Intrinsic and acquired resistances are major obstacles in cancer therapy. Genetic characterization is commonly used to identify predictive or prognostic biomarker signatures and potential cancer targets in samples from therapy-naïve patients. By far less common are such investigations to identify specific, predictive and/or prognostic gene signatures in patients or cancer cells refractory to a specific molecular-targeted intervention. This, however, might have a great value to foster the development of tailored, personalized cancer therapy. Based on our identification of a differential radiosensitization by single and combined β1 integrin (AIIB2) and EGFR (Cetuximab) targeting in more physiological, three-dimensional head and neck squamous cell carcinoma (HNSCC) cell cultures, we performed comparative whole exome sequencing, phosphoproteome analyses and RNAi knockdown screens in responder and nonresponder cell lines. We found a higher rate of gene mutations with putative proteinchanging characteristics in non-responders and different mutational profiles of responders and non-responders. These profiles allow stratification of HNSCC patients and identification of potential targets to address treatment resistance. Consecutively, pharmacological inhibition of mTOR and KEAP1 effectively diminished non-responder insusceptibility to β1 integrin and EGFR targeting for radiosensitization. Our data pinpoint the added value of genetic biomarker identification after selection for cancer subgroup responsiveness to targeted therapies.