Tumor-draining pulmonary vein–derived small extracellular vesicle GCC2 as a prognostic biomarker in resected lung adenocarcinoma

Accurate prognostic biomarkers are needed to improve postoperative risk stratification in patients with lung adenocarcinoma (LUAC). Because peripheral blood (PB) contains heterogeneous small extracellular vesicles (sEVs) derived from multiple tissues, tumor-draining pulmonary vein (TDPV) blood may provide a tumor-proximal source of sEV-based biomarkers. We investigated whether TDPV-derived sEV-GRIP and coiled-coil domain-containing 2 (GCC2) were associated with postoperative outcomes, compared with PB-derived biomarkers, in patients undergoing surgical resection. This single-center retrospective exploratory study analyzed paired TDPV and PB samples from 70 patients with LUAC who underwent curative resection between June 2015 and October 2021, together with PB samples from 50 cancer-free reference participants. sEV-GCC2, sEV-CD63, serum carcinoembryonic antigen (CEA), and circulating GCC2 were measured. The analyses evaluated cancer-versus-control discrimination, pathological stage migration within the LUAC cohort, and recurrence-related outcomes, including time to recurrence (TTR) and recurrence-free survival (RFS). The mean follow-up duration was 33.31 months. Patients with LUAC had higher sEV-GCC2 concentrations in the TDPV than in paired PB samples ( p = .0004). In the analysis comparing patients with LUAC and cancer-free reference participants, the area under the curve (AUC) for TDPV sEV-GCC2 was 0.899, compared with 0.819 for PB sEV-GCC2 ( p = .0068). For pathological upstaging within the LUAC cohort, the AUC was 0.583 for TDPV and 0.460 for PB, with no significant difference between the two sampling sites (ΔAUC, 0.123; 95% confidence interval [CI], − 0.048 to 0.294; p = .158). High TDPV sEV-GCC2 concentration was associated with shorter TTR (log-rank p = .022) and RFS (univariable hazard ratio [HR], 5.653; 95% CI, 1.863–17.155; p = .002). Neither PB nor TDPV sEV-GCC2 remained significant after adjustment for pTNM stage and postoperative adjuvant therapy, whereas pTNM stage and adjuvant therapy were associated with both TTR and RFS in the adjusted models. TDPV-derived sEV-GCC2 showed greater discrimination between patients with LUAC and cancer-free reference participants than PB-derived sEV-GCC2, suggesting a potential advantage of tumor-proximal sampling. Although associated with TTR and RFS in univariable analyses, TDPV sEV-GCC2 did not outperform paired PB for pathological upstaging or add prognostic information beyond pTNM stage. Prospective studies are needed to determine whether TDPV sEV-GCC2 levels can provide clinically useful information in patients with resected LUAC. ClinicalTrials.gov identifier NCT05587114 (registered on October 19, 2022, retrospectively registered).