Abstract
The Ki-67 labeling index (Ki-67 LI) and microvascular invasion (MVI) are critical prognostic biomarkers in hepatocellular carcinoma (HCC). Preoperative, noninvasive prediction of their dual positivity status remains challenging. This study aimed to develop and validate a combined model integrating radiomic features from gadoxetic acid disodium (Gd-EOB-DTPA)-enhanced magnetic resonance imaging (MRI) with clinical data. It also aimed to assess the effectiveness of combining Gd-EOB-DTPA-enhanced MRI radiomics with clinical features for the preoperative prediction of Ki-67/MVI dual positivity in HCC. A total of 142 pathologically confirmed HCC patients were categorized into dual-positive (Ki-67 LI > 20% and MVI-positive) and non-dual-positive (Ki-67 LI ≤ 20% and/or MVI-negative) groups. Clinical variables (sex, age, hepatitis status, tumor diameter, alpha-fetoprotein [AFP], liver function, inflammatory indices, and tumor differentiation), along with MRI data, were analyzed. Radiomic features were extracted from hepatobiliary-phase regions of interest. Key predictors were selected using the least absolute shrinkage and selection operator and multivariate logistic regression to construct the nomogram. The model was evaluated using the receiver operating characteristic curves, calibration plots, and decision curve analysis. Tumor diameter, AFP, gamma-glutamyl transferase, and differentiation grade significantly differed between the groups (P < .05), and 6 radiomic features were selected to generate a radiomics score. Multivariate analysis identified tumor diameter, AFP, and radiomics score as independent predictors of dual positivity for Ki-67/MVI. The combined model demonstrated excellent calibration and superior predictive performance, achieving an area under the curve of 0.879, sensitivity of 70.0%, specificity of 92.3%, accuracy of 78.8%, precision of 93.3%, and F1-score of 80%. Follow-up after surgery showed a significantly higher early recurrence rate in the dual-positive HCC (Ki-67/MVI) group than that in the non-dual-positive group (P < .05). The Gd-EOB-DTPA-enhanced MRI radiomics-clinical nomogram effectively predicted preoperative Ki-67/MVI dual positivity in HCC. This combined method surpassed the individual-modality models, providing significant assistance in risk assessment and tailored treatment planning for patients with HCC.