MOSAIC: An Artificial Intelligence–Based Framework for Multimodal Analysis, Classification, and Personalized Prognostic Assessment in Rare Cancers

作者：S. D'amico, Lorenzo Dall'olio, C. Rollo, P. Alonso, Iñigo Prada-Luengo, Daniele Dall’Olio, C. Sala, Elisabetta Sauta, G. Asti, L. Lanino, G. Maggioni, Alessia Campagna, Elena Zazzetti, Mattia Delleani, M. Bicchieri, Pierandrea Morandini, V. Savevski, Borja Arroyo, J. Parras, Lin-Pierre Zhao, U. Platzbecker, M. Díez-Campelo, V. Santini, Pierre Fenaux, T. Haferlach, Anders Krogh, S. Zazo, P. Fariselli, T. Sanavia, M. D. Della Porta, G. Castellani · 发表于：JCO Clinical Cancer Informatics · 年份：2024 · DOI：10.1200/CCI.24.00008 · 被引用次数：26 · 研究领域：Medicine

PURPOSE Rare cancers constitute over 20% of human neoplasms, often affecting patients with unmet medical needs. The development of effective classification and prognostication systems is crucial to improve the decision-making process and drive innovative treatment strategies. We have created and implemented MOSAIC, an artificial intelligence (AI)–based framework designed for multimodal analysis, classification, and personalized prognostic assessment in rare cancers. Clinical validation was performed on myelodysplastic syndrome (MDS), a rare hematologic cancer with clinical and genomic heterogeneities. METHODS We analyzed 4,427 patients with MDS divided into training and validation cohorts. Deep learning methods were applied to integrate and impute clinical/genomic features. Clustering was performed by combining Uniform Manifold Approximation and Projection for Dimension Reduction + Hierarchical Density-Based Spatial Clustering of Applications with Noise (UMAP + HDBSCAN) methods, compared with the conventional Hierarchical Dirichlet Process (HDP). Linear and AI-based nonlinear approaches were compared for survival prediction. Explainable AI (Shapley Additive Explanations approach [SHAP]) and federated learning were used to improve the interpretation and the performance of the clinical models, integrating them into distributed infrastructure. RESULTS UMAP + HDBSCAN clustering obtained a more granular patient stratification, achieving a higher average silhouette coefficient (0.1...