BACKGROUND: The ratio of serum apolipoprotein B (apoB) to apolipoprotein A-I (apoAI) had been reported as a prognostic factor in colorectal cancer. This retrospective study aimed to assess the implication of apoB-to-apoAI ratio in predicting liver metastasis from rectal cancer (RC). METHODS: The clinical data of 599 locally advanced RC patients treated with chemoradiotherapy followed by surgery were reviewed. Serum apoAI, apoB and apoB-to-apoAI ratio were analyzed for their correlation with the liver-metastasis-free, other-metastasis-free and overall survivals, together with the pretreatment and postsurgical pathoclinical features of the patients. Univariate and multivariate survival analyses were realized through the Kaplan-Meier approach and Cox model, respectively. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated for independent predictors. RESULTS: Carbohydrate antigen 19 - 9 ≥ 26.3 U/ml, apoB-to-apoAI ratio ≥ 0.63, tumor regression grade 5 - 3, pT4 and pN + stage emerged as independent predictors of poorer liver-metastasis-free survival. The hazard ratios were 1.656 (95% CI, 1.094-2.506), 1.919 (95% CI, 1.174-3.145), 1.686 (95% CI, 1.053-2.703), 1.890 (95% CI, 1.110-3.226) and 2.012 (95% CI, 1.314-2.077), respectively. Except apoB-to-apoAI ratio, the other 4 factors were also independent predictors of poorer other-metastasis-free and overall survivals. And the independent predictors of poorer overall survival also included age ≥ 67 years old, distance to anal verge < 5 cm. CONCLUSIONS: Serum apoB-to-apoAI ratio could be used as a biomarker for prediction of liver metastasis risk in locally advanced RC.