Highlights
The Hyve and AstraZeneca recently co-developed and released new features for Survival KM curves to the cBioPortal community. The enhanced survival KM curve will offer new standard KM plots in the portal that can be readily used in publications or reports while eliminating manual figure manipulation, reducing data transcription, and driving survival analysis standardization. These enhancements exemplify the collaborative spirit within the cBioPortal open-source community, where contributions from collaborators, including AstraZeneca, propel the platform’s capabilities forward. The latest features are now available to a broader audience through cBioPortal's Release v5.4.6 as of October 17, 2023, marking another milestone in our shared commitment to advancing cancer research.
Introduction
One common question in cancer research is to determine the time between response to a treatment, progression-free survival time (PFS), or overall survival time (OS). Kaplan-Meier (KM) curves are used to show these outcomes by plotting the estimated survival probabilities/percentages (on the Y-axis) and time elapsed after entry into the study (on the X-axis). While cBioPortal provides KM curve features on both the group comparison and results pages, it falls short in allowing annotations such as hazard ratios, landmark values, and the number at risk that are usually reported in most standard clinical trial publications (Figure 1A, old survival chart). These annotations are standard in reporting clinical trials and can add additional information that cBioPortal’s KM plot currently lacks.
The Hyve extended the Kaplan-Meier (KM) survival chart functionality by adding three new elements: the Number-at-risk table, Hazard ratio table, and Landmarks. Continue reading to learn more about these elements:
The Number-at-Risk table
The Number-at-Risk table displays the total number of subjects at any specific time point during the study. It provides insight into how many subjects remain in the study at a given time and have not experienced an event (e.g., progression or death)
Image 1: Number-at-risk table and Hazard ratio Hazard ratio
The hazard ratio (HR) describes the relative risk of the complication based on a comparison of event rates. The new feature now allows users to calculate the HR based on the control group. This will populate the legend with the HR to the control group and create a new table containing pairwise HRs between all groups, along with their 95% confidence interval (CI). Note that the HR is calculated with no stratification.Landmarks
In survival analysis, landmark analysis refers to designating a time point during the follow-up period, known as the landmark time, and analyzing only those subjects who have survived until the landmark time. We have added a new feature to visualize landmark time points, displaying the percentage of subjects at that specific time point. This allows an understanding of the percentage of subjects who responded by the landmark time.
Image 2: Landmark lines “The new KM plot can be readily used in publications with all necessary calculations, such as hazard ratios. It will avoid the unnecessary repeated analysis and greatly save time for data scientists supporting clinical trials,” said Zhongwu Lai, Senior Director of Bioinformatics at AstraZeneca.
“Survival analysis is one of cBioPortal's most popular features. These new enhancements enable users to explore differences in survival patterns more easily and interactively, helping them generate stronger hypotheses that might ultimately lead to better outcomes for patients”, says Ino de Bruijn, Data Visualization Team Lead at Memorial Sloan Kettering Cancer Center. In addition, he comments on the importance of collaboration, “cBioPortal is used all over the world at academic institutions, hospitals and in industry. Developing novel functionality that’s useful for everyone requires broad expertise, including in cancer biology, statistics, engineering, and design. This collaborative effort to improve the survival analysis is a great example of how everyone benefits when we work together”.
Try it out yourself.
The new survival chart features are now available for all studies containing survival data, such as the TMB and Immunotherapy (MSK, Nat Genet 2019) study.
To view the survival chart on the Group Comparison page, select a clinical variable of interest and click ‘Compare Groups.’ For instance, let’s pick ‘Drug Type’ and investigate whether there is a difference in survival.
By clicking on the Survival tab, you will see the overall survival for the three selected groups.
Landmark lines can be added by selecting the ‘Add Landmarks’ checkbox and entering the desired values, separated by commas or spaces. Note that the landmark lines will shift accordingly when the X-axis slider is adjusted.
To calculate the hazard ratio, check the ‘Calculate Hazard Ratios’ box. This will add the hazard ratio to the legend and generate a hazard ratio table below the figure.
Closing remarks
The Hyve provides services to develop, extend, and improve features in cBioPortal, such as the new survival analysis plot described in this article. Implemented features are shared with the community via releases in the cBioPortal repository on GitHub. We hope you enjoy these new features and find them helpful in your work.
Our heartfelt appreciation goes to AstraZeneca for sponsoring and collaborating on these new features and advancing cBioPortal’s capabilities. Such partnerships not only enhance the platform but also help the broader open-source software and cancer researchers’ community, fostering a collaborative environment that benefits researchers worldwide.
For inquiries on cBioPortal feature development projects or other cBioPortal-related services, feel free to contact us.
This article was written by Tim Kuijpers, Data Engineer @ The Hyve and Fengjian Shi, Product Owner @ AstraZeneca.
cBioPortal