Between May 1992 and October 1994, 114 patients visited our department for treatment of primary lung cancer. Among them, we prospectively evaluated mediastinal lymph nodes in 40 patients using CT and TEE; these patients later underwent thoracotomy. There were 30 men and 10 women aged 40 to 79 years (mean, 64 years). In preoperative staging, numbers of patients with stage I, stage II, and stage IIIA were 24 patients, 3 patients, and 13 patients, respectively. Eighteen of the remaining 74 patients underwent thoracotomy because of preoperative stage I to IIIA. However, these patients were not contained in this study, because TEE examinations were not performed. We could not obtain informed consent from these patients. The remaining 56 patients did not undergo thoracotomy because of stage IIIB or stage IV or advanced age. These patients were treated with irradiation and chemotherapy. Selection criteria for thoracotomy were as follows: no visible direct mediastinal invasion by the tumor on CT; no enlarged contralateral mediastinal nodes seen on CT; and no known distant metastases.
This study did not include stage IIIB and stage IV in preoperative staging. TEE was performed to evaluate mediastinal lymph node metastases. Informed consent was obtained from all patients before TEE. Both CT and TEE were performed during 3 weeks immediately before thoracotomy. CT scans were performed with a machine (TCT-900S; Toshiba; Tokyo). Contiguous 1-cm-thick sections were obtained at 1-cm intervals from the lung apices to at least the adrenal glands. CT scanning was initiated about 30 s after the start of IV power injection of 100 mL of contrast material (iopamidol 300; Bracco Industria Chemica; Milan, Italy) at the rate of 2 mL/s. Standard scanning parameters (120 kVp; 120 to 250 mA; 512×512 matrix; 1-s scanning time) were used, depending on the size of patients. CT scans were reviewed at levels appropriate for the mediastinum (level, 50 HU; window width, 300 HU).
A 7.5-MHz B-scan echoscope (Machida EPE-703FL; Tokyo, Japan) with a color-coded Doppler unit and an ultrasonograph SSA-260A (Toshiba) were used to measure the size of mediastinal lymph nodes. CT scans and TEE findings in each case were interpreted by three radiologists. When the analysis disagreed, a conclusion was reached with a consensus after discussion. Mediastinal lymph nodes were localized according to the lymph nodes mapping scheme of the Japanese Lung Cancer Society (JLCS) derived from the proposition by Naruke et al, which is in international use. That is to say, station 1 (JLCS) contains the greater part of station 2 (American Thoracic Society [ATS] lymph node mapping system) and a part of station 3 (ATS). Station 2 (JLCS) contains the greater part of station 4 (ATS). Station 10R (ATS) contains station 4R (JLCS). The other mediastinal stations of JLCS correspond to those of ATS.
Lymph nodes were measured on the long-axis diameter and the short-axis diameter of the maximal cut surface with TEE, and on the short-axis diameter with CT before thoracotomy. After fixation (10% neutral buffered formaldehyde solution), every lymph node was bisected on a plane parallel to the large axis, long- and short-axis diameters of the maximal cut surface were measured, respectively. All specimens were routinely processed with hematoxylin-eosin staining. All patients underwent thoracotomy with dissection of the mediastinal lymph nodes. Mediastinoscopy was not performed before thoracotomy, but performed at thoracotomy to evaluate lymph nodes in the contralateral side of the mediastinum. If lymph nodes were accessible, biopsy was performed. R2b resection was performed in our institution. During a right thoracotomy in right-sided lung cancer, all accessible lymph nodes in the superior (ie, superior mediastinal, paratracheal, anterior mediastinal, pretracheal and retrotracheal, and tracheobronchial nodes) and inferior (ie, subcarinal, paraesophageal, and pulmonary ligament nodes) mediastinum were resected together with the surrounding fat tissue. During a left thoracotomy in left-sided lung cancer, all accessible lymph nodes in the superior (ie, superior mediastinal, paratracheal, anterior mediastinal, pretracheal and retrotracheal, tracheobronchial, and subaortic and paraaortic nodes) and inferior (ie, subcarinal, paraesophageal, and pulmonary ligament nodes) mediastinum were resected. On both sides, hilar, interlobar, and lobar nodes were resected. Only resected nodes were considered for analysis of measurements. Of the biopsies of lymph nodes that were done, those not resected in total were excluded from analysis. We can assure correspondence between nodal stations on imaging and at surgery according to the lymph nodes mapping scheme of the JLCS. However, precise correspondence between lymph nodes on imaging and at surgery is a difficult problem if several lymph nodes were sampled in the same station. So, we did the following judgment: for example, 3 nodes were resected in station 4L. One contained malignancy, and the others were benign. Two nodes were detected by TEE, and one node was detected by CT in the same station. In this case, we judged that a lymph node on CT is malignant. Furthermore, we judged that a larger lymph node on TEE is malignant, and a smaller lymph node on TEE is benign. That is to say, larger nodes on imaging were judged malignant. Lymph nodes were considered abnormal if they were 10 mm or greater in short-axis diameter on CT and TEE. The diameters of malignant nodes were compared with those of benign nodes by means of the Student’s t test.