🤖 AI Summary
Clinical data are frequently compromised by missing values, leading to machine learning models with unstable features, poor interpretability, and insufficient robustness—limitations that hinder their deployment in high-stakes clinical decision-making. To address this challenge, this work proposes CACTUS, a novel framework that uniquely prioritizes feature stability as a core evaluation criterion. By integrating feature abstraction, interpretable classification, and systematic stability analysis, CACTUS enables trustworthy predictions even with small-scale, heterogeneous, and incomplete clinical datasets. Evaluated on a cohort of 568 hematuria patients, CACTUS achieves competitive or superior predictive performance while substantially enhancing the stability of key features under missing data conditions. Notably, it demonstrates robustness in sex-stratified analyses, thereby improving the model’s clinical credibility and reproducibility.
📝 Abstract
Machine learning models are increasingly applied to biomedical data, yet their adoption in high stakes domains remains limited by poor robustness, limited interpretability, and instability of learned features under realistic data perturbations, such as missingness. In particular, models that achieve high predictive performance may still fail to inspire trust if their key features fluctuate when data completeness changes, undermining reproducibility and downstream decision-making. Here, we present CACTUS (Comprehensive Abstraction and Classification Tool for Uncovering Structures), an explainable machine learning framework explicitly designed to address these challenges in small, heterogeneous, and incomplete clinical datasets. CACTUS integrates feature abstraction, interpretable classification, and systematic feature stability analysis to quantify how consistently informative features are preserved as data quality degrades. Using a real-world haematuria cohort comprising 568 patients evaluated for bladder cancer, we benchmark CACTUS against widely used machine learning approaches, including random forests and gradient boosting methods, under controlled levels of randomly introduced missing data. We demonstrate that CACTUS achieves competitive or superior predictive performance while maintaining markedly higher stability of top-ranked features as missingness increases, including in sex-stratified analyses. Our results show that feature stability provides information complementary to conventional performance metrics and is essential for assessing the trustworthiness of machine learning models applied to biomedical data. By explicitly quantifying robustness to missing data and prioritising interpretable, stable features, CACTUS offers a generalizable framework for trustworthy data-driven decision support.