Abstract Early-stage laryngeal cancer has cure rates above 80% with exclusive radiotherapy. However, when salvage treatment is needed, surgery is the main option and despite efforts to conservative procedures to preserve laryngeal function, up to 50% will require total laryngectomy at diagnosis of recurrence or future. So, the use of microRNAs as molecular markers is a method to be tested with the aim of predicting the chances of success with radiotherapy, lowering tumor relapses and turning possible to reduce the need for salvage procedures that would affect speech and impact in quality of life. Objectives: To identify microRNAs capable of segregating radioresistant from radiosensitive tumors, to determine prognosis in pre-treatment tumors and to predict failure with radiotherapy. Methods: Molecular Screening of 667 microRNAs using taqman low density array (TLDA) for target selection and then validation in 34 patients, being 20 radioresistant and 14 radiosensitive by RT-PCR with individual assays. Statistical analysis was performed by Chi-square and Fisher exact test for associations between variables, and log-rank method of Kaplan-Meier for recurrence free survival between molecular, clinical, demographic and treatment variables. Also to verify differences in distribution between the expression levels in radioresistant and radiosensitive the Mann-Whitney test was performed. Results: Four microRNAs were selected for validation, miR-296, miR-452, miR-183*, miR-200c based on fold-change and literature review and those were analyzed with respect to age, gender, TNM, laryngeal subsite, radiation dose and treatment interval. Among clinical, demographic, treatment and molecular variables, only miR-296 was shown to present different distribution with statistical significance between radioresistant and radiosensitive (p = 0.002), also demonstrated association with recurrence (p = 0.01) with an odds ratio (OR) of 8.6, confidence interval (CI) between 1.7 and 42.2. Conclusions: In our study, only miR-296 demonstrated the ability to predict patients most likely to fail radiotherapy, also showing a worse prognosis for tumors with high expression level. miR-296 is a molecular predictive and prognostic marker to be considered prior to treatment selection for early-stage laryngeal tumors. Citation Format: Danielle C C Maia, Ana Carolina de Carvalho, Maria Aderuza Horst, Cristovam Scapulatempo Neto, Andre Lopes Carvalho, Andre L. Vettore. miR-296 as prognostic and predictive molecular marker for recurrence in early-stage laryngeal carcinoma treated with definitive radiotherapy. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 1481. doi:10.1158/1538-7445.AM2014-1481