Since cancer is the second cause of death worldwide, the search for effective therapy without side effects is pertinent. To this end, selective targeting of cancer cells compared to surrounding healthy tissue is a main strategy. We have recently demonstrated that the metabolism and accumulation pattern of phenolics is deviating in dysfunctional compared to healthy cells. In this scope, the distinct cellular fate in dysfunctional (cancer) cells may offer an alternative approach for selective anticancer targeting. Curcumin is a polyphenol extracted from Curcuma longa with several beneficial properties including anticancer activity. However, its low bioavailability, stability and PAINS classification hamper application in health care. Previously, 90 curcumin analogues have been synthesized in our lab, including thiazepane derivatives which deviate from the curcumin structure by their out-of-plane configuration. In this project, the most promising analogues will be investigated using 3D in vitro models for colon cancer, more specifically tumour spheroids and a newly developed embedded tumour model. This way, the analogues can be screened for selectivity by examining accumulation, biotransformation and -activity in one model using microscopic, analytical (UPLC-QTOF-MS), cellular and molecular techniques, followed by in vivo validation. The new compounds and mechanistic insights may therefore open perspectives for product development in food (supplement) and pharma industry.