This paper introduces a framework for disease pre-diction from multimodal genetic and imaging data. We proposea multilevel survival model which allows predicting the time ofoccurrence of a future disease state in patients initially exhibitingmild symptoms. This new multilevel setting allows modeling theinteractions between genetic and imaging variables. This is incontrast with classical additive models which treat all modalitiesin the same manner and can result in undesirable eliminationof specific modalities when their contributions are unbalanced.Moreover, the use of a survival model allows overcoming thelimitations of previous approaches based on classification whichconsider a fixed time frame. Furthermore, we introduce specificpenalties taking into account the structure of the different types ofdata, such as a group lasso penalty over the genetic modality a a ℓ2-penalty over the imaging modality. Finally, we propose a fastoptimization algorithm, based on a proximal gradient method.The approach was applied to the prediction of Alzheimer’sdisease (AD) among patients with mild cognitive impairment(MCI) based on genetic (single nucleotide polymorphisms - SNP)and imaging (anatomical MRI measures) data from the ADNIdatabase. The experiments demonstrate the effectiveness of themethod for predicting the time of conversion to AD. It revealedhow genetic variants and brain imaging alterations interact in theprediction of future disease status. The approach is generic andcould potentially be useful for the prediction of other diseases