Transcriptional analysis of small cell lung cancer transformation in epidermal growth factor receptor mutated lung adenocarcinomas.

Abstract

e21100 Background: Epidermal growth factor receptor ( EGFR)-mutated lung adenocarcinoma (LUAD) could benefit from EGFR-TKIs (tyrosine kinase inhibitors) treatment, but drug resistance seems to be inevitable. Small cell lung cancer (SCLC) transformation counts for 3-15% of the resistance mechanism, and was amply studied at genomics level but rarely at transcriptional level. Methods: The expression of 730 mRNAs were investigated by Nanostring nCounter Pancancer Pathway Panel on 72 formalin-fixed and paraffin-embedded (FFPE) samples from 27 EGFR-mutated LUAD patients with SCLC transformation after EGFR-TKIs therapy (19 LUAD samples before transformation, LUAD-BT; 21 SCLC samples after transformation, SCLC-AT), 12 EGFR-mutated LUAD patients never SCLC transformed after EGFR-TKIs therapy (12 samples, LUAD-NT) and 20 stage IIĨIV primary SCLC patients (20 samples, SCLC-P). For patients enrolled in LUAD-NT group, tissue biopsies were performed at least twice, which were all diagnosed as pure LUAD, and the overall survival (OS) since first line therapy was longer than the median transformation time (mTt) of SCLC-transformed patients in our study. mRNA expression patterns and biological pathway scores were compared among four groups. The candidate predictive biomarkers from mRNA expression pattern analysis were validated by area under curve (AUC) of receiver operating characteristic curve (ROC) and the logarithm of fold change to the base 2 (log2FC). Results: On the last day of follow up (1st February 2022), the shortest OS of LUAD-NT patients was 28.4 months, whereas the mTt was 27.5 months. Among four groups, LUAD-NT and LUAD-BT showed the most similar mRNA expression patterns, and SCLC-P were significantly different from the others. 8.6% (63/730) mRNA showed significant downregulation after SCLC transformation, while 3.6% (26/730) showed significant upregulation ( p value adjusted by Benjamin & Hochberg’s method < 0.05, SCLC-AT vs LUAD-BT). In pathway enrichment analysis, the score of RAS and TGF-β pathways were significantly lower in SCLC-AT than LUAD-BT. Compared with SCLC-P, 6 upregulated mRNAs in SCLC-AT were observed (log2FC > 2, AUC > 0.85 and each raw p value < 0.05), including AR, COL5A1, GHR, HMGA2, IGFBP3 and IL6R, which could be further validated as diagnostic markers in a larger cohort. Moreover, compared with LUAD-NT, 4 mRNAs ( BRIP1, CCNE2, CDKN2A and MCM2) were found to be significantly upregulated (log2FC > 1, AUC > 0.75 and each raw p value < 0.05) in LUAD-BT, indicating the predictive value of SCLC transformation. Conclusions: The transformation of LUAD to SCLC may be promoted by transcriptional events. We also described some significantly different expressed mRNAs that could candidate as predictive or diagnostic markers for SCLC transformation, which should be further validated in a larger cohort.