Vascular invasion-associated gene expression is detectable in pre-surgical biopsies of stage I lung adenocarcinoma [Spatial Transcriptomics]
Microscopic vascular invasion (VI) predicts recurrence and benefit from lobectomy in stage I lung adenocarcinoma (LUAD) but cannot be accurately predicted before surgery. Thus, biomarkers that identify this aggressive tumor subset are needed. Here, we show that VI in stage I LUAD is associated with reproducible gene expression programs detectable beyond the invasive focus. Using bulk RNA sequencing of 162 resected tumors and spatial transcriptomics in a subset, we identify and characterize a VI-associated gene signature. A predictor derived from this signature is associated with VI and recurrence in an independent validation cohort and is robust to intra-tumor heterogeneity in multiregional sampling data. In a cohort of pre-surgical biopsies, predictor scores correlate with matched resections and show promising discrimination of VI. These findings indicate that VI-associated transcriptional changes extend across the tumor and are detectable in limited biopsy material, supporting further validation for preoperative risk stratification in stage I LUAD.
import lamindb as ln
import spatialdata as sd
coll = ln.Collection.get(key="geo-gse273378")
coll.ordered_artifacts.df() # the 16 members, ordered
# download every member .zarr to the local cache:
paths = [a.cache() for a in coll.ordered_artifacts.all()]
# …or load each straight into memory as a SpatialData object:
sdatas = {a.key: sd.read_zarr(a.cache()) for a in coll.ordered_artifacts.all()} Cite this collection
No linked publication found — please cite the data source.