Quantitative Analysis of Cartilage Invasion by Laryngeal Carcinoma with Dual-Energy CT Iodine Overlay Images and Spectrum Curve

Background: Laryngeal cancer stage is essential for determining the appropriate treatment strategies and evaluating prognosis for laryngeal carcinoma, which is based primarily on laryngeal cartilage invasion or not. However, there was no optimal means to confirm whether cartilage was being invaded before operation. With dual-energy CT development, it provided a new potential approach for laryngeal cartilage invasion diagnosis. Objectives: To evaluate the clinical value of dual-energy computed tomography with iodine overlay images and spectrum curve in the quantitative analysis of laryngeal cartilage invasion in patients with laryngeal carcinoma. Patients and Methods: Dual-source dual-energy CT was performed for one hundred and twenty patients with confirmed diagnosis of laryngeal carcinoma by surgical pathology (56 cases with laryngeal cartilage invasion and 64 cases without invasion). The mean CT attenuation values of laryngeal cartilage and the tumor in routine scan, arterial and venous phases was measured and cartilage invasion was considered to be present with asymmetric sclerotic changes, erosion, lysis and extralaryngeal tumor spread with iodine overlay images and routine scan images. Moreover, the sensitivity and specificity were analyzed. The slope of spectral curve and iodine concentration of the tumors, invaded laryngeal cartilage and non-invaded laryngeal cartilage were compared between arterial phase and venous phase. Furthermore, the receiver operating characteristic (ROC) curve of mean CT attenuation value, the slope of spectral curve and iodine concentration were made comparative analysis. Results: The mean CT attenuation values of the tumors, invaded laryngeal cartilage and non-invaded laryngeal cartilage had significant difference among the routine scan, arterial phase and venous phase (P < 0.05). The sign of asymmetric sclerotic changes had no significant difference between iodine overlay images and routine scan images (the same specificity and sensitivity, 77.8% and 52.6%, respectively). However, the signs of erosion, lysis and extralaryngeal tumor spread had significant difference. The specificity and sensitivity of iodine overlay images were superior to that of the routine scan images in the three signs (specificity and sensitivity of erosion: 92.3% vs. 61.5%, and 86.7% vs. 60.0%, respectively; lysis: 82.4% vs. 69.2%, and 81.8% vs. 66.7%, respectively; extralaryngeal tumor spread: 90.9% vs. 68.3%, and 100% vs. 63.5% respectively). The slope of spectral curve and iodine concentration of the tumors, invaded laryngeal cartilage and non-invaded laryngeal cartilage showed significant difference (P < 0.05). The iodine concentration of the tumor was higher than invaded laryngeal cartilage. The area under the curve (AUC) value of mean CT attenuation value in routine scan, arterial phase and venous phase were 0.064, 0.265, and 0.192, respectively. The AUC value of the slope of spectral curve in arterial phase was 0.970 and its selected cutoff was considered as 1.09. In this cutoff point, the diagnostic indices including sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were 88.9%, 100%, 100%, and 82.6%, respectively, while the AUC value of the slope of spectral curve in venous phase was 0.944 and its selected cutoff was considered as 0.95. In this cutoff point, the diagnostic indices including sensitivity, specificity, PPV, NPV, positive likelihood ratio (PLR), negative likelihood ration (NLR) were 81.5%, 96.3%, 91.7%, 77.3%, 7.76, and 19.30, respectively. The AUC value of the slope of iodine concentration in arterial phase was 0.995 and its selected cutoff was considered as 1.05. In this cutoff point, the diagnostic indices including sensitivity, specificity, PPV and NPV were 92.6%, 100%, 95.8%, and 100%, respectively. Moreover, the AUC value of the slope of iodine concentration in venous phase was 0.905 and its selected cutoff was considered as 0.75. In this cutoff point, the diagnostic indices including sensitivity, specificity, PPV, NPV, PLR, and NLR were 77.8%, 88.9%, 91.3%, 82.6%, 7.00, and 24.5, respectively. Conclusion: Dual-energy CT can provided a new potential approach for laryngeal cartilage invasion diagnosis in order to guide the selection of optimal operation plan and improve the prognosis.