6082 Background: The present study was undertaken to evaluate the possibility of using the expression of a panel of proteins involved in apoptosis and/or growth control as predictive markers for radiotherapy response. Methods: Nine head and neck squamous cell carcinoma (HNSCC) cell lines were selected to represent different parts of the spectrum of in vitro radiosensitivity. These cell lines have an intrinsic radiosensitivity (IR) between 1.4 and 2.6 measured with a 96 well plate clonogenic assay. Results: Fourteen proteins e.g. Bcl-2, Bcl-XL, Bax, Bad, Bak, Puma, survivin, COX-2, Hsp70, MDM2, p53, EGFR, cyklin D1 and SMAD4 were investigated in above mentioned cell lines and in normal oral keratinocytes (NOK). Each protein was quantified using Western blot analyse and the relative densiometric value from each blot was adjusted for actin and standardized to NOK. The correlation between each protein and IR was analyzed but none significant correlation was found. When combine different proteins, EGFR, Survivin, Bcl-2, Bak, Hsp70 and Smad4 together show the highest significant correlation to the IR. As Western blot is a semi quantitative method we decided that a division into groups depending on the level of protein expression would give a more realistic description of the prognostic value of the detected changes The protein expression was classified into four groups (0–4 point; no, small, middle or large changes compared to NOK) and each p53 mutation got one point. We named this grouping Number of Negative Points (NNP). The NNP for all 15 factors together show a significant correlations to IR, R=0.826 and P= 0.006. A multivariate statistical calculation was carried out to select the combination of factors that had the strongest correlation to IR and the best combination was EGFR, survivin and p53 mutations, R=0.935 and P<0.001. Conclusions: Our results show that it is only possible to predict the IR of tumor cells if combinations of prognostic factors can be analysed. Taken together, we here present a novel method that gives the possibility to combine different prognostic factors and which also enhances the chance to predict the radiosensitivity of HNSCC cells. No significant financial relationships to disclose.
Integrated Data Analysis Implicates PLAU, SERPINE1, SPP1, and MMP1 as Prognostic Factors in Head and Neck Squamous Cell Carcinoma
