Chinese Hepatolgy ›› 2019, Vol. 24 ›› Issue (1): 20-23.

• Original Articles • Previous Articles     Next Articles

Transient elastography combined with alpha fetoprotein as an early-warning model for hepatocellular carcinoma

CHEN Song-hai1, JI Dong1, HU Zhi-jun3, LU Yin-ying2, WANG Chun-yan1, SHAO Qing1, CHEN Guo-feng1   

  1. 1. The Second Liver Cirrhosis Diagnosis and Treatment Center, The fifth medical center of the Chinese PLA General Hospital, Beijing 100039, China;
    2. Comprehaensive Liver Cancer Center, The fifth medical center of the Chinese PLA General Hospital, Beijing 100039, China;
    3. Department Ⅱ of spleen-stomach liver disease, Baoji traditional Chinese medicine hospital, Shanxi 721000, China
  • Received:2018-07-16 Online:2019-01-31 Published:2020-04-09
  • Contact: CHEN Guofeng, E-mail:guofengchen302@163.com; JI dong, E-mail: jidg302@126.com

PDF (PC)

384

Abstract: Objective To establish a noninvasive early-warning model for hepatocellular carcinoma (HCC) and improve the diagnosis accuracy of high-risk individuals as early as possible.Methods Patients with hepatitis B virus related-cirrhosis or HCC, who were admitted to our hospital from 2008 to 2017, were enrolled. They were followed up at least 3 times with an interval of 3 to 6 months. Liver stiffness measurement (LSM) by transient elastography was performed on the day enrolled. Alpha-fetoprotein (AFP) and other serum parameters were detected at each visit. HCC was diagnosed using multiphase-dynamic-contrast-enhanced magnetic resonance imaging scan. The noninvasive predictors for HCC were identified using logistic regression, and an early-warning model named ALA was constructed with age, LSM and AFP. The receiver operating characteristic (ROC) curve analysis was applied to evaluate the early warning effectiveness of ALA on HCC. Results A total of 321 consecutive patients were included in the study, who were randomly divided into 2 groups at the ratio of 2∶1 (200 cases in training group and 121 cases in validation group). The clinical parameters (age, sex, Child-Pugh score, liver function, AFP, LSM) showed no differences between the 2 groups. The age, percentage of HCC patients, percentage of Child-A patients, median of AFP and median of LSM in training and validation group was 51.42±10.10 and 49.96±10.23 years old, 34% and 35.5%, 65% and 57%, 12 ng/ml and 6 ng/ml, 25 kPa and 27kPa, respectively. Multivariable logistic regression showed that age, LSM and AFP were the independent risk factors for HCC. The ALA model was calculated using 0.103×age (years) +0.116×LSM (kPa) +0.026×AFP (ng/ml). The area under ROC curve of ALA was 0.91 (95% confidence interval, 0.86-0.95) in the training group. Using ALA of 9.61 as the cut-off value, the accuracy was 84.3% in the validation group.Conclusion ALA, a noninvasive early-warning model consisting of age, LSM and AFP, is worthy of clinical application, which might improve the early diagnosis rate of HCC and provide early curative interventions for HCC patients.

Key words: Hepatocellular carcinoma, Alpha-fetoprotein, Liver stiffness measurement, Transient elastography, Noninvasive, Early warning model