No prospective study was conducted in elderly patients with cancer to assess the relative value of disease-related and patient-related prognosis factors. Patient-related prognostic factors have been highlighted in elderly patients with cancer resulting in the necessity of a geriatric assessment. The impact on overall survival of all of these factors was recognized in elderly people with cancer but remains unknown in High Risk Myelodysplastic Syndromes (HR-MDS). Therefore this information could be crucial to better select geriatric assessment domains relevant for the prediction and to recommend simplified tool after stratification of geriatric assessment domains thanks to their predictive value.

The main hypothesis is that patient-related factors will have a better capacity to predict survival and treatment tolerance than disease-related factors in HR-MDS aged 75 and over and that the predictive value will be different among assessment tools which allows a selection of reduced number of tools for clinical use.

To best knowledge estimation of predictive value of geriatric assessment tools remains unknown and explains why no standardization of practice exists. In testing all tools at the same cohort of patients allows to compare different tools and to define minimal and optimal geriatric assessment for HR-MDS. To determine the best strategy of geriatric assessment will allow in a second time to measure the impact of the use of this geriatric standardized evaluation by comparing patients'care and prognosis according to the use or not by the doctors of the new scores.

Research outcomes are various medical, economic and ethic. Medical because decision-making will be improved with simplified geriatric assessment; economic because a better knowledge of geriatric assessment will improve treatment toxicity prevention and decrease treatment costs. Ethic will be associated with this project because a better knowledge of geriatric assessment tools to predict survival and tolerance treatment could improve the choice of best supportive care if prognosis markers are not favorable to active therapy. This project could induce important modification of practice in this area to an improved personalized treatment and simplification of geriatric assessment allowing a large diffusion in hospitals and clinics.

inclusion criteria:

• Diagnosis of HR-MDS inferior to 6 weeks,
• No treatment received before inclusion,
• IPSS superior to 1 (intermediate 2 and high risk IPSS categories),
• Social system affiliation,
• Informed consent signed.

exclusion criteria:

• Patient eligibility to stem cell transplantation,
• IPSS inferior or egal 1 (low or intermediate 1 IPSS),
• Concomitant investigational trial participation, which could interfere,
• Patients under legal protection.