Lin Ouyang1, Zhao-yin Shi, Zhi-gang Lin. 1. Department of Imaging Medicine, PLA 175 Hospital (Affiliated Southeast Hospital of Xiamen University), Zhangzhou, Fujian 363000, China.
Abstract
BACKGROUND: The specificity and precision of lymphadenopathy assessment using US, CT and MRI are generally unsatisfactory, while fluorodeoxyglucose-positron emission tomography/computed tomography ((18)F-FDG PET/CT) can support this process by providing additional information about the lymph node features. However, which image features of (18)F-FDG PET/CT play the key role in the diagnosis and cutoffs of malignant cervical lymphadenopathy still needs to be determined by further studies. Our study aimed to identify (18)F-FDG PET/CT abnormalities that would assist in making a reliable diagnosis of malignant cervical lymphadenopathy in enlarged cervical lymph nodes of patients with unknown primary diseases. METHODS: One hundred and ninety-one consecutive patients of cervical lymphadenopathy with unknown primary causes were examined by (18)F-FDG PET/CT from May 2007 to October 2011 and a definite diagnosis was established by pathologic biopsy. (18)F-FDG PET/CT images were evaluated to identify the relevant abnormalities. All image features were analyzed by optimal scale regression tests to determine the important factors that were predictive for the diagnosis of malignant cervical lymphadenopathy and the cutoffs. RESULTS: The factors studied in (18)F-FDG PET/CT images for predicting malignant cervical lymphadenopathy were sex, age, node location, size, shape, margins, maximum standard uptake value (SUV), mean SUV, FDG uptake pattern and number of nodes. It was found that mean SUV, maximum SUV, FDG uptake pattern, location, size and margins were the important risk factors of cervical lymph nodes that could predict malignant cervical lymphadenopathy. Signs of mean SUV ≥ 2.5 (or maximum SUV ≥ 3.5), nodular FDG uptake pattern, location of IIA, III, IV, VB, VI and VII regions, size ≥ 1.5 cm and vague margins had their optimal diagnostic accuracy (Ac) and Youden index (YI), further, combination of any three factors of these six important risk factors would led to the best diagnostic Ac of 96% and YI of 0.93. CONCLUSIONS: Signs of mean SUV, maximum SUV, FDG uptake pattern, location, size and margins of node in (18)F-FDG PET/CT imaging are important predictive factors of malignant cervical lymphadenopathy. A combination of multiple factors may yield a higher diagnostic efficacy.
BACKGROUND: The specificity and precision of lymphadenopathy assessment using US, CT and MRI are generally unsatisfactory, while fluorodeoxyglucose-positron emission tomography/computed tomography ((18)F-FDG PET/CT) can support this process by providing additional information about the lymph node features. However, which image features of (18)F-FDG PET/CT play the key role in the diagnosis and cutoffs of malignant cervical lymphadenopathy still needs to be determined by further studies. Our study aimed to identify (18)F-FDG PET/CT abnormalities that would assist in making a reliable diagnosis of malignant cervical lymphadenopathy in enlarged cervical lymph nodes of patients with unknown primary diseases. METHODS: One hundred and ninety-one consecutive patients of cervical lymphadenopathy with unknown primary causes were examined by (18)F-FDG PET/CT from May 2007 to October 2011 and a definite diagnosis was established by pathologic biopsy. (18)F-FDG PET/CT images were evaluated to identify the relevant abnormalities. All image features were analyzed by optimal scale regression tests to determine the important factors that were predictive for the diagnosis of malignant cervical lymphadenopathy and the cutoffs. RESULTS: The factors studied in (18)F-FDG PET/CT images for predicting malignant cervical lymphadenopathy were sex, age, node location, size, shape, margins, maximum standard uptake value (SUV), mean SUV, FDG uptake pattern and number of nodes. It was found that mean SUV, maximum SUV, FDG uptake pattern, location, size and margins were the important risk factors of cervical lymph nodes that could predict malignant cervical lymphadenopathy. Signs of mean SUV ≥ 2.5 (or maximum SUV ≥ 3.5), nodular FDG uptake pattern, location of IIA, III, IV, VB, VI and VII regions, size ≥ 1.5 cm and vague margins had their optimal diagnostic accuracy (Ac) and Youden index (YI), further, combination of any three factors of these six important risk factors would led to the best diagnostic Ac of 96% and YI of 0.93. CONCLUSIONS: Signs of mean SUV, maximum SUV, FDG uptake pattern, location, size and margins of node in (18)F-FDG PET/CT imaging are important predictive factors of malignant cervical lymphadenopathy. A combination of multiple factors may yield a higher diagnostic efficacy.