AB008. Activated pathways of Thymic Epithelial Tumors: a RYTHMIC study

BACKGROUND: Thymic epithelial tumors (TETs) are rare malignancies of the anterior mediastinum with a high histopathological diversity from thymoma A to thymic carcinoma (TC). The biology of TETs is poorly understood and knowledge of the transcriptomic fingerprint of thymoma and TC is limited. Up to 30% of patients will develop associated autoimmune disorders, mainly myasthenia gravis (MG). We aimed to characterize main cancer activation pathways of TET subgroups. METHODS: We selected a representative balanced set of thymoma and TCs to analyze 24 main cancer activation pathways using gene expression throughout Oncology biomarker panel (2,562 genes). Tumor representative paraffin-embedded blocks were macrodissected. Then, we merged data with The Cancer Genome Atlas (TCGA) data (profiles with >30% tumor cellularity kept). We correlated epidemiologic, clinical and pathological characteristics of patients with genes expression based on cancer Hallmarks and immunedeconv (v2.0.4). RESULTS: Three hundred and fourteen patients were included, including 120 from TCGA. Median-age at diagnosis was 52 (10–84). Fifty two percent were women. Eighty four out of 314 (26.7%) reported MG, mostly in thymoma B2 (11,4%) and B3 (8%) but none for TC. AB was the most frequent thymoma subtype (n=70, 22.3%), followed by B2, B1, B3, A and TC. RNA expression analysis identified 3 main molecular subgroups or clusters, distribution of histological subtypes among them was diverse (P<0.0001). Cluster 1 was represented meanly by thymic carcinoma, cluster 2 was associated to thymoma type B and Cluster 3 to thymoma type A and AB. Activated pathways of histological subtypes were as follows: thymoma A showed activation of angiogenesis, Hedgehog and Notch hallmarks, as for thymoma AB; thymoma B1 and B2 showed cell cycle checkpoint factors activated pathway; thymoma B3 protein secretion pathway and; TC Epithelial to mesenchymal transition (EMT), MTOR1 and MYC pathways. Then, we analyzed activated pathways of the 3 molecular subgroups: cluster 1, with the worst prognostic, was associated to inflammatory signaling, MTOR1, KRAS and EMT pathways; cluster 2, with the best prognostic, showed activated cell control transcription factors hallmark and; cluster 3, showed cell differentiation activated pathway. We found a difference in the presence of B and T-cells among clusters and thymoma subtypes. Cluster 1, thymoma A and TC showed higher representation of B-cells (P<0.0001, respectively) and regulation T-cells (Treg) (P<0.0001, respectively); in contrast, cluster 2 and thymomas AB and B a higher proportion of CD8+ T-cells (P<0.0001, respectively). Interestingly, non-regulatory CD4+ T-cells did not show significantly results in any subset. Of note, Macrophages M1 were presented in cluster 1 and M2 in cluster 3 (P<0.0001, respectively). Median follow-up was 35 months [95% confidence interval (CI): 27.03–42.96 months]. Median-OS was 350 months (NR- NR). Cluster 1 showed a poorer prognostic (median-OS of 74 months vs. NR and NR; P<0.0001) comparing to cluster 2 and 3, respectively. CONCLUSIONS: We describe differential molecular characteristics among histological subgroups in 3 molecular subgroups. Clusters were significantly associated to survival outcomes and showed distinguish activated cancer pathways. The analysis suggests new therapeutic venues.