The present disclosure describes a non-linear compartmental model using PET-derived data to predict, on a patient-specific basis, the optimal therapeutic dose of cargo carrying antibody (e.g., huA33) such as radiolabeled antibody, the antigen occupancy, residency times in normal and malignant tissues, and the cancer-to-normal tissue (e.g., colorectal cancer-to-normal colon tissue) therapeutic index. In addition, the non-linear compartmental model can be readily applied to the development of strategies such as multi-step targeting (MST) designed to further improve the therapeutic indices of RIT.
