RNA-sequencing on EUS-acquired tissue of pancreatic cancer is feasible and allows identification of biomarkers of prognosis and of response to intensified chemotherapy

Aims There are limited molecular features to allow treatment personalization and outcome prediction of Pancreatic-Ductal-Adenocarcinoma(PDAC) patients. Two transcriptome subtypes have been identified: classical and basal-like. Classical PDACs show better prognosis and response to FOLFIRINOX, while basal-like display an opposite behaviour. This transcriptome-based classification is not straightforward and relies mainly on surgical series, with limited data on EUS-FNA samples.We have recently showed that the splicing factor QKI is a determinant of basal-like signature and that its evaluation together with GATA6 has strong prognostic capacity. Our aim was to investigate transcriptome subtypes and gene expression of specific biomarkers on baseline EUS-FNA samples of PDAC patients and their association with prognosis and response to chemotherapy.

Methods 85 PDAC patients who underwent EUS at diagnosis were enrolled in a prospective study and FNA samples investigated by RNA-seq. PURIST score was applied to define subtypes, Differential Gene Expression DGE analysis was performed and the prognostic value of QKI and GATA6 expression was investigated. The outcomes were overall survival(OS) and progression-free survival(PFS) corrected for age and stage.

Results RNA-seq was successful in 84/85 (98.8%) samples. According with PURIST, 4 patients were classified as basal-like (4.8%), 80 being classical. OS was lower in Basal-like vs classical patients(5.5 vs 13 months; p=0.01), with increased risk of death (HR 5.6; 95%CI:1.5-20.9; p=0.01). PURIST score also correlated inversely with OS (r=-0.4; p=0.0002). In 64 patients treated with first-line chemotherapy (FOLFIRINOX n=24; Abraxane-Gemcitabine, AG n=30; PAXG n=10) QKI expression correlated with OS(r=-0,36: p=0.0038).Among patients treated with FOLFIRINOX, median PFS was 7 months (IQR 2-9) in patients with a QKI/Gata6 ratio>1.13 and not reached in patients with ratio≤1.13 (HR=5.1;95%CI:1.2-21.5); QKI/GATA6 ratio was not associated with PFS in patients treated with the other regimens. At DGE analysis REG4 was the most enriched gene in patients with PFS>6 months; a cut-off for REG4 expression>111 was associated with response to AG (HR=0.2; 95%CI: 0,090-0,59; p=0.002).

Conclusions Transcriptome analysis of EUS-FNA samples is feasible and allows stratification of PDAC patients in terms of prognosis and chemosensitivity. Few selected classifiers such as QKI, GATA6 and REG4 may be sufficient to predict response to certain treatments. An additional wider cohort of samples is currently undergoing sequencing.

Publikationsverlauf

Artikel online veröffentlicht:
27. März 2025