TY - JOUR
T1 - Machine learning algorithm improves the detection of NASH (NAS-based) and at-risk NASH
T2 - A development and validation study
AU - Lee, Jenny
AU - Westphal, Max
AU - Vali, Yasaman
AU - Boursier, Jerome
AU - Petta, Salvatorre
AU - Ostroff, Rachel
AU - Alexander, Leigh
AU - Chen, Yu
AU - Fournier, Celine
AU - Geier, Andreas
AU - Francque, Sven
AU - Wonders, Kristy
AU - Tiniakos, Dina
AU - Bedossa, Pierre
AU - Allison, Mike
AU - Papatheodoridis, Georgios
AU - Cortez-Pinto, Helena
AU - Pais, Raluca
AU - Dufour, Jean-Francois
AU - Leeming, Diana Julie
AU - Harrison, Stephen
AU - Cobbold, Jeremy
AU - Holleboom, Adriaan G.
AU - Yki-Järvinen, Hannele
AU - Crespo, Javier
AU - Ekstedt, Mattias
AU - Aithal, Guruprasad P.
AU - Bugianesi, Elisabetta
AU - Romero-Gomez, Manuel
AU - Torstenson, Richard
AU - Karsdal, Morten
AU - Yunis, Carla
AU - Schattenberg, J. rn M.
AU - Schuppan, Detlef
AU - Ratziu, Vlad
AU - Brass, Clifford
AU - Duffin, Kevin
AU - Zwinderman, Koos
AU - Pavlides, Michael
AU - Anstee, Quentin M.
AU - Bossuyt, Patrick M.
N1 - Funding Information: The LITMUS project has received funding from the Innovative Medicines Initiative 2 Joint Undertaking under grant agreement No. 777377. This Joint Undertaking receives support from the European Union’s Horizon 2020 research and innovation program and EFPIA. Sven Francque holds a senior clinical investigator fellowship from the Research Foundation Flanders (FWO) (1802154N). Publisher Copyright: © 2023 John Wiley and Sons Inc.. All rights reserved.
PY - 2023/7/1
Y1 - 2023/7/1
N2 - Background and Aims: Detecting NASH remains challenging, while at-risk NASH (steatohepatitis and F≥ 2) tends to progress and is of interest for drug development and clinical application. We developed prediction models by supervised machine learning techniques, with clinical data and biomarkers to stage and grade patients with NAFLD. Approach and Results: Learning data were collected in the Liver Investigation: Testing Marker Utility in Steatohepatitis metacohort (966 biopsy-proven NAFLD adults), staged and graded according to NASH CRN. Conditions of interest were the clinical trial definition of NASH (NAS ≥ 4;53%), at-risk NASH (NASH with F ≥ 2;35%), significant (F ≥ 2;47%), and advanced fibrosis (F ≥ 3;28%). Thirty-five predictors were included. Missing data were handled by multiple imputations. Data were randomly split into training/validation (75/25) sets. A gradient boosting machine was applied to develop 2 models for each condition: clinical versus extended (clinical and biomarkers). Two variants of the NASH and at-risk NASH models were constructed: direct and composite models. Clinical gradient boosting machine models for steatosis/inflammation/ballooning had AUCs of 0.94/0.79/0.72. There were no improvements when biomarkers were included. The direct NASH model produced AUCs (clinical/extended) of 0.61/0.65. The composite NASH model performed significantly better (0.71) for both variants. The composite at-risk NASH model had an AUC of 0.83 (clinical and extended), an improvement over the direct model. Significant fibrosis models had AUCs (clinical/extended) of 0.76/0.78. The extended advanced fibrosis model (0.86) performed significantly better than the clinical version (0.82). Conclusions: Detection of NASH and at-risk NASH can be improved by constructing independent machine learning models for each component, using only clinical predictors. Adding biomarkers only improved the accuracy of fibrosis.
AB - Background and Aims: Detecting NASH remains challenging, while at-risk NASH (steatohepatitis and F≥ 2) tends to progress and is of interest for drug development and clinical application. We developed prediction models by supervised machine learning techniques, with clinical data and biomarkers to stage and grade patients with NAFLD. Approach and Results: Learning data were collected in the Liver Investigation: Testing Marker Utility in Steatohepatitis metacohort (966 biopsy-proven NAFLD adults), staged and graded according to NASH CRN. Conditions of interest were the clinical trial definition of NASH (NAS ≥ 4;53%), at-risk NASH (NASH with F ≥ 2;35%), significant (F ≥ 2;47%), and advanced fibrosis (F ≥ 3;28%). Thirty-five predictors were included. Missing data were handled by multiple imputations. Data were randomly split into training/validation (75/25) sets. A gradient boosting machine was applied to develop 2 models for each condition: clinical versus extended (clinical and biomarkers). Two variants of the NASH and at-risk NASH models were constructed: direct and composite models. Clinical gradient boosting machine models for steatosis/inflammation/ballooning had AUCs of 0.94/0.79/0.72. There were no improvements when biomarkers were included. The direct NASH model produced AUCs (clinical/extended) of 0.61/0.65. The composite NASH model performed significantly better (0.71) for both variants. The composite at-risk NASH model had an AUC of 0.83 (clinical and extended), an improvement over the direct model. Significant fibrosis models had AUCs (clinical/extended) of 0.76/0.78. The extended advanced fibrosis model (0.86) performed significantly better than the clinical version (0.82). Conclusions: Detection of NASH and at-risk NASH can be improved by constructing independent machine learning models for each component, using only clinical predictors. Adding biomarkers only improved the accuracy of fibrosis.
UR - http://www.scopus.com/inward/record.url?scp=85163490526&partnerID=8YFLogxK
U2 - https://doi.org/10.1097/HEP.0000000000000364
DO - https://doi.org/10.1097/HEP.0000000000000364
M3 - Article
C2 - 36994719
SN - 0270-9139
VL - 78
SP - 258
EP - 271
JO - Hepatology (Baltimore, Md.)
JF - Hepatology (Baltimore, Md.)
IS - 1
ER -