Value of combined model based on FSIP1 gene methylation in early diagnosis of hepatocellular carcinoma
10.3760/cma.j.cn114452-20250102-00005
- VernacularTitle:基于FSIP1基因甲基化的联合模型在肝细胞癌早期诊断中的价值
- Author:
Suli YANG
1
;
Juan LI
;
Qiuchen QI
;
Peilong LI
;
Yan XIE
;
Dong SUN
;
Chuanxin WANG
;
Lutao DU
Author Information
1. 山东大学第二医院检验医学中心，济南 250033
- Publication Type:Journal Article
- Keywords:
Hepatocellular carcinoma;
Peripheral blood mononuclear cell;
DNA methylation;
Early diagnosis
- From:
Chinese Journal of Laboratory Medicine
2025;48(7):908-916
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To analyze the changes of DNA methylation in peripheral blood mononuclear cells (PBMC) of patients with hepatocellular carcinoma (HCC) and to evaluate the clinical value of a combined model based on FSIP1 gene methylation on the early diagnosis of HCC.Methods:This is a case-control study. From May 2023 to September 2024, 183 HCC patients and 155 healthy controls were collected in Qilu Hospital of Shandong University. The selected study subjects were divided into three cohorts: 14 HCC patients and 39 healthy controls formed the discovery cohort, a screening cohort consisted of 36 HCC patients and 39 healthy controls, 133 HCC patients and 77 healthy controls were included in the model construction cohort. 935k methylation chip analysis was used to identify specific differentially methylated sites in peripheral blood PBMC of the discovery cohort. The absolute value of the average methylation level difference between HCC group and healthy control group (|Δβ|) and P value were calculated. Then targeted bisulfite sequencing was used to verify the differentially methylated sites in the screening cohort. Finally, based on MethylTarget methylation sequencing technology, differential methylation sites were further verified in model construction cohort (divided into training set and validation set, training set consisted of 99 HCC patients and 57 healthy controls; validation set consisted of 34 HCC patients and 20 healthy controls). HCC early diagnosis model was constructed by random forest algorithm combined with clinical parameters and the diagnostic performance of the model was evaluated by receiver operating characteristic (ROC) curve in the validation set. Results:The total of 7 249 differentially methylated sites between HCC patients and healthy controls in discovery cohort were selected under the rule of |Δβ|≥0.06 and P<0.01. Among them, the cg02155073 site located on FSIP1 was hypermethylated in PBMC of HCC patients in the screening cohort and model cohort ( P<0.001). The AUC of HCC early diagnosis model (FmAP) based on FSIPI in the validation set was 0.967 (95% CI 0.924-1.000); sensitivity was 88%, specificity was 95%. The model had good diagnostic efficacy for patients with early HCC, stage Ⅰ-Ⅱ HCC AUC was 0.958 (95% CI 0.898-1.000). The FmAP model also had diagnostic value for tumor size <2 cm HCC and AFP negative HCC, with AUC of 0.955 (95% CI 0.898-1.000) and 0.964 (95% CI 0.934-0.994).The sensitivity were 92% and 93% and specificity both were 84%. Conclusion:The FmAP model based on FSIP1 gene methylation has good clinical value for the early diagnosis of hepatocellular carcinoma.
