Diagnostic and prognostic performance of urinary metabolomic biomarkers in lung cancer: A systematic review and meta-analysis.

e20551 Background: Urinary biomarkers offer a non-invasive strategy for the detection and prognostication of lung cancer; however, reported diagnostic and prognostic performance varies substantially across biomarker classes, analytical platforms, and modeling strategies. This systematic review and meta-analysis aimed to synthesize the evidence on the diagnostic accuracy and prognostic value of urinary biomarkers in lung cancer. Methods: A systematic literature search was conducted in PubMed, Embase, and the Cochrane Library (2000–2026). Eligibility criteria and data extraction were guided by the PICO framework, encompassing adult lung cancer populations (NSCLC/SCLC) and appropriate control groups, urinary metabolomic profiling using liquid chromatography–mass spectrometry (LC-MS), gas chromatography–mass spectrometry (GC-MS), or nuclear magnetic resonance (NMR) platforms, and outcomes including diagnostic accuracy metrics and prognostic hazard ratios. Study screening was conducted using Rayyan, and random-effects meta-analyses were performed in R (version 4.4.2). Pooled diagnostic estimates were calculated using logit-transformed proportions, with heterogeneity quantified using the I² statistic. Prognostic hazard ratios were pooled on the logarithmic scale. Results: Eleven studies comprising 4,589 participants, including lung cancer cases (n = 2,019) and controls (n = 2,570), were included; histology was predominantly NSCLC or mixed populations, and staging information (available in nine studies) most commonly encompassed stage I–IV disease. Eight studies were diagnostic, two assessed both diagnostic and prognostic performance, and one was exclusively prognostic. Pooled diagnostic estimates showed a sensitivity of 89.2% (95% CI: 85.8–91.8%; I² = 26.2%), specificity of 86.6% (95% CI: 70.8–94.5%; I² = 97.1%), and overall accuracy of 86.0% (95% CI: 74.7–92.7%; I² = 90.8%). The pooled area under the curve (AUC) was 0.89 (95% CI: 0.85–0.94; I² = 93.8%), indicating strong discriminative performance. Prognostic meta-analysis of overall survival, based on two studies, yielded a pooled hazard ratio (HR) of 2.04, suggesting poorer outcomes in biomarker-defined high-risk groups, although confidence intervals were wide due to limited data. One additional study reported progression-free survival with similar adverse prognostic associations. Internal validation was performed in all studies, while normalization was reported in 94% of studies. Conclusions: Urinary biomarkers demonstrate clinically relevant diagnostic performance for lung cancer and emerging prognostic value for overall and progression-free survival. Harmonized analytical pipelines, predefined thresholds, and independent external validation are essential for successful clinical translation.