Using Quantitative Parameters Derived from Pretreatment Dual-Energy Computed Tomography to Predict Histopathologic Features in Head and Neck Squamous Cell Carcinoma.
Hesong Shen,Yuanying Huang,Xiaoqian Yuan,Daihong Liu,Chunrong Tu,Yu Wang,Xiaoqin Li,Xiaoxia Wang,Qiuzhi Chen,Jiuquan Zhang
DOI: https://doi.org/10.21037/qims-21-650
2022-01-01
Quantitative Imaging in Medicine and Surgery
Abstract:BACKGROUND:Head and neck squamous cell carcinoma (HNSCC) patients with a high tumor grade, lymphovascular invasion (LVI), or perineural invasion (PNI) tend to demonstrate a poor prognosis in clinical series. Thus, the identification of histopathological features, including tumor grade, LVI, and PNI, before treatment could be used to stratify the prognosis of patients with HNSCC. This study aimed to assess whether quantitative parameters derived from pretreatment dual-energy computed tomography (DECT) can predict the histopathological features of patients with HNSCC.METHODS:In this study, 72 consecutive patients with pathologically confirmed HNSCC were enrolled and underwent dual-phase (noncontrast-enhanced phase and contrast-enhanced phase) DECT examinations. Normalized iodine concentration (NIC), the slope of the spectral Hounsfield unit curve (λHU), and normalized effective atomic number (NZeff) were calculated. The attenuation values on 40-140 keV noise-optimized virtual monoenergetic images [VMIs (+)] in the contrast-enhanced phase were recorded. The diagnostic performance of the quantitative parameters for predicting histopathological features, including tumor grade, LVI, and PNI, was assessed by receiver operating characteristic curves.RESULTS:The NIC, λHU, NZeff, and attenuation value on the VMIs (+) at 40 keV (A40) in the grade III group, LVI-positive group, and PNI-positive group were significantly higher than those in the grade I and II groups, the LVI-negative group, and the PNI-negative group (all P values <0.05). A multivariate logistic regression model combining these 4 quantitative parameters improved the diagnostic performance of the model in predicting tumor grade, LVI, and PNI (areas under the curve: 0.969, 0.944, and 0.931, respectively).CONCLUSIONS:Quantitative parameters derived from pretreatment DECT, including NIC, λHU, NZeff, and A4,0 were found to be imaging markers for predicting the histopathological characteristics of HNSCC. Combining all these characteristics improved the predictive performance of the model.