This study aimed to quantify the full clinical courses characterized by alternating transitions between remission and relapses for small hepatocellular carcinomas treated by percutaneous ethanol injection, and to ascertain the significant predictors for the remission, relapse, and finally to metastasis or death.PATIENTS AND METHODS
A three-state Markov process was used to depict full clinical course. A total of 108 patients who underwent nonsurgical therapy as the first choice of treatment were derived from consecutive clinical series of patients registered in one medical center renowned for their use of percutaneous ethanol injection.RESULTS
We found that approximately 57.19% (95% confidence interval [CI]: 39.82%, 74.56%) patients were promptly responsive to initial treatment, whereas 43.81% had delayed response or were resistant to treatment. The rate of relapse (per month) was higher than the rate of remission (19.20% [95% CI: 15.36%, 23.04%] vs. 13.82% [95% CI: 10.72%, 16.93%]). The results from the multivariate analysis indicate that the significant predictors for the full clinical courses are total bilirubin, a-fetoprotein, prothrombin time, globulin, tumor morphology, and alkaline phosphatase.CONCLUSIONS
A three-state remission-relapse-death model was proposed to quantify and ascertain the predictors for the multistate disease progression of small hepatocellular carcinomas treated by percutaneous ethanol injection. This model captures the risk of recurrence of tumor as well as the primary endpoint of death.