Experimental investigations of tumors often result in data reflecting very complex underlying mechanisms. Computer
models of such phenomena enable their analysis and may lead to novel and more efficient therapy strategies. We present
a generalized finite-element mechano-chemical model of a solid tumor and assess its suitability for predicting therapy outcome. The model includes hosting tissue, tumor cells (vital and necrotic), nutrient (oxygen), blood vessels, and a growth inhibitor. At a certain time instant of the tumor development virtual therapies are performed and their outcomes are presented. The model parameters are obtained either directly from the available literature or estimated using multiscale modeling. First results indicate the usefulness of multi-physics tumor models for predicting therapy response. In the proposed model a regression of a manifest tumor after therapy may be observed.