Diagnostic accuracy of non‐invasive tests to screen for at‐risk MASH—An individual participant data meta‐analysis
Ferenc E. Mózes,Jenny A. Lee,Yasaman Vali,Emmanuel A. Selvaraj,Arjun N. A. Jayaswal,Jérôme Boursier,Victor de Lédinghen,Monica Lupșor‐Platon,Yusuf Yilmaz,Wah‐Kheong Chan,Sanjiv Mahadeva,Thomas Karlas,Johannes Wiegand,Shalimar,Emmanouil Tsochatzis,Antonio Liguori,Vincent Wai‐Sun Wong,Dae Ho Lee,Adriaan G. Holleboom,Anne‐Marieke van Dijk,Anne Linde Mak,Hannes Hagström,Camilla Akbari,Masashi Hirooka,Dong Hyeon Lee,Won Kim,Takeshi Okanoue,Toshihide Shima,Atsushi Nakajima,Masato Yoneda,Paul J. Thuluvath,Feng Li,Annalisa Berzigotti,Yuly P. Mendoza,Mazen Noureddin,Emily Truong,Céline Fournier‐Poizat,Andreas Geier,Theresa Tuthill,Carla Yunis,Quentin M. Anstee,Stephen A. Harrison,Patrick M. Bossuyt,Michael Pavlides
DOI: https://doi.org/10.1111/liv.15914
IF: 8.754
2024-04-05
Liver International
Abstract:Background & Aims There is a need to reduce the screen failure rate (SFR) in metabolic dysfunction‐associated steatohepatitis (MASH) clinical trials (MASH+F2‐3; MASH+F4) and identify people with high‐risk MASH (MASH+F2‐4) in clinical practice. We aimed to evaluate non‐invasive tests (NITs) screening approaches for these target conditions. Methods This was an individual participant data meta‐analysis for the performance of NITs against liver biopsy for MASH+F2‐4, MASH+F2‐3 and MASH+F4. Index tests were the FibroScan‐AST (FAST) score, liver stiffness measured using vibration‐controlled transient elastography (LSM‐VCTE), the fibrosis‐4 score (FIB‐4) and the NAFLD fibrosis score (NFS). Area under the receiver operating characteristics curve (AUROC) and thresholds including those that achieved 34% SFR were reported. Results We included 2281 unique cases. The prevalence of MASH+F2‐4, MASH+F2‐3 and MASH+F4 was 31%, 24% and 7%, respectively. Area under the receiver operating characteristics curves for MASH+F2‐4 were .78, .75, .68 and .57 for FAST, LSM‐VCTE, FIB‐4 and NFS. Area under the receiver operating characteristics curves for MASH+F2‐3 were .73, .67, .60, .58 for FAST, LSM‐VCTE, FIB‐4 and NFS. Area under the receiver operating characteristics curves for MASH+F4 were .79, .84, .81, .76 for FAST, LSM‐VCTE, FIB‐4 and NFS. The sequential combination of FIB‐4 and LSM‐VCTE for the detection of MASH+F2‐3 with threshold of .7 and 3.48, and 5.9 and 20 kPa achieved SFR of 67% and sensitivity of 60%, detecting 15 true positive cases from a theoretical group of 100 participants at the prevalence of 24%. Conclusions Sequential combinations of NITs do not compromise diagnostic performance and may reduce resource utilisation through the need of fewer LSM‐VCTE examinations.
gastroenterology & hepatology