TY  - JOUR
T1  - Clinical–Inflammatory Phenotypes and Circulating hsa_circ_101555 Along the Cirrhosis–Hepatocellular Carcinoma Continuum: A Cross-sectional Study 
AU  - Badwei, Nourhan
AU  - Moez, Amal Tohamy Abdel
AU  - Salem, Houssam El-Deen M.
AU  - El-Khazragy, Nashwa
AU  - Gado, Mohammed Soliman
JF  - Gene Expression
VL  - 25
IS  - 3
SN  - 1555-3884
SP  - e00023
EP  - e00023
Y1  - 2026-07-30
DO  - 10.14218/GE.2026.00023
UR  - https://www.xiahepublishing.com/1555-3884/GE-2026-00023
AB  - Background and objectives 
The clinical spectrum from cirrhosis to hepatocellular carcinoma (HCC) reflects interactions among hepatic dysfunction, systemic inflammation, and tumor burden. Whether circulating biomarkers add value beyond clinical parameters remains uncertain. This study aimed to evaluate an exploratory Model for End-Stage Liver Disease (MELD)–neutrophil-to-lymphocyte ratio (NLR)-based clinical–inflammatory phenotyping framework for discriminating advanced HCC features and to determine whether circulating hsa_circ_101555 provides incremental discriminatory value beyond this framework.
Methods
This single-center cross-sectional study included 92 consecutive patients (30 with cirrhosis without HCC and 62 with HCC). Patients were classified into three exploratory clinical–inflammatory phenotypes using a hierarchical MELD–NLR algorithm (Phenotype I, n = 25; II, n = 33; III, n = 34). Circulating hsa_circ_101555 was quantified by reverse transcription quantitative polymerase chain reaction. Receiver operating characteristic analysis evaluated Barcelona Clinic Liver Cancer stage C among patients with HCC (n = 62; events = 28). Internal validation used bootstrap resampling.
Results
Higher-risk phenotypes included progressively larger proportions of patients with HCC and greater frequencies of advanced tumor characteristics. The combined MELD–NLR model showed the highest discrimination (the area under the receiver operating characteristic curve (AUC) 0.90; 95% confidence interval 0.82–0.97), with 85.7% sensitivity, 82.4% specificity, and 83.9% accuracy. This performance exceeded that of NLR alone (AUC, 0.80) and MELD alone (AUC, 0.77). Circulating hsa_circ_101555 was associated with smaller tumors and an earlier Barcelona Clinic Liver Cancer stage but showed modest discrimination (AUC, 0.69) and did not improve the MELD–NLR model (ΔAUC = 0.002; DeLong P = 0.79).
Conclusions
The exploratory MELD–NLR-based clinical–inflammatory framework identifies patient groups with differing frequencies of advanced HCC features. Circulating hsa_circ_101555 provides no incremental discriminatory value beyond routinely available clinical–inflammatory parameters and requires external validation before clinical use.