The value of dual-layer detector spectral CT multiparameter imaging in distinguishing between benign lesions and carcinomas of the larynx and hypopharynx

Background: Accurately distinguishing between benign and malignant lesions in the larynx and hypopharynx is crucial for formulating individualized and precise treatment plans. Due to the low soft-tissue resolution of conventional computed tomography (CT), the diagnostic accuracy of laryngeal and hypopharyngeal lesions is not satisfactory in clinical practice. Therefore, this study aimed to assess the value of dual-layer detector spectral computed tomography (DLSCT) multiparameter imaging for distinguishing between laryngeal and hypopharyngeal benign lesions (LHBLs) and laryngeal and hypopharyngeal carcinomas (LHCs) and to identify the optimal DLSCT parameters.

Methods: The data of 141 patients were retrospectively analyzed. Spectral CT parameters, including those derived from the arterial phase (AP) and venous phase (VP), as well as the maximum lesion diameter (MLD), were measured and analyzed. The diagnostic performance of each spectral parameter was evaluated using receiver operating characteristic (ROC) curve analysis and the area under the curve (AUC). The DeLong test was then used to compare differences in diagnostic performance. The parameters demonstrating the highest AUC in the AP and VP were integrated via logistic regression. Clinical utility was assessed by decision curve analysis (DCA).

Results: With the exception of the virtual non-contrast CT values in the AP and the standardized effective atomic number (sZeff) in both the VP and AP, all the measured spectral CT parameters differed significantly between the LHBLs and LHCs (P<0.002). Among all the single parameters analyzed, the 40-kiloelectron volt (keV) virtual monochromatic image (VMI) in the VP achieved the highest AUC of 0.932, which was significantly higher than that of the conventional image (CI) (P<0.05). The AUC values for the CIs in the AP and VP were 0.837 and 0.866, respectively. When the 40 keV-AP VMI, 40 keV-VP VMI, and MLD were combined in a model, the AUC increased to 0.970, which was significantly higher than the AUC of the model that combined the 40 keV-AP VMI and MLD (P=0.009), but did not differ significantly from the AUC of the model that combined the 40 keV-VP VMI and MLD (P=0.403). The model that combined the 40 keV-AP VMI, 40 keV-VP VMI, and MLD had the greatest net clinical benefit, but it was only superior to the model that combined the 40 keV-VP VMI and MLD within certain risk threshold ranges.

Conclusions: DLSCT exhibited better diagnostic performance for distinguishing between LHBLs and LHCs than conventional CT. Notably, 40 keV-VP VMI demonstrated favorable diagnostic value and may serve as a potential auxiliary imaging tool for this differential diagnosis.