Evidence on relative treatment effects concerning OS, progression-free survival (PFS) and discontinuation due to any reason (treatment persistence) and adverse events (tolerability) was estimated using a mixed treatment comparison following a systematic review of randomized clinical trials enrolling post-menopausal women with hormone-sensitive ABC. Health service costs were included and a lifetime perspective adopted (5% annual discount rate). Results: Everolimus+exemestane is estimated to significantly delay progression or death (HR PFS = 0.53; 95% CI: [0.37; 0.76]) and to increment life expectancy by 6.8 months in comparison to fulvestrant (HR OS = 0.82; 95% CI: [0.50; 1.36]), resulting in a 0.45 discounted life year (LY) gain. Corresponding incremental health service costs amount to 16,544€ /patient starting everolimus+exemestane. This results in an incremental cost-effectiveness ratio of 36,703€ /LY gained with everolimus+exemestane. Probabilistic sensitivity analysis showed a greater than 60% probability of everolimus+exemestane being costeffective against fulvestrant, at a willingness to pay of 50,000€ /LY. ConClusions: We evidence how valuable information from clinical trials can be pooled and used to inform about the therapeutic and economic value of guideline recommended therapies for advanced breast cancer.

A153a first-line therapy to treat metastatic gastric cancer was found to have an incremental cost-effectiveness ratio of $150,452 per QALY gained when compared to chemotherapy alone. The one-way sensitivity analysis at +/-20% for all parameters found that the most sensitive parameters were median overall survival and median progression free survival of both the trastuzumab + chemotherapy and chemotherapy alone treatments, as well as utility of progression free survival and progressed disease. Overall, the model is robust as it is not sensitive to any of the other parameters. ConClusions: Trastuzumab in combination with chemotherapy as a first-line treatment is likely to be cost-effective for adult metastatic gastric cancer patients with an overexpression of HER2 with a cost per QALY gained just

Objectives: Erlotinib, a representative of targeted therapy, has been approved by FDA for first-line therapy in EGFR-mutations positive NSCLC. However the high price of drug created a large barrier in practice especially in Vietnam. The aim of this study is to estimate the cost-utility of erlotinib versus the standard chemotherapy as firstline therapy in advanced EGFR mutation-positive NSCLC in Vietnam. MethOds: A 3-state Markov model was developed to evaluate the cost-utility of erlotinib versus the standard chemotherapy over life-time horizon. The discount rate of 3% has been evaluated for both cost and QALY. One-way sensitivity analysis was performed to evaluate uncertainties of parameters. Results: Compared with standard therapy, erlotinib regimen in first-line therapy in EGFR-mutations positive NSCLC resulted in the increase of 439.02 million VND treatment cost (534.16 million and 95.14 million VND, respectively) with the increase of 0,11 QALY (1.38 vs 1.27, respectively). The ICUR of erlotinib versus standard chemotherapy was 4.1 billion VND/QALY, which is 34 times higher than the willingness-to-pay of Vietnam (around 120 million VND). One-way sensitivity analysis showed such influencing factors on ICUR as drug price, price of medical services, discount rate, from which price of erlotinib was the most influencing factor. cOnclusiOns: Erlotinib was not cost-effective compared with standard chemotherapy in first-line therapy in EGFR-mutations positive NSCLC.

A157Objectives: Utility values are a key component of economic evaluations and a stated requirement of reimbursement bodies when evaluating a new treatment. The current study estimated health-state utilities associated with metastatic breast cancer (mBC) and adverse events (AE) of potential new treatments from a societal perspective in the UK. MethOds: In-depth telephone interviews were conducted in 2015 in the UK with 12 mBC patients (age≥ 60), 3 oncologists and 1 nurse practitioner to draft and validate health-state descriptions relevant to mBC patients. Ten mBC health-states were developed to be further assessed in TTO interviews: stable disease, progressive disease, mood change/anxiety, rash, nausea/vomiting, fatigue, stomatitis, hyperglycemia, neutropenia, and elevated liver functioning. One hundred individuals from the UK were recruited to complete a socio-demographic form, EQ-5D and Time-trade-off (TTO) interviews where time (in years) was "traded" to avoid mBC health-states, thereby determining preferences for developed healthstates. A mixed model was also developed to predict utility decrements for healthstates and assess the impact of patient-level variables (eg age, gender) on predicted utilities. Results: The predicted utility value for stable disease was 0.83 (CI: 0.79-0.86) based on mixed model analyses. The lowest predicted utility value was (0.39 [CI: 0.33-0.45]) for progressive disease. Predicted utilities for other health states ranged from 0.77 (hyperglycemia [CI: 0.72-0.80]) to 0.46 (mood change/depression and fatigue [CI: 0.40-0.52]). Further analyses revealed that age and gender did not significantly affect predicted utility values in mBC. cOnclusiOns: The current study shows the impact of treatment-related AEs on patient's HRQL. Health-state utility values in mBC were highest for stable disease followed by hyperglycemia. Symptomatic side effects and progressive disease in mBC had the lowest values suggesting societal preference to avoid these AEs if possible. Additional research is needed to further validate these findings.

A 3 4 7 -A 7 6 6 A729 oncologists in a two-round Delphi survey both, from the NHS and the society perspective. Unit costs were obtained from national databases. Direct costs included medication and healthcare resources costs (including visits, tests, hospitalizations, surgery, adverse events management and treatment of specific metastases). Indirect costs were estimated using the human-capital approach. Direct and total costs were described for patients with LR and/or MR and for an average patient from a hypothetical cohort of individuals with a BC recurrence. Uncertainties were explored in an univariate sensitivity-analysis. Results: Within a 100-patients'cohort, 7 would have an LR, 90 an MR, and 3 an LR followed by an MR. All patients with an LR would receive targeted therapies combinations. Patients estimated to have an MR received up to 5 treatment lines (5% no treatment; 10%:1L; 13%:2L; 20%:3L; 21%:4L and 31%:5L). Tthe most common treatments in each line were pertuzumabtrastuzumab-taxane (1L), trastuzumab emtansine (2L), lapatinib-capecitabine (3L) and trastuzumab combinations (4-5L). Depending on the type of recurrence and the number of lines received, estimated direct and total costs per patient ranged between € 38,511 (1L) and € 308,869 (5L); total costs between € 50,908 and € 358,000. Direct and total costs for an average patient were € 208,682 and € 235,138 respectively. ConClusions: Direct and indirect costs of treating recurrences need to be integrated in the cost-analysis involving the incorporation or novel targeted agents in the early BC setting to prevent recurrence. Strategies reducing the recurrences risk such as targeted therapies in earlier stages of HER2+ BC would result in significant economic savings for the Spanish NHS and the society.objeCtives: The aim of this study was to describe the estimated costs of recurrent breast cancer (BC) (either locoregional recurrences [LR] or unresectable/metastatic recurrences [MR]) in patients with HER2+ BC in Spain. Methods: The patients'flow, use of healthcare resources and loss of workdays were described by 15 Spanish