Tissue Is the Issue: Is Endoscopic Ultrasonography with or without Fine-Needle Aspiration Biopsy in the Staging of Non-Small-Cell Lung Cancer an Advance?

Non-small-cell lung cancer is the leading cause of death from malignancy in the United States [1]. The TNM (T = primary tumor, N = regional lymph nodes, M = distant metastasis) staging system for this type of cancer subdivides patients into anatomic groups. This system is central to determining treatment options and prognosis and to comparing survival among therapeutic trials [1]. In the absence of distant metastasis, mediastinal lymph node metastasis in non-small-cell lung cancer is a critical issue. Almost half of patients with this type of cancer present with metastasis to the mediastinal lymph nodes, a characteristic that is the major obstacle to cure. Ipsilateral mediastinal and subcarinal lymph nodes are designated as N2 nodes, and contralateral mediastinal lymph nodes are considered N3 nodes. Patients with T1 (<3 cm) or T2 (>3 cm) non-small-cell lung cancer and N2 metastasis (stage IIIA) are potential candidates for surgery. Surgical resection is generally not curative, however, and most patients with N2 disease receive multimodality treatment (for example, surgery, neoadjuvant chemotherapy, adjuvant radiation therapy, and chemotherapy) [1]. Optimal treatment for N3 disease (stage IIIB) is combination chemotherapy and radiation therapy [1]. The 5-year survival rate is approximately 30% in patients with stage N2 disease documented at surgical resection and approximately 5% in patients with N3 disease determined on radiography [2]. Thus, distinguishing between N2 and N3 disease in non-small-cell lung cancer has major implications for both treatment and prognosis.

To evaluate the mediastinal lymph nodes, computed tomography of the thorax with intravenous contrast scanning should be done in patients suspected of having lung cancer [1]. The sensitivity and specificity of computed tomography for detecting N2 or N3 metastasis in lung cancer are both 60% to 70% [3, 4]. In non-small-cell lung cancer, the risk for metastasis to the mediastinal lymph nodes increases with lymph node size. However, normal-sized lymph nodes can harbor cancer and, conversely, one study showed that 37% of mediastinal lymph nodes with a diameter of 2 to 4 cm were benign [4]. The sensitivity and specificity of magnetic resonance imaging for detecting N2 or N3 metastasis in lung cancer are similar to those of computed tomography [5, 6].

In this issue, Gress and colleagues [7] report on the ability of esophageal endoscopic ultrasonography, fine-needle aspiration biopsy guided by endoscopic ultrasonography, and computed tomography to detect metastasis to posterior mediastinal lymph nodes in 52 patients with non-small-cell lung cancer and enlarged mediastinal lymph nodes on computed tomography. The sensitivity and specificity of computed tomography were 64% (14 of 22 patients) and 35% (8 of 23 patients), respectively. In contrast, the respective sensitivity and specificity of endoscopic ultrasonography were 86% (19 of 22 patients) and 83% (19 of 23 patients); these values improved to 93% (14 of 15 patients) and 100% (9 of 9 patients) with the addition of guided fine-needle aspiration biopsy. These findings are similar to those reported by Giovannini and colleagues [8] and Silvestri and associates [9] (sensitivities of 81% and 89%, respectively, and specificities of 100% each) for ultrasonography-guided aspiration biopsy of posterior mediastinal lymph nodes in lung cancer. Endoscopic ultrasonography can locate masses and lymph nodes adjacent to the esophagus, and ultrasonography-guided aspiration biopsy can provide cytologic diagnosis of gastrointestinal lesions [8, 10]. The latter technique can sample posterior mediastinal lymph nodes but not lymph nodes that are lateral and anterior to the trachea [7-9].

Although these results are promising, the sensitivity and specificity of ultrasonography-guided aspiration biopsy reported by Gress and colleagues may be spuriously inflated. Thirty-one patients had thoracotomy with lung resection and dissection of ipsilateral and subcarinal lymph nodes (N2). At the same time, 11 of these patients did not have aspiration biopsy of the subcarinal or N3 lymph nodes because the nodes appeared normal by endoscopic ultrasonography, and 10 patients had negative results on ultrasonography-guided aspiration biopsy of enlarged subcarinal or N3 lymph nodes. Because the N3 lymph nodes were never surgically sampled, we are uncomfortable assuming that all of them were negative for cancer. The authors' data show false-positive rates of 10% for ultrasonography-guided aspiration biopsy and 16% for endoscopic ultrasonography alone.

Given the limitations of the anatomic criteria used with computed tomography and magnetic resonance imaging, the interest in alternative noninvasive methods for staging the mediastinum in lung cancer has been growing. Positron emission tomography, which exploits the increased uptake of glucose by malignant cells, has been used to stage the mediastinum in non-small-cell lung cancer. The sensitivity and specificity of this method were 78% to 100% and 52% to 81%, respectively [11, 12]. Dextran-coated iron oxide particles, which are phagocytized by the reticuloendothelial system, have been used as a contrast agent in magnetic resonance imaging to increase detection of lymph node metastasis. The initial experience in patients with head and neck cancer was encouraging (sensitivity, 95%; specificity, 84%) [13]. It seems unlikely, however, that these methods will be able to reliably detect micrometastasis in normal-sized mediastinal lymph nodes.

Definitive staging of mediastinal lymph nodes in lung cancer still requires obtaining a histologic or cytologic specimen. The mediastinal lymph nodes can be surgically staged by using mediastinoscopy, mediastinotomy, thoracoscopy, or thoracotomy. For lung cancer, mediastinoscopy has a sensitivity of 70% to 95% and a specificity of 100% [14, 15]. Hilar and mediastinal adenopathy can be sampled by computed tomography-guided transthoracic needle aspiration, although this technique is generally not used in lymph nodes less than 2 cm in diameter. Fiberoptic bronchoscopy with transbronchial needle aspiration biopsy can sample hilar, anterior, and posterior mediastinal lymph nodes. The sensitivity and specificity of transbronchial needle aspiration biopsy of mediastinal lymph nodes for lung cancer are 80% to 90% and 100%, respectively [16, 17]. Fiberoptic bronchoscopy with ultrasonography-guided transbronchial needle aspiration biopsy is also now available. However, endobronchial ultrasonography, which is not done in real time, did not increase the diagnostic yield of transbronchial needle aspiration biopsy with onsite cytologic evaluation. Integrating an ultrasonographic transducer in a flexible bronchoscope, which would make real-time endobronchial ultrasonography possible during transbronchial needle aspiration biopsy, may improve the sensitivity of this method [17]. Endobronchial ultrasonography has not yet been widely adopted. Transbronchial needle aspiration biopsy of an enlarged subcarinal lymph node using real-time computed tomographic fluoroscopy has been reported, but the effect of this technology is unknown [18].

Endoscopic ultrasonography-guided fine-needle aspiration biopsy can safely and effectively evaluate both benign and malignant mediastinal disease. As promising as it may seem, however, its role in clinical practice has not yet been fully defined [7-9]. Should this method replace computed tomography for evaluating posterior mediastinal lymph nodes in patients suspected of having lung cancer? Not quite yet, in our view; a clinical trial that directly compares the diagnostic yield, cost, and safety of computed tomography, endoscopic ultrasonography-guided fine-needle aspiration biopsy, bronchoscopic transbronchial needle aspiration biopsy, and mediastinoscopy or mediastinotomy is required at this point. We suspect that endoscopic ultrasonography-guided aspiration biopsy may prove to be technically easier than transbronchial needle aspiration biopsy for some lymph nodes on the left side but not for nodes on the right side. However, this idea also needs to be studied in a clinical trial. As an alternative, patients with lung cancer may be scheduled for combined ultrasonography-guided aspiration biopsy and bronchoscopic transbronchial needle aspiration biopsy for diagnosis or staging of mediastinal lymph nodes. The cost of these two procedures will need to be compared with that of mediastinoscopy or comparable surgical procedures.

In 1997, tissue remains the issue (however we obtain it) for patients with lung cancer and enlarged mediastinal lymph nodes. Endoscopic ultrasonography-guided fine-needle aspiration biopsy seems to improve the ability to stage posterior mediastinal lymph nodes in non-small-cell lung cancer, but it is only one of many rapidly emerging techniques that are transforming our ability to sample that tissue safely and effectively.

• Copyright ©2004 by the American College of Physicians