One of the most important goals of cancer treatment is the ability to select the best therapy for each individual cancer patient and to predict the therapy response as accurate as possible.  

For this purpose, strong efforts were made to establish preclinical models such as 2D primary cell culture or patient-derived xenografts (PDX) in mice. Primary 2D cell culture are inexpensive and easy to generate, whereas mouse models are very cost intensive and time-consuming until results can be generated. However, both lack the ability to maintain the tumors original microenvironment. Considering these limitations, in terms of personalized medicine, these models are not feasible to be used for therapy recommendations.

Protocols for the generation of patient-derived 3D cell cultures of tumor tissue have been established for several solid tumor entities, including very recently also head and neck squamous cell carcinoma (HNSCC). These so-called “patient-derived organotypic tumorspheres” (PDOTS) are embedded in an extracellular matrix, thereby imitating cell-cell and cell-matrix interactions. In addition, PDOTS can be generated and expanded within three to four weeks, thus allowing molecular tumor profiling and ex vivo drug screening within a clinically acceptable time frame. For prediction of therapy response, concerning personalized medicine, 3D PDOTS culture system might thus represent a more suitable preclinical model.