The Role of MRI in the Diagnosis of Papillary Thyroid Carcinoma with Hobnail Features: A Case-Control Study.
Lanyun Wang,Bin Song,Hao Wang,Yongqi Chen,Weiyan Liu,Ran Wei,Zedong Dai,Wenjuan Hu,Yi Ding
DOI: https://doi.org/10.1166/jmihi.2020.2983
2020-01-01
Journal of Medical Imaging and Health Informatics
Abstract:Purpose—To evaluate whether Magnetic Resonance Imaging (MRI) preoperatively could be used as a potentially routine tool to recognize papillary thyroid carcinoma with hobnail features (HPTC). Methods—This retrospective study included 97 patients with confirmed papillary thyroid carcinoma (PTC) who underwent thyroid MRI examination one week before surgery between January 2014 and June 2017. The differences in age, sex, tumor size, each MRI feature (lesion shape; signal heterogeneity on T1WI, T2WI and diffusion weighted imaging (DWI); significantly high signal on T2WI, segmentation linear low signal on T2WI, T2WI lobulation, T2WI ripple-like performance; and early enhancement degree, delayed enhancement pattern) and apparent diffusion coefficient (ADC) values of HPTC and PTC without hobnail features (NHPTC) were determined. The evaluation indicators of significant MRI features were determined. Independent-samples t test, chi-square test, Fisher's exact test, nonparametric test, and analyses of receiver operating characteristic (ROC) curve were performed. Results—There were 10 patients with HPTCs and 81 patients with NHPTCs included in the study. Compared to the NHPTCs, the HPTCs had larger volumes (P = 0.025), more irregular shapes (P = 0.004), more heterogeneous T2WI and DWI signals (P = 0.028, 0.006, respectively), higher frequencies of significantly high signals on T2WI, segmentation linear low signals, lobulation and ripple-like performances (P = 0 000 all), and higher ADC values (P = 0 036). In addition, the delayed mass-like enhancement was more common in HPTCs (P = 0.046). Characteristics including significantly high signal on T2WI, segmentation linear low signal, lobulation, ripplelike performance and a size ≥2.15 cm have significant power to predict HPTCs. Characteristics including the absence of irregular shape, a heterogeneous signal on T2WI or a significantly high signal on T2WI have significant power to predict NHPTCs. Conclusions—HPTC lesions have unique MRI features distinct from NHPTCs and recognizing these features may help to assess the invasiveness of PTC preoperatively.