OBJECTIVES: The aim of this study is to determine a set of MRI lymph node characteristics predictive for extranodal tumor spread (ENS) in head and neck cancer patients. METHODS: In 39 patients, 60 lymph nodes with on MRI a minimal axial diameter of more than 1cm or an inhomogeneous enhancement were studied. Two radiologists evaluated all MR-images for findings potentially indicative of the presence of ENS. Sensitivity, specificity and odds ratios based on logistic regression were calculated. RESULTS: On MR-imaging, 20 lymph nodes were staged positive for ENS. On histopathology, 30 nodes were positive for ENS. In total, 14 nodes (23%) were scored differently on MR-imaging and histopathology. The MR-finding "infiltration of adjacent planes" established a specificity of 100% (lower 90% confidence bound: 91%) and sensitivity of 50% (95% confidence interval [CI]: 28-72%). CONCLUSION: The MRI finding "infiltration of adjacent planes" may be high enough (100% in our study) to be used for treatment planning.
OBJECTIVES: The aim of this study is to determine a set of MRI lymph node characteristics predictive for extranodal tumor spread (ENS) in head and neck cancerpatients. METHODS: In 39 patients, 60 lymph nodes with on MRI a minimal axial diameter of more than 1cm or an inhomogeneous enhancement were studied. Two radiologists evaluated all MR-images for findings potentially indicative of the presence of ENS. Sensitivity, specificity and odds ratios based on logistic regression were calculated. RESULTS: On MR-imaging, 20 lymph nodes were staged positive for ENS. On histopathology, 30 nodes were positive for ENS. In total, 14 nodes (23%) were scored differently on MR-imaging and histopathology. The MR-finding "infiltration of adjacent planes" established a specificity of 100% (lower 90% confidence bound: 91%) and sensitivity of 50% (95% confidence interval [CI]: 28-72%). CONCLUSION: The MRI finding "infiltration of adjacent planes" may be high enough (100% in our study) to be used for treatment planning.