166 Xenium-based in situ spatial transcriptomics identifies poor prognostic markers in relapsed angioimmunoblastic T-cell lymphoma

Introduction/Objective Angioimmunoblastic T-cell lymphoma (AITL) is a distinct subtype of peripheral T-cell lymphoma (PTCL), accounting for approximately 20% of PTCL cases. Most newly diagnosed AITL patients present with advanced-stage disease and exhibit poor prognosis, with a median survival of less than three years. Currently, no standardized treatment strategies exist for AITL, particularly for relapsed/refractory AITL (RR-AITL). Therefore, elucidating the underlying molecular mechanisms and identifying novel therapeutic targets for AITL, especially in RR-AITL, is critically needed. Methods/Case Report This study included ten formalin-fixed, paraffin-embedded (FFPE) samples, comprising five non- relapsed/refractory AITL (NR-AITL, defined as those without recurrence or progression and who are still alive within three years) and five RR-AITL specimens. RR-AITL was defined as either disease relapse within one year after achieving complete response to first-line therapy or failure to attain partial response after standard first-line treatment. We applied Xenium in situ technology to profile the tumor microenvironment (TME) while preserving spatial architecture. Cell type annotation was performed using the Seurat package. Proliferation scores were computed via the AUCell package based on the gene signature proposed by Yang et al., while stemness scores were evaluated using CytoTRACE to assess T-cell differentiation status. Differentially expressed genes (DEGs) in each follicular helper T-cell (TFH) cluster were identified using the FindAllMarkers function in R software(v4.1.2). Results We delineated ten T-cell clusters, including four TFH subsets (TFH-1, TFH-2, TFH-3, and TFH-4). Among these, TFH-3 exhibited the highest proliferation and stemness scores. Notably, the proportion of TFH-3 cells showed a marked increase in ND-AITL samples compared to RR-AITL, implying its possible association with disease progression or treatment resistance. Comparative analysis of TFH-3 against other TFH clusters revealed NEIL3 as a uniquely upregulated gene. To validate the clinical relevance of our findings, we interrogated publicly available PTCL datasets (GSE58445). Consistent with our results, high NEIL3 expression correlated with significantly worse overall survival (OS, p = 0.027, log-rank test). Conclusion Our in situ spatial transcriptomic analysis highlights the heterogeneity of the TME between NR-AITL and RR-AITL. The TFH-3 subset, which is predominant in RR-AITL, demonstrates enhanced proliferative capacity and stem-like properties, suggesting a role in disease progression. Furthermore, NEIL3 emerges as a potential prognostic marker and therapeutic target for RR-AITL.