Objective To investigate the prognostic factors of hepatocellular carcinoma (HCC) and develop a Nomogram prediction model based on clinical data and miR-497-5p expression levels to improve disease-free survival (DFS) after surgery.

Methods This study utilized the GSE31384 dataset (comprising miRNA expression profiles and clinical data from 166 HCC patients), followed by quantile normalization. Key miRNAs were identified through Kaplan-Meier survival analysis. In this retrospective study, 231 patients with pathologically-confirmed HCC treated at Sun Yat-sen University Cancer Center from 2001 to 2009 were included. They were randomly divided into a training cohort (n = 162) and an internal validation cohort (n = 69) at a 7∶3 ratio. Additionally, 328 HCC patients admitted to Shanghai Tenth People’s Hospital from 2004 to 2014 were included as an external validation cohort. All patients underwent hepatectomy. Baseline clinical data, biochemical test results and 24-month follow-up outcomes (until April 30, 2025) were collected. The expression levels of miR-497-5p were measured using quantitative reverse transcription polymerase chain reaction (qRT-PCR). Univariate and multivariate Cox’s analyses were conducted to identify prognostic factors, and a Nomogram prediction model was subsequently constructed. The performance of the model was evaluated using calibration curves, decision curve analysis (DCA), receiver operating characteristic (ROC) curves, and Kaplan-Meier survival curves.

Results  Forty-nine differentially-expressed miRNAs (17 upregulated, 32 downregulated) were identified in the GSE31384 dataset (166 HCC cases) using quantile normalization and Kaplan-Meier survival analysis (P ＜ 0.01). The low expression of miR-497-5p was significantly associated with poor prognosis in HCC patients (P ＜ 0.001). qRT-PCR further confirmed its expression level in HCC tissues was significantly lower than that in paired adjacent non-tumor tissues (P ＜ 0.01). Multivariate Cox’s analyses identified multiple tumors (HR = 1.939, 95%CI: 1.258-2.989, P = 0.003) and maximum tumor diameter ≥ 5 cm (HR = 2.219, 95%CI: 1.125-4.379, P = 0.021) as the independent risk factors for DFS in HCC patients, while high expression of miR-497-5p (HR = 0.644, 95%CI: 0.432−0.959, P = 0.030) was a protective factor for DFS. Prediction model Y = −0.105 × log₂ (miR-497-5p relative expression) + 0.767 × number of tumors + 0.961 × maximum tumor diameter. Calibration curves and DCA demonstrated good calibration performance and clinical utility. DeLong's test with Bonferroni correction results demonstrated that the model exhibited significantly superior predictive performance. Using the median risk score (1.176) of the training cohort as the cutoff value, Kaplan-Meier analysis demonstrated that 5-year DFS and OS rates of high-risk patients in the training cohort, internal and external validation cohorts were significantly lower compared with those in low-risk counterparts.

Conclusions The expression level of miR-497-5p, number of tumors, and maximum tumor diameter are the independent predictive factors for DFS in HCC patients. The Nomogram prediction model based on these indicators can effectively assess clinical prognosis of HCC patients.