Colorectal cancer (CRC) is worldwide the third leading cause of cancer-related dead in men and the second in women, with over 1.8 million new cases diagnosed in 2018. CRC is a heterogeneous cancer type that usually arises from precancerous lesions, called adenomas, which progress to an invasive cancer through the adenoma-carcinoma pathway. Despite the effective screening programs leading to the timely removal of precancerous lesion and the standardization of preoperative and postoperative care, the incidence and mortality rates of CRC are still high. The gold standard for CRC treatment includes radiotherapy and adjuvant chemotherapy for stage III, and neoadjuvant therapy with the addition of targeted drugs (depending on the mutational status) for metastatic stage IV. However, the prognosis and drug responses of CRC patients are very heterogeneous due to the onset of resistance mechanisms to both conventional and targeted agents, resulting in a restricted efficacy of therapy and reduced survival time. Patient-specific resistance to therapy in CRC is the major obstacle in the conception of effective cancer treatment. The classical tumor classification (UICC) based on primary tumor size (T category), regional lymph nodes involvement (N category) and presence of distant metastatic spread (M category) is not sufficient to take optimal therapeutic decisions, since there is an extensive inter-tumor heterogeneity of CRC at the genomic, epigenomic, transcriptomic and immune levels resulting in heterogeneous drug response. Therefore, there is a strong need to develop an effective patient-specific treatment prediction system able to produce and integrate comprehensive data obtained from individual patients and allow the translation of the patient-specific tumor features into personalized therapeutic decisions. In our work, we are developing a precision oncology platform that integrates tumor organoids, with high-throughput and high-content data for making therapeutic recommendations for individual patients. We have established a living biobank of tumor organoids derived from histologically verified CRC tissue samples from biopsy material of CRC patients subjected to surgery. We have fully characterized the tumor organoids via exome sequencing, RNA sequencing and proteomic analysis and we have established and optimized the experimental procedure to perform perturbation experiments on organoids using drugs and prepare samples for phosphoproteomics measurements. CRC organoids from different patients were selected according to their characteristics and patient-specific perturbation data sets were generated by challenging organoids with targeted drugs. The perturbation data were acquired with RNA sequencing, proteomics and phosphoproteomics approaches. We are developing a literature-based signaling network model for CRC that will be adapted based on the generated data into patient-specific network models to be used to simulate drug combinations and identify optimal therapeutic strategies for single patient. The best combinatorial agents predicted will be experimentally validate by testing the drug combination specific for each patient on the patient-specific organoids. Thus, with our work we are establishing a predictive strategy aiming at identifying effective combination therapies for CRC patients. Importantly, our work will be able to account for changes of the tumor features as a result of disease progression and treatment-related selection, allowing a dynamic design of therapeutic recommendation that will translate into an optimal personalized patient care.