Chang Devon, Truong Emily, Mena Edward A, Pacheco Fabiana, Wong Micaela, Guindi Maha, Todo Tsuyoshi T, Noureddin Nabil, Ayoub Walid, Yang Ju Dong, Kim Irene K, Kohli Anita, Alkhouri Naim, Harrison Stephen, Noureddin Mazen
Arnold O. Beckman High School , Irvine , California , USA.
Department of Medicine , Cedars Sinai Medical Center , Los Angeles , California , USA.
Hepatology. 2023 Feb 1;77(2):546-557. doi: 10.1002/hep.32655. Epub 2022 Aug 9.
We assessed the performance of machine learning (ML) models in identifying clinically significant NAFLD-associated liver fibrosis and cirrhosis.
We implemented ML models including logistic regression (LR), random forest (RF), and artificial neural network to predict histological stages of fibrosis using 17 demographic/clinical features in 1370 patients with NAFLD who underwent liver biopsy, FibroScan, and labs within a 6-month period at multiple U.S. centers. Histological stages of fibrosis (≥F2, ≥F3, and F4) were predicted using ML, FibroScan liver stiffness measurements, and Fibrosis-4 index (FIB-4). NASH with significant fibrosis (NAS ≥ 4 + ≥F2) was assessed using ML, FibroScan-AST (FAST) score, FIB-4, and NAFLD fibrosis score (NFS). We used 80% of the cohort to train and 20% to test the ML models. For ≥F2, ≥F3, F4, and NASH + NAS ≥ 4 + ≥F2, all ML models, especially RF, had primarily higher accuracy and AUC compared with FibroScan, FIB-4, FAST, and NFS. AUC for RF versus FibroScan and FIB-4 for ≥F2, ≥F3, and F4 were (0.86 vs. 0.81, 0.78), (0.89 vs. 0.83, 0.82), and (0.89 vs. 0.86, 0.85), respectively. AUC for RF versus FAST, FIB-4, and NFS for NASH + NAS ≥ 4 + ≥F2 were (0.80 vs. 0.77, 0.66, 0.63). For NASH + NAS ≥ 4 + ≥F2, all ML models had lower/similar percentages within the indeterminate zone compared with FIB-4 and NFS. Overall, ML models performed better in sensitivity, specificity, positive predictive value, and negative predictive value compared with traditional noninvasive tests.
ML models performed better overall than FibroScan, FIB-4, FAST, and NFS. ML could be an effective tool for identifying clinically significant liver fibrosis and cirrhosis in patients with NAFLD.
我们评估了机器学习（ML）模型在识别具有临床意义的非酒精性脂肪性肝病（NAFLD）相关肝纤维化和肝硬化方面的性能。
我们实施了包括逻辑回归（LR）、随机森林（RF）和人工神经网络在内的ML模型，使用1370例在美国多个中心接受肝活检、FibroScan检查和实验室检查的NAFLD患者的17种人口统计学/临床特征来预测纤维化的组织学阶段。使用ML、FibroScan肝脏硬度测量值和Fibrosis-4指数（FIB-4）预测纤维化的组织学阶段（≥F2、≥F3和F4）。使用ML、FibroScan-天门冬氨酸氨基转移酶（FAST）评分、FIB-4和NAFLD纤维化评分（NFS）评估伴有显著纤维化的非酒精性脂肪性肝炎（NAS≥4+≥F2）。我们使用队列的80%来训练ML模型，20%来测试模型。对于≥F2、≥F3、F4以及NASH+NAS≥4+≥F2，与FibroScan、FIB-4、FAST和NFS相比，所有ML模型，尤其是RF，主要具有更高的准确性和曲线下面积（AUC）。对于≥F2、≥F3和F4，RF相对于FibroScan和FIB-4的AUC分别为（0.86对0.81、0.78），（0.89对0.83、0.82），以及（0.89对0.86、0.85）。对于NASH+NAS≥4+≥F2，RF相对于FAST、FIB-4和NFS的AUC为（0.80对0.77、0.
