The latest medical research on Transplant Hepatology

Cut-points for non-invasive tests (NITs) for risk stratification in metabolic dysfunction-associated steatotic liver disease (MASLD) were derived from predominantly non-Hispanic populations. It is unknown if these cut-points perform adequately in Hispanic individuals. We assessed the performance characteristics of current NIT cut-points among Hispanic patients and determined whether they could be further optimized.

We prospectively enrolled 244 adults with biopsy-proven MASLD. Participants underwent a research visit with magnetic resonance elastography (MRE) and vibration controlled transient elastography (VCTE). Histology and imaging assessments were conducted centrally. Diagnostic performance was evaluated by area under the receiver-operating curve (AUROC) and optimal cut-points were identified by Youden J analysis. The mean (±SD) age and body mass index were 52.6 (±13) and 31.6 (±4.6) kg/m2. Overall, 40% had diabetes, 31% (N=75) were Hispanic. 40% of Hispanic and 28.4% of non-Hispanic patients had significant fibrosis. To detect significant fibrosis, MRE and VCTE exhibited significantly lower accuracy in Hispanic versus non-Hispanic participants (AUROC: MRE, 0.87 vs. 0.98, p=0.01; VCTE, 0.78 vs. 0.92, p=0.02). Clinical care algorithms yielded high false-negative rates among Hispanic participants (14% with low-risk FIB-4 and 21% with low-risk VCTE had advanced fibrosis on biopsy). Cut-points of 2.73 kPa for MRE and 6.9 kPa for VCTE were optimal to detect significant fibrosis in Hispanic individuals. Findings were validated in a Latin American cohort.

Lower NIT cut-points may be needed to optimize surveillance for significant fibrosis due to MASLD in Hispanic populations commensurate with their higher burden and severity of disease.

Chronic hepatitis C virus (HCV) infection afflicts around 50 million people globally, causing ~250,000 deaths yearly. An effective vaccine needs to overcome high viral diversity and HCV's ability to evade neutralizing antibodies (NAbs). Rapid antigenic drift in the N-terminal motif of envelope protein E2, named hypervariable region 1 (HVR1), is critically involved in NAb evasion via an incompletely understood mechanism involving viral entry factors. The canonical length of HVR1 is 27 amino acids, but insertions of 2-4 amino acids was described in patients infected with genotype 1b. We aimed at determining whether HVR1 insertions may be underreported due to extreme HVR1 variability.

We observed a 0.7% HVR1 insertion prevalence in routine NGS patient contigs. Thus, we performed direct sequence analysis of E1E2 sequences from 131 HCV infected patients. Interestingly, we observed that 3% of patients harbored viruses (genotype 1a, 2b, 3a) with dominant HVR1 insertions. Insertion of longer non-canonical HVR1s into HCV cell culture recombinants frequently caused loss of fitness. However, culture-viable viruses with HVR1 insertions were fully viable in vivo. Interestingly, in adapted genotype 1b recombinants with HVR1 insertions, we found internal HVR1 deletions, that increased antibody sensitivity, which surprisingly correlated more with reduced LDLr than reduced SR-BI dependency, indicating a role of LDLr in NAb evasion. Conversely, HVR1 insertions had no effect on receptor dependency, however, they modulated epitope-specific NAb sensitivity.

HVR1 insertion prevalence and NAb sensitivity modulation indicate they represent a mechanism by which HCV evades emerging NAbs during infection.

Diagnosis code classification is a common method for cohort identification in cirrhosis research, but it is often inaccurate and augmented by labor-intensive chart review. Natural language processing (NLP) using large language models (LLMs) is a potentially more accurate method. To assess LLMs' potential for cirrhosis cohort identification, we compared code-based versus LLM-based classification with chart review as a "gold standard."

We extracted and conducted a limited chart review of 3,788 discharge summaries of cirrhosis admissions. We engineered zero-shot prompts using Generative Pre-trained Transformer (GPT)-4 to determine whether cirrhosis and its complications were active hospitalization problems. We calculated positive predictive values (PPVs) of LLM-based classification versus limited chart review, and PPVs of code-based versus LLM-based classification as a "silver standard" in all 3,788 summaries.

Versus gold standard chart review, code-based classification achieved PPVs of 82.2% for identifying cirrhosis, 41.7% hepatic encephalopathy, 72.8% ascites, 59.8% gastrointestinal bleeding, and 48.8% spontaneous bacterial peritonitis. Compared to chart review, GPT-4 achieved 87.8-98.8% accuracies for identifying cirrhosis and its complications. Using LLM as a silver standard, code-based classification achieved PPVs of 79.8% for identifying cirrhosis, 53.9% hepatic encephalopathy, 55.3% ascites, 67.6% gastrointestinal bleeding, and 65.5% spontaneous bacterial peritonitis.

LLM-based classification was highly accurate versus manual chart review in identifying cirrhosis and its complications - this allowed us to assess the performance of code-based classification at scale using LLMs as a silver standard. These results suggest LLMs could augment or replace code-based cohort classification and raise questions regarding the necessity of chart review.