Accuracy of helical computed tomography for the identification of lymph node metastasis in resectable non-small cell lung cancer

Abstract

The criteria for the diagnosis of lymph node metastasis (LNM) in non-small cell lung cancer were investigated using helical computed tomography (hCT). The conventional criterion (1-cm short axis threshold) is generally accepted; however, this criterion is based on conventional CT. New criteria for LNM were investigated because the resolution of hCT is better than that of conventional CT. Ninety-seven NSCLC patients examined with hCT were enrolled. Both the long axis (LA) and short axis (SA) of the nodes were measured using hCT. Based on the receiver operating characteristic curves, the thresholds that gave optimal sensitivity and specificity for LNM were 13 mm for LA and 9 mm for SA. The LNM diagnosis was re-evaluated using the combination of cutoff values. When the LA was > or =13 mm and the SA was > or =9 mm, the sensitivity, specificity, and accuracy were 56.3%, 92.1%, and 88.1%, respectively. When the LA was > or =13 mm or SA was > or =9 mm, sensitivity, specificity, and accuracy were 75.0%, 74.7%, and 74.7%, respectively. These values were not so different from the conventional criterion recalculated from these data. The new criteria are considered to be useful for making a LNM diagnosis. The conventional criteria for the LNM diagnosis might therefore be applicable even for hCT.