Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Unsupervised risk factor identification across cancer types and data modalities via explainable artificial

Maximilian Ferle1,2,3,4, Jonas Ader5,6, Thomas Wiemers6,7

  • 1Center for Scalable Data Analytics and Artificial Intelligence (ScaDS.AI) Dresden/Leipzig, Universität Leipzig, Leipzig, Germany. maximilian.ferle@uni-leipzig.de.

NPJ Digital Medicine
|May 11, 2026
PubMed

Related Concept Videos

Cancer Survival Analysis01:21

Cancer Survival Analysis

Cancer survival analysis focuses on quantifying and interpreting the time from a key starting point, such as diagnosis or the initiation of treatment, to a specific endpoint, such as remission or death. This analysis provides critical insights into treatment effectiveness and factors that influence patient outcomes, helping to shape clinical decisions and guide prognostic evaluations. A cornerstone of oncology research, survival analysis tackles the challenges of skewed, non-normally...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Author Correction: Teclistamab-based induction treatment in transplant-eligible, newly diagnosed multiple myeloma: a phase 2 trial.

Nature medicine·2026
Same author

Interpretable Bayesian Modeling for Multireader Multicase Studies: Addressing Overdispersion and Limited Sample Size in Diagnostic Enhancement Evaluation.

Statistics in medicine·2026
Same author

Teclistamab-based induction treatment in transplant-eligible, newly diagnosed multiple myeloma: a phase 2 trial.

Nature medicine·2026
Same author

Evaluating CMV Risk Stratification, Donor Characteristics, and Post-Transplant Outcomes in Kidney Transplant Recipients: A Retrospective Eurotransplant Center Analysis.

Transplantation proceedings·2026
Same author

Publisher Correction: Unsupervised risk factor identification across cancer types and data modalities via explainable artificial intelligence.

NPJ digital medicine·2026
Same author

CT-defined sarcopenia as a predictive factor in patients undergoing portal vein embolization: a systematic review and meta-analysis.

Translational gastroenterology and hepatology·2026
Summary

This study introduces a new AI method to find distinct patient groups for better cancer risk stratification. The approach uses survival heterogeneity to identify prognostically significant features for personalized treatment.

Area of Science:

  • Oncology
  • Bioinformatics
  • Artificial Intelligence

Background:

  • Risk stratification is crucial for clinical decisions but current methods struggle to translate complex survival data into practical criteria.
  • Sophisticated survival analysis often lacks direct application in patient group identification.

Purpose of the Study:

  • To develop a novel, model-agnostic method for training neural networks to identify prognostically distinct patient groups.
  • To directly optimize for survival heterogeneity across patient clusters for improved risk stratification.

Main Methods:

  • A new method was developed to train any neural network on any data modality.
  • The method directly optimizes for survival heterogeneity across patient clusters.
  • Applied to multiple myeloma (MM) and non-small cell lung cancer (NSCLC) datasets, including laboratory and imaging data.

Related Experiment Videos

Main Results:

  • The method successfully identified prognostically distinct patient groups in both MM and NSCLC.
  • Post-hoc explainability analyses revealed clinically meaningful features aligning with established risk factors.
  • Findings were externally validated in MM (GMMG-MM5) and NSCLC (institutional data).

Conclusions:

  • This pan-cancer, model-agnostic approach discovers novel prognostic signatures across diverse data types.
  • The interpretable results complement treatment personalization and clinical decision-making in oncology.
  • The method offers a powerful tool for enhancing risk stratification and patient management.