Caussy Cyrielle, Vergès Bruno, Leleu Damien, Duvillard Laurence, Subtil Fabien, Abichou-Klich Amna, Hervieu Valérie, Milot Laurent, Ségrestin Bérénice, Bin Sylvie, Rouland Alexia, Delaunay Dominique, Morcel Pierre, Hadjadj Samy, Primot Claire, Petit Jean-Michel, Charrière Sybil, Moulin Philippe, Levrero Massimo, Cariou Bertrand, Disse Emmanuel
Département Endocrinologie, Diabète et Nutrition, Hôpital Lyon Sud, Hospices Civils de Lyon, Pierre-Bénite, France.
CarMeN Laboratory, INSERM U1060, INRA U1397, Institut National des Sciences Appliquées de Lyon, Université Claude Bernard Lyon 1, Pierre-Bénite, France.
Diabetes Care. 2025 Jun 1;48(6):877-886. doi: 10.2337/dc24-2075.
Screening for advanced fibrosis (AF) resulting from metabolic dysfunction-associated steatotic liver disease (MASLD) is recommended in diabetology. This study aimed to compare the performance of noninvasive tests (NITs) with that of two-step algorithms for detecting patients at high risk of AF requiring referral to hepatologists.
We conducted a planned interim analysis of a prospective multicenter study including participants with type 2 diabetes and/or obesity and MASLD with comprehensive liver assessment comprising blood-based NITs, vibration-controlled transient elastography (VCTE), and two-dimensional shear-wave elastography (2D-SWE). AF risk stratification was determined by a composite criterion of liver biopsy, magnetic resonance elastography, or VCTE ≥12 kPa depending on availability.
Of 654 patients (87% with type 2 diabetes, 56% male, 74% with obesity), 17.6% had an intermediate/high risk of AF, and 9.3% had a high risk of AF. The area under the empirical receiver operating characteristic curves of NITs for detection of high risk of AF were as follows: fibrosis-4 index (FIB-4) score, 0.78 (95% CI 0.72-0.84); FibroMeter, 0.74 (0.66-0.83); FibroTest, 0.78 (0.72-0.85); Enhanced Liver Fibrosis (ELF) test, 0.82 (0.76-0.87); and SWE, 0.84 (0.78-0.89). Algorithms with FIB-4 score/VCTE showed good diagnostic performance for referral of patients at intermediate/high risk of AF to specialized care in hepatology. An alternative FIB-4 score/ELF test strategy showed a high negative predictive value (NPV; 88-89%) and a lower positive predictive value (PPV; 39-46%) at a threshold of 9.8. The FIB-4 score/2D-SWE strategy had an NPV of 91% and a PPV of 58-62%. The age-adapted FIB-4 score threshold resulted in lower NPVs and PPVs in all algorithms.
The FIB-4 score/VCTE algorithm showed excellent diagnostic performance, demonstrating its applicability for routine screening in diabetology. The ELF test using an adapted low threshold at 9.8 may be used as an alternative to VCTE.
糖尿病学领域建议对代谢功能障碍相关脂肪性肝病(MASLD)导致的晚期肝纤维化(AF)进行筛查。本研究旨在比较非侵入性检测(NITs)与两步法算法在检测需要转诊给肝病专家的AF高危患者方面的性能。
我们对一项前瞻性多中心研究进行了计划中的中期分析,研究对象为2型糖尿病和/或肥胖且患有MASLD的参与者,采用包括基于血液的NITs、振动控制瞬时弹性成像(VCTE)和二维剪切波弹性成像(2D-SWE)在内的综合肝脏评估。根据可获得性,通过肝活检、磁共振弹性成像或VCTE≥12 kPa的综合标准确定AF风险分层。
在654例患者中(87%为2型糖尿病患者,56%为男性,74%为肥胖患者),17.6%有AF中/高风险,9.3%有AF高风险。NITs检测AF高风险的经验性受试者操作特征曲线下面积如下:纤维化-4指数(FIB-4)评分,0.78(95%CI 0.72-0.84);FibroMeter,0.74(0.66-0.83);FibroTest,0.78(0.72-0.85);增强肝纤维化(ELF)检测,0.82(0.76-0.87);以及SWE,0.84(0.78-0.89)。FIB-4评分/VCTE算法在将AF中/高风险患者转诊至肝病专科护理方面显示出良好的诊断性能。另一种FIB-4评分/ELF检测策略在阈值为9.8时显示出较高的阴性预测值(NPV;88-89%)和较低的阳性预测值(PPV;39-46%)。FIB-4评分/2D-SWE策略的NPV为91%,PPV为58-62%。年龄调整后的FIB-4评分阈值在所有算法中导致较低的NPV和PPV。
FIB-4评分/VCTE算法显示出优异的诊断性能,证明其适用于糖尿病学的常规筛查。使用9.8的调整后低阈值的ELF检测可作为VCTE的替代方法。