Bo-Gyeom Kim, Gakyung Kim, Yoshinari Abe, Pino Alonso, Stephanie Ameis, Alan Anticevic, Paul D. Arnold, Srinivas Balachander, Nerisa Banaj, Nuria Bargalló, Marcelo C. Batistuzzo, Francesco Benedetti, Sara Bertolín, Jan Carl Beucke, Irene Bollettini, Silvia Brem, Brian P. Brennan, Jan K. Buitelaar, Rosa Calvo, Miguel Castelo-Branco, Yuqi Cheng, Ritu Bhusal Chhatkuli, Valentina Ciullo, Ana Coelho, Beatriz Couto, Sara Dallaspezia, Benjamin A. Ely, Sónia Ferreira, Martine Fontaine, Jean-Paul Fouche, Rachael Grazioplene, Patricia Gruner, Kristen Hagen, Bjarne Hansen, Gregory L. Hanna, Yoshiyuki Hirano, Marcelo Q. Höxter, Morgan Hough, Hao Hu, Chaim Huyser, Toshikazu Ikuta, Neda Jahanshad, Anthony James, Fern Jaspers-Fayer, Selina Kasprzak, Norbert Kathmann, Christian Kaufmann, Minah Kim, Kathrin Koch, Gerd Kvale, Jun Soo Kwon, Luisa Lazaro, Junhee Lee, Christine Lochner, Jin Lu, Daniela Rodriguez Manrique, Ignacio Martínez-Zalacaín, Yoshitada Masuda, Koji Matsumoto, Maria Paula Maziero, Jose M. Menchón, Luciano Minuzzi, Pedro Silva Moreira, Pedro Morgado, Janardhanan C. Narayanaswamy, Jin Narumoto, Ana E. Ortiz, Junko Ota, Jose C. Pariente, Chris Perriello, Maria Picó-Pérez, Christopher Pittenger, Sara Poletti, Eva Real, Y. C. Janardhan Reddy, Daan van Rooij, Yuki Sakai, João Ricardo Sato, Cinto Segalas, Roseli G. Shavitt, Zonglin Shen, Eiji Shimizu, Venkataram Shivakumar, Carles Soriano-Mas, Nuno Sousa, Mafalda Machado Sousa, Gianfranco Spalletta, Emily R. Stern, S. Evelyn Stewart, Philip R. Szeszko, Rajat Thomas, Sophia I. Thomopoulos, Daniela Vecchio, Ganesan Venkatasubramanian, Chris Vriend, Susanne Walitza, Zhen Wang, Anri Watanabe, Lidewij Wolters, Jian Xu, Kei Yamada, Je-Yeon Yun, Mojtaba Zarei, Qing Zhao, Xi Zhu, Honami Arai, Ana Isabel Araújo, Kentaro Araki, Paul D. Arnold, Justin T. Baker, Núria Bargalló, Sara Bertolín, John R. Best, Premika S. W. Boedhoe, Sven Bölte, Vilde Brecke, Jan K. Buitelaar, Rosa Calvo, Carolina Cappi, Joao Castelhano, Wei Chen, Sutoh Chihiro, Kang Ik Kevin Cho, Sunah Choi, Daniel Costa, Nan Dai, Shareefa Dalvie, Damiaan Denys, Juliana B. Diniz, Isabel C. Duarte, Calesella Federico, Jamie D. Feusner, Kate D. Fitzgerald, Egill Axfjord Fridgeirsson, Edna Grünblatt, Sayo Hamatani, Gregory Hanna, Mengxin He, Odile A. van den Heuvel, Marcelo Q. Höxter, Morgan Hough, Keisuke Ikari, Jonathan Ipser, Hongyan Jiang, Linling Jiang, Niels T. de Joode, Norbert Kathmann, Taekwan Kim, Hitomi Kitagawa, Masaru Kuno, Yoo Bin Kwak, Jun Soo Kwon, Wieke van Leeuwen, Chiang-shan Ray Li, Na Li, Yanni Liu, Fang liu, Antonio Carlos Lopes, Jin Lu, Yuri Milaneschi, Hein van Marle, Sergi Mas, David Mataix-Cols, Maria Alice de Mathis, Maria Paula Mazieiro, Sarah Medland, Renata Melo, Euripedes C. Miguel, Astrid Morer, Alessandro S. De Nadai, Tomohiro Nakao, Masato Nihei, Luke Norman, Erika L. Nurmi, Joseph O’Neil, Sanghoon Oh, Sho Okawa, John C. Piacentini, Maria Picó-Pérez, Natalia Rodriguez, Daan van Rooij, João R. Sato, Cinto Segalas, Renata Silva, Noam Soreni, Michael Stevens, Anouk van der Straten, Jumpei Takahashi, Tais Tanamatis, Jinsong Tang, Anders Lillevik Thorsen, David Tolin, Anne Uhlmann, Benedetta Vai, Ysbrand D. van der Werf, Dick J. Veltman, Nora Vetter, Jicai Wang, Cees J. Weeland, Guido A. van Wingen, Stella J. de Wit, Nicole Wolff, Xiufeng Xu, Tokiko Yoshida, Fengrui Zhang, Paul M. Thompson, Willem B. Bruin, Guido A. van Wingen, Federica Piras, Fabrizio Piras, Dan J. Stein, Odile A. van den Heuvel, Helen Blair Simpson, Rachel Marsh, Jiook Cha
Molecular Psychiatry 2024年2月7日
Abstract
White matter pathways, typically studied with diffusion tensor imaging (DTI), have been implicated in the neurobiology of obsessive-compulsive disorder (OCD). However, due to limited sample sizes and the predominance of single-site studies, the generalizability of OCD classification based on diffusion white matter estimates remains unclear. Here, we tested classification accuracy using the largest OCD DTI dataset to date, involving 1336 adult participants (690 OCD patients and 646 healthy controls) and 317 pediatric participants (175 OCD patients and 142 healthy controls) from 18 international sites within the ENIGMA OCD Working Group. We used an automatic machine learning pipeline (with feature engineering and selection, and model optimization) and examined the cross-site generalizability of the OCD classification models using leave-one-site-out cross-validation. Our models showed low-to-moderate accuracy in classifying (1) “OCD vs. healthy controls” (Adults, receiver operator characteristic-area under the curve = 57.19 ± 3.47 in the replication set; Children, 59.8 ± 7.39), (2) “unmedicated OCD vs. healthy controls” (Adults, 62.67 ± 3.84; Children, 48.51 ± 10.14), and (3) “medicated OCD vs. unmedicated OCD” (Adults, 76.72 ± 3.97; Children, 72.45 ± 8.87). There was significant site variability in model performance (cross-validated ROC AUC ranges 51.6–79.1 in adults; 35.9–63.2 in children). Machine learning interpretation showed that diffusivity measures of the corpus callosum, internal capsule, and posterior thalamic radiation contributed to the classification of OCD from HC. The classification performance appeared greater than the model trained on grey matter morphometry in the prior ENIGMA OCD study (our study includes subsamples from the morphometry study). Taken together, this study points to the meaningful multivariate patterns of white matter features relevant to the neurobiology of OCD, but with low-to-moderate classification accuracy. The OCD classification performance may be constrained by site variability and medication effects on the white matter integrity, indicating room for improvement for future research.