Clinical usability of an explainable AI decision support tool and evaluation of multimodal models in NSCLC

Abstract Despite a decade in, immunotherapy (IO) treatment selection in non-small cell lung cancer (NSCLC) remains largely guided by subgroup analyses and imperfect programmed death ligand 1 (PD-L1) and clinical scores. To our knowledge, I 3 LUNG ( NCT05537922 ) is currently the largest international, real-world, multimodal, artificial intelligence (AI)-based study, enrolling 2,396 patients. We integrated real-world clinical and blood (CB) data, computed tomography (CT) images, digital pathology (DP), and genomics into machine learning early fusion (MLEF) and deep learning intermediate fusion (DLIF) models. Machine learning (ML) and deep learning (DL) CB-only models achieved consistent performance across outcomes with area under the curve (AUC) up to 0.77 in the test (TEST) set. Performance drop in external validation (EXVAL) likely reflects population differences (AUC range: 0.55–0.72). AI models significantly surpassed PD-L1, Eastern Cooperative Oncology Group performance status (ECOG PS), neutrophil-to-lymphocyte ratio (NLR), lactate dehydrogenase (LDH) and Lung Immune Prognostic Index (LIPI) score in the independent TEST set. The clinical usability study showed that lung expert and nonexpert physicians improved their prediction with the explainable AI (XAI) ML CB-only based tool. Although multimodal integration with MLEF (CB+CT+DP) was associated with higher performance, its incremental benefit remains uncertain, not translated in TEST and EXVAL. The I 3 LUNG project is a pioneering framework showing the clinical usefulness of AI tools. A prospective validation of the decision support system (both CB and multimodal) is currently undergoing in more than 2,000 patients.

Authors

Institutions

Publication Details

Journal
Nature Medicine
Published
2026-09-13
DOI
https://doi.org/10.1038/s41591-026-04488-2
Citations
1
Primary Topic
Radiomics and Machine Learning in Medical Imaging
Type
article
Field-Weighted Citation Impact
5.61

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
article

Clinical usability of an explainable AI decision support tool and evaluation of multimodal models in NSCLC

Eleonora Gusmaroli, Helena Linardou, Bogot R. Naama, Marlen Szewczyk et al.
1 citations
Nature Medicine
Radiomics and Machine Learning in Medical Imaging
5.61
article

Clinical usability of an explainable AI decision support tool and evaluation of multimodal models in NSCLC

Eleonora Gusmaroli, Helena Linardou, Bogot R. Naama, Marlen Szewczyk, Andra Diana Dumitrascu, Laila C. Roisman, Margherita Ruggirello, Luca Agnelli, Michele Zanitti, Marcello Restelli, Dario Monzani, Iris Watermann, Giulia Montelatici, Alessandra Pedrocchi, Chiara Cavalli, Claudia Proto, Sabina Sangaletti, Francesco Trovò, Federica Corso, Patricia Iranzo, Francesca Piazza, Christine M. Bestvina, Laura Mazzeo, Alessandra Russo, Monica Ganzinelli, Valentina Bartolomeo, Eliana Rulli, Aina Arbusà, V. Miskovic, Roberta Serino, C. Agosta, Rocio Caro-Consuegra, José Rodríguez-Morató, A. Esposito, Claudia Giani, Rebecca Romanò, Paolo Ambrosini, Reham Basheer, Daniele Lorenzini, Arsela Prelaj, Giancarlo Pruneri, Alessandro Chiesi, Till Olchers, Giulia Corrao, Emilia Ambrosini, Melissa Fernández-Pinto, Ronald Simon, Teresa Beninato, Elena Fountzilas, Enriqueta Felip, Erica Pietroluongo, Matteo Sacco, Stefano Natangelo, A. Ferrarin, Chiara Giangregorio, Cecilia Silvestri, Giuseppe Leone, Marco Meazza Prina, A. Zec, Anna Di Lello, Rosa Maria Di Mauro, Giorgia Di Liberti, Maria Spector, Margherita Favali, Giuseppe Nardo, C. Bonalume, B. Guirges, Teresa Arangoa, A. Lahiani Hafzadi, Ludovica Lerma, Costanza Siniscalchi, Luca Invernizzi, Constantin Blanke-Roeser, Moreno Marino, Michael Willis, Heinz Richter, Ghazal Farhikhteh, Michele Borraccino, Filippo de Braud, Marta De Ponti, Mario Occhipinti, Alexander T. Pearson, Nikolaos Spathas, Marina Chiara Garassino, Michele Pio Di Palma, Davide Macocchi, Sokol Kosta, Simone Rota, Evangelos Sarris, Cristina Maria Licciardello, Giuseppe Lo Russo, Miriam Fink, Martin Reck, Leonardo Provenzano, Marta Brambilla, Stefan Steurer, Samuel G. Armato, Gabriella Pravettoni, Nir Peled, Andrea Spagnoletti
article en
1 citations

Abstract

Abstract Despite a decade in, immunotherapy (IO) treatment selection in non-small cell lung cancer (NSCLC) remains largely guided by subgroup analyses and imperfect programmed death ligand 1 (PD-L1) and clinical scores. To our knowledge, I 3 LUNG ( NCT05537922 ) is currently the largest international, real-world, multimodal, artificial intelligence (AI)-based study, enrolling 2,396 patients. We integrated real-world clinical and blood (CB) data, computed tomography (CT) images, digital pathology (DP), and genomics into machine learning early fusion (MLEF) and deep learning intermediate fusion (DLIF) models. Machine learning (ML) and deep learning (DL) CB-only models achieved consistent performance across outcomes with area under the curve (AUC) up to 0.77 in the test (TEST) set. Performance drop in external validation (EXVAL) likely reflects population differences (AUC range: 0.55–0.72). AI models significantly surpassed PD-L1, Eastern Cooperative Oncology Group performance status (ECOG PS), neutrophil-to-lymphocyte ratio (NLR), lactate dehydrogenase (LDH) and Lung Immune Prognostic Index (LIPI) score in the independent TEST set. The clinical usability study showed that lung expert and nonexpert physicians improved their prediction with the explainable AI (XAI) ML CB-only based tool. Although multimodal integration with MLEF (CB+CT+DP) was associated with higher performance, its incremental benefit remains uncertain, not translated in TEST and EXVAL. The I 3 LUNG project is a pioneering framework showing the clinical usefulness of AI tools. A prospective validation of the decision support system (both CB and multimodal) is currently undergoing in more than 2,000 patients.

Nature Medicine
Hebron University (PS), Universität Hamburg (DE), University of Milan (IT), Mario Negri Institute for Pharmacological Research (IT), Shaare Zedek Medical Center (IL), University of Chicago (US), St. Luke's Hospital (GR), MedSIR (Spain) (ES), Vall d'Hebron Institut de Recerca (ES), University Medical Center Hamburg-Eppendorf (DE), Swedish Institute for Health Economics (SE), Chan Zuckerberg Initiative (United States) (US), Vall d'Hebron Hospital Universitari (ES), German Center for Lung Research (DE), Hospital Universitario de La Princesa (ES), LungenClinic Grosshansdorf (DE), Metropolitan Hospital (GR), Fondazione IRCCS Istituto Nazionale dei Tumori (IT), European Institute of Oncology (IT), Aalborg University (DK), University of Palermo (IT), Politecnico di Milano (IT)
European Commission
Openalex Percentile: Top 3%
Radiomics and Machine Learning in Medical Imaging
5.61
DOIOpen Access PDFOpenAlex
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.