Development of a post-treatment prognostic model for hepatocellular carcinoma based on nutritional, immune, and inflammatory scoring systems and REDCap-enabled follow-up

TL;DR

This study developed a prognostic model for hepatocellular carcinoma patients using pre-treatment inflammation, nutrition, and immune scores to predict survival after treatment.

Contribution

The novel contribution is identifying PALBI and CRAFITY as independent prognostic factors and constructing a risk score model for hepatocellular carcinoma survival prediction.

Findings

PALBI and CRAFITY scores were identified as independent predictors of disease-free survival in HCC patients.

The constructed risk score model stratified patients into high- and low-risk groups with significantly different survival outcomes.

Inflammation-related scores showed better predictive value for disease-free survival compared to metabolism- and immune-related scores.

Abstract

This study examined the association between pre-treatment inflammation, immune cell- and nutrition/metabolism-related scores, and prognosis of patients with hepatocellular carcinoma (HCC) post-treatment.

This study collected clinical data on demographics, pretreatment blood tests, pathology, and follow-up. Key markers included C-reactive protein, albumin, neutrophil and lymphocyte counts, creatinine, bilirubin, international normalized ratio, tumor size and number, alpha-fetoprotein, platelet count, and CD4+/CD8+ T-cell levels. Disease-free survival (DFS) was calculated from treatment to recurrence. Twelve scores were derived. Kaplan–Meier and univariate Cox analyses identified significant predictors, followed by multivariate Cox models to determine independent risk factors. Logistic regression and receiver operating characteristic (ROC) analyses assessed predictive performance. Scores were grouped as inflammation-, metabolism-, or immune-related to construct nomograms and evaluate C-index values using R software.

Except for Gender (p = 0.019), all other clinical characteristics showed no statistically significant differences between the training and validation sets (p > 0.05).Univariate Cox regression showed that pre-albumin (P = 0.01), PNI (P < 0.001), TBS (P = 0.01), ALBI (P < 0.001), PALBI (P < 0.001), and CRAFITY (P < 0.001) were significantly associated with DFS. Multivariate analysis identified PALBI (P = 0.03) and CRAFITY (P = 0.04) as independent predictors. A prognostic model was constructed: Risk score = 0.03903 × TBS + 0.79809 × PALBI + 0.40881 × CRAFITY, stratifying patients into high- and low-risk groups. Kaplan–Meier analysis showed significantly better DFS in the low-risk group (P = 0.001). ROC analysis for 1- and 2-year DFS yielded AUCs of 0.69 and 0.75. Logistic regression confirmed the risk score as a predictor of mortality (P = 0.002, AUC = 0.644). Excluding TBS, the remaining scores were grouped into inflammation-related, nutrition/metabolism-related, and immune-related categories. Corresponding nomograms showed good calibration, with C-index values of 0.610, 0.581, and 0.575, respectively.

Pre-treatment PALBI and CRAFITY scores are independent prognostic factors for post-treatment survival among patients with HCC, with inflammation-related scores providing superior predictive value for DFS compared to metabolism- and immune-related scores.

Figures