Background and Objectives: Serum carcinoembryonic antigen (CEA) is a widely used biomarker in non-small cell lung cancer (NSCLC). However, its association with oncogenic driver alterations and prognostic significance across molecular subtypes in metastatic disease remains insufficiently defined. Materials and Methods: This retrospective multicenter study included 332 patients with metastatic NSCLC harboring oncogenic alterations (EGFR, ALK, ROS1, KRAS, and others) from eight oncology centers in Türkiye. Baseline serum CEA levels measured at metastatic diagnosis were analyzed on the natural logarithmic scale. Associations between CEA levels, molecular subtypes, clinical features, and overall survival (OS) were evaluated using generalized linear models and Cox proportional hazards regression. Results: Baseline CEA levels differed significantly across molecular subtypes (p = 0.001), with EGFR-mutant tumors showing the highest median levels. Multivariable analysis identified driver alteration, histology, and metastatic burden as independent determinants of baseline CEA. Higher baseline CEA and metastatic site count were independently associated with increased mortality risk (HR 1.151 and 1.279 per unit increase, respectively; p < 0.001), while female sex was protective (HR 0.626; p = 0.004). KRAS mutations were associated with poorer survival compared with EGFR (HR 2.370; p < 0.001). Kaplan–Meier analyses showed a consistent trend toward longer OS in patients with CEA < 5 ng/mL, with significance only in the rare alteration subgroup. Conclusions: Baseline CEA may reflect underlying tumor biology across molecular subtypes and are associated with survival outcomes in metastatic NSCLC. However, given the variability across subgroups and modest effect sizes, these findings should be interpreted with caution. Prospective studies evaluating longitudinal CEA dynamics are warranted.

Background: National trends in sepsis-associated cancer mortality in the USA remain incompletely characterised, particularly regarding long-term structural shifts and their variation across oncologic subpopulations.Methods: We conducted a population-based time-series analysis using the CDC WONDER Multiple Cause of Death database (1999–2020). Cases were defined by an underlying cause of death of malignant neoplasm (ICD-10 codes C00–C97) and a contributing cause of sepsis (A40–A41). We calculated age-standardised mortality rates (ASMR) indexed to the 2000 US Standard Population. Temporal dynamics were assessed using segmented Poisson regression to estimate annual percent changes (APC) and identify structural breakpoints. Furthermore, we employed counterfactual modelling to quantify excess mortality by projecting pre-2010 trends forward, and performed stratified analyses by cancer type (solid vs haematological), sex, and geography.Findings: Of 14 713 914 cancer deaths, 455 369 (3·10%) involved sepsis. Crude mortality increased from 6·26 to 7·94 per 100 000. Segmented regression identified a structural acceleration in the early 2010s, with APCs increasing from 0·47% (95% CI 0·36–0·57) to 2·84% (2·54–3·13). This inflection corresponded to an estimated 31 896 excess deaths compared with historical trends. Crucially, trends diverged by cancer type: mortality declined for haematological malignancies (APC −1·97%; 95% CI −2·21 to −1·73) but surged for solid tumours (APC 2·25%; 95% CI 2·10–2·39). Population ageing accounted for 77·1% of the post-inflection increase; COVID-19 contributions were transient.Interpretation: A structural acceleration in sepsis-associated cancer mortality began in the early 2010s, driven largely by solid tumours and exceeding demographic expectations. This identifies sepsis as a critical, structural determinant of cancer outcomes, necessitating prioritisation within oncologic survivorship guidelines and health system infrastructure.