JMIR Publications

ABSTRACT

Background:

Prognostic genes or gene signatures have been widely used to predict patients’ survival and aid the decision of therapeutic options. Although some web-based survival analysis tools have been developed, they have several limitations.

Objective:

To overcome these, we developed ESurv, an easy, effective, and excellent tool that can perform advanced survival analyses of data from The Cancer Genome Atlas (TCGA) or directly from users. Users can conduct univariate analyses and grouped variable selections using multi-omics data in TCGA.

Methods:

We used R program to code survival analyses based on multi-omics data from TCGA. To perform analyses, we excluded patients and genes that have insufficient information. Clinical variables were classified as ‘0’ and ‘1’ when there were two categories (for example, chemotherapy: ‘no’ or ‘yes’), and dummy variables were applied when the categories were three or more (for example, laterality: ‘right’, ‘left’, or ‘bilateral’).

Results:

In univariate analyses, they can identify prognostic significances of single gene with survival curve (median or optimal cut-off), area under the curve (AUC) with C statistics, and receiver operating characteristics (ROC). They can obtain prognostic variable signatures based on multi-omics data with clinical variables by using grouped variable selections (LASSO, Elastic Net, and Net) with the above results. In addition, users can make custom gene signatures for specific cancers with genes of interest. One of the most important functions is that users can perform all survival analyses using their own data.

Conclusions:

With advanced statistical techniques suitable for high dimensional data such as genetic data and integrated survival analysis, ESurv (https://easysurv.net) overcomes all the limitations of previous web-based tools. ESurv will help biomedical researchers easily perform survival analysis.