Interactive Web Application for Plotting Personalized Prognosis Prediction Curves in Allogeneic Hematopoietic Cell...

Area of Science:

• Hematology
• Oncology
• Medical Informatics

Background:

• Allogeneic hematopoietic cell transplantation (allo-HCT) offers a curative treatment for hematological malignancies.
• Current prognosis estimation relies on empirical methods, lacking objective data.
• There is a need for precise, data-driven tools to predict patient outcomes post-allo-HCT.

Purpose of the Study:

• To develop an objective tool for personalized prognosis prediction following allo-HCT.
• To create an interactive web application for visualizing patient-specific survival and incidence curves.
• To enhance clinical decision-making by providing data-driven prognostic insights.

Main Methods:

• Developed an interactive web application using a random survival forest model.
• Incorporated 8 key patient-specific prognostic factors into the prediction model.
• Utilized a cohort of 363 patients who underwent allo-HCT for model development and validation.

Main Results:

• The web application interactively displays personalized prediction curves for 1-year overall survival, progression-free survival, relapse/progression, and nonrelapse mortality (NRM).
• The model demonstrated strong predictive performance, with areas under the receiver-operating characteristic curves ranging from 0.70 to 0.77 in the test cohort.
• Validation was performed on a time-sequentially split cohort (70% training, 30% test).

Conclusions:

• The developed web application provides objective, personalized prognosis predictions for allo-HCT candidates.
• This tool can significantly aid transplant clinicians in informing patients and facilitating treatment decisions.
• The application represents a valuable advancement in personalized medicine for hematological disorders requiring transplantation.