In the present study, the size of the lymph nodes was measured on TEE and nodal specimens in long- and short-axis diameter. There were no differences between malignant and benign nodes in both diameters. Furthermore, normal-sized N2 nodes were demonstrated in 73% of N2 nodes on TEE, and mediastinal lymph node metastases on CT were misdiagnosed in 64% of patients with N2 nodes because of normal-sized N2 nodes. These findings indicate that we are not able to reliably differentiate malignant mediastinal lymph nodes from benign nodes by size because of high frequency of normal-sized N2 nodes. Kerr et al reported that there were no significant differences between malignant nodes and benign nodes in diameter, and concluded that malignant mediastinal nodes were not larger than benign nodes and small lymph nodes were not infrequently malignant. This report supports our data in relationship of size between malignant and benign nodes.
The measurements of nodal diameters by TEE are different from the measurements taken from the CT image, which are usually the diameters of lymph nodes in the horizontal plane, because lymph nodes are rarely spherical. The image of the entire lymph node can be demonstrated by moving the probe at TEE, and we can measure nodal diameters of the maximal cut surface without partial volume effect and influence of window setting like CT. Therefore, we can measure nodal diameters more accurately by TEE than by CT. Furthermore, the detection rate of mediastinal lymph nodes by TEE (98/208, 47%) was superior to that by CT (51/208, 25%). However, lymph nodes were detected only in the left side of the mediastinum by TEE because of ultrasound reflection due to air in bronchi. This finding seems to be a limitation of the examination by TEE. Although not the purpose of our study, in our experience, the detection rate of lymph nodes in the left side of the mediastinum by TEE was superior to that by CT, and smaller lymph nodes were demonstrated by TEE than by CT. Therefore, it seems to be useful for evaluation of contralateral nodal stations in patients with right-sided lung cancer. That is to say, if enlarged lymph nodes are demonstrated by TEE, the presence of N3 nodes is suspected. If normal-sized lymph nodes are seen, we can recommend that surgeons perform careful nodal sampling. Furthermore, although we consider that TEE is useful for evaluation of mediastinal invasion, we have not obtained sufficient data. Several limitations of this study must be stressed.
Our data are derived from a small and selected group of patients. At present, patients with N3 involvement are considered to have nonoperable stage IIIB disease. Therefore, patients with stage IIIB or IV in preoperative staging did not come to thoracotomy, and then, all patients with bulky N2 nodes or extranodal disease belonged to stage IIIB or stage IV. This appeared to be the reason that large lymph nodes were few in our study. Furthermore, the mean diameters of the normal-nodal specimens in our study were smaller than those noted in other reports because of shrinkage after fixation. That is to say, the measurements of nodal specimens in our study were different from the measurements of fresh lymph nodes. However, our harvesting of lymph nodes is more extensive than in other studies.
In conclusion, it is not possible to reliably differentiate malignant mediastinal nodes from benign nodes by size alone, and we should be aware of high frequency of normal-sized N2 nodes in patients with operable stage of lung cancer.