AASLD Practice Guidance on Acute-on-chronic liver failure and the management of critically ill patients with cirrhosis

Geographic disparities in access to liver transplant for advanced cirrhosis: Time to ring the alarm!

Decompensated cirrhosis and hepatocellular cancer are major risk factors for mortality worldwide. Liver transplantation (LT), both live-donor LT or deceased-donor LT, are lifesaving, but there are several barriers toward equitable access. These barriers are exacerbated in the setting of critical illness or acute-on-chronic liver failure. Rates of LT vary widely worldwide but are lowest in lower-income countries owing to lack of resources, infrastructure, late disease presentation, and limited donor awareness. A recent experience by the Chronic Liver Disease Evolution and Registry for Events and Decompensation consortium defined these barriers toward LT as critical in determining overall survival in hospitalized cirrhosis patients. A major focus should be on appropriate, affordable, and early cirrhosis and hepatocellular cancer care to prevent the need for LT. Live-donor LT is predominant across Asian countries, whereas deceased-donor LT is more common in Western countries; both approaches have unique challenges that add to the access disparities. There are many challenges toward equitable access but uniform definitions of acute-on-chronic liver failure, improving transplant expertise, enhancing availability of resources and encouraging knowledge between centers, and preventing disease progression are critical to reduce LT disparities.

Bridging the critically ill patient with acute to chronic liver failure to liver transplantation

Liver transplantation (LT) has emerged as an effective therapy for severe forms of acute-on-chronic liver failure (ACLF), an entity characterized by the development of multiorgan failure and high short-term mortality. The aim of critical care management of ACLF patients is to rapidly treat precipitating events and aggressively support failing organs to ensure that patients may successfully undergo LT or, less frequently, recover. Malnutrition and sarcopenia are frequently present, adversely impacting the prognosis of these patients. Management of critical care patients with ACLF is complex and requires the participation of different specialties. Once the patient is stabilized, a rapid evaluation for salvage LT should be performed because the time window for LT is often narrow. The development of sepsis and prolonged organ support may preclude LT or diminish its chances of success. The current review describes strategies to bridge severe ACLF patients to LT, highlights the minimal evaluation required for listing and the currently suggested contraindications to proceed with LT, and addresses different aspects of management during the perioperative and early posttransplant period.

Development of a Liver Disease-Specific Large Language Model Chat Interface using Retrieval Augmented Generation

Background

Large language models (LLMs) have significant capabilities in clinical information processing tasks. Commercially available LLMs, however, are not optimized for clinical uses and are prone to generating incorrect or hallucinatory information. Retrieval-augmented generation (RAG) is an enterprise architecture that allows embedding of customized data into LLMs. This approach "specializes" the LLMs and is thought to reduce hallucinations.

Methods

We developed "LiVersa," a liver disease-specific LLM, by using our institution's protected health information (PHI)-complaint text embedding and LLM platform, "Versa." We conducted RAG on 30 publicly available American Association for the Study of Liver Diseases (AASLD) guidelines and guidance documents to be incorporated into LiVersa. We evaluated LiVersa's performance by comparing its responses versus those of trainees from a previously published knowledge assessment study regarding hepatitis B (HBV) treatment and hepatocellular carcinoma (HCC) surveillance.

Results

LiVersa answered all 10 questions correctly when forced to provide a "yes" or "no" answer. Full detailed responses with justifications and rationales, however, were not completely correct for three of the questions.

Discussions

In this study, we demonstrated the ability to build disease-specific and PHI-compliant LLMs using RAG. While our LLM, LiVersa, demonstrated more specificity in answering questions related to clinical hepatology - there were some knowledge deficiencies due to limitations set by the number and types of documents used for RAG. The LiVersa prototype, however, is a proof of concept for utilizing RAG to customize LLMs for clinical uses and a potential strategy to realize personalized medicine in the future.