Multimodality Stacking with Blockwise missing values and application to the PIONeeR biomarkers study for prediction of resistance to immunotherapy
arXiv:2605.25050v1 Announce Type: cross Abstract: Integrating multimodal datasets in clinical oncology is frequently hindered by high dimensionality and blockwise missingness, where entire data sources are unavailable for specific patient subsets. Standard survival models often struggle with these gaps, leading to biased results or patient exclusion. We introduce Multimodality Stacking with Blockwise missing values (MSB), a late-fusion framework for survival analysis that independently models modality-specific features before aggregating predictions via a cross-validated stacking meta-learner.
The increasing availability and complexity of multimodal clinical data, along with advancements in AI/ML techniques, enable more sophisticated approaches to data integration and missing value handling in medical research.
This development can significantly improve the accuracy and robustness of predictive models in oncology, leading to more personalized and effective treatment strategies for patients.
The ability to integrate incomplete multimodal datasets reduces the need for patient exclusion and addresses a significant challenge in real-world clinical data, enhancing the utility of AI in medical diagnostics and prognostics.
- · Oncology researchers
- · Pharmaceutical companies developing immunotherapies
- · AI in healthcare developers
- · Cancer patients
- · Traditional statistical modeling approaches
- · Clinical trials with strict data completeness requirements
Improved prediction of immunotherapy response in cancer patients using AI models.
Accelerated development and adoption of AI-driven personalized medicine in oncology.
Potential for broader application of similar blockwise missing data techniques across other complex multimodal biomedical datasets, transforming R&D efficiency and patient outcomes.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG