Objective: Malignant melanoma with brain metastases is associated with a higher risk of death. No specific treatments were demonstrated to be useful in a such situation. Drugs as temozolomide orally or new targeted treatments showed significant objective response rates, even complete regression. Such responses could be obtained using new strategies based on dynamic changes over time of some molecular markers. Elements of ethics should be taken into account in order to adapt treatment and avoid resistance. Methods:The case of a 59-year old male with a primary cutaneous melanoma of the trunk, treated at Cancer Institute "Ion Chiricuta" from August 2001 is presented. After multiple loco-regional relapses, the patient developed brain metastases and started temozolomide 150 mg once daily, five days, every 4 weeks, 6 cycles with concurrent whole brain external radiotherapy. Comparative immunostaining including proliferation and pro-apoptotic molecular markers between the initial diagnosis (2001), before (2005) and after (2008) temozolomide treatment was performed.Result: Nine months after the start of temozolomide treatment, complete response was confirmed by magnetic resonance imaging. Overall cancer specific survival was 41 months. Ki-67, cyclin E, HMB-45 expression and Bax/ Bcl-2 ratio increased during almost 10 years of treatment and follow-up. Bcl-2 staining was absent at the last analysis. Only p53 and Bax expression doesn't changed during treatment. Conclusion:It seems that metastatic melanoma cells lost some of pro-apoptotic markers and overexpressed markers of proliferation. Predictive markers of response and resistance were actively identified; their combination and dynamic over time could help the oncologist to select those metastatic patients with highest chances of response. Dynamic changes of these molecular markers would guide treatment and the overall core strategy. Serial biopsies and tissue analyses become a challenging ethical issue. Empiric treatments based on a unique tumor signature should be modified using an adaptive approach.