Deep hierarchical subtyping of multi-organ systemic sclerosis trajectories - a EUSTAR study

作者：Cécile Trottet, Manuel Schürch, Ahmed E. Allam, L. Petelytska, Ivan Castellvi, Radim Bečvář, Jeska de Vries‐Bouwstra, Florenzo Iannone, Patricia Esmeralda Carreira, Marie‐Elise Truchetet, Giovanna Cuomo, Elena Rezuş, Francesco Paolo Cantatore, Carmen Pilar Simeón‐Aznar, Magda Ileana Pârvu, Marta Dzhus, Oliver Distler, Anna‐Maria Hoffmann‐Vold, Michael O. Krauthammer, EUSTAR Collaborators, Ivan Castellvi, Jeska de Vries‐Bouwstra, Silvia Bellando-Randone, Ulrich Andreas Walker, Maurizio Cutolo, Simona Rednic, Yannick Allanore, Carlomaurizio Montecucco, Srdjan Novak, Gabór Kumánovics, Przemysław Jacek Kotyla, Elisabetta Zanatta, Katja Perdan Pirkmajer, Gianluca Moroncini, Paolo Airó, Mislav Radić, Alexandra Balbir‐Gurman, N Hunzelmann, Luca Idolazzi, Joško Mitrović, Christopher P. Denton, Madelon C Vonk, Jelena Colic, Jörg Henes, Ivan Foeldvari, Gianluigi Bajocchi, Tânia Santiago, Bojana N. Stamenkovic, Maria De Santis, Claudia Ickinger, Lidia P Ananieva, Klaus Henrik Søndergaard, Gabriella Szűcs, David Launay, Valeria Riccieri, Andra Rodica Balanescu, Ana Maria Gheorghiu, Christina Bergmann, Luc Mouthon, Vanessa M. Smith, Mette Mogensen, Marie Vanthuyne, Juan José Alegre Sancho, Brigitte Granel, Carolina de Souza Müller, Svetlana Agachi, Alberto Cauli, Kamal Solanki, Eiman Soliman, Edoardo Rosato, Rosario Foti, Britta Maurer, Marzena Olesińska, Nihal Awad, S. Blaise, Patricia Senet, Emmanuel Chatelus, Ira Litinsky, Francesco Del Galdo, Eduardo Mario Kerzberg, Jasminka Milas-Ahic, Massimiliano Limonta, Antonella Marcoccia, Thierry Martin, Anna Wojteczek, Gabriela Riemekasten, Lélita da Conceição Santos, Yair M. Levy, Daniel Brito de Araújo, Marek Brzosko, Oscar Massimiliano Epis, Petros Paul Sfikakis, Ana-Maria Ramazan, Alain Lescoat, Marco Matucci‐Cerinic, Julia Spierings, Fabiola Atzeni, Masataka Kuwana, Arsène MEKINIAN, Mickaël Martin, Gonçalo Boleto, Nicoletta Del Papa, Enrico Selvi, Marta Mosca, Ulrich Gerth, Duygu Temiz Karadağ, Anastas Zgurov Batalov, K. Ginosyan, Nune Manukyan, Mohammad Ebrahim Naffaa, Cristina Maglio, Miriam Retuerto, Futoshi Iwata, Monique Hinchcliff, Roberto Giacomelli, Francesco Benvenuti, Helena Santos Carneiro, Esther Vicente Rabaneda, Andrea‐Hermina Györfi, Lilian Maria Lopez Nunez, Rossella De Angelis, Irene Carrión-Barberà, Alejandro J. Brigante, Yasser M El Miedany, Rong Mu, Alexandra Daniel, Amato de Paulis, Chris T. Derk, Lijun Zhang, Bogdan Piotr Batko, Ivette Casafont Solé, Anna Lewandowska‐Polak, Qingran Yan, Tuncay Duruöz, Seda Çolak, Janeth Edelina Villegas Guzmán, Claudia S Mora-Trujillo, Maria Sole Chimenti, Samah A. El-Bakry, Fatma Alıbaz-Öner · 发表于：npj Digital Medicine · 年份：2025 · DOI：10.1038/s41746-025-01962-y · 被引用次数：5 · 研究领域：Systemic Sclerosis and Related Diseases、Mast cells and histamine、Inflammatory Myopathies and Dermatomyositis

Systemic sclerosis (SSc) is a chronic autoimmune disease with multi-organ involvement. Historically, SSc classification has focused on the type of skin involvement (limited versus diffuse); however, a growing evidence of organ-specific variability suggests the presence of more than two distinct subtypes. We propose a semi-supervised generative deep learning framework leveraging expert-driven definitions of organ-specific involvement and severity. We model SSc disease trajectories in the European Scleroderma Trials and Research (EUSTAR) database, containing 14,000 patients across 67,000 medical visits, and identify clinically meaningful subtypes to enhance patient stratification and prognosis. We systematically evaluate the model's predictive accuracy, robustness to missing data, and clinical interpretability. We identified five patient clusters, separating patients based on the degree of organ involvement. Notably, a subset with limited skin involvement still showed high risks of lung and heart complications, underscoring the importance of data-driven methods and multi-organ models to complement established insights from clinical practice.