Global performance of non-invasive tests in MASLD: Insights from the G-MASLD study
Zobair M. Younossi, Leyla de Avila, Salvatore Petta, Hannes Hagström, Seung Up Kim, Atsushi Nakajima, Javier Crespo, Laurent Castéra, Naim Alkhouri, Ming‐Hua Zheng, Sombat Treeprasertsuk, Prooksa Ananchuensook, S. Shalimar, Emmanuel Tsochatzis, Shenoy Kotacherry Trivikrama, Leena Kondarappassery Balakumaran, Jian‐Gao Fan, Stuart K. Roberts, Khalid Alswat, Vincent Wai‐Sun Wong, Yusuf Yılmaz, Winston Dunn, Sven Francque, Ahmed Cordie, Ming‐Lung Yu, Mattias Ekstedt, George Boon‐Bee Goh, Cláudia P. Oliveira, Mário Guimarães Pessôa, Wah‐Kheong Chan, Marlén Ivón Castellanos Fernández, Ajay Duseja, Juan Pablo Arab, George Papatheodoridis, Giada Sebastiani, Cristiane Alves Villela‐Nogueira, Roberta D’Ambrosio, Pietro Lampertico, Khalid Al‐Naamani, Adriaan G. Holleboom, Arun Valsan, Arathi Venu, Mohamed El‐Kassas, Grazia Pennisi, Ying Shang, Wen‐Yue Liu, Hye Won Lee, Takashi Kobayashi, Satoru Kakizaki, Cyrielle Caussy, Brian L. Pearlman, Paula Iruzubieta, Rida Nadeem, Felice Cinque, Antonia Neonaki, Mirko Zoncapè, Rui-Xu Yang, Sherlot Juan Song, Nicholas Dunn, Zouhir Gadi, Ming-Lun Yeh, Kevin Kim-Jun the, Sanjiv Mahadeva, Licet Gonzalez Fabian, Ahmed Almohsen, Nathalie C. Leite, Nicola Pugliese, Johan Vessby, Chencheng Xie, Narendra Singh Choudhary, Ethan Friend, María A. Poca, Takumi Kawaguchi, Francesco Paolo Russo, Adrian Gadano, Luis Antonio Diaz, Ashwani K. Singal, Bérénice Segrestin, Nadege Gunn, Dı́dac Mauricio, Marco Arrese, Anna Ludovica Fracanzani, Rosa Lombardi, Brian Lam, Andrei Racila, Saleh A. Alqahtani, Maria Stepanova
Abstract
BACKGROUND: Metabolic dysfunction-associated steatotic liver disease (MASLD) is highly prevalent worldwide. Performance of non-invasive tests (NITs) in patients with MASLD recruited from different regions of the world was evaluated. METHODS: MASLD patients with liver biopsies and NIT data [fibrosis-4 index (FIB-4), enhanced liver fibrosis (ELF), and liver stiffness measurement (LSM)] were enrolled through the Global NASH Council collaboration (G-MASLD). FibroScan-AST (FAST) and Agile-3+/Agile-4 were calculated. NITs' performance for predicting ≥F2 (significant fibrosis), ≥F3 (advanced fibrosis), or cirrhosis (F4) was determined in patients from different regions. RESULTS: A total of 17,792 MASLD patients from 41 countries were included: 14% had F0, 32% F1, 18% F2, 22% F3, 13% F4 (cirrhosis); 48% NAS ≥5. Advanced fibrosis prediction by NITs was variable across regions for FIB-4 [pooled AUC (95% CI)=0.80 (0.79-0.81)], the lowest in Latin America [0.75 (0.71-0.79)], the highest in MENA [0.84 (0.82-0.87)], and ELF [pooled AUC=0.77 (0.76-0.79)], the lowest in Europe [0.72 (0.69-0.76)], the highest in North America [0.80 (0.78-0.82)]. Prediction of advanced fibrosis by LSM [pooled AUC=0.84 (0.83-0.85)] was similar across regions except North America [0.78 (0.76-0.81)]. In addition, FAST [AUC=0.75 (0.74-0.76)] and Agile-3+ [AUC=0.87 (0.86-0.88)] performed similarly across regions. Similar trends were observed for the NITs predicting significant fibrosis. Finally, the accuracy of Agile-4 for predicting cirrhosis [AUC=0.90 (0.89-0.91)] was the lowest in North America [0.85 (0.83-0.87)], the highest in MENA [0.96 (0.94-0.98)]. CONCLUSIONS: The diagnostic performance of common NITs for fibrosis in MASLD varies across the world. In the large multinational G-MASLD sample, the most accurate NITs were Agile-3+ and Agile-4 composite scores.