nTerim Application Example

A study is being carried out to compare two-year survival after surgery for a brain tumour for patients treated with a new combination of chemotherapy and radiation, versus those treated with standard methods. We will use nTerim to figure out what sample size is required to have 90% power for comparing the two groups.

Data will be analysed using a log-rank test with a two-sided 5% significance level, assuming that the hazards are proportional. In previous studies with the standard method about 20% of patients survived for two years following surgery. The new therapy would be considered markedly better if it increases survival by about 15% from 20% to 35%.

The study is planned for four interim analyses, a six monthly check. The Pocock spending function is to be used.

nTerim provides graphical output for the boundary values used over the interim analysis (Figure 1) along with a plot of the estimated sample size for a given power (Figure 2). A summary statement is also automatically generated at the end of the analysis (Figure 3).

Figure 1: Plotted Boundaries Values

Figure 2: Power vs. Sample Size

Figure 3: Output Statement from nTerim