MR Imaging Of Thyroid Nodules And Correlation With Pathology

Part I MR Imaging Characteristics of Nodular Goiter and Correlation with PathologyObjective To investigate the magnetic resonance imaging (MRI) features of nodular goiter (NG) and correlate with histopathological results. Materials and Methods Fifty five patients with NG proven by surgery and pathology underwent the MR examination. The number, size, shape, margin, signal characteristics, and enhancement patterns of NG were evaluated. MR imaging and pathological correlation were performed. Results One hundred and ten nodules were found in 55 patients with single nodule in 19 (34.5%) cases, and multiple nodules in 36 (65.5%) cases on MR images. One hundred and one (91.8%) nodules were well-defined and had regular margin. The remaining 9 (8.2%) noduleswere obscure and had irregular margin. Sixty four (58.2%) nodules were solid and other 46 (41.8%) nodules were cystic. Four types of signal pattern were identified as follows:1. the isointense or slight hyperintense signal on T1-weighted imaging (TIWI) and hyperintense on T2-weighted imaging (T2WI) in 58 (52.7%) nodues,2. the hyperintense signal on both T1WI and T2WI in 37 (33.6%) nodules,3. the hypo-, iso-, and hyperintense on signal T1WI and hypointense on T2WI in 9 (8.2%) nodules, and 4. the isointense signal on both T1WI and T2WI in 6 (4.5%) nodules. After administration of contrast media,54 of 64 solid NG nodules manifested marked enhancement compared with backgrounding thyroid tissue. The remaining 10 nodules showed moderate enhancement compared with backgrounding thyroid tissue. Among 46 cystic NGs,31 nodules showed ring enhancement,11 NGs had intracystic mural nodules with marked enhancement in 5 and moderate enhancement in 6 mural nodules, and 10 NGs demonstrated no enhancement. Grossly, the multiple nodules pattern of NG manifested as different size of nodules, translucent appearance with whitish gray or brownish color, tender or hard architecture, and well-circumscribed character. The incomplete thin capsule was seen in 19 nodules, while slightly thick fibrous capsule was seen in 91 nodules. The single nodule pattern of NG was whitish gray. The hemorrhagic or necrotic degeneration presented in 8 nodules with viscous bloody fluid outflowing on dissected section. The calcifications in fibrous scar or necrotic area were seen in 11 nodules. Microscopically, the solid nodule manifestes as the adenomatous hyperplastic or mixed with colloid nodule. The cystic nodule was the colloid nodule, or hemorrhagic or necrotic nodule, while the enhanced intracystic mural nodule was adenomatous hyperplastic nodule.Fifty five patients with NG accompanied thyroid adenoma diagnosed correctly by MRI in 4 cases, thyroid carcinoma misdiagnosed by MRI in 5 cases, and Hashimoto’s thyroiditis all missed by MRI in 5 cases. Conclusion NG has MR imaging characteristics which are consistent with the corresponding histopathological changes.Part II MR Imaging Characteristics of Thyroid Carcinoma and Correlation with PathologyObjecive To investigate the magnetic resonance imaging (MRI) features of thyroid carcinoma and correlate with histopathological results. Materials and Methods Twenty-seven patients with surgically and pathologically proven thyroid carcinoma underwent MRI examination, including 23 with papillary carcinomas, two with follicular carcinoma, one with undifferentiated carcinoma and one with metastasis. The number, size, shape, margin, signal characteristics, and enhancement patterns of nodules were evaluated and correlated with pathological findings. Results Thirty-three nodules were found with solitary nodule in 23 (85%) cases and multiple nodules in 4 (15%) cases, solid in 27 (82%) nodules and cystic in 6 (18%) nodules. Thirty (91%) nodules were ill-defined and irregular, and the remaining 3 (9%) nodules were well-defined and regular. On T1WI,31 nodules appeared the isointense, one nodule was slight hypointense, and one nodules was mixed iso-, and hyperintense signal. On T2WI,12 nodules were homogenous hyperintense,4 nodules were hyperintense with isointense,15 nodules appeared heterogeneous isointense and slight hyperintense signal, and the remaining 2 nodules were isointense signal. Incomplete low signal ring was identified in 9 nodules. After administration of contrast media, all nodules demonstrated mild to marked enhancement. The markedly enhanced rim and moderately enhanced center was seen in 8 nodules. Six nodules showed residual ring sign which appeared as incomplete enhanced ring and pseudocapsule proven by pathology. Six nodules showed homogeneous enhancement with marked enhancement in 4 cases and lower enhancement in 2 cases compared with background thyroid tissue. Nineteen nodules showed heterogeneous enhancement with marked enhancement in 11 and lower enhancement in 8 cases compared with background thyroid tissue. Seven patients appeared as metastatic lympadenopathy in the neck with heterogeneous signal and enhancement in 5 cases. Histopathologically, thyroid carcinoma are often whitish gray lesions with irregular shape, representing with a pure papillary or apapillary with follicular patterns, with intratumoural sclerosis. The proportion of above-mentioned tumor components is usually variabl. The concomitant diseases included NGs in 6 cases and Hashimoto’s thyroiditis in one case. Conclusion MRI can demonstrate valuable imaging characteristics of thyroid carcinoma which are helpful for imaging diagnosis and differentiation.PartⅢLogistic Regression Analysis of Combining Conventional and Functional MRI Features of Solitary Thyroid NoduleObjective To create a solitary thyroid nodule estimation based on conventional and functional MRI using Logistic regression that can screen out the specific features for distinguishing malignan from benign nodules. Materials and Methods One hundred thirty-two patients with solitary thyroid nodules confirmed by pathology were enrolled in the study. MRI features were evaluated as follows:shape, margin, texture, signal intensity (SI) on T1 WI and T2WI, lymphadenectasis, cystic degeneration of lymph node, SI on DW images, apparent diffusion coefficient (ADC) value and time-intensity curve (TIC). Logistic regression model was established to screen out significant MRI features for differentiating malignant from benign nodules. A receiver operating characteristics (ROC) curve was made to assess diagnostic value of the Logistic regression model. Correlation was analyzed between the Logistic regression model and pathological results. Results Logistic regression model:Logit (P)=-3.254+1.096×margin+5.731×cystic degeneration of lymph node+2.165× ADC+1.417×TIC. The diagnostic sensitivity, specificity and accuracy were 94.7%, 91.5% and 92.4%, respectively. The area under ROC curve was 0.935. Conclusion The margin, cystic degeneration of lymph node, ADC value and TIC pattern play an important role for the differentiation of benign and malignant solitary thyroid nodules. The Logistic regression model of combining conventional and functional MRI features could effectively differentiate malignant from benign solitary thyroid nodules.