A dynamic data visualization tool provides nuanced clinical trial insights

For rare diseases, such as pediatric cancer, enrolling enough patients on a clinical trial to perform statistical analysis can be challenging. Therefore, scientists seek ways to gather more insights from the sparse data they do have. Traditional clinical trial reporting frequently includes a hazard ratio, which summarizes the relationship between two groups in the context of a time-to-event outcome, such as survival.

Hazard ratios are useful in analysis, but static. Contextualizing the numbers can be difficult, especially in understanding the complex interactions for patients belonging to multiple demographic groups. However, a common statistical model for the hazard ratio, the Cox regression model, also contains that complex information, though it has historically been too difficult to effectively show in published studies or share between investigators.

To make such analyses easier, St. Jude scientists developed a statistical package that automatically runs these analyses and presents them to users in a dynamic and clinically useful format. The software, which they named shinyCox, is designed for a standard statistical program, R, which has a framework for building interactive applications, called R Shiny, and which uses Cox regression models.

In research published in JCO Precision Oncology, the scientists illustrated the benefits of the software by analyzing data from two randomized clinical trials of pediatric patients with acute myeloid leukemia for predictors of treatment efficacy.

Building on previous work, the researchers assessed how baseline factors and the ACS10 pharmacogenomics score, which uses a patient’s DNA sequence to determine their likely response to different treatments, could impact predicted survival. They examined the predicted overall and event-free survival between patients assigned to introductory treatment regimens of clofarabine plus cytarabine or daunorubicin and etoposide combined with low-dose cytarabine or high-dose cytarabine.

While ACS10 performs well alone in predicting outcomes, shinyCox illustrates how age and sex may also be important predictors that can add to the pharmacogenomics score, improving projections and altering some conclusions about the best treatments for certain subgroups.

“shinyCox is not built to replace hazard ratios or traditional analysis, but rather to augment them,” said senior author Stanley Pounds, PhD, Department of Biostatistics. “We have given the biomedical community a tool to better see how patient groups might differ, providing investigators a far more nuanced understanding of their clinical trials and ultimately empowering them to make discoveries that will push forward personalized therapy research that creates better treatment outcomes for patients.”