<p>Background: High-dose methotrexate (HD-MTX) serves as a cornerstone treatment for central nervous system lymphoma. However, its clinical utility is limited by delayed drug excretion and associated adverse events. The methylenetetrahydrofolate reductase (MTHFR) C677T polymorphism, a critical genetic variant influencing MTX metabolism, has been proposed by some experts as an indicator for HD-MTX dose reduction due to its potential association with delayed excretion. Objective: This study aimed to identify clinical risk factors for delayed MTX excretion and evaluate the predictive value of MTHFR C677T polymorphism in HD-MTX treatment cycles for CNS lymphoma patients. Methods: We conducted a retrospective analysis of 213 HD-MTX treatment cycles in CNS lymphoma patients (derived from 98 patients) treated at our institution between May 2015 and December 2023. Each patient received a median of 3 HD-MTX cycles (range 1–9 cycles). Treatment cycles were stratified into delayed excretion (n = 66) and normal excretion (n = 147) groups based on international consensus guidelines for pharmacokinetic monitoring of serum MTX concentrations. MTHFR C677T genotypes were analyzed using PCR-RFLP with Sanger sequencing validation (10% random sample). Clinical parameters were compared between groups; univariate logistic regression (retention threshold: P &lt; 0.10) was used for initial risk factor screening, and a mixed-effects logistic regression model with a random intercept for individual patients (retention: P &lt; 0.05) was applied for identifying independent risk factors to correct for non-independence of repeated measures. Results: Univariate analysis revealed four potential risk factors (P &lt; 0.1): advanced age (OR = 1.03), MTHFR C677T CT/TT (OR = 1.73), hypoalbuminemia (&lt; 35 g/L) (OR = 3.42), and acidic urine pH (&lt; 7.0) (OR = 4.15). Mixed-effects multivariate analysis identified hypoalbuminemia (adjusted OR = 3.02, 95%CI 1.56–5.85) and acidic urine pH (adjusted OR = 4.15, 95%CI 2.17–7.94) as independent predictors of delayed excretion (P &lt; 0.05), with a random effect variance of σ²=0.21 indicating mild individual intervariability. The MTHFR C677T CT/TT genotype showed no statistically significant association in the adjusted model (P = 0.095). A total of 35.1% (13/37) of patients with CT/TT genotype exhibited delayed clearance in initial cycles (1–2) but normalized excretion in subsequent administrations (cycles 3–6, involving 42 treatment cycles). The median number of HD-MTX cycles administered was 4 (range 1–8) in patients with at least one delayed excretion cycle and 5 (range 1–9) in patients with all normal excretion cycles.Conclusion: Our findings demonstrate that urine acidification and hypoalbuminemia, rather than MTHFR C677T polymorphism, constitute primary risk factors for HD-MTX excretion delay in treatment cycles. These results suggest that optimal management should prioritize urinary alkalinization (pH ≥ 7.0) and albumin level maintenance over genotype-guided dose reduction. These modifiable factors present actionable targets for improving HD-MTX safety profiles in CNS lymphoma treatment.</p>