FFPE肿瘤的高灵敏度全转录组测序可实现免疫谱分析和预测性生物标志物发现

Background: Immune-based biomarkers are essential for predicting therapeutic response in cancer patients, yet transcriptomic profiling of formalin-fixed paraffin-embedded (FFPE) tissues remains limited by RNA degradation and low input availability. PredicineWTS is a high-sensitivity whole-transcriptome sequencing (WTS) assay optimized for FFPE specimens, designed to support immuno-oncology research and precision therapy development. Methods: PredicineWTS underwent analytical validation for FFPE inputs, assessing accuracy, precision, linearity, and reproducibility at low RNA quantities. The assay maintains high accuracy and precision at a 30 ng RNA input, enabling reliable transcriptome-wide profiling of challenging FFPE samples. A comprehensive bioinformatics analysis framework was established to perform quality control and characterize gene expression. Differential expression analysis, immune cell infiltration estimation, and HLA typing were performed to profile the tumor immune microenvironment and patient-specific immunogenomic features. Results: The assay was applied to a clinical cohort of 24 urothelial carcinoma patients with paired pre- and post-treatment FFPE samples to identify transcriptional changes associated with therapeutic response. With 44 out of 48 samples achieving a unique alignment rate >80% and an average exonic rate of 86.9%, the assay demonstrates robust capture of high-quality RNA molecules for expression estimation. Comparative expression analysis revealed five immune-related genes-CD80, OASL, IDO1, HAVCR2 (TIM-3), and GZMB-that were significantly downregulated in responders (p < 0.01). These genes have been reported to be immune-associated regulators that reduce immune activation and shift immune contexture during effective therapy. We observed a significant 10% decrease in tumor purity following treatment in responders (p = 0.03), while no significant changes were observed in non-responders. These findings implicate a role for the tumor microenvironment in treatment efficacy. The patient-specific HLA typing analysis demonstrated substantial inter-individual variability in antigen presentation capacity, underscoring its relevance for guiding personalized immunotherapy strategies. Conclusions: PredicineWTS enables highly sensitive, low-input whole-transcriptome profiling of FFPE tumor tissues, supporting robust immunogenomic characterization in clinical research settings. In this clinical cohort, decreased expression of five key immune genes and differential immune infiltration patterns were associated with treatment response. Integration of transcriptomic profiling, immune landscape assessment, and HLA typing provides a comprehensive framework for immunotherapy biomarker discovery and personalized treatment planning.