| pubmed-article:2247348 | pubmed:abstractText | We investigated and compared the ability to diagnose metastasis of lung cancer to the mediastinum and hilar lymph nodes using CT and EUS (endoscopic ultrasonography by radial scanning method) in 27 patients undergoing resection of primary lung cancer and 6 autopsy cases. We also determined the relationship between the presence or absence of metastasis and the size of each lymph node based on the lymph node size measured at the time of resection and its histopathological findings, and we then set up a standard value that was the most accurate in evaluating the presence or absence of metastasis using a receiver operating characteristic (ROC) curve. When lymph node sizes appearing as images were compared with their actual sizes measured on resected specimens before formalin fixation, the short axis measured by either method was found to generally agree with the actual values, while the long axis was slightly smaller than the actual values, although EUS gave more accurate values. When the ability to diagnose metastasis was compared between CT and EUS using the standard value obtained from the ROC curve (a more than 8 mm short axis was defined as positive for metastasis), there were no differences in the ability to delineate the entire area of the mediastinum, including hilar lymph nodes. With respect to individual sites, although there was some difficulty delineating some regions in the mediastinum (pretracheal lymph node) with EUS, more lymph nodes in the mediastinum that were delineated by EUS histopathologically had metastatic lesions than those delineated by CT. However, both methods often failed to delineate hilar lymph nodes, with no difference shown between these two methods. | lld:pubmed |
