AASLD访谈 | Laurie Deleve教授:纵览会议全景,人工智能离肝病临床越来越近

2021/11/19 10:22:25 国际肝病网
 
第72届美国肝病研究学会年会(AASLD2021)已于2021年11月12日至15日在美国加利福尼亚州阿纳海姆以在线的形式召开。本次会议亮点十足,美国南加利福尼大学凯克医学中心Laurie DeLeve教授在接受《国际肝病》采访时,总体介绍了本次会议上的一些值得关注的领域以及亮点主题,其中也提到了本次会议上关于人工智能在肝脏疾病中的应用研究专题。以下将带来Laurie DeLeve教授的采访报道以及筛选的关于人工智能在评估肝纤维化方面的研究报道。
 
AASLD2021会议聚焦
 
Laurie DeLeve教授:本届会议上有太多太多可以推进肝病学可发展的专场讨论,涉及十分广泛的范围。在未来的新研究方法方面,“Meet the Expert”环节的“单细胞分析在基础和临床研究中的应用”、在肝脏研究中人工智能导航基础知识相关的临床研究研讨会以及Hans Popper教授在“State of the Art Lecture”环节的关于“使用单细胞基因组学方法解码人类肝脏疾病”的讲座值得大家关注。
 
此外,在疾病领域,NAFLD、酒精相关性肝病和病毒性肝炎三者的重要性均不容忽视。鉴于酒精和非酒精性脂肪性肝病的重要性,今年的“Postgraduate Courses”环节特别关注了酒精和代谢综合征引起的脂肪肝。AASLD2021上报道了丰富的脂肪肝相关研究,我强烈鼓励大家参与这些会议。消除病毒性肝炎是AASLD会议关注的一个重点,我们不仅在年会议程中对其做了专门的设计,而且它仍然会是AASLD全年的关注重点,我们参与了许多相关的会议活动,以推动肝病科学的发展。
 
Laurie Deleve: This AASLD conference includes academic content on multiple topics. Among these content, what do you pay most attention to or think is the most valuable and promising for research? There are so many sessions this year that are advancing the science of liver disease that it is very difficult to narrow it down. We have seen advances in a wide swath of areas so I would not and could not pick out any one particular area. However, to move us into the future with new research methods, I would draw attention to Meet the Expert Session 16 on Single Cell Analysis in Basic and Clinical Research, as well as the Clinical Research Workshop on the basics of navigating artificial intelligence in liver research. I would also highly recommend the Hans Popper State of the Art Lecture on Using Single Cell Genomics Approaches to Decode Human Liver Disease.  
 
In the field of disease, I can’t single out one, because the trio of NAFLD, alcohol associated liver disease and viral hepatitis are too important to overlook. We have not done disease-focused postgraduate courses in recent years. However, given the importance of alcohol and NAFLD, this year’s postgraduate course is specifically focused on fatty liver disease from both alcohol and metabolic syndrome. Research and MTEs on fatty liver disease are well represented throughout TLMdX and I highly encourage participation in these events. The elimination of viral hepatitis is a key focus of AASLD, and we devote programming to it at TLMdX, but it remains an important focus for AASLD year-round and we engage in many events and opportunities to move the science forward.
 
Oral-78:采用人工智能分析的数字病理学(DP-AI)克服了当前评分系统在评估NASH F3患者纤维化消退方面的局限性
 
在NASH试验中,纤维化消退且无肝纤维化恶化是既定的主要终点。在评估纤维化消退时应用该终点的公认局限性——当前评分系统的分类性质,不能捕捉晚期纤维化或肝硬化患者(F3~F4)在同一阶段的胶原纤维变化;F3期胶原蛋白沉积范围广等对NASH患者的药物开发提出了挑战。DP-AI是一种新的方法,可在线性范围内提供敏感、可重复的纤维化参数定量。本研究的目的是根据传统评分方法,应用DP-AI评估F3纤维化“稳定”患者治疗干预后的纤维化动态(AASLD2021, oral-78.)[1]
 
研究者对来自FLIGHT-FXR研究(NCT02855164)的99名NASH患者(F2=42, F3=57)的198次配对肝活检数据进行了分析,其中这些患者分别接受安慰剂(PBO;n=34)、tropifexor 140 μg(n=37)或tropifexor 200 μg(n=28)治疗了48周。使用二次谐波(SHG)和双光子激发荧光(TPEF)显微成像在基线检查(BL)和治疗结束(EOT)时对肝纤维化(qFibrosis)进行定量。
 
结果显示,与常规NASH CRN评分相比,BL和EOT的qFibrosis显示更多患者出现纤维化消退,尤其是在F3病例中。总的来说,qFibrosis连续值随着纤维化分期的增加而增加,随着纤维化分期的降低而减少。在那些没有纤维化分期变化的患者中,接受tropifexor治疗的患者qFibrosis评分降低,表明其能够捕捉到传统分期未识别的变化。tropifexor治疗组患者中隔宽度的减小进一步证明了抗纤维化作用的有效性。
 
综上所述,在分类NASH-CRN系统中,DP-AI可以以完全定量的方式量化纤维化量和间隔参数(例如间隔厚度)的变化。因此,DP-AI可以为NASH试验中的纤维化消退和干预效果提供更详细的评估。
 
参考来源:
 
[1]Dean Tai, Yayun Ren, Nikolai V Naoumov, et al. DIGITAL PATHOLOGY WITH ARTIFICIAL INTELLIGENCE ANALYSES (DP-AI) OVERCOMES THE LIMITATIONS OF CURRENT SCORING SYSTEMS IN ASSESSING FIBROSIS REGRESSION FOR NASH F3 PATIENTS. AASLD2021, oral-78.