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

作者：Saverio D’Amico, Lorenzo Dall’Olio, Cesare Rollo, Patricia Alonso, Iñigo Prada-Luengo, Daniele Dall’Olio, Claudia Sala, Elisabetta Sauta, Gianluca Asti, Luca Lanino, Giulia Maggioni, Alessia Campagna, Elena Zazzetti, Mattia Delleani, Maria Elena Bicchieri, Pierandrea Morandini, Victor Savevski, Borja Arroyo, Juan Parras, Lin‐Pierre Zhao, Uwe Platzbecker, María Díez‐Campelo, Valeria Santini, Pierre Fenaux, Torsten Haferlach, Anders Krogh, Santiago Zazo, Piero Fariselli, Tiziana Sanavia, Matteo Giovanni Della Porta, Gastone Castellani · 发表于：JCO Clinical Cancer Informatics · 年份：2024 · DOI：10.1200/cci.24.00008 · 被引用次数：21 · 研究领域：Acute Myeloid Leukemia Research、Artificial Intelligence in Healthcare and Education、Digital Imaging for Blood Diseases

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 (...