Patients with metastatic cancer have a substantial symptom burden and may receive aggressive care at the end of life. There is evidence that the use of chemotherapy near the end of life is not related to its likelihood of providing benefit and the overuse of aggressive anticancer therapies near the end of life may result in more toxicity than clinical benefit. Moreover, proposing new lines of treatment after successive therapeutic failures may be a way of avoiding discussion of prognosis and advance care planning. It has been proposed that systems not providing overly aggressive care near the end of life would be the ones in which less than 10% of patients receive chemotherapy in the last 14 days of life. Presently the first consultation between patient and oncologist is ruled in France by the first "Plan Cancer" and the "Dispositif d'annonce" (announcement planning). Oncologists are supposed to explain the diagnosis of cancer and to present a treatment plan. In routine practice for metastatic non curable cancer patients, chemotherapy is presented as the leading therapy and its side effects are explained. The use of chemotherapy has been associated with the worsening of two major competitive life-threatening conditions for cancer patients: cachexia and thrombo-embolic events. Nevertheless the risk of worsening both those conditions is hardly explained in routine practice.

This study proposes to examine in a monocentric interventional prospective randomized trial, the impact of a particular way for the oncologist to present chemotherapy at the diagnosis stage on the easiness of timely chemotherapy interruption at the end of life. The main objective is to determine whether or not the explanation of the potential role of anticancer chemotherapy in worsening life-threatening conditions impacts the proportion of patients receiving chemotherapy in the last 30 days of life compared with usual presentation. Secondary objectives are to determine the impact of this communication strategy on overall survival and other indicators of aggressiveness of care and palliative care resources use.

The hypothesis is that the intervention will allow 15% of patients receiving anticancer therapy during the last 30 days of life, as compared to 30% in the control group.

The investigators expect that the intervention evaluated in this study will reduce the rate of patients receiving chemotherapy during the last 30 days of life hopefully improving the quality of end of life care. A secondary objective is overall survival and this study will therefore verify that the intervention arm is not associated with poorer overall survival. But more probably investigators expect patients in the intervention arm to have an improved overall survival mainly link to a decrease in harms due to chemotherapy given near the end of life and to better palliative care. In effect the hypothesis is that showing the life-threatening risks associated with chemotherapy and thus reducing for patients the importance of this treatment will leave room for improved palliative care as shown notably by earlier and more frequent referral to palliative care specialists. If this trial is positive, it will prove the capital role of patient-doctor communication in cancer care and that few differences in communication strategy could improve end of life care and maybe even survival. The impact on the oncology community would be major since the intervention could be easily transposed in all practices at no additional cost. It would also emphasize the importance of communication skills and human relationship in the very technical field of medical oncology.

Inclusion Criteria:

• stage IV lung cancer, histologically proven
• > 18 years old
• diagnosed during the last 8 weeks
• systemic therapy indicated

Exclusion Criteria:

• prior systemic therapy for metastatic disease
• comorbidity not compatible with the inclusion according to the investigator
• patient has not given its non-objection or not being able to give
• patient unaffiliated or not beneficiary of a social security scheme