Haybittle J L
MRC Cancer Trials Office, Cambridge, UK.
Clin Oncol (R Coll Radiol). 1998;10(2):92-4. doi: 10.1016/s0936-6555(05)80485-6.
Improvements in life expectancy could be a more readily appreciated measure of benefit from a clinical trial than relative risks, odds ratios or increases in survival rate at some arbitrary point in time. Parametric models of survival experience can be used to determine differences in life expectancy. Using the log-normal model, it is shown that the increases in 10-year survival rate found by the overviews of adjuvant systemic therapy trials in early breast cancer are consistent with only small overall increases in life expectancy of about 1 year for Stage I and about 2 years for Stage II. However, if adjuvant therapy transfers a patient from being not cured to being cured, then her life expectancy will have been improved by 16 years for Stage I disease and by 21 years for Stage II. Model analyses on large data sets, such as are available in overviews, could possibly provide some evidence on whether the effect of adjuvant systemic therapy is to increase the cure rate, with the consequent considerable increase in life expectancy for some patients, or whether the effect is only a small increase in life expectancy for those who are not cured.
与相对风险、比值比或某个任意时间点的生存率增加相比，预期寿命的改善可能是一种更容易理解的衡量临床试验益处的指标。生存经验的参数模型可用于确定预期寿命的差异。使用对数正态模型表明，早期乳腺癌辅助全身治疗试验综述中发现的10年生存率增加仅与预期寿命总体小幅增加一致，I期约为1年，II期约为2年。然而，如果辅助治疗将一名患者从不治愈转变为治愈，那么对于I期疾病，她的预期寿命将提高16年，对于II期疾病将提高21年。对大型数据集进行模型分析（如综述中可获得的数据）可能会提供一些证据，证明辅助全身治疗的效果是提高治愈率，从而使一些患者的预期寿命大幅增加，还是仅使那些未治愈患者的预期寿命小幅增加。
