Decreased Circulating Glutamine and Myo-Inositol in Cervical Cancer: Exploratory NMR Metabolomics and Internal Discrimination Using Their Additive Feature (MQ)

Background: Cervical cancer is accompanied by metabolic reprogramming, but circulating NMR metabolite signatures require rigorous statistical assessment before clinical interpretation. Methods: Serum metabolite concentrations were analyzed in 33 cervical cancer (CC) cases and 31 normal controls (NC). Group differences across 35 primary metabolites were assessed using Welch’s t-test with Benjamini–Hochberg false-discovery-rate (FDR) control; Mann–Whitney tests were used as distribution-robust sensitivity analyses. Age-adjusted linear models with HC3 robust standard errors were fitted for each primary metabolite (primary analysis: raw concentrations; sensitivity analysis: log2-transformed concentrations), with FDR correction across the 35 models. MQ was defined as [glutamine] + [myo-inositol]. ROC discrimination was summarized using stratified-bootstrap 95% confidence intervals, and paired DeLong testing compared MQ with glutamine. Results: Glutamine (CC 229.7 ± 102.2 μM; NC 434.8 ± 123.8 μM) and myo-inositol (CC 18.3 ± 8.8 μM; NC 35.8 ± 11.0 μM) were lower in CC, and both remained significant after age adjustment and FDR correction. Within CC, serum glutamine was strongly inversely correlated with tumor length (Spearman ρ = −0.94, n = 31; exploratory, requires independent replication). Apparent ROC AUCs were 0.906 for glutamine, 0.889 for myo-inositol, and 0.919 for MQ. The paired DeLong comparison of MQ versus glutamine was not statistically significant (AUC difference 0.0132; approximate 95% CI −0.0031 to 0.0295; p = 0.113). Repeated cross-validation is reported as within-cohort resampling performance and not independent validation. Conclusions: The findings support lower circulating glutamine and myo-inositol in this cohort. MQ is an exploratory additive feature without demonstrated superiority over glutamine alone; external validation and adjustment for additional clinical confounders are required.