OBJECTIVE: To assess the diagnostic certainty of CT images to identify regional spread of head and neck tumours. PATIENTS AND METHOD: A CT study was performed on 86 patients with neoplasms in the larynx and/or pharynx. After this, surgical dissection was performed for 142 neck sides. In the imaging study the following parameters were considered for all patients before suspecting malignant nodes: size of adenopathy bigger than 10-11 mm, irregular borders, central necrosis, spherical shape, capsular enhancement and presence of groups with three or more lymphadenopathies. RESULTS: 48.5 % of dissections were N+. Sensitivity and specificity for physical examination were 59 % and 82 %, respectively, against 73 % and 86 % for CT. Lymph node necrosis was the pattern with the greatest sensitivity and specificity (35.8 % and 100 %, respectively) and its accuracy was 69.7 %. Accuracy was 60.5 % for spherical shape, 59.8 % for node size, and between 54 % and 58 % for the other three criteria. In the histopathological findings, 25 % of neck dissections were N+ when only one pattern had been detected on CT, whereas those neck sides in which four patterns were identified simultaneously showed regional spread in 100 %. CONCLUSIONS: The presence of specific morphological imaging criteria for head and neck tumours in cervical lymph nodes and their concomitance increase the accuracy of imaging to predict regional spread.
OBJECTIVE: To assess the diagnostic certainty of CT images to identify regional spread of head and neck tumours. PATIENTS AND METHOD: A CT study was performed on 86 patients with neoplasms in the larynx and/or pharynx. After this, surgical dissection was performed for 142 neck sides. In the imaging study the following parameters were considered for all patients before suspecting malignant nodes: size of adenopathy bigger than 10-11 mm, irregular borders, central necrosis, spherical shape, capsular enhancement and presence of groups with three or more lymphadenopathies. RESULTS: 48.5 % of dissections were N+. Sensitivity and specificity for physical examination were 59 % and 82 %, respectively, against 73 % and 86 % for CT. Lymph node necrosis was the pattern with the greatest sensitivity and specificity (35.8 % and 100 %, respectively) and its accuracy was 69.7 %. Accuracy was 60.5 % for spherical shape, 59.8 % for node size, and between 54 % and 58 % for the other three criteria. In the histopathological findings, 25 % of neck dissections were N+ when only one pattern had been detected on CT, whereas those neck sides in which four patterns were identified simultaneously showed regional spread in 100 %. CONCLUSIONS: The presence of specific morphological imaging criteria for head and neck tumours in cervical lymph nodes and their concomitance increase the accuracy of imaging to predict regional spread.