American journal of hematology · 2025
International Consensus Histopathological Criteria for Subtyping Idiopathic Multicentric Castleman Disease Based on Machine Learning Analysis.
СтатьяConsensus StatementПоддержка службы здравоохранения СШАПоддержка NIH: внешняя программаПоддержка исследования: вне правительства СШАНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Nishimura Midori Filiz, Haratake Tomoka, Nishimura Yoshito, Nishikori Asami, Sumiyoshi Remi, Ujiie Hideki, Kawahara Yuri, Koga Tomohiro, Ueki Masao, Laczko Dorottya, Oksenhendler Eric, Fajgenbaum David C, van Rhee Frits, Kawakami Atsushi, Sato Yasuharu
Аннотация
Idiopathic multicentric Castleman disease (iMCD) is a rare lymphoproliferative disorder classified into three recognized clinical subtypes-idiopathic plasmacytic lymphadenopathy (IPL), TAFRO, and NOS. Although clinical criteria are available for subtyping, diagnostically challenging cases with overlapping histopathological features highlight the need for an improved classification system integrating clinical and histopathological findings. We aimed to develop an objective histopathological subtyping system for iMCD that closely correlates with the clinical subtypes. Excisional lymph node specimens from 94 Japanese iMCD patients (54 IPL, 28 TAFRO, 12 NOS) were analyzed for five key histopathological parameters: germinal center (GC) status, plasmacytosis, vascularity, hemosiderin deposition, and "whirlpool" vessel formation in GC. Using hierarchical clustering, we visualized subgroups and developed a machine learning-based decision tree to differentiate the clinical subtypes and validated it in an external cohort of 12 patients with iMCD. Hierarchical cluster analysis separated the IPL and TAFRO cases into mutually exclusive clusters, whereas the NOS cases were interspersed between them. Decision tree modeling identified plasmacytosis, vascularity, and whirlpool vessel formation as key features distinguishing IPL from TAFRO, achieving 91% and 92% accuracy in the training and test sets, respectively. External validation correctly classified all IPL and TAFRO cases, confirming the reproducibility of the system. Our histopathological classification system closely aligns with the clinical subtypes, offering a more precise approach to iMCD subtyping. It may enhance diagnostic accuracy, guide clinical decision-making for predicting treatment response in challenging cases, and improve patient selection for future research. Further validation of its versatility and clinical utility is required.