BACKGROUND/AIM: A new scoring system [albumin-bilirubin-platelet (ALBI-PLT) score] was reported for identifying cirrhotic patients without high-risk varices (HRV), and patients with ALBI grade 1 (</=-2.60) and a platelet count over 150x10(9)/l were shown to have a low risk of having HRV. The present study modified the cut-off values of the variables in the ALBI-PLT score. PATIENTS AND METHODS: Among a total of 338 patients with chronic liver diseases, possible cut-off values of the ALBI score and the platelet count were determined by analyzing the first-half group (training cohort: N=169) with the receiver operating characteristic (ROC) method. The utility of the determined values was evaluated in the second-half group (validation cohort: N=169) and total cohort (N=338). In addition, the utility of the modified cut-off values was evaluated in patients with compensated cirrhosis (cirrhotic cohort: N=87). RESULTS: Possible cut-off values of the ALBI score and platelet count were found to be -2.36 and 114x10(9)/l, respectively. In the training cohort, these cut-off values provided a higher ratio of avoiding esophagogastroduodenoscopy than the original ALBI-PLT score (53.3% vs. 25.4%, p<0.01). Consistent results were observed in the validation cohort (28.4% vs. 15.4%, p<0.01), total cohort (40.8% vs. 20.4%, p<0.01), and cirrhotic cohort (32.2% vs. 11.5%, p<0.01). However, the missing ratio of patients with the HRV was not significantly increased in any cohort studied. CONCLUSION: Modification of the ALBI-PLT score may be useful for predicting patients without HRV.