area corrected to height and dose intensity, or infection. During the course of chemotherapy, we observed a loss of muscle mass in most patients. Those findings comply with previous studies on younger patients. Further research is needed to clarify causality of low muscle mass in men on shorter survival. Such a study should be prospective and include interventions aimed at improving muscle mass pre- and during the treatment. Disclosure of interest: None declared Keywords: Elderly, lymphoma, muscle mass, psoas area, sarcopenia

S22

ABSTRACTS

O08 TRANSCRIPTION FACTORS (BACH2 AND PRDM1) AND CHECKPOINT INHIBITORS EXPRESSION IN T-LYMPHOCYTES FROM CHRONIC LYMPHOCYTIC LEUKEMIA PATIENTS AND AGE-MATCHED HEALTHY DONORS D. Luan1, K. Willard-Gallo2, J.-N. Lodewyckx2, C. Solinas2, C. Gu-Trantien2, S. Garaud2, H. Duvillier1, B. Stamatopoulos1, C. Sibille3, D. Bron1,* 1 Clinical Hematology Dpt, 2Molecular Immunology Unit, 3Anatomopathology Unit, Institut Jules Bordet, ULB, Brussels, Belgium

Introduction: Aging is characterized by a progressive decline in immune surveillance that favors tumor development in older patients. We previously reported BACH2 gene as a candidate tumor suppressive gene (TSG). Objectives: We thus examined the expression of specific transcription factors (BACH2 and PRDM1) and checkpoint inhibitors (PD-1 and PD-L1) in the major lymphocytes subsets for their potential role in immunosenescence. Methods: Peripheral blood mononuclear cells were isolated from whole blood using Lymphoprep density gradient centrifugation. Lymphocyte subsets (CD19+, CD3+CD4+; CD3+CD8+) were isolated for subsequent molecular analyses using the MACS Technology (Miltenyi), with the purity of each lymphocyte subpopulation between 95–99%. PD-1 (PDCD1), PD-L1 (CD274), IL4, IFNG, BACH2 and PRDM1 mRNA transcripts were quantified using qRT-PCR. BACH2 and BLIMP1 (PRDM1) protein expression were examined by Western blotting. Results: Blood samples were obtained from 60 healthy volunteers and 41 untreated B-cell chronic lymphocytic leukemia (B-CLL) patients (median: 67 years). Healthy donors (HD) between the ages of 20 to 90 years subdivided into <50 years (median: 36 years) and 50 years (median: 65 years). BACH2 mRNA expression in the HD groups is significantly down-regulated in CD4+, CD8+ T cells and CD19+ B cells from the older HD group (p=0.0012, 0.0045, and 0.0367, respectively). BACH2 expression was further reduced in CD4+, CD8+ T cells and CD19+ B cells from CLL patients compared to HD well balanced for age (p=0.001, <0.0001, and 0.0043, respectively). PRDM1 mRNA expression was inversely correlated with BACH2 in CD4+, CD8+ T cells and CD19+ B cells (r=0.61, 0.71, and 0.65, respectively). Curiously, PRDM1 was, as expected, significantly up-regulated in CD4+ and CD8+ T cells (p=0.0034, and p=0.0017) from B-CLL patients but not in their leukemic B cells. Western blotting analysis demonstrated that BACH2 and BLIMP1 (PRDM1) protein expressions in the T and B cell subpopulations were significantly correlated with transcript expression. BACH2-deficient mice have been shown to have an increased numbers of IL4-producing CD4+ T cells. We also observed that BACH2 down-regulation is correlated with increased IL-4 mRNA expression (r=0.67) but not IFN in CD4+ T cells. These observations suggest that BACH2 downregulation in CD4+ T cells could enhance the expression of effector memory-related genes, particularly Th2, such as IL-4 and PRDM1. PD-1 mRNA expression was up-regulated in CD4+, CD8+ T cells (p=0.0153 and 0.0214) in the older HD group and also up-regulated in the T cells from B-CLL patients (p=0.0014

and 0.0023) when compared to age-matched HD population. High PD-L1 mRNA expression was correlated with increased age in HD B cells (p=0.04) with a further increase detected in leukemic B cells (p=0.001). We also observed an inverse correlation between BACH2 and PD-1 in CD4+, CD8+ T cells (r=0.62 and 0.68); and between BACH2 and PD-L1 in CD19+ B cells (r=0.66). Conclusion: These data suggest that down-regulation of BACH2/PRDM1 and up-regulation of PD1/PD-L1 mRNA expression in major lymphocyte subsets from CLL patients and older healthy controls are significantly correlated with the aging immune cells and could be part of the immunosenescence process. Disclosure of interest: None declared Keywords: Aging, CLL, elderly

O09 FACTORS ASSOCIATED WITH ADVANCE DIRECTIVES COMPLETION AMONG OLDER ADULTS WITH CANCER E. Soto Perez De Celis1,*, H. Kim2, M. Koczywas1, M. Fakih1, D. Li1, J. Chao1, V. Chung1, P. Twardowski1, M. Cristea1, S. Pal1, Y. Yuan1, C. Sun2, L. Chien1, C. Kelly1, M. Trent1, K. Charles1, H. Tan1, E. Roberts1, D. Rivera3, G. Varatkar4, S. Hite4, D. Bhatt1, K. W. Yu1, D. Mitani5, B. Ferrell6, A. Hurria1 1 Department of Medical Oncology and Therapeutics Research, 2 Survey Research Core, 3Supportive Care Medicine, 4Department of Rehabilitation Services, 5Patient, Family and Community Education, 6 Division of Nursing Research and Education, City of Hope, Duarte, United States

Introduction: The completion of Advance Directives (AD) increases access to palliative care, and improves quality of life. Despite this evidence, AD completion by patients with cancer in the United States has been reported to be as low as 37%. Previous studies in patients with cancer have identified older age, advanced stage and race/ethnicity as predictive factors for higher AD completion rates, but little is known about which factors are related to a higher likelihood of having AD among older patients with cancer. Objectives: To identify the sociodemographic, clinical and geriatric assessment characteristics associated with a higher likelihood of having AD among older adults with cancer starting a new line of chemotherapy. Methods: This is a secondary analysis of a prospective randomized controlled trial aimed at identifying vulnerabilities in older adults (age 65) starting a new line of chemotherapy for solid tumors through the use of a geriatric assessment (NCT02517034). The relationships between not having AD at the time of enrollment and sociodemographic, clinical and geriatric assessment variables (including functional status, cognition, psychological status, social support, comorbidity, nutritional status and quality of life [QOL]) were evaluated using univariate and multivariate logistic regression. Results: 335 older adults (median age 71, 57% female) were included. Patients had gastrointestinal (31%, n=103), breast (20%, n=68), genitourinary (GU) (19%, n=62), lung (17%, N=58)

ABSTRACTS

and gynecologic (7%, n=24) tumors. 77% of patients had Stage IV disease (n=257). 57% of patients (n=191) were Non-Hispanic White, 21% (n=69) were Hispanic, 11% (n=37) Asian and 5% (n=16) African American. 42% (n=142) of patients had never completed AD at the time of study enrollment. On univariate analysis the following factors were associated with higher odds of not having AD at a p value of <0.05: younger age; African American, Hispanic or Asian race/ethnicity; lower educational level; poor emotional support (defined using the MOS Social Support Survey); non-GU malignancies; Karnofsky Performance Status (KPS) <70%; anxiety/depression (defined using the Mental Health Inventory-17 questionnaire); visual impairment; dependence in independent activities of daily living (IADL) and activities of daily living (ADL); and poorer QOL (defined using the FACT-G questionnaire). On multivariate analysis, the following factors remained significantly associated with higher odds of not having AD: African American or Hispanic race/ethnicity (OR 2.8, 95% CI 1.4–5.8); high school or lower educational level (OR 2.4, 95% CI 1.2–4.5); and having a non-GU malignancy (OR 2.5, 95% CI 1.1–5.57). The following factors were associated with lower odds of not having AD: older age (OR 0.9, 95% CI 0.89–0.98), higher KPS score (OR 0.97, 95% CI 0.95–0.99) and better QOL (0.98, 95% CI 0.96–0.99). Conclusion: The proportion of older patients with AD in our patient population was higher than that reported in previous studies. Among older adults with cancer, those with lower educational levels, Hispanic or African American race/ ethnicity and worse self-reported QOL are less likely to have AD, regardless of their geriatric assessment results. This information stresses the importance of designing culturally tailored interventions aimed at improving the understanding of AD and increasing their acceptability and completion among older adults with cancer with diverse educational, cultural and ethnic backgrounds. Disclosure of interest: E. Soto Perez De Celis: None declared, H. Kim: None declared, M. Koczywas: None declared, M. Fakih: None declared, D. Li Consultant for: Novartis, J. Chao: None declared, V. Chung Consultant for: Perthera, Speakers bureau: Celgene, P. Twardowski: None declared, M. Cristea: None declared, S. Pal Consultant for: Pfizer, Novartis, Genentech, Aveo, Eisai, Ipson, Exelixis, Y. Yuan Grant / Research Support from: Eisa, Novartis, Merck, Pfizer, Genentech, Puma, Consultant for: Novartis, C. Sun: None declared, L. Chien: None declared, C. Kelly: None declared, M. Trent: None declared, K. Charles: None declared, H. Tan: None declared, E. Roberts: None declared, D. Rivera: None declared, G. Varatkar: None declared, S. Hite: None declared, D. Bhatt: None declared, K. W. Yu: None declared, D. Mitani: None declared, B. Ferrell: None declared, A. Hurria Grant / Research Support from: Celgene Corporation, Novartis, Consultant for: Boehringer Ingelheim, Pieiran Biosciences Keywords: Advance care planning, decision making, healthcare disparities, palliative care, supportive care

S23

O10 CHARACTERISTICS ASSOCIATED WITH PHYSICAL FUNCTION TRAJECTORIES IN OLDER ADULTS WITH CANCER RECEIVING CHEMOTHERAPY: A MULTICENTER PROSPECTIVE COHORT STUDY M. L. Wong1,*, S. M. Paul2, L. C. Walter3, C. Miaskowski2 1 Divisions of Hematology/Oncology and Geriatrics, 2School of Nursing, 3Division of Geriatrics, University of California, San Francisco, San Francisco, California, United States

Introduction: Functional decline during cancer treatment is associated with poor quality of life and decreased survival. Prior studies of physical function in older adults with cancer evaluated associations of demographic and clinical characteristics with functional decline during chemotherapy (CTX), but the impact of symptoms on functional decline remains unknown. Additionally, prior studies compared physical function at only two time points and did not include multiple assessments to map the full trajectory of physical function. Objectives: To determine which demographic, clinical, and symptom characteristics are associated with initial levels as well as trajectories of physical function at six time points over two cycles of CTX in adults age 65 with breast, gastrointestinal, gynecologic, or lung cancer. Methods: A total of 363 older adults with cancer were recruited from two Comprehensive Cancer Centers, one Veterans Affairs hospital, and four community oncology programs in the US. All patients received CTX within the preceding four weeks and were scheduled to receive at least two additional cycles. For each CTX cycle, changes in physical function were assessed prior to CTX administration and at one and two weeks after CTX administration using the Medical Outcomes Study Short Form-12 Physical Component Summary (PCS) score. PCS scores can range from 0 (lowest) to 100 (highest level of functioning). To evaluate common symptoms at enrollment, patients completed questionnaires on morning and evening fatigue, morning and evening energy, pain, depression, trait and state anxiety, attentional function, and sleep disturbance. Multivariable hierarchical linear modeling was used to evaluate for inter-individual variability in initial levels and trajectories of PCS score. Results: Mean age was 71.4 years (SD 5.5, range 65–90). Patients were predominately female (68.3%). Gastrointestinal was the most common cancer type (32.8%) followed by breast (23.1%), lung (22.3%), and gynecological cancer (21.8%). Interindividual variability in PCS scores (spaghetti plot) was found with mean initial PCS score of 40.5 (SD 0.45). On average, PCS scores declined 0.21 points linearly at each subsequent time point. Lower initial PCS scores were associated with older age, higher comorbidity, lack of active employment, lack of regular exercise, lack of trait anxiety, lower attentional function, presence of pain, morning fatigue, and lack of evening energy. Only morning fatigue (p=0.04) and initial PCS score (p=0.01) were associated with functional decline over time. Of note, clinical characteristics including Karnofsky Performance Status score, cancer type, number of prior cancer treatments, and presence of metastatic disease were not associated with either initial levels or trajectories of PCS scores.

S24

ABSTRACTS

Conclusion: While several symptoms were associated with initial PCS scores in older adults with cancer receiving CTX, morning fatigue was the only symptom associated with functional decline. Regular assessment of fatigue, specifically morning fatigue, and evidence-based interventions should be considered to maintain physical function in older adults during CTX. Disclosure of interest: None declared

Methods: We used an interpretive descriptive methodology to conduct an in-depth exploration of the perspectives of both older adults (n=17) and healthcare professionals (n=21) regarding older adults’ use of CRII. We conducted interviews (n=40) and focus groups (n=3), and analyzed the data using thematic analysis. Results: Patients and healthcare providers felt CRII was a beneficial resource to address gaps in information, and to supplement information from their healthcare professionals. Regarding older adults, healthcare professionals felt they were less likely to use CRII, or to want to seek information from the internet, and felt that most users of CRII were young and often female. Conversely, older adults described unique strategies to seek, find, and validate CRII, in spite of their technological unfamiliarity. Older adults also described mobilizing the support of family and friends to find accurate and reliable information. Conclusion: The perceptions shared by healthcare professionals about older adults’ CRII use and capabilities, conflicted with patient accounts. These are important findings for practice, pointing to the need for greater awareness of older adults CRII use and how to integrate this awareness into patient-healthcare professional encounters. Future research is needed to explore interventions to support CRII use amongst older adults, due to their unique information needs and preferences, the high likelihood of comorbidities, and the interplay with family and caregiver support. Disclosure of interest: None declared

Keywords: Chemotherapy, functional decline, morning fatigue, physical function

Keywords: Communication, older adults, self-management, technology

O11 OLDER ADULTS’ USE OF CANCER-RELATED INTERNET INFORMATION: CONCLUSIONS AND FUTURE DIRECTIONS FROM A QUALITATIVE STUDY K. R. Haase1,*, L. Holtslander1, R. Thomas2 1 University of Saskatchewan, Saskatoon, 2University of Ottawa, Ottawa, Canada

O12 HEALTH-RELATED QUALITY OF LIFE AS A PREDICTOR OF 10-YEAR MORTALITY IN OLDER PATIENTS WITH EARLY STAGE BREAST CANCER C. Dumontier1,*, R. A. Silliman2, K. Clough-Gorr3, A. E. Stuck4, A. Moser5 1 Internal Medicine Residency Program, 2Section of Geriatrics, Boston University Medical Center, Boston, United States, 3National Cancer Registry of Ireland, Cork, Ireland, 4Bern University Hospital and University of Bern, 5Institut für Sozial- und Präventivmedizin, Bern, Switzerland

Fig. 1 (abstract O10) – Spaghetti plot of individual PCS score trajectories over two cycles of CTX.

Introduction: Two in five Canadians will be diagnosed with cancer in their lifetimes, and the majority of those diagnosed will be over age 50. Whereas, all individuals with cancer face physical and psychosocial distress related to their illness, older adults may also face alterations in cognition, functional status, and age-related bias, particularly related to technology use. Across all ages, people with cancer increasingly use Cancer-Related Internet Information (CRII) to manage their patient experience. Despite widespread biases regarding the technological capabilities of older adults, recent studies suggest older adults are becoming avid users of CRII to manage their physical and psychosocial concerns when diagnosed with cancer. Objectives: The purpose of this presentation is to offer insights into older adult use of CRII, the role it plays in interactions with the healthcare system, and how healthcare professionals view older adults, and their use of CRII.

Introduction: Health-related quality of life (HRQOL) is emerging not only as an important outcome in survivorship care but also as a factor thought to influence mortality. However, prognostic models that include HRQOL alongside traditional breast cancer (BC) prognostic measures are lacking. Objectives: Our objective was to develop a 10-year mortality risk score based on selected treatment and HRQOL variables. Methods: We studied 660 women 65-years old diagnosed with stage I-IIIA primary breast cancer in years 1997–1999. Over 10 years, medical and psychosocial data were collected from interviews, medical records, and death indexes. Treatment variables included receipt of definitive locoregional surgery±radiation, chemotherapy, and tamoxifen. HRQOL

ABSTRACTS

variables included physical function [10-item Physical Function Index (PFI-10) from the Medical Outcomes Study Short Form-36 (MOS SF-36)]; mental health [5-item Mental Health Index (MHI-5) from the MOS SF-36]; and social support [8-item modified MOS Social Support Scale (mMOS-SSS)]. We used penalized logistic regression models to develop a 10year mortality risk score. We then assessed its discrimination (c-statistic) and calibration (observed versus predicted mortality using the Hosmer-Lemeshow (HL) test). Results: 230 of 660 women (34.8%) died though 10-years of follow-up. Mental health and physical function had strong independent associations with mortality in adjusted analyses (women with high MHI-5: OR 0.57, 95% CI 0.39, 0.85; women with high PFI-10: OR 0.63, 95% CI 0.43, 0.92). The c-statistic of a risk score using only age, number of comorbidities, stage of BC, and BC treatment was 0.71. The c-statistic increased to 0.74 with the addition of HRQOL measures and showed good calibration (p=0.72 from HL test). Conclusion: HRQOL is independently associated with 10year mortality in older patients with early stage breast cancer and adds predictive ability with good discrimination and calibration to age, comorbidity, stage of BC, and BC treatment. These findings suggest that interventions to improve mental health, physical function, and social support might benefit not only a patient’s HRQOL but also her survival. Disclosure of interest: None declared Keywords: Early stage breast cancer, health-related quality of life, prognosis, survivorship

O13 FATIGUE DURING ONCOLOGIC TREATMENTS IN ELDERLY VERSUS YOUNG PATIENTS: A CALL FOR SIOG RECOMMENDATIONS C. Falandry1,*, E. Carola2, A. Raynaud-Simon3, C. Cuvier4, V. Mari5, J.-M. Turpin6, P. Soubeyran7, M. Benoit8, F. Retornaz9, R. Boulahssass10 1 Geriatrics, Hospices Civils de Lyon, Lyon University, CarMEN laboratory, FHU ONCOAGE, Lyon, 2Oncology,, Groupe Hospitalier Public du Sud de l’Oise, Senlis, 3Geriatrics, Paris Nord Val de Seine University Hospitals AP-HP; Denis Diderot Faculty of Medicine, 4 Centre des maladies du sein (Sénopôle Saint-Louis), Hôpitaux universitaires Saint-Louis, Lariboisière, Fernand-Widal, Paris, 5 Oncology, Centre Antoine Lacassagne, 6Geriatrics, Hôpital Cimiez, Nice, 7Oncology, Institut Bergonié, Bordeaux, 8Psychiatrie d’Adultes et de Psychologie Médicale Pole Neurosciences Cliniques, CHU Pasteur 1, Nice, 9Gériatrie Polyvalente, Centre Gérontologique Départemental, Hôpital européen, Marseille, 10Geriatrics, Coordination Unit for Geriatric Oncology (UCOG) PACA East FHU ONCOAGE, Nice, France

Introduction: Older cancer patients lay at the crossroads of several fatigue providers. Cancer treatment-related fatigue (CTRF) represents the most common treatmentassociated side effect while cancer-related fatigue appears in advanced cancer, anaemia or respiratory tract obstruction and may be difficult to distinguish from CTRF. In the geriatric

S25

field independently of cancer, fatigue is proposed as one of the features characterizing the frailty syndrome. More generally fatigue is strongly associated with negative healthrelated events and may be the marker of the depletion of the body’s homeostatic reserves. Despite that, current NCI recommendations on fatigue management do not consider age as a risk factor for increased CTRF [1]. Objectives: To identify reports of relevant clinical studies to answer the following question: “Is there evidence that age increases the risk of CTRF?” for the majority of targeted therapies and chemotherapy regimen performed in the 7 most frequent cancer indications in the elderly. Methods: Computerised literature searches of MEDLINE, EMBASE and The Cochrane Library were performed. Since CRF is unsystematically reported and alternatively called “fatigue”, “asthenia”, ”lethargy”, ”weakness” no systematic analysis could be performed. Were included subgroup analyses as well as older patients’ specific studies. Results: Among 172 regimen analysed, data on all grade and/or grade3+ fatigue were available for 124 (72%) and differed according the tumour type and the setting. The reporting was high (>80%) in colorectal cancer, metastatic breast and prostate cancer and low (<60%) in adjuvant breast cancer, small cell lung cancer, non Hodgkin lymphoma, urothelial and ovarian cancers. A classification of the CTRF risk was performed according the published rates of grade 3+ CRF with the following thresholds: weak (<5%), moderate (5–10%), high (10–20%) and very high (20%). As rates of fatigue differed largely between trials for the same regimen, the higher scores were considered for this classification. Specific data on elderly patients, with different age thresholds, were found for only 10 regimens (6%). When compared to younger patients, the frequency and severity of fatigue was significantly increased. Conclusion: CRF is insufficiently reported in clinical trials and, when evaluated, is poorly reproducible from one trial to the other. As the advent of targeted therapies provides the opportunity to lower the intensity of adverse events, future work is needed to standardise CTRF evaluation – according for example patients’ reported outcomes questionnaires – in order to include it in the benefit/risk balance. Elderly specific data on fatigue are reported for only 6% of the regimen. Given the significant impact of fatigue on negative health-related events, SIOG recommendations should be proposed to standardize fatigue assessment and propose specific interventions. Reference: [1] Board PS and PCE. Fatigue (PDQ®). 2014 Aug 28 [cited 2016 Feb 19]; Available from: http://www.ncbi.nlm.nih.gov.gate2. inist.fr/books/NBK66049/ Disclosure of interest: C. Falandry Consultant for: TEVA, HOSPIRA, ASTELLAS, JANSSEN ONCOLOGY, E. Carola Consultant for: TEVA, A. Raynaud-Simon Consultant for: TEVA, C. Cuvier Consultant for: TEVA, V. Mari Consultant for: TEVA, J.-M. Turpin Consultant for: TEVA, P. Soubeyran Consultant for: TEVA, M. Benoit Consultant for: TEVA, F. Retornaz Consultant for: TEVA, R. Boulahssass Consultant for: TEVA Keywords: Fatigue, treatment toxicity

S26

ABSTRACTS

O14 BEVACIZUMAB + CHEMOTHERAPY VERSUS CHEMOTHERAPY ALONE IN OLDER PATIENTS WITH UNTREATED METASTATIC COLORECTAL CANCER: A RANDOMIZED PHASE II TRIAL – PRODIGE 20 STUDY RESULTS T. Aparicio1,*, O. Bouché2, E. Francois3, F. Retornaz4, E. Maillard5, J. Taieb6, S. Kirscher7, P. L. Etienne8, R. Faroux9, F. Khemissa Akouz10, F. El Hajbi11, C. Locher12, Y. Rinaldi13, T. Lecomte14, S. Lavau-Denes15, M. Baconnier16, A. Oden-Gangloff17, D. Genet18, L. Bedenne19, E. Paillaud20 and PRODIGE – GERICO 1 Gastro Enterologie, Hopital Saint Louis, Paris, 2Gastro Enterologie, CHU Robert Debré, Reims, 3Oncology, Centre Antoine Lacassagne, Nice, 4Geriatry, Hôpital Européen, Marseille, 5Statistic, FFCD, Dijon, 6Digestive oncology, HEGP, Paris, 7Oncology, Institut Sainte Catherine, Avignon, 8Oncology, Centre CARIO - HPCA, Pleven, 9 Gastro Enterologie, CH Les Oudairies, La Roche sur Yon, 10Gastro Enterologie, CH Saint Jean, Perpignan, 11Oncology, Centre Oscar Lambret, Lille, 12 Gastro Enterologie, CH de Meaux, Meaux, 13 Gastro Enterologie, Hôpital Européen, Marseille, 14 Gastro Enterologie, CHU Trousseau, Tours, 15Oncology, CHU Dupuytren, Limoges, 16 Gastro Enterologie, CH Annecy Genevois, Annecy, 17 Gastro Enterologie, CHU Charles Nicolle, Rouen, 18Oncology, Clinique François Chénieux, Limoges, 19 Gastro Enterologie, CHU Le Bocage, Dijon, 20Geriatry, CHU Henri Mondor, Créteil, France

Introduction: Metastatic colorectal cancer frequently occurs in older patients. Bevacizumab in combination with front line chemotherapy is a standard treatment but some concerns were raised about tolerance of bevacizumab in the older patients. Objectives: The purpose of PRODIGE 20 was to evaluate tolerance and efficacy of bevacizumab according to specific endpoints in this population. Methods: Patients aged 75 and over were randomly assigned to bevacizumab + chemotherapy (BEV) versus chemotherapy (CT). LV5FU2, FOLFOX and FOLFIRI regimen were prescribed according to investigator’s choice. The co-primary composite endpoint, assessed 4 months after randomization, was based on efficacy: tumor control and absence of decrease >2 points of the Spitzer QoL index and safety: absence of severe cardiovascular toxicities and unexpected hospitalization. Efficacy in more than 20% and safety in more than 40% of the patient were the lower thresholds expected at 4 months. A geriatric assessment was performed at randomization and at each evaluation. The predictive value of geriatric and oncologic factors was determined for the combined coprimary composite endpoint assessing safety and efficacy of treatment (BEV and/or CT) simultaneously and also progression-free survival (PFS) and overall survival (OS). Results: 102 pts were randomized in this trial (51 BEV and 51 CT), median age was 80 years (range 75–91). The primary endpoint was met for efficacy in 50.0% [90% CI: 37.1–62.9] and 57.8% [90% CI: 44.4–70.3] and for safety in 60.9% [90% CI: 47.7–73.0] and 71.1% [90% CI: 58.0–82.0] of patients in BEV and CT respectively. Median PFS was 9.7 months in BEV and 7.8 months in CT. Median OS was 21.7 months in BEV and 19.8 months in CT. The 36-month OS rate was 27.0% [95% CI: 15.7–39.7] in BEV and 10.1% [95% CI: 3.1–22.0] in CT. Severe

toxicities grade 3/4 were observed in 80.4% and 63.3% in BEV and CT respectively. On multivariate analysis, baseline normal independent activity of daily living (IADL) score and no previous cardiovascular disease predicted the combined co-primary endpoint. High (vs low) baseline Köhne score predicted short PFS and baseline Spitzer quality of life (QoL) score <8, albumin level 35 g/L, CA19.9 >2 ULN and high baseline Köhne score predicted short OS. Survival without deteriorated QoL and autonomy were similar with BEV and CT treatment. On subgroup analyses, the trend for a better PFS and OS with BEV was maintained in patients with baseline impaired IADL or nutritional status. Conclusion: Both arms met the primary safety and efficacy criteria. Normal IADL score was associated with a good efficacy and safety of both BEV and CT treatments. Köhne criteria may be relevant prognostic factors in older patients. Adding BEV to CT did not impair patient autonomy or QoL. Disclosure of interest: T. Aparicio Consultant for: Ipsen, HalioDX, Pierre Fabre, Speakers bureau: Roche, Sanofi, Novartis, O. Bouché Grant / Research Support from: Boehringer Ingelheim, Consultant for: Roche, Merck, Amgen, Novartis, Lilly, Bayer, E. Francois: None declared, F. Retornaz Grant / Research Support from: Gilead, Consultant for: Gilead, Teva, E. Maillard: None declared, J. Taieb Consultant for: Merck,, S. Kirscher: None declared, P. L. Etienne: None declared, R. Faroux: None declared, F. Khemissa Akouz: None declared, F. El Hajbi : None declared, C. Locher Consultant for: Amgen, Roche, Sanofi, Novartis, Ipsen, Celgene, Y. Rinaldi: None declared, T. Lecomte: None declared, S. Lavau-Denes: None declared, M. Baconnier: None declared, A. Oden-Gangloff: None declared, D. Genet: None declared, L. Bedenne: None declared, E. Paillaud: None declared Keywords: Bevacizumab, colorectal cancer, geriatric evaluation, prognostic factors, randomized trial

O15 ASSOCIATIONS OF PRE-CHEMOTHERAPY INFLAMMATION WITH POST-CHEMOTHERAPY FRAILTY IN PATIENTS WITH BREAST CANCER AGED 50+: LONGITUDINAL DATA FROM THE UNIVERSITY OF ROCHESTER NCI COMMUNITY ONCOLOGY RESEARCH PROGRAM (UR NCORP) NETWORK N. Gilmore1,*, S. Kadambi2, L. Lei3, A. Magnuson1, S. Mohile1, M. Janelsins4 1 Cancer Center, 2Internal Medicine, 3Public Health, 4Cancer Control, University of Rochester, Rochester, United States

Introduction: Frailty is an age-related syndrome characterized by weakness and fatigue that negatively affects quality of life. Previous studies have shown that chronic inflammation is a significant physiologic feature of frailty; as individuals age, there is immune dysregulation characterized by low grade inflammation (increased circulating interleukin-6 (IL-6) and soluble tumor necrosis factor receptor (sTNFR) I&II). Whether these inflammatory markers are associated with frailty in patients who receive chemotherapy remains unknown.

ABSTRACTS

Objectives: To determine if inflammatory cytokine levels pre-chemotherapy (pre-chemo) are predictive of postchemotherapy (post-chemo) frailty in patients with breast cancer. Methods: Female breast cancer patients aged 50+ scheduled to receive adjuvant/neoadjuvant chemotherapy (n=144) and age matched controls (n=142) were recruited from the UR NCORP network. Frailty was assessed by a modified frailty score (0–4) using self-reported weakness (4 on MD Anderson Symptom Inventory (SI)), exhaustion (4 on SI), physical activity (<150 minutes/week on Aerobic Center Longitudinal Study Physical Activity Questionnaire (ACLS)) and walking speed (<2mph ACLS). Fasting blood was drawn at pre-chemo and post-chemo time points and sent to UR NCORP where serum cytokines were measured via Luminex multiplex assays. Age matched controls were assessed at equivalent time points. Chi-square tests were used to compare frailty between cancer patients and controls and between prechemo and post-chemo. T-tests were used to evaluate the associations between pre-chemo cytokine level (median as cutoff) and post-chemo frailty for patients and controls. Results: Patients with stage I-IIIc breast cancer (mean age 60; range 50–76) and controls (mean age 59; range 50–81) were evaluated. Pre-chemo, a significant percentage of cancer patients were weaker (20% vs 6%, p<0.01), more exhausted (33% vs 18%, p<0.01), had lower physical activity (8% vs 3%, p=0.04) and walked slower (56% vs 40%, p<0.01) compared to controls. Comparing post-chemo to pre-chemo, a greater percentage of cancer patients were weaker (53% vs 20%, p<0.01), more exhausted (63% vs 33%, p< 0.01), had lower physical activity (16% vs 12%, p<0.05) and walked slower (77% vs 56%, p<0.01), while controls did not have significant changes over time. Overall, cancer patients were more frail pre-chemo compared to controls (with frailty scores of 1.17 vs 0.65, p<0.01) and were more frail post-chemo compared to pre-chemo (2.08 vs 1.17, p<0.01). Cancer patients with prechemo levels of IL-6, sTNFRI and sTNFRII above the median were more frail after receiving chemotherapy than those with values below the median whereas this association was not seen in controls. (2.31 vs 1.86, p=0.03; 2.30 vs 1.88, p=0.04; 2.29 vs 1.87, p=0.04). Conclusion: Cancer patients 50+ are significantly more frail than age matched controls before treatment and chemotherapy exacerbates frailty in these patients. Higher than median pre-chemo levels of studied cytokines were significantly associated with frailty post-chemo in women aged 50+ with breast cancer. These results suggest that prechemo levels of inflammatory cytokines may help serve as a biomarker to determine which cancer patients will have greater frailty after chemotherapy. Future studies need to validate and expand these findings. Disclosure of interest: None declared Keywords: Breast cancer, frailty, inflammation

S27

O16 NORDIC9: A NORDIC RANDOMIZED PHASE II TRIAL EXPLORING TREATMENT STRATEGIES OF OLDER PATIENTS WITH METASTATIC COLORECTAL CANCER; RESULTS OF A PREPLANNED SAFETY ANALYSIS S. B. Winther1,2,*, H. Sorbye3, Å. Berglund4, P. Österlund5–8, B. Glimelius4, C. Qvortrup9, P. Pfeiffer1,2 and Nordic Biomodulator Group, AgeCare 1 Department of Oncology, Odense University Hospital, 2University of Southern Denmark, Odense, Denmark, 3Department of Oncology, Haukeland University Hospital, Bergen, Norway, 4Department of Immunology, Genetics and Pathology, Uppsala University, Uppsala, Sweden, 5Department of Oncology, Tampere University Hospital, 6 Tampere University, Tampere, 7Department of Oncology, Helsinki University Hospital, 8Helsinki University, Helsinki, Finland, 9 Department of Oncology, Copenhagen University Hospital, Copenhagen, Denmark

Introduction: Colorectal cancer is a disease of the older but knowledge about treatment strategies among these patients (pts) is sparse as this population is underrepresented in clinical trials [1]. S-1 is a well-tolerated oral 5-FU prodrug [2] and a good alternative in older mCRC pts. Attention may though be focused on kidney function, as a previous study (SOFT) [3] has shown an increased incidence of grade 3–4 diarrhea in patients with a creatinine clearance <70 ml/min compared to patients with creatinine clearance 70 ml/min (21% vs. 6%). Objectives: The overall aim of the NORDIC9 trial is to add knowledge on how to select the optimal treatment for older mCRC pts, who are not candidates for standard combination therapy. Methods: Older (70 years) mCRC pts, who are not candidates to full-dose combination therapy, are randomized to: Arm A: full dose monotherapy (S-1 30 mg/m2 po bid day 1–14 q3w, followed by second line irinotecan); or Arm B: reduced dose (80%) combination therapy (S-1 20 mg/m2 po bid day 1–14 + oxaliplatin 100 mg/m2 iv day 1 q3w, followed by reduced S-1 + irinotecan). Bevacizumab (7.5 mg/kg iv) may be added at the discretion of the treating clinician. Geriatric screening tools (G8, VES-13, Timed-Up-and-Go, and Hand Grip strength), Charlson Comorbidity Index, and Quality of Life are evaluated at baseline. Blood samples and tumor tissue are prospectively collected. Primary endpoint is progression-free survival. Secondary endpoints are overall survival, response rate and correlations between the geriatric screening tools and toxicity as well as efficacy. Results: A preplanned safety analysis was performed when 50 pts had received 3 cycles. 12 pts received bevacizumab. Median age was 79 (range 70–88) years, 24 pts were female, performance status was 0 (43%), 1 (38%) or 2 (19%). Five (10%) pts discontinued therapy after only 1 cycle due to toxicity (n=2) or PD/clinical deterioration (n=3); 45 pts (90%) continued therapy beyond 3 cycles. Pts receiving only 1 cycle had numerically a worse G8 and VES-13 score. Grade 3–4 non-hematological toxicity was fatigue (6%), diarrhea (10%), nausea (4%) and vomiting (6%), and was experienced by 10 pts, 6 of them received 3 cycles of treatment. Grade 3–4 diarrhea was more frequent (16% vs. 4%) in pts with slightly reduced

S28

ABSTRACTS

kidney function (calculated GFR 70 ml/min) compared to pts with normal kidney function (calculated GFR > 70 ml/min). There was no hand-foot-syndrome, cardiac or hematological grade 3–4 toxicity. Conclusion: The safety analysis shows acceptable toxicity. 10% of the pts received only 1 cycle of treatment and these pts had worse G8 and VES-13 score. The NORDIC9 trial continues according to the original design as recommend by the safety committee. Enrollment was initiated March 2015. June 2017, 142/160 pts are included. EudraCT nr. 2014-000394-39. References: [1] Papamichael et al. Treatment of colorectal cancer in older patients: International Society of Geriatric Oncology (SIOG) consensus recommendations 2013. Ann Oncol 2015;26:463– 76 [2] Winther et al. Experience with S-1 in older Caucasian patients with metastatic colorectal cancer (mCRC): Findings from an observational chart review. Acta Oncol 2016;55:881–5 [3] Yamada et al. Leucovorin, fluorouracil, and oxaliplatin plus bevacizumab versus S-1 and oxaliplatin plus bevacizumab in patients with metastatic colorectal cancer (SOFT): an open-label, non-inferiority, randomised phase 3 trial. Lancet Oncol 2013;14:1278–86 Disclosure of interest: S. Winther Grant / Research Support from: Taiho, Nordic Drugs, H. Sorbye: None declared, Å. Berglund: None declared, P. Österlund: None declared, B. Glimelius Consultant for: PledPharma AB, C. Qvortrup: None declared, P. Pfeiffer Grant / Research Support from: Taiho, Nordic Drugs, Consultant for: Taiho, Nordic Drugs Keywords: Geriatric screening tools, metastatic colorectal cancer, older, S-1 (Teysuno)

O17 TOWARD PERSONALIZING CARE FOR OLDER MEN WITH MCRPC (TOPCOP): THE EFFECT ON PHYSICAL FUNCTION FROM TREATMENT FOR MCRPC H. Breunis1,*, U. Emmenegger2, S. Berry2, S. Hotte3, A. Hansen1, A. Joshua1, P. Warde4, N. Fleshner5, G. Tomlinson6, N. Timilshina6, S. Cheng7, K. Akilan1, S. Garuba1, S. Khan1, D. Samaroo1, O. Samadi1, S. Baig1, J. Tavalis1, M. Rathore1, M. Klowak1, S. Alibhai1 1 Medicine, UHN, 2Medical Oncology, Sunnybrook Health Sciences Centre, Toronto, 3Medical Oncology, Juravinski Cancer Centre, Hamilton, 4Radiation Medicine Program, 5Surgical Oncology, 6UHN, 7 Medicine, Sunnybrook Health Sciences Centre, Toronto, Canada

Introduction: The physical effects of treatments in metastatic castrate-resistant prostate cancer (mCRPC) in older men, who are already suffering musculoskeletal toxicities of long-term androgen deprivation, are not well known. Objectives: To evaluate the effects of treatment with chemotherapy (CHEMO), abiraterone (ABI), or enzalutamide (ENZA) on physical function in men with mCRPC.

Methods: Men age 65 or older were enrolled at two participating academic centres, Princess Margaret Cancer Centre and Sunnybrook Health Sciences Centre in Toronto, Canada. Participants were enrolled in three cohorts in this observational study: mCRPC patients starting CHEMO, ABI, or ENZA. Physical Function (PF) was evaluated objectively with grip strength (using a Jamar dynamometer) and the Short Physical Performance Battery (SPPB, which includes gait speed in 4 metres and 5 timed chair stands). PF tests were administered at baseline and final visit in the 3 cohorts. In the CHEMO cohort PF tests were also administered at alternate cycles, in the ABI and ENZA every 3 to 4 months during treatment. Changes in physical performance over time were analyzed using linear mixed effects regression. The primary comparison time point was at 6 months across all cohorts. Results: To date 68 participants have been enrolled (CHEMO n=30; ABI n=11, and ENZA n=27) of whom 24 have completed or discontinued treatment. PF data were collected for 24 participants who completed 6 months of treatment or had a final visit and 44 who had PF administered at 6 months on treatment. The mean age of participants receiving CHEMO was 72 years (range 64–90); ABI 77 years (range 64–84), and ENZA 75 years (range 65–90). The median VES-13 score at baseline in the CHEMO cohort was 2 (IQR 1–7), 0 (IQR 0–1) in the ABI cohort, and 1 (IQR 0–3) in the ENZA cohort. Reasons for discontinuation in the CHEMO cohort were completion of 9 to 10 cycles (n=3), disease progression (n=8), and toxicity or intolerability (n=7). In the ENZA cohort discontinuation was due to disease progression (n=5). Grip strength measured at final or 6 months visit declined in the CHEMO cohort by 2.9% (–8.8%, 1.9%), increased by 0.8% (–5.9%, 4.7%) in the ABI cohort, and declined 10.7% (–17.3%, –1.6%) in the ENZA cohort (p=0.11). Timed chair stands increased 1.2% (–14.9%, 42.6%) in the CHEMO cohort, increased 4.1% (–17.5%, 10.6%) in the ABI cohort, and declined 5.5% (–21.8%, 10.6%) in the ENZA cohort (p=0.53). Gait speed declined 30% (–36.3%, 6.1%) in the CHEMO cohort, 0% (–6.9%, 14.6%) in the ABI cohort, and by 2.6% (–12.7%, 22.3%) in the ENZA cohort (p=0.031). Conclusion: Although our sample size is limited, Physical Function (particularly grip strength and gait speed) appeared to decline with both CHEMO and ENZA but not with ABI. Gait speed was most affected by CHEMO whereas grip strength was most affected by ENZA. A larger sample is needed to confirm these findings. Study accrual and follow-up continue. Disclosure of interest: None declared Keywords: Chemotherapy, gait speed, metastatic castrate resistant prostate cancer, older adults, physical function

O18 FACTORS ASSOCIATED WITH THE INFLAMMATORY RESPONSE TO SURGERY IN ELDERLY ONCOLOGICAL PATIENTS M. Plas1,*, H. van der Wal-Huisman1, J. J. de Haan2, A. Rutgers3, A. R. Absalom4, G. H. de Bock5, B. L. van Leeuwen1 1 Surgical Oncology, 2Medical Oncology, 3Immunology, 4 Anesthesiology, 5Epidemiology, University Medical Center Groningen, Groningen, Netherlands

ABSTRACTS

Introduction: The inflammatory response to surgery is a complex mechanism and influenced by patient and disease characteristics, as well as by characteristics related to the surgical procedure itself. Although intended to be protective, an inflammatory response can cause collateral tissue damage and lead to pathology. Especially in the elderly patient this can lead to functional loss. Objectives: To identify pre- and peroperative factors associated with the extent of the inflammatory response following surgery in the elderly. Methods: Patients of 65 years and older undergoing a surgical procedure for a solid malignant tumour were prospectively included in an observational cohort study. Inflammatory factors were measured in blood serum samples pre- and postoperatively: C-reactive protein (CRP), Interleukin (IL)-1b, IL-6, IL-10, (IL-12, and tumour necrosis factor-alpha (TNF-). To measure the surgery-evoked inflammatory response, preoperative results of the inflammatory factor assay, were compared with the postoperative results. Results: Between July 2010 and April 2014, blood serum samples of 224 patients were obtained. One-hundred-eight patients (48.2%) were male, median age of patients 72 (65–89) years. The predominant diagnosis was carcinoma, 157 (70.1%). Patient or tumour characteristics were not associated with baseline serum concentrations of the inflammatory markers. Multivariate analysis identified anaesthesia duration and blood loss as independent predictors for the peroperative change in CRP ((B: 0.026; 95% 0.01–0.04)/(B: –0.013; 95% –0.016 to –0.01)) and IL-1b ((B: 0.002; 95% 0.0–0.004)/(B: –0.001; 95% –0.002 to –0.001)). Furthermore, the change in IL-6 identified anaesthesia duration (B: 1.21; 95% 0.9–1.5), intracavitary surgery (B: 116.60; 95% 14.9–218.3) and blood loss (B: 0.08; 95% 0.0–0.1) as predictors. For the change in Il-10, intracavitary surgery (B: 63.40; 95% 9.5–117.3) and major surgery (B: 126.15; 95% 75.8–176.5) were identified. Conclusion: The inflammatory response following surgery is influenced by surgical characteristics rather than by preoperative factors, including disease stage, neo-adjuvant treatment and comorbidities. Patients undergoing longer surgical procedures, intracavitary surgery or with more blood loss, show the greatest inflammatory response to surgery. Disclosure of interest: None declared Keywords: Inflammatory response, surgery

O19 RADIOTHERAPY FOR ELDERLY PATIENTS WITH GLIOBLASTOMA MULTIFORME, PATIENT OUTCOMES IN ROUTINE CLINICAL PRACTICE E. Smith1,*, S. Iqbal1, E. Lethbridge1, S. Lawless2, N. Wadd2, J. Lewis1 1 Clinical Oncology, Northern Centre for Cancer care, Newcastle, 2 Clinical Oncology, James Cook Hospital, Middlesbrough, United Kingdom

Introduction: Clinical trials evaluating management of Glioblastoma multiforme (GBM) commonly exclude elderly

S29

patients. We reviewed patients 70 or older, with a diagnosis of GBM, who were treated with radiotherapy. Objectives: The aim was to compare our routine clinical practice outcomes for elderly patients with a diagnosis of GBM compared to that in clinical trials. We compared our outcomes with the Nordic randomised phase 3 trial in elderly patients with GBM [1]. Methods: Northern Centre for Cancer Care and James Cook University Hospital are the two cancer centres for patients with brain tumours in the North East of England, catering for a population of 3 million people. Patients were identified through an electronic referral for radiotherapy. Data from electronic and medical records was obtained Results: 87 patients were treated between January 2009 and June 2014. Median age was 74 (range 70–85). 71% had palliative radiotherapy with 30Gy in six fractions being the commonest schedule. 24% of palliative patients had a radiological diagnosis only. 9/61 patients did not complete palliative radiotherapy due to clinical deterioration or disease progression. 3 palliative patients had further treatment on disease progression, one patient had further surgery and 2 patients had palliative temozolamide. 29% (25 patients) were treated with 60Gy in 30 fractions, 7 patients had concurrent temozolamide. All radical treated patients had a histological diagnosis and were performance status 0 or 1. All patients treated with radical intent completed the radiotherapy course as planned. Average time from diagnosis to start of radiotherapy was 44 days (range 28–67). Median survival for palliative radiotherapy and radical radiotherapy was 6 months and 17 respectively. In the Nordic study median overall survival was 7 months, for hypofractionated radiotherapy, and 5.2 months for standard radiotherapy [1]. Conclusion: Patient selection is the critical factor in achieving survival benefit in elderly patients. All patients treated were good performance status 0-2 in concordance with trial criteria. Our survival outcome for patients with palliative treatment is similar to that in the trial population although 14% did not complete the prescribed treatment. Our review shows that elderly patients can be treated with radical intent with good survival however numbers are small. Reference: [1] Annika et al. Temozolomide versus standard 6-week radiotherapy versus hypofractionated radiotherapy in patients older than 60 years with glioblastoma: the Nordic randomised, phase 3 trial. Lancet 13;9:916–926. Disclosure of interest: None declared Keywords: Brain tumour, radiotherapy

O20 INCLUDING PATIENT PREFERENCES AND GERIATRIC ASSESSMENT IN THE EVERY DAY DECISION MAKING PROCESS S. Festen1,*, M. Kok2, A. K. Reyners2, M. Beernink3, L. Olsder3, H. van der Wal3, A. H. van der Leest4, P. De Graeff5, B. L. Van Leeuwen3 1 Department of Internal Medicine, University Center of Geriatric Medicine, 2Department of medical oncology, 3Department of Surgery,

S30

ABSTRACTS

4

Department of Radiotherapy, University Medical Center Groningen, Department of Internal Medicine, University Center of Geriatric Medicine, UMCG, Groningen, Netherlands

5

Introduction: There is increased awareness of the added value of geriatric assessment (GA) to the treatment decision process in older cancer patients. In addition, knowledge of personal goals and preferences can guide treatment choices and enhance patient centered care. The incorporation of this information into multidisciplinary decision making is not established. Objectives: To incorporate assessment of geriatric domains and patient preferences into a structured decision making process for older cancer patients. Methods: This multidisciplinary collaboration project was conducted at the University Medical Center Groningen, the Netherlands and funded by the Dutch Cancer Society. Patients of 70 years or older with a solid malignancy were included and interviewed by trained oncology nurses, during their visit to either the surgical or medical oncology outpatient clinic. During this interview a short GA was performed. In addition, the patients’ treatment goals and preferences were assessed using an option tool. Treatment goals offered were: prolonging life, maintaining independence, reducing/ eliminating pain, or other symptoms. These goals were rated by the patient in order of importance. In addition to the tumor specific multidisciplinary team meetings (MDT), patients were discussed in an onco-geriatric MDT attended by a surgeon, radiotherapist, medical oncologist, geriatrician, and nurses. All patients were discussed using information concerning disease characteristics, comorbidities, GA outcomes and patient preferences. This was done in a systematic, stepwise fashion. The nurses presented the patient’s narrative and results of GA. Benefits and risks of different treatment options based on both medical technical and patient specific information, were discussed (Table 1). Time to benefit was related to the estimated life expectancy and patients’ preferences were taken into account. This was done by translating treatment effects into what this would mean to a patient on a day to day level. After carefully balancing the expected risks and benefits of different treatment options, a treatment proposal was formulated. This proposal was compared to the earlier treatment advice given by the tumor specific MDT. Table 1 (abstract 020) – Stepwise approach to multidisciplinary decision making 1. Indication for treatment and alternative treatment options 2. Presence of physical complaints likely to be resolved by treatment 3. Estimated risk of complications 4. Life expectancy without the current disease 5. Life expectancy with the disease with and without treatment 6. Patients goals, preferences and expectations 7. Effect of treatment on these goals and preferences

Results: From September 2014 until November 2016, 250 patients were included. The mean age was 78 years. Ninety percent lived independently. Fifty-two percent were ADL dependent and 54% considered frail. Treatment goals

considered most important by patients were; prolonging life (31%) and maintaining independence (29%). Compared to the treatment advice given by the tumor specific MDT an adjusted treatment advice was given by the oncogeriatric MDT in 48% of cases. Thirteen percent of the patients were additionally assessed in the geriatrics outpatient clinic. Conclusion: In this patient centered method, both medical technical and patient specific information was discussed in a standardized, stepwise manner, as described, in an oncogeriatric MDT. As a result an adjusted treatment advice was formulated in 48% of cases. In only 13% additional assessment by a geriatrician proved necessary. Disclosure of interest: None declared Keywords: Decision making, elderly oncological patients, geriatric assessment, nurses, patient preferences

O21 OPENING UP THE BLACK BOX OF CGA IN GERIATRIC ONCOLOGY CLINIC – UNDERSTANDING ENHANCEMENTS TO CARE BEYOND TREATMENT-DECISION MAKING S. M. Alibhai1,*, R. Jin2, A. Loucks2, D. Yokom1, S. Watt3, N. Timilshina1, M. T. Puts4, A. Berger1 1 Medicine, 2Nursing, 3Supportive Care, University Health Network, 4 Nursing, University of Toronto, Toronto, Canada

Introduction: Comprehensive geriatric assessment (CGA) in older adults with cancer aids treatment decision-making, determining frailty status, and prognostication. Much less is known about the supportive care elements or enhancements to care afforded by the CGA, and whether these enhancements vary by the indication for patient referral (e.g. pre-treatment, on active treatment, etc.). Objectives: To examine enhancements to care provided by a geriatric oncology clinic and to determine whether these vary by indication for referral. Methods: All patients age 65 or older referred to the Older Adults with Cancer Clinic at the Princess Margaret Cancer Centre, Toronto, Canada between July 2015 (clinic opening) and December 2016 were included. Treatment enhancements were recorded prospectively in our clinical database and categorized in 5 categories (educational support, comorbidity management, disease/symptom management, oncologic treatment delivery, peri-op management recommendations) based on a consensus among team members. Indications for referral were categorized into 3 groups: pre-treatment (n=51, 38%), on active treatment (n=70, 51%), survivorship phase (n=15, 11%). Descriptive statistics and logistic regression (adjusted for age and gender) were used to analyze the data. Results: 136 patients were seen during the study period (mean age 79.5 years, 75% male). Overall, educational support (99%) and comorbidity management (94%) were the most common enhancements, whereas peri-op management (7%) was the least common. When stratified by indication for referral, unsurprisingly peri-op management was only offered to pre-treatment patients. Enhancements to disease/ symptom management were offered more often to patients

ABSTRACTS

on active treatment than pre-treatment (54% versus 29%, odds ratio 3.14, p=0.005). Other enhancements to care did not vary by indication for referral. Table 1 (abstract O21) – Frequency of each enhancement to care by indication of referral

Intervention type

Active PreTreatment treatment (n=70) (n=51)

Others (n=15)

p-value

Enhanced treatment delivery, n (%)

24 (34.3)

18 (35.3)

1 (6.7)

0.08

Comorbidity mgmt, n (%)

68 (97.1)

45 (88.2)

15 (100.0)

0.10

Educ. support, n (%)

69 (98.5)

50 (98.1)

15 (100.0)

0.86

Disease/Symptom mgmt, n (%)

38 (54.3)

15 (29.4)

3 (20.0)

0.005

0

9 (17.6)

0

0.002

Peri-op mgmt, n (%)

Conclusion: Educational support and comorbidity management are nearly universally offered by our geriatric oncology clinic. Most enhancements to care do not vary by indication for referral. Enhancements to peri-op management were provided in almost 1 in 5 pre-treatment patients. Enhancements to disease/symptom management were more common in patients on active treatment than pre-treatment patients, likely due to a combination of greater symptom burden and treatment-related toxicities. Understanding the enhancements to care provided by geriatric oncology clinics can help with resource planning and program design/ justification. Disclosure of interest: None declared Keywords: Comorbidity, comprehensive geriatric assessment, outpatient, patient education, symptom management

O22 IMPACT OF A SPECIFIC MULTIDISCIPLINARY GERIATRIC ONCOLOGY CLINICAL PATHWAY ON THE NUMBER AND TYPES OF IMPLEMENTED GERIATRIC INTERVENTIONS R. Moor1,*, P. Cornette2, V. Verschaeve3, G. Debugne4, A. van Maanen5, Y. Humblet6, I. Gilard2, I. Clement Corral7, P. Betomvuko7, R. Poletto8, N. Nols9, F. Blancke10, F. Cornelis6 1 Geriatric Oncology, 2Geriatrics, Cliniques Universitaires Saint Luc, Bruxelles, 3Oncology, Grand Hopital de Charleroi, Charleroi, 4 Geriatrics, Centre Hospitalier de Mouscron, Mouscron, 5Statistics, 6 Oncology, Cliniques Universitaires Saint Luc, Bruxelles, 7Geriatrics, 8 Geriatric Oncology, Grand Hopital de Charleroi, Charleroi, 9 Oncology, 10Geriatric Oncology, Centre Hospitalier de Mouscron, Mouscron, Belgium

Introduction: Geriatric interventions may improve several aspects of older cancer patient’s condition. Therefore we created a multidisciplinary geriatric oncology (GO) clinical pathway with active implementation of geriatric interventions. Objectives: The aim of this study was to assess the impact of a specific multidisciplinary GO clinical pathway on the number and types of implemented geriatric interventions.

S31

Methods: Multicenter (N=3) observational study including patients 70 years with cancer for whom anticancer therapy was considered. Patients with a G8 score 14 underwent a multidimensional Geriatric Assessment (GA). Each patient with GA was discussed at a weekly Multidisciplinary Geriatric Oncology Meeting (MGOM) gathering at least a geriatrician, an oncologist and a GO care coordinator. A personalized GO care plan including opinion on the proposed anticancer therapy and recommended geriatric interventions was addressed to treating physicians. Implementation of the geriatric interventions was actively coordinated by the GO care coordinator: for each patient, the appropriate healthcare workers were personally contacted and appointments were made. About three months after MGOM, a follow-up was performed and the implementation of the geriatric interventions was checked. At each step of this clinical pathway data were prospectively collected and statistical analysis was performed. Results: From March 2013 until February 2015, 1310 patients were screened. G8 score was 14 in 901 (68.8%) of them. Among these patients mean age was 81.4±6.0 years, 56.7% were female, 88.6% lived at home and 54.8% had professional help/care at home at baseline. Cancer was newly diagnosed in 83.1% and stage-IV disease in 35.5% of patients. Chemotherapy was the most frequently proposed treatment (43.8% of patients). Patients took on average 6.4±3.5 different drugs per day. Mean ADL- and IADL-scores were 9.2±4.5 (Katz scale, /24) and 4.6±2.6 (Lawton scale, /8) respectively. Two-fifths (39.2%) of patients experienced at least one fall during the past year. Mean Timed Up and Go was 16.6±9.4 seconds. Mean pain- and fatigue-scores (both measured by visual analogical scale, /10) were 3.0±3.2 and 4.5±3.2 respectively. MMSE and GDS-15 scores were abnormal in 18.2% and 28.1% of patients respectively. Malnutrition or risk of malnutrition was present in 84.4% of patients. Mean ZBI-12 score was 9.7±7.7. Geriatric problems were detected in 96.1% of patients by GA. At baseline, 4.1±1.9 geriatric interventions on average per patient were recommended. At follow-up, 86.2% of patients had at least one suggested intervention implemented. The mean number of implemented geriatric interventions per patient at follow-up time point was 3.5±1.7. Most frequent implemented geriatric interventions were: nutritional interventions in 79.6% of patients (incl nutritional supplements in 62.8%), functional support in 40.4% of patients (incl physiotherapy sessions in 22.9% and nursing care in 15.2%), medical treatment adaptation in 28.4% of patients, psychological support in 27.9% of patients, fatigue monitoring in 21.2% of patients and pain management in 20.4% of patients. Conclusion: In a population of old and frail cancer patients, a multidisciplinary GO clinical pathway with active implementation of GA-based geriatric interventions is feasible and leads to a high level of recommended and implemented interventions. The most frequent implemented geriatric interventions were nutritional and functional supports. Disclosure of interest: None declared Keywords: Geriatric assessment, geriatric interventions, geriatric oncology clinical pathway, multidisciplinarity

S32

ABSTRACTS

O23 THE RELATIONSHIPS BETWEEN GERIATRIC IMPAIRMENTS, SYMPTOM SEVERITY, AND QUALITY OF LIFE IN 342 OLDER ADULTS WITH ADVANCED CANCER: A UNIVERSITY OF ROCHESTER NCI COMMUNITY ONCOLOGY RESEARCH PROGRAM (UR NCORP) MULTISITE TRIAL S. Mohile1,*, E. Culakova2, H. Xu3, K. P. Loh1, M. Wells2, J. Giguere4, A. Conlin5, E. Dib6, P. Duberstein7, W. Dale8, M. Flannery9 1 Medicine, 2Surgery, 3Public Health Sciences, University of Rochester Medical Center, Rochester, 4Oncology, NCORP of the Carolinas, Greenville, South Carolina, 5Oncology, Pacific Cancer Research Consortium, Portland, Oregon, 6Oncology, Michigan Cancer Research Consortium NCORP, Ypsilanti, MI, 7Psychiatry, University of Rochester, Rochester, 8Supportive Care Medicine, City of Hope, Duarte, California, USA, 9School of Nursing, University of Rochester, Rochester

Introduction: Quality of life (QoL) is a critical outcome for older adults with advanced cancer. QoL is a multidimensional construct that includes physical, functional, psychological, and social domains. Objectives: To determine if symptom burden explains variability in QoL, above and beyond impairment in geriatric assessment (GA) measures for older adults with advanced cancer. Methods: Baseline data from a cluster RCT conducted by the UR NCORP at 68 oncology practices were analyzed. Inclusion criteria were: 70 years old, advanced solid tumor diagnosis, impairment in at least 1 GA domain (e.g., function, cognition). Multiple reliable and valid self-report measures and objective assessments for each GA domain were completed. The relationships between QoL domains (Functional Assessment Cancer Therapy (FACT) subscale scores), impairment in GA measures (using established cut points for validated tools), and symptoms [MD Anderson Symptom Inventory (MDASI)], were analyzed controlling for demographic and disease characteristics. Variables were carefully examined and selected to minimize collinearity. The 4 dependent variables were FACT sub scale scores. The independent variables were entered in 3 hierarchical steps: Step 1) gender, race, income, marital status, cancer type (Gastrointestinal, Lung, or other), receiving chemotherapy (Yes, No); Step 2) Impairments in GA measures (Polypharmacy, Mini Cog, Weight Loss, Timed Up and Go, Instrumental Activities of Daily Living (IADL) Comorbidity, Older Adult Resource Survey (OARS) Medical Social Support, Generalized Anxiety Disorder (GAD)-7 added; Step 3) Symptoms (MDASI total of 13 severity scores) added. The 3 hierarchical steps were conducted for each of the 4 subscales of FACT. Results: 342 patients completed the baseline assessment (mean age 77 years (SD 5.39), 43% female, 90% white, 48% income <$50,000 US, 72% partnered, 30% GI cancer, 25% lung cancer, 68% receiving chemotherapy). The mean MDASI symptom severity score was 29.68 (SD 21.56) (range 0–111 of possible range 0–130). The mean total FACT score was 80.74 (SD 14.73) (range 35–108 of possible range 0–108). The pattern of results was similar for all 4 domains of QoL with demographic and disease characteristics accounting for only

2-5% of the variance, impairments in GA measures explaining more variance, and symptoms explaining additional unique variance. Table 1 illustrates that explanation for variance (R2) increases with each step. Table 1 (abstract O23)

Predictors

Physical well being

Functional Social Emotional well being well being well being

Step1 Demographics and Disease

P=0.046 R2=0.042

P=0.009 R2=0.054

P =0.314 R2=0.024

P=0.134 R2=0.033

Step2 GA Impairments

P<0.001 R2=0.233

P<0.001 R2=0.291

P <0.001 R2=0.106

P<0.001 R2=0.174

Step3 Symptom Severity

P<0.001 R2=0.462

P<0.001 R2 =0.378

P <0.001 R2 =0.163

P<0.001 R2=0.248

Conclusion: In this vulnerable older adult population with advanced solid tumors there was a wide range in QoL scores. Both impairments in GA measures and symptom severity explained unique differences in all FACT sub scores. The clinical implication is that both GA and comprehensive symptom assessment are necessary to understand differences in quality of life for this population. Disclosure of interest: None declared Keywords: Geriatric assessment, quality of life, symptom burden

O24 THE IMPACT OF GERIATRIC ASSESSMENT MEASURES ON CAREGIVER EMOTIONAL HEALTH: DATA FROM A MULTICENTER GERIATRIC ASSESSMENT (GA) STUDY IN THE UNIVERSITY OF ROCHESTER NCI COMMUNITY ONCOLOGY RESEARCH PROGRAM (UR NCORP) NETWORK S. Mohile1,*, H. Xu1, E. Culakova1, B. Canin1, F. Marie1, N. Gilmore1, C. Kamen1, A. Hurria2, W. Dale2, C. Heckler1, N. Vogelzang3, M. Thomas4, E. Dib5, M. Karen1, P. Duberstein1 1 university of Rochester Medical Center, Rochester, 2City of Hope, Duarte, 3Nevada Cancer Research Foundation, Nevada, 4Southeast Clinical Research Consortium, South Carolina, 5Michigan Cancer Research Consortium, Ann Arbor, United States

Introduction: Depression is common in caregivers, and patient and caregiver factors influence caregiver depression. However, little is known about the associations of patient health status with caregiver emotional health. Impairments in GA measures of older patients with cancer may influence caregiver anxiety and depression. Objectives: To examine the associations between patient and caregiver factors with caregiver anxiety and depression. Methods: This is an analysis of baseline data from a GA intervention study conducted at 68 individual oncology practices in UR NCORP. Pts aged 70 with one or more GA impairment and with an advanced, incurable solid tumor cancers/lymphoma were enrolled; patients could enroll with one caregiver. Relationships between the patient’s baseline impairments in GA measures (using 12 validated

ABSTRACTS

tools) with caregiver anxiety (5 on Generalized Anxiety Disorder-7, GAD-7) and caregiver depression (2 on Patient Health Questionnaire-2, PHQ-2) were assessed in separate multivariable models adjusted for significant caregiver (cg) and patient (pt) characteristics (e.g., cg age, cg sex, cg education, cg income, cg comorbidity (3 or 1 that significantly affects quality of life), pt cancer type, pt treatment status, pt distress). In addition, separate multivariable models were used to investigate the independent association of number of GA impairments and impairments in individual GA measures with outcomes. Results: Among 349 pts (mean age 77, range 70–96), >50% had 6 impaired GA measures. Of the 349 caregivers (mean age 66, range 26–92), 26% screened positively for anxiety and 19% for depression. In bivariate analyses, number of impaired GA measures, patient anxiety (5 on GAD-7) and patient depression (5 on Geriatric Depression Scale, GDS) were associated caregiver anxiety at p<0.10. In bivariate analyses, number of impaired GA measures, impaired patient nutritional status (significant weight loss, low body mass index, or impaired Mini Nutritional Assessment screening score), patient assistance with activities of daily living (ADLs) and instrumental ADLs, patient falls, patient comorbidity, and patient depression were associated with caregiver depression at p<0.10. In multivariate analyses, caregiver comorbidity (adjusted odds ratio (AOR) 2.67; 95% Confidence Interval (CI): 1.49–4.78, p=0.001) and patient distress (AOR 1.93; 95% CI: 1.09–3.41, p=0.025) were associated with caregiver anxiety. The number of impaired GA measures (AOR 1.19; 95% CI: 1.04– 1.38, p=0.015) and caregiver comorbidity (AOR 2.37; 95% CI: 1.26–4.47, p=0.007) were associated with caregiver depression. In a separate model investigating individual GA measures, impaired patient ADL (AOR 2.17; 95%CI: 1.16–4.06, p=0.015), impaired patient nutritional status (AOR 2.17; 95%CI: 1.10– 4.26, p=0.025), and caregiver comorbidity (AOR 2.53; 95%CI: 1.34–4.77, p=0.004) were associated with caregiver depression. Conclusion: A high proportion of caregivers of older patients with cancer report anxiety and depression. Interventions to improve caregiver emotional health should include attention to caregiver comorbidity and patient distress, function, and nutritional status. Disclosure of interest: None declared Keywords: Anxiety, caregiver, depression, distress, geriatric assessment

O25 UNMET SOCIAL SUPPORT NEEDS AMONG OLDER ADULTS WITH CANCER G. R. Williams1,*, M. Pisu1, G. Rocque1, C. Williams1, R. Taylor1, E. Kvale1, E. Patridge1, S. Bhatia1, K. Kenzik1 1 University of Alabama at Birmingham, Birmingham, United States

Introduction: Cancer is a disease of aging, disproportionately affecting older adults. Adequate social support for older adults is necessary across the cancer continuum to maintain quality of life and reduce mortality and morbidity. However,

S33

little is known about the social support needs of older adults with cancer. Objectives: Our objective was to examine the social support needs, specifically the unmet needs, of older adults with cancer in order to better target support interventions. Methods: Medicare beneficiaries (age 65) with cancer were identified from the University of Alabama at Birmingham Health System Cancer Community Network (CCN) from 12 cancer centers across 5 states in the southeastern United States. Participants underwent a telephone-based survey assessing social support needs using a modified version of the Medical Outcomes Study Social Support Survey. Twentythree social support items were assessed across five subdomains including physical (5 items), informational (3 items), emotional (5 items), practical (6 items), and medical support (4 items). Participants were asked if they had support available for each item (all, most, some, a little, or none of the time) and whether they required support for that support item (yes/no). We defined an “unmet need” if participants reported some/a little/none availability to the first question, and yes to the second question. If there was any unmet need within a domain, then the domain was counted as having “unmet needs”. We assessed the frequencies of unmet needs across domains, examined the association with patient characteristics as well self-reported symptoms using chisquare statistic and estimated relative risk (RR) and 95% confidence intervals (CI) using modified Poisson regression. Results: Of the 1,460 respondents, the average age was 74 (IQR 70–78) and 40% female. The most common cancers included breast (24%), prostate (13%), gastrointestinal (12%), and lung cancers (12%) and predominately consisted of early stage malignancies (stage 0/I [41%], stage II/III [41%]). Respondents ranged from within 12 months from diagnosis (32%), 12–36 months (29%), and more than 36 months (39%) with 28% on active treatment. The majority of patients had at least one unmet need (92%) with the greatest proportion of unmet needs in the medical (39%) and informational (36%) sub-domains. Multivariable analyses showed minority patients were at greater risk for unmet emotional (RR 1.3, 95% CI 1.1, 1.7), informational (RR 1.5, 95% CI 1.1, 1.9), medical (RR 1.4, 95% CI 1.1, 1.8), and practical (RR 1.9, 95% CI 1.3, 2.7) needs (all p<0.05). Patients who were divorced/never married had greater unmet physical (RR 1.5, 95% CI 1.2, 1.8), emotional (RR 1.4, 95% CI 1.1, 1.8), and practical (RR 2.2, 95% CI 1.5, 3.3) needs (all p<0.01). Patients reporting more symptoms had greater risk of unmet physical needs (RR 1.1, 95% CI 1.0,1.1) and those with lower income were also at risk of reporting unmet informational needs (RR 1.5, 95% CI 1.1,1.9) (all p<0.01). No significant differences were identified by age at diagnosis, gender, comorbidity score, education, time since diagnosis, or active treatment status. Conclusion: In this population of older adults with cancer we found high levels of unmet social support needs, particularly in the physical and emotional support sub-domains with minority and divorced/never married respondents at greatest risk for unmet needs. Social support interventions are needed across the cancer continuum in older adults that target minority patients and those without a spouse or partner. Disclosure of interest: None declared

S34

ABSTRACTS

Keywords: Aging, geriatric oncology, social support

Results: The study successfully completed accrual during the funding period. Over 32 months, 547 patients were accrued to the study. Peak enrolment was 34 patients/month. As of March 15, 2017, 1438 [MM1] patients were identified as being approached about the study as per screening log review. Of these, 472 (33%) were deemed ineligible by research staff; 15 (1%) declined and were ineligible; 389 (27%) declined prior to consent; 215 (15%) had missing data. After pre-screening, 349 (24%) patients enrolled. However, 59 (4%) patients consented but ultimately did not proceed on study. Of these, 23 (39%) patients were SF. Six SF patients (26%) did not have 1 GA impairment beyond polypharmacy. However, more non-enrolled patients were SW (N=36, 61%), mostly due to supervening health problems such as being too ill, going on hospice, or having died (N=17; 47%). Of the SF/SW patients, 37 (63%) had some baseline information available; 92% completed 1 GA domain measure. Among demographic/ clinical characteristics, the only significant difference was in cancer types between the 2 groups (p<0.05). However, among baseline GA measures, a higher mean number of impairments was seen in the enrolled group (4.50 vs. 3.78, p<0.05). Moreover, among specific GA measures, there were more patients with nutritional and physical performance impairment in the enrolled group (See Table). Conclusion: Our study showed that it was feasible to enroll vulnerable older adults onto a cancer clinical trial by using a GA as part of routine study screening. Additional studies analyzing enrollment facilitators/barriers will better inform the design and conduct of future cancer studies targeting older adults. Disclosure of interest: None declared

O26 FEASIBILITY OF ENROLLING VULNERABLE OLDER ADULTS WITH CANCER IN A GERIATRIC ASSESSMENT MODEL OF CARE CLUSTER RANDOMIZED TRIAL R. Maggiore1,*, M. Wells1, H. Xu1, S. Mohile1 1 University of Rochester, Rochester, United States

Introduction: Older adults represent the majority of patients with cancer, yet they are under-represented in cancer clinical trials. COACH, the Improving Communication in Older Cancer Patients and Their Caregivers study (NCT02107443), compared whether a geriatric assessment (GA)-based model of care vs. usual care for vulnerable older adults with cancer will improve patient and caregiver satisfaction regarding communication about aging-related issues. This nationwide cluster randomized trial funded by the Patient-Centered Outcome Research Institute (PCORI) included 68 communitybased oncology practices affiliated with the University of Rochester NCI Community Oncology Research Program (NCORP). Patients were age 70 years and had 1 baseline GA impairment excluding polypharmacy. Objectives: To assess the feasibility of enrolling vulnerable older patients onto this trial. Methods: This study is a descriptive analysis of procedures and metrics related enrollment onto a recently completed multi-center, community-based clinical trial. We reviewed screening logs from the study opening April 3, 2014 through March 15, 2017, which was the last data extraction for interim analyses. Screening log information reflects patients who were approached prior to informed consent. Our analysis focused on patient retention after informed consent. After consent, patients who chose not to remain in the study due to volitional reasons or extenuating circumstances were labeled “screening withdrawal” (SW). On the other hand, patients not meeting study inclusion criteria were labeled “screening failure” (SF). We also examined the differences in demographic, clinical, and baseline GA measures between enrolled and SF/SW patients.

Keywords: Clinical trial, enrollment, feasibility, geriatric assessment, study design

Table 1 (abstract O26) – Comparison of geriatric assessment impairments between enrolled vs. non-enrolled study patients Geriatric Assessment Domain Impaired Cognitive Function Study Enrollment Status

N

N (%)

P

Functional Status N (%)

P

Mental Health N (%)

Comorbidity P

N (%)

P

Social Support N (%)

P

Nutritional Status N (%)

P

Physical Performance N (%)

P

Polypharmacy N (%)

P

Enrolled

349 123 (35) 0.73 208 (60) 0.28 90 (26) 0.12 218 (64) 0.45 82 (23)

SF/SW*

34

Enrolled

349

223 (64) <0.05

SF/SW

35

12 (34)

Enrolled

349

331 (95) <0.05

SF/SW

36

27 (75)

Enrolled

349

294 (84) 0.62

SF/SW

37

30 (81)

13 (38)

17 (50)

13 (38)

19 (56)

1.0

8 (23)

Abbreviations: SF/SW: Screening Failure/Screening Withdrawal *SW/SF subtotals may not be consistent across domains due to patients not completing some or all of the assessment.

ABSTRACTS

O27 GENETIC PROFILE AND CLINICAL CHARACTERISTICS OF OLDER WOMEN WITH BREAST CANCER IN THE CLINICAL CANCER GENOMICS COMMUNITY RESEARCH NETWORK Y. Chavarri-Guerra1,2,*, C. B. Hendricks3, S. Brown4, K. E. Weeman5, M. Hander6, Z. E. Segota7, C. Hake8, S. Sand2, T. P. Slavin2, A. Hurria2, E. Soto-Perez-de-Celis2, B. Nehoray2, I. Solomon2, L. Kuzmich2, K. Blazer2, L. Van Tongeren2, K. Blankstein9, J. Weitzel2 1 Instituto Nacional de Ciencias Medicas y Nutricion, Salvador Zubiran, Mexico, Mexico, 2City of Hope, Duarte, 3Maryland Oncology Hematology, Bethesda, 4St. Joseph Hospital and Mission Hospital, Orange, 5Aultman Hosp, Wooster, 6Kootenai Clinic Cancer Serivces, Coeur d’Alene, 7Holy Cross Medical Group, Fort Lauderdale, 8Waukesha Memorial Hospital-ProHealth Care Research Institute, Waukesha, 9Hunter don Hematology Oncology, Flemington, United States

Introduction: The role of breast cancer (BC) susceptibility gene mutations is known in younger women with BC, and young age at onset is a recognized criterion in the NCCN guidelines for genetic cancer risk assessment (GCRA). However, the potential role of germline cancer susceptibility mutations in older women with BC is not well studied, in part because they are less likely to undergo GCRA. Objectives: We aimed to characterize and compare the germline mutation profile of women with a history of BC aged 65 years and <65 years at the time of GCRA in the Clinical Cancer Genomics Community Research Network (CCGCRN), a large consortium of over 40 collaborating clinics across the United States and Latin America. Methods: Genetic testing results from all women with breast cancer (invasive or ductal carcinoma in situ) enrolled in the CCGCRN registry from 1997 to 2016 were included in this analysis. Socio-demographic characteristics, clinical variables and genetic profiles were compared between women aged 65 and <65 years at the time of GCRA. Data was analyzed using Fisher’s test and 2 statistics, with a two-sided p value of <0.05 considered significant. Results: We identified 1,035 women aged 65 years and 8,900 women aged <65 years who were diagnosed with BC. 10.4% of women 65 (n=108) and 11.6% of their younger counterparts (n=1,033) were found to carry a BC-associated germline mutation (p=0.27). Germline mutations in high-risk genes (e.g. BRCA) represented 85% of carriers among women 65 years (n=92), and 96% of those <65 (n=987), (p<0.01). BRCA2 was the most frequently mutated high-risk gene among women 65 years, representing 41.6% of cases (n=45), followed by BRCA1 (37%, n=40), PALB2 (6%, n=6) and TP53 (<1%, n=1). The most frequently identified mutations in high-risk genes for women <65 were in BRCA1 (53%, n=547), BRCA2 (37%, n=386), PALB2 (3%, n=26) and TP53 (2%, n=23). Notably, mutations in moderate-risk genes were found in 17% of women 65 (n=18) and in 4% of those <65 (n=45) (p<0.01). Among women 65 with a mutation in in a moderate-risk gene, the most prevalent was CHEK2 (14%, n=15), followed by ATM (2%, n=2) and NF1 (<1%, n=1). While CHEK2 was also the most common mutated moderate-risk gene among women <65 years (3%, n=32), it represented a smaller proportion of cases compared to the

S35

older group (3% vs 14%, p< 0.01). ATM (1%, n=14) and NF1 (<1%, n=7) were observed in a similar proportion as the older group. Among women aged 65, mean age at first BC diagnosis was 56.4 years (range 29–84), compared to 40.8 years (range 19–64) in women <65. 25.9% of older women (n=28) had their first BC diagnosed at age 65, of which 60.7% (n=17) were associated with BRCA2 and 21.4% (n=6) with BRCA1 mutations. BRCA2 mutations were more frequent among women diagnosed with BC at age 65 than in those <65 (p<0.01). Older women with germline mutations were more likely to present with stage 0-II BC (93% vs 77%, p<0.01) and with hormone receptor (HR) positive tumors (68% vs 53%, p=0.04). Conclusion: Older women were more likely than their younger counterparts to have mutations in moderate-risk genes or in the BRCA2 high-risk gene, which could explain why older women carrying BC susceptibility mutations presented more frequently with HR-positive tumors. Our results show that older women should not be excluded from GCRA, and clinicians should be aware of the NCCN criteria for their inclusion. Disclosure of interest: Y. Chavarri-Guerra: None declared, C. Hendricks: None declared, S. Brown: None declared, K. Weeman: None declared, M. Hander: None declared, Z. Segota: None declared, C. Hake: None declared, S. Sand: None declared, T. Slavin: None declared, A. Hurria Grant / Research Support from: Celgene, Novartis and GSK, Consultant for: Carevive, Sanofi, Gtx Inc. and Boehringer Ingelheim., E. Soto-Perez-deCelis: None declared, B. Nehoray: None declared, I. Solomon: None declared, L. Kuzmich: None declared, K. Blazer: None declared, L. Van Tongeren: None declared, K. Blankstein: None declared, J. Weitzel: None declared Keywords: Breast cancer, germline mutations

O28 COMPREHENSIVE ANALYSIS OF IMMUNE-RELATED TOXICITIES AMONG OLDER ADULTS TREATED ON NOVEL IMMUNOTHERAPIES ON PHASE I CLINICAL TRIALS I. M. Subbiah1,*, T. Fujii2, A. Lui2, S. Roy2, R. Khanji2, K. R. Hess2, V. Subbiah2, A. Naing2, D. S. Hong2 1 Cancer Medicine, 2MD Anderson Cancer Center, Houston, United States

Introduction: Cancer remains a disease that disproportionately affects older adults. Despite accounting for the greatest proportion of those with cancer, older patients (age 65 and above) represent a significantly lower proportion of patients included in clinical trials, particularly early phase trials with novel therapeutics. An oft-cited barrier to enrolling an older adult is the trepidation over a higher incidence of toxicities than the middle age subset. Objectives: To that end, we investigated the participation and incidence of immune related toxicities among older adults receiving treatment on phase 1 immunotherapy trials. Methods: We queried a prospectively maintained department database of patients with advanced cancers and identified 422 patients treated on immunotherapy-based

S36

ABSTRACTS

phase I trials bw 04/2009–09/2015. We gathered baseline clinical characteristics and immune-related adverse events (irAE) such as endocrinopathies, diarrhea/colitis, pneumonitis, constitutional (eg fatigue, fever, anorexia), myalgia, and dermatitis. Results: Overall, 116 of the 422 patients treated were older adults aged 65 years and above (27%, median 70 years), 50 were adolescent/young adults (AYA) aged 15–39 years (12%, median 30 years), 256 mid age aged 40–64 years (61%, median 56 years). The primary cancers were GI (n=108, 26%), thoracic/ head/neck (n=84, 20%), GU (n=54, 13%), and GYN (n=47, 11%). Among the three age groups, the median PFS was comparable (2.4 months older adults, 2.1 months AYA, 2.1 months mid age). Overall, older adults had a higher incidence of irAE than mid age or AYA (low grade [G1/2] 49% vs 34% vs 34%, p 0.02; high grade [G3/4] 19% vs 11% vs 12%; p=0.14). When assessing the irAE rates of older adults to AYA and mid age pts, the odds ratio of high grade irAEs was 1.81 (95% CI 1.01, 3.24; p=0.05) and low grade irAEs was 1.85 (95% CI 1.20, 2.85; p=0.0055). Most common low grade (grades 1 and 2) irAEs among all patients were fatigue (n=76, 18%), dermatitis (n=59, 14%), fever (n=29, 7%) and anorexia (n=28, 7%) – older adults had a higher incidence of low grade fatigue (25% vs 15%, OR 1.84, 95% CI 1.09, 3.10; p=0.025). Conclusion: Older adults remained underrepresented on immunotherapy-based phase I trials. Older adults did have a higher likelihood of experiencing an immune-related toxicity than the mid age and AYA pts. These findings warrant additional studies on patient selection and vigilant monitoring of senior adults while receiving novel immunotherapies. Disclosure of interest: None declared Keywords: Clinical trial, clinical trial participation, immunerelated toxicity, immunotherapy, toxicity

O29 TOLERABILITY AND OUTCOMES OF RADIATION THERAPY FOR BREAST CANCER IN OLDER WOMEN: A RETROSPECTIVE STUDY IN 817 PATIENTS K. I. Cao1,*, F. Salviat2, A. Fourquet1, M.-C. Falcou2, F. Laki3, P. Beuzeboc4, A. Savignoni2, P. Poortmans1, Y. M. Kirova1 1 Department of Radiation Oncology, 2Department of Biostatistics, 3 Department of Surgical Oncology, 4Department of Medical Oncology, Institut Curie, Paris, France

Introduction: Breast cancer (BC) management in older women requires an individual approach and is becoming increasingly topical given the aging population. Postoperative radiation therapy (RT) is a standard treatment of BC after breast-conserving-surgery in most patients but its relative benefit may be counteracted by potential side-effects, especially in elderly. Objectives: The aim of this study was to assess acute and long-term radiation-induced toxicities and the impact of comorbidities on outcomes in the older women treated by RT for non-metastatic breast cancer.

Methods: Women aged 70 years at diagnosis, who received exclusive or postoperative RT for primary non-metastatic breast cancer, including carcinoma in situ, between 2003 and 2009 were retrieved from the Institut Curie registry. We calculated the Charlson comorbidity index (CCI) for each patient and collected the cardiovascular risk factors other than age (hypertension, dyslipidemia, smoking status). We analyzed overall survival (OS), progression free survival (PFS) and acute and late toxicities according to the CTCAE (Common Terminology Criteria for Adverse Events) v3.0. Results: A total of 817 patients were included in this study. Median age at diagnosis was 76.6 years [70–93.3]. Most patients had HR+ (hormone-receptor positive) HER2– breast cancer (83.9 %). 517 patients (62.7%) had at least one cardiovascular risk factor. With a median follow-up of 6.7 years [0.5–13], OS at 5 years was 86.3%, 95% CI 83.8–88.8, and PFS was 84.5%, 95% CI 81.9–87.1. OS at 5 years was statistically different according to the Charlson index: 90.2%, 95% CI 87.2–93.3, for a CCI of 0, 84.6%, 95% CI 80.5–88.8, for a CCI of 1, and 78%, 95% CI 70.5– 86.2, for a CCI  2 (p<0.001, log-rank test), respectively. Similar results were found for PFS (p<0.001, log-rank test). 22.6% of patients had no toxicity; of those who experienced toxicity, most was limited to grade I or II. Only five cases (0.6%) of radiation-induced pneumonitis were reported after a median time of 16.4 months (grade I, n=1; grade II, n=2). One case (0.1%) of myocardial ischemia was described 14.5 months after RT. Women older than 80 years were less likely to have acute dermatitis (OR=0.62; 95% CI 0.45–0.85), long-term breast pain (OR=0.31; 95% CI 0.14–0.62), and long-term breast deformation (OR=0.63; 95% CI 0.42–0.93) compared to patients younger than 80 years. There was no significant association found between other cardiovascular risk factors and toxicities. Conclusion: Radiation therapy for breast cancer in the older women is well-tolerated. An extended follow-up is planned in order to assess toxicities at a longer time horizon. Further studies could be envisaged to assess the quality-of-life during and after RT for breast cancer in the older patient population. Disclosure of interest: None declared Keywords: Breast cancer, older women, radiotherapy

O30 EVALUATING THE EFFECTS OF EXERCISE ON ANXIETY AND DEPRESSION IN 198 OLDER PATIENTS WITH CANCER RECEIVING ACTIVE CHEMOTHERAPY: DATA FROM A RANDOMIZED CONTROLLED TRIAL IN THE UNIVERSITY OF ROCHESTER CANCER CENTER NCI COMMUNITY ONCOLOGY RESEARCH PROGRAM K. P. M. Loh1,*, S. G. Mohile1, B. E. Canin1, J. Bautista1, P.-J. Lin1, M. Flannery1, L. J. Peppone1, S. Kasbari2, B. T. Esparaz3, J. P. Kuebler4, K. Mustian1 1 University of Rochester, Rochester, 2Southeast Clinical Oncology Research Consortium (SCOR), 3Heartland NCORP, 4Columbus NCORP, United States

Introduction: Anxiety and depression are common symptoms experienced by older patients with cancer. In prior

ABSTRACTS

studies, both symptoms have been associated with elevated inflammatory cytokines. Pharmacologic interventions for these symptoms may or can lead to significant side effects such as drowsiness, confusion and falls in older adults. Exercise has been shown to improve anxiety and depression as well as to decrease inflammation in the general cancer population. However, exercise studies in older adults receiving treatment for cancer are scarce and research that evaluates the mechanisms by which exercise may decrease anxiety or depression in older patients with cancer is even more limited. Objectives: We conducted a secondary analysis of a nationwide randomized controlled trial to assess the effects of exercise on anxiety and depression in older patients with cancer receiving active chemotherapy. We also investigated mechanisms that could explain the anxiolytic and antidepressive effects of exercise. Methods: The study was conducted within the University of Rochester Cancer Center NCI Community Oncology Research Program (URCC NCORP). A total of 198 older patients with cancer (aged 60 years) were enrolled and included in our analysis. Patients in the study were randomized to receive chemotherapy alone (Control) or to a 6-week exercise program (Exercise for Cancer Patients; EXCAP). EXCAP is a homebased progressive aerobic and anaerobic exercise program in which patients were instructed to walk as many steps everyday as they can and the number of steps was recorded using a pedometer (aerobic). In addition, they were educated and provided with the instructions to perform therapeutic band training (anaerobic). Analysis of covariance (ANCOVA) was used to evaluate the effects of EXCAP on anxiety and depression, which were assessed using the State Trait Anxiety Inventory and Profile of Mood States-Depression subscale at baseline prior to initiation of chemotherapy and 6 weeks after. We adjusted for gender, chemotherapy cycle duration (every 2-week or 3-week) and baseline values for anxiety or depression. We also assessed the associations of changes in inflammatory cytokines (IL-1, IL-6, IL-8, IL-10, IFN- and TNFr1) with changes in anxiety and depression using Pearson correlation.

S37

Fig. 1 (abstract O30) – State Trait Anxiety Inventory scores in older patients receiving chemotherapy with or without EXCAP at baseline and 6 weeks post-intervention. Smaller values indicate lower anxiety levels. Results: The median age of our population was 66.7±2.3 years; 92% were female and 77% had breast cancer. Seventytwo percent received every 3-week and 28% received every 2-week chemotherapy regimens. Compared to the control group, the EXCAP group had significantly decreased level of anxiety (mean change in score; EXCAP: -3.29, Control: –1.20, p=0.04) (Figure 1). There was also a non-significant decrease in depression (mean change in score; EXCAP: −1.83, Control: 0.03, p=0.08). In the EXCAP group, positive correlation was noted between TNFr1 and changes in anxiety (r=0.37, p=0.009). No associations were noted in other inflammatory cytokines and changes in anxiety. Similarly, changes in inflammatory cytokines were not associated with changes in depression. Conclusion: Our analysis has shown that exercise decreased anxiety and potentially depression in older patients with cancer receiving active chemotherapy. The effect of exercise on anxiety may be mediated by reducing inflammation, specifically through TNFr1. Disclosure of interest: None declared Keywords: Anxiety, depression, exercise