Temporal FCA for Longitudinal Patient Trajectories

Incorporate temporal Formal Concept Analysis to model longitudinal patient trajectories and their evolution over time, thereby extending the FCA-guided counterfactual framework beyond static cross-sectional data.

Background

The FCA-Guided Counterfactual framework is evaluated on static, cross-sectional TCGA-BRCA patient data. Its lattice construction and counterfactual generation therefore do not represent how clinical features, disease states, or patient concepts change over time.

The authors identify temporal FCA—such as tracking concept-lattice evolution over time—as an unresolved research challenge. Addressing it would enable counterfactual explanations for longitudinal patient monitoring and disease progression, a direction also listed in the paper’s future-work agenda.

References

Incorporating temporal FCA (e.g., concept lattice evolution over time) is an open research challenge.

— FCA-Guided Counterfactual Explanations for Multi-Modal Breast Cancer Diagnosis: A Framework Achieving Perfect Validity with Emergent Sparsity  (2609.20067 - Isa et al., 17 Sep 2026) in Section 5.1.1, Limitations, item 4 (Temporal dynamics)