| Summary: | 博士 === 國立臺灣大學 === 流行病學研究所 === 92 === Background： Rarely has been addressed as to whether risk of death for small hepatocellular carcinoma (HCC) treated by percutaneous ethanol injection increases with time and how prognosis of these small tumours is affected by pretreatment variables and a constellation of time-varying predictors. In addition to using death as primary endpoint, intermediate endpoint based on relapse of tumor after the application of PEI has become the core of subject for clinical surveillance of small HCCs treated by PEI. The disease process in terms of remission-relapse may also increase the cost of PEI, which calls attention to cost-effectiveness analysis for the comparison between PEI and surgery.
Aims： We aimed to develop a comprehensive clinical surveillance system for evaluation of primary endpoint in mortality in association with pre-treatment variables allowing for non-constant hazard rate and dynamic change of time-varying predictors, for evaluation of intermediate endpoint based on relapse of tumour, and for cost-effectiveness analysis of comparing two treatment modalities.
Methods: A total of 108 patients with hepatocellular carcinoma smaller than five cm in diameter treated by percutaneous ethanol injection with or without transcatheter arterial chemo/embolization were recruited. All patients’ liver function reserve were classified as Child-Pugh A (n=84) or B (n=24) at diagnosis. Pre-treatment variables and series of laboratory data and clinical assessments were retrieved from medical records of surveillance and treatments. Logarithm of hazard rate (per month) with time since therapy was assessed. The Weibull model was used to elucidate the effect of pre-treatment clinico-pathological variables on prognosis.The time-dependent scoring system for the prediction of risk for death was developed on the basis of time-dependent Cox regression model. A three-state continuous Markov model was constructed to estimate the transition parameters.
Results:
Part I: The rate of death increased by 4.7% (95% CI: 3.7-5.7%) per month since treatment. Child-Pugh B status was associated with a 2.8-fold risk (95% CI: 1.52-5.16) of death. Those with a high level of AST or alcoholic cirrhotics had a two-fold risk (95% CI: 1.14-3.42) for death from HCC.
Part II: The time trends of some predictors parallels cumulative survival of small hepatocellular carcinoma cases. Higher serum alpha-fetoprotein level was identified the most significant time-dependent predictor. Other significant predictors included asparatate transminase, bilirubin, alkaline phosphatase, albumin, and prothrombin time. Time-dependent surveillance scoring system shows the cutoff points of scores for 6-month, 1-year, and 2-year survival were 48, 39, and 36, respectively; with the corresponding estimates of sensitivity, 100%, 85%, 74%, and estimates of specificity, 97%, 96%, 88%, respectively.
Part III: Prompt initial remission after curative treatment was achieved in 57.19% (95% CI: 39.82%-74.56%), treatment modality was significant associated with initial remission. Age and tumour morphology were two significant predictors. An increased one unit of age (in year) led to 3% extra risk for relapse (OR=1.03, 95% CI:1.00-1.06), which was slightly larger than 2% additional chance of returning to remission. Multiple nodules or large tumour had 3.7-fold (95% CI: 1.99-6.86) risk for relapse, which was also greater than 2.22-fold (95% CI: 1.06-4.64) chance of remission as compared with solitary nodule or small tumour. Significant time-varying factors included level of AFP, level of albumin, level of globulin, level of AST greater than 100 U/L, high level of bilirubin, high level of ALKP, and longer prothrombin time.
Part IV: Patients treated by PEI have better survival than those treated with surgery before 45 months post treatment. Cumulative survival of PEI treated patients become worse later with 5-year survival of 19% in comparison with 33% in surgical treated patients. Average cost-effectiveness ratios for surgery and non-surgery were 18,748.45 NTD and 19,972.22 NTD per life-month saved. The incremental cost-effectiveness ratio was 33,274 NTD.
Conclusion: Systematic evaluation of clinical surveillance of small HCCs treated by percutaneous ethanl injection was developed in this thesis. This system is very useful for clinical management of patients treated by percutaneous ethanl injection.
|
|---|
