Abstract
BACKGROUND: Solitary pulmonary nodules (SPNs) are common incidental findings for which accurate characterization remains challenging on conventional computed tomography (CT), where phase-to-phase attenuation differences can yield equivocal results. Subtraction CT (SCT) generates motion-compensated iodine maps that quantify enhancement more robustly and can offer higher contrast-to-noise ratio (CNR) and lower dose than dual-energy CT (DECT). However, its value for differentiating SPNs has not been established. The aim of this study was to evaluate the diagnostic performance of iodine map metrics derived from SCT for differentiating SPNs and to assess their incremental value over that of conventional morphological features. METHODS: A total of 91 patients with pathologically confirmed SPNs (44 malignant, 25 benign, 22 inflammatory) were retrospectively analyzed. All underwent triple-phase SCT [non-contrast-enhanced phase (NC), arterial enhanced phase (AP), and venous enhanced phase (VP)] to generate subtraction iodine map of AP (SubA) and subtraction iodine map of VP (SubV) iodine maps. Nine CT value metrics were assessed. Interobserver agreement was evaluated using intraclass correlation coefficients (ICC). Diagnostic performance was analyzed using receiver operating characteristic (ROC) curves and DeLong tests. RESULTS: Five core metrics (NC, AP, VP, SubA, SubV) demonstrated excellent interobserver agreement (ICC >0.89). Hounsfield units (HU) of SubA showed the highest stability and significant differences across all three groups (benign: 21.2±11.1 HU, malignant: 34.5±16.6 HU, inflammatory: 43.9±10.7 HU, all P<0.05). SubA achieved the highest area under the curve (AUC) in distinguishing benign from inflammatory nodules (AUC =0.95, sensitivity =88%, specificity =95%). Combined models incorporating both SubA and morphological features improved diagnostic accuracy for all comparisons, significantly increasing AUCs for benign vs. malignant (0.82 to 0.92), inflammatory vs. benign (0.80 to 0.95), and inflammatory vs. malignant (0.83 to 0.88) (all P<0.05). CONCLUSIONS: Quantitative iodine values from SCT, particularly SubA, enhance the differentiation of SPNs with higher stability and diagnostic performance than conventional methods. The integration of SCT iodine maps with morphological assessment significantly improves diagnostic accuracy and supports clinical decision-making.