The value of multi-parameters MRI of tongue carcinoma in predicting cervical node metastases
10.3760/cma.j.issn.1005-1201.2016.05.008
- VernacularTitle:多参数MRI预测舌癌颈部淋巴结转移的价值
- Author:
Wei GUO
;
Dehong LUO
;
Lin LI
;
Meng LIN
;
Yanfeng ZHAO
;
Liang YANG
;
Lei HU
;
Xinming ZHAO
;
Chunwu ZHOU
- Publication Type:Journal Article
- Keywords:
Tongue neoplasms;
Lymphatic metastasis;
Magnetic resonance imaging
- From:
Chinese Journal of Radiology
2016;50(5):353-356
- CountryChina
- Language:Chinese
-
Abstract:
Objective The aim of this study was to determine whether multi?parameters MRI of tongue carcinoma have the potential to predict cervical lymph node metastases. Methods A total of 46 patients with tongue carcinoma, who underwent MRI scan preoperatively, were investigated retrospectively and were divided into cervical lymph node (LN) metastases group (unilateral LN+, n=16;bilateral LN+, n=14) and no cervical lymph node metastases group (LN-, n=16) according to their pathological grading. Of the 40 patients with tongue carcinoma underwent plain and contrast MRI scan, 6 patients have plain MRI scan, and 32 have DWI examination.The ADC value, tumor length, tumor thickness, sublingual distance between tumor and sublingual space, and para?lingual distance between tumor and tongue midlinedetermined from MRI, were preoperatively estimated and compared with the pathological findings of cervical lymph nodes. A unpaired t test was used to analyze normal distributed continuous data, and a Mann?Whitney U test was used to analyze abnormally distributeddata. The ROC was used to evaluate the efficacy of MRI in predicting the metastasis of cervical lymph nodes. Results The indexes of ADC value, tumor length, tumor thickness, and para?lingual distance between tumor and tongue midline, which all showed significant difference between LN+group and LN-group (all P<0.05), and the index of sublingual distance between tumor and sublingual space showed no significantly association with LN+ (P>0.05). The index of ADC value showed significant difference between unilateral LN+group and bilateral LN+group (P<0.05), and the other indexes, which all showed no significantly association with bilateral LN+ (all P>0.05). The ROC curve analysis of the ADC value, tumor length, tumor thickness, and para?lingual distance between tumor and tongue midline of the neck lymph node metastasis were carried out, with the cutoff set as 1.13×10?3 mm2/s, 31.08 mm, 17.33 mm and-2.26 mm. The corresponding area under curve(AUC), sensitivity, and specificity were 0.878, 90.9%and 90.0%; 0.822, 83.3% and 81.3%; 0.834, 86.7% and 81.3%; 0.794, 86.7% and 75.0%, respectively. The ROC curve analysis of the ADC of the bilateral neck lymph node metastasis was also carried out, with the cutoff of ADC value set as 1.07×10?3 mm2/s, the corresponding AUC, sensitivity, and specificity were 0.806, 80.0%and 75.0%. Conclusion The ADC value, tumor length ,tumor thickness and para?lingual distance between tumor and tongue midline,determined from MR imaging, all can be used as independent factors in predicting cervical lymph node metastasis, where ADC value may be helpful to predict bilateral neck lymph node metastasis.
