Spread through air spaces (STAS) is defined as tumour cells identified beyond the main tumour margin within adjacent alveolar spaces, appearing as micropapillary clusters, solid nests or single cells. First described in 2015, STAS is a reproducible histologic feature with established adverse prognostic impact in lung adenocarcinoma (LUAD). Its prognostic relevance has prompted the International Association for the Study of Lung Cancer (IASLC) to recommend incorporation of STAS as a histologic descriptor in the proposed ninth edition of the TNM classification. Prior large-stage analyses showed STAS associated with worse survival after adjustment for age, sex, grade and lymphovascular invasion, but these prior series lacked comprehensive molecular data. Reported relationships between STAS and oncogene status, including EGFR, have been inconsistent in previous studies.
This study focused on patients with AJCC8 stage 1 LUAD from a single centre (National Cancer Centre Singapore) with at least 2 years of follow-up after surgical resection and known EGFR mutation and STAS status. All cases were confirmed adenocarcinoma on surgical resection by dedicated thoracic pathologists. STAS assessment followed the 2015 WHO classification criteria and was performed within routine institutional histopathologic reporting; pathologists were not blinded to clinical information. EGFR mutation testing was performed prospectively using Cobas, Sanger sequencing and/or next-generation sequencing per institutional standards. PD-L1 expression was assessed by immunohistochemistry (SP263) and reported as tumour proportion score (TPS).
Fresh-frozen tumour samples and matched normal controls underwent whole-exome sequencing (WES) and RNA sequencing. WES was performed at a mean coverage depth of approximately 400× for tumour samples and 100× for matched normals. RNA sequencing targeted a depth of 50 million paired-end reads per sample. Transcriptomic subtypes were assigned as terminal respiratory unit (TRU), proximal proliferative (PP), or proximal inflammatory (PI) using previously described classification schemes. Sequencing resources were prioritised for EGFR-mutant tumours as part of an ongoing translational programme, contributing to a higher proportion of EGFR-mutant cases in the sequenced subset. Further technical details are available in the article’s supplementary appendix.
Among 378 patients in the cohort, 54.2% had STAS+ tumours. The incidence of STAS was similar between EGFR-mutant and EGFR-wildtype LUAD. STAS+ status was significantly associated with several adverse clinicopathologic features: higher frequency of stage 1B disease, presence of lymphovascular invasion (LVI), higher histologic grade, micropapillary or solid predominant adenocarcinoma histologic subtypes, and PD-L1 TPS ≥1%.
Disease-free survival (DFS) was analysed using Kaplan–Meier estimation and Cox proportional hazards modelling, with DFS defined from surgery to recurrence or death. In multivariable analysis adjusting for relevant covariates, STAS+ status remained independently associated with inferior DFS (hazard ratio 2.32, 95% CI 1.16–4.63; p = 0.017), indicating that presence of STAS confers prognostic information beyond standard clinicopathologic factors in stage 1 LUAD.
At the genomic level, STAS+ tumours were enriched for TP53 co-mutations and features of genomic instability, including whole genome doubling. These observations indicate a molecular phenotype in STAS+ tumours characterised by increased genomic alteration burden and instability. The enrichment of TP53 alterations aligns with prior reports in broader stage cohorts, although earlier series showed variable results depending on gene panels and stage composition.
Transcriptomic profiling showed that STAS+ tumours were more frequently classified into non-TRU transcriptional subtypes (PP or PI) and demonstrated upregulation of cell cycle–related pathways. These transcriptomic signals were directionally consistent when analyses were stratified by EGFR mutation status, suggesting conserved biological processes associated with STAS across oncogenic subsets.
Integrated genomic and transcriptomic data from this stage 1–restricted LUAD cohort support that STAS+ tumours represent a distinct biological subset characterised by genomic instability, TP53 co-mutations and cell cycle dysregulation. These molecular features provide a plausible biological basis for the observed adverse prognosis and higher recurrence risk associated with STAS, and they may inform risk stratification and decisions about adjuvant therapy selection in early-stage disease.
The study reports sequencing and analytic approaches and references supplementary materials for additional technical detail. Pathologists assessing STAS were not blinded to clinical information, which may introduce assessment bias. The article notes sequencing was prioritised for EGFR-mutant tumours, accounting for higher EGFR representation among sequenced samples. The source text states that further details are provided in the Supplementary Appendix; any additional specific numerical breakdowns of genomic alterations beyond those reported in the abstract or full text excerpt (for example per-subgroup mutation rates or gene-level frequencies) were not reported in the provided source text.