Differentiation of Benign and Malignant Liver Tumors Using Non-Enhanced Spectral CT Quantitative Parameters Combined with Inflammatory Indicators

Purpose This study aims to evaluate the effectiveness of using quantitative parameters from non-contrast dual-layer spectral CT combined with inflammatory markers to differentiate between benign and malignant hepatic tumors and to assess their diagnostic performance. Materials and Methods A retrospective analysis was conducted on 119 patients with hepatic lesions who underwent dual-layer spectral CT scanning. Clinical data, including gender, age, albumin-to-total bilirubin ratio (ABR), neutrophil-to-lymphocyte ratio (NLR), lymphocyte-to-monocyte ratio (LMR), and platelet-to-lymphocyte ratio (PLR), were collected. Measurement of spectral CT parameters in lesions on plain scan images, including CT values of virtual monoenergetic images at 40 keV, 70 keV, and 140 keV, effective atomic number, and electron density. Clinical and spectral CT parameters were screened using logistic regression, and the discriminative efficacy of these parameters in differentiating hepatic vascular carcinoma from hepatocellular carcinoma was evaluated using the area under the ROC curve alongside logistic regression models. Results Multivariable logistic regression identified the lymphocyte-to-monocyte ratio (LMR), Zeff-range, and electron density as independent predictors for distinguishing between benign and malignant liver tumors. Receiver operating characteristic (ROC) analysis revealed no significant difference in the area under the curve (AUC) between quantitative spectral CT parameters (AUC= 0.79, 95% CI: 0.70-0.86) and clinical parameters (AUC=0.75, 95% CI: 0.67-0.83; P=0.551). However, an integrated model combining both parameter sets achieved superior diagnostic performance (AUC=0.85, 95% CI: 0.78-0.91). Conclusion Both non-enhanced dual-layer spectral CT quantitative parameters and clinical parameters can effectively distinguish between benign and malignant liver tumors, and their combination can further enhance diagnostic performance.