This is the authors' abstract. We don't add key points for this paper.
Ovarian cancer diagnosis is driven by biomarkers. We introduce biomedical Shapley behavioural spaces, a framework that transforms biomarker values into representations of how each patient contributes to population-level statistical properties. Using Shapley value decomposition, we computed each sample's contribution to variance, skewness, kurtosis, and entropy across all biomarkers, revealing hidden structure invisible to conventional analysis. Applied to two ovarian cancer data sets, the PLCO screening trial (1,007 samples, 42 biomarkers) and AZOC clinical cohort (309 samples, 45 biomarkers), behavioural transformation identified patient subgroups exhibiting striking post hoc malignancy enrichment. A kurtosis-defined subgroup in AZOC achieved 100% positive predictive value (34/34 cancerous), while a variance-defined subgroup in PLCO showed 3.32-fold cancer enrichment (p < 0.001). These findings are presented as exploratory subgroup characterisations rather than validated predictive performance. The identified subgroups exhibited statistically coherent biomarker patterns, including marked tumour marker elevation and systemic features, that are consistent with known disease biology but require external validation with clinical outcomes.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Liu et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: