Advanced Colorectal Cancer: A Meta-Analysis of 25 Trials

This dataset comprises 3,791 patients enrolled in 25 randomized trials. The data were collected and checked by the Meta-Analysis Group in Cancer between 1990 and 1996 to confirm the benefits of experimental fluoropyrimidine treatments in advanced (metastatic) colorectal cancer (Buyse et al., 2000). The variables collected for every patient randomized in each of these 25 trials consisted of baseline clinical characteristics (patient identification, eligibility, date of randomization, age, sex, performance status, primary tumor site, site of metastases), treatment allocated by randomization, tumor response, duration of response (if applicable), date of death or last visit, survival status, and cause of death (if applicable). Four meta-analyses were carried out and published using these data (Buyse et al., 2000). In all four meta-analyses, the comparison was between a control treatment (Z = 0) and an experimental treatment (Z = 1). The control treatments were similar across the four meta-analyses and consisted of fluorouracil or floxuridine given as a bolus intravenous injection. The experimental treatments differed across the four meta-analyses: it was either fluorouracil modulated by leucovorin, fluorouracil modulated by methotrexate, fluorouracil given by continuous infusion,


Five Clinical Trials in Schizophrenia. Scatter plots of the change in the CGI scores against the change in the BPRS scores (upper left figure), the change in the CGI scores against the change in the PANSS scores (upper right figure), and the change in the BPRS scores against the change in the PANSS scores (bottom figure). The full and dashed lines result from regressing the endpoint on the Y-axis on the endpoint on the X-axis in the groups that were administered risperidone and the active control, respectively. Jitter (a small amount of random noise) was added to the change in the CGI scores to prevent overplotting of the data points.

Note: CGI = Clinical Global Impression, BPRS = Brief Psychiatric Rating Scale, and PANSS = Positive and Negative Syndrome Scale.


Advanced Ovarian Cancer Trials. Progression-free survival (S) and overall survival (T) curves (left and right panel, respectively) combining the data of the four available studies.

or hepatic-arterial infusion of floxuridine for patients with metastases confined to the liver.

In the meta-analysis of 25 trials, the response rate was 22% with the experimental treatments vs. 12% with the control treatments, a highly statistically significant benefit (odds ratio = 0.48, 95% CI [0.40; 0.57], p < 0.0001). The experimental treatments also had a significant effect on overall survival, though the magnitude of the benefit was modest, with an absolute difference of a few percentage points at all times between the Kaplan-Meier survival estimates (hazard ratio = 0.90, 95% CI [0.84; 0.97], p = 0.003).

For the surrogacy analyses (Buyse et al., 2000; Burzykowski, Molenberghs, and Buyse, 2004), the true endpoint (T) is survival time, calculated from the day of randomization to the day of death irrespective of the cause of death. The candidate surrogate endpoint (S) is tumor response. Tumor response is commonly defined as a categorical variable with four ordered categories: complete response (CR), partial response (PR), stable disease (SD) and progressive disease (PD). CR is defined as the disappearance of all detectable tumor. PR is defined as a decrease of 50% or more in the tumor surface area measured by CT scan (sum of the products of the largest perpendicular diameters of all measurable disease), without appearance of new lesions and a duration of at least 4 weeks. SD is defined as a decrease of less than 50% or an increase of less than 25% in the tumor surface area, without new lesions. PD is defined as an increase of more than 25% in the tumor surface area or the appearance of any new lesion. The surrogacy analyses are carried out with


Advanced Ovarian Cancer Trials. Overview of the variables in the dataset.

Variable name



The identification number of a patient


The center in which a patient was treated


The treatment indicator, coded as 0=CP and 1=CAP


Progression-free survival (the candidate surrogate)


Censoring indicator for progression-free survival


Survival time (the true endpoint)


Censoring indicator for survival time

response considered a binary outcome: S = 1 for patients with CR or PR and S = 0 for patients with SD or PD.

In the remainder of this book, this dataset will be referred to as the Col- orectal25 dataset. The dataset is included in the R library Surrogate (where it can be accessed using the command data(Colorectal25)) and it can be downloaded from (file Colorectal25.txt). Table 2.5 provides an overview of the variables that are included in the Col- orectal25 dataset. The data were provided by the Meta-Analysis Group In Cancer (see Buyse et al., 2000).

< Prev   CONTENTS   Source   Next >