A method for predicting an effect of a medication or a treatment regimen to a subject suffering from a cancer, the method comprises: (A) obtaining a tissue from the subject; (B) dissociating the tissue to obtain a multicellular cluster, wherein the multicellular cluster comprises the cancer cell; (C) culturing the multicellular cluster on a cellulose sponge; (D) exposing the cultured multicellular cluster to the medication or the treatment regimen; and (E) measuring a first survival rate of the cancer cell before exposing to the medication or the treatment regimen and a second survival rate of the cancer cell after exposing to the medication or the treatment regimen, when the second survival rate is lower than the first survival rate, the method predicts positive effect of the medication or the treatment regimen to the subject.