Lung cancer invasion into the aortic wall (T4/stage

Transcription

1 Transesophageal Echographic Determination of Aortic Invasion by Lung Cancer* Carsten Schröder, MD; Bernd Schönhofer, MD, FCCP; and Bernd Vogel, MD, FCCP Background: In planning lung cancer therapy, the possibility of mediastinal invasion merits attention. The results of CT and MRI in this respect are unsatisfactory, especially in determining aortic involvement. Study objectives: To determine the validity of transesophageal echography in proving the invasion of lung cancer into the aortic wall. Patients: Two hundred one patients with lung cancer abutting against the aorta were examined using transesophageal echography and CT. In 97 patients, the results of both imaging techniques were compared with the surgical/pathologic results. Results: In a vast majority, transesophageal echography leads to a definitive result while CT remains equivocal. Controlled by surgical/pathologic results in 97 patients, transesophageal echography yielded a diagnostic accuracy of 91.8%. Conclusions: In lung cancer abutting against the aorta, the diagnostic procedure should be complemented by transesophageal echography if the therapeutic management depends on whether the aortic wall is invaded by the tumor or not. (CHEST 2005; 127: ) Key words: aortic invasion; lung cancer; transesophageal echography Abbreviations: CI confidence interval; pt2 pathologic T2; pt4 pathologic T4 Lung cancer invasion into the aortic wall (T4/stage IIIb) often means local unresectability or at least gives reasons for a modification of the surgical approach or a decision for neoadjuvant combined modality therapy. To prevent ineffective and unnecessary thoracotomies, a valid imaging procedure is urgently needed. Furthermore, this procedure should direct patients to a possible curative intervention presuming the tumor is resectable. For editorial comment see page 420 T-staging of lung cancer is mostly based on CT. However, reliable CT criteria of mediastinal tumor invasion 1 6 and the potential advantages of MRI 7,8 have not been defined. *From the Department of Pulmonary Medicine, Klinikum Hannover Heidehaus, Hannover, Germany. Manuscript received March 15, 2004; revision accepted September 23, Reproduction of this article is prohibited without written permission from the American College of Chest Physicians ( permissions@chestnet.org). Correspondence to: Bernd Vogel, MD, FCCP, Department of Pulmonary Medicine, Klinikum Hannover Heidehaus, Am Leineufer 70, Hannover, Germany; bernd.vogel. heidehaus@klinikum-hannover.de Only a few reports 9,10 about transesophageal echography in lung cancer staging mention the question of aortic invasion on the basis of individual cases. Since 1993, we have tried to clarify the validity of transesophageal echography in detecting and ruling out aortic invasion by lung cancer, compared to the results of CT and MRI. Materials and Methods Between 1993 and 2003, 5,000 referred patients with lung cancer were treated in our specialized thoracic institution; of these, 201 patients (median age, 65.4 years; 42 women and 159 men) were examined with bronchogenic cancer of the left lung abutting against the aorta. The tumors were located in the upper lobe (n 93), the lower lobe (n 61), or the left main bronchus (n 47). Histologic distribution was as follows: squamous cell carcinoma (n 106), adenocarcinoma (n 71), small cell carcinoma (n 20), carcinoid (n 3), and carcinosarcoma (n 1). Following institutional review board approval and informed written consent, transesophageal echography was performed using a 5-MHz biplane probe (ultramark IX; ATL; Bothell, WA) in 47 patients, a 5-MHz monoplane probe (UST-936 5, Aloka SSD 630; Aloka Company; Tokyo, Japan) in 139 patients, and a multiplane probe (UST , Aloka SSD 5000; Aloka Company) in 15 patients. Forty-nine patients received pharyngeal anesthesia with lidocaine, and 152 patients were examined during general anesthesia using rigid bronchoscopy. No complications were observed during the procedure. 438 Clinical Investigations

2 All patients underwent biplane chest radiography and CT, and six patients underwent MRI as well. Most patients underwent CT as outpatients in 40 different radiology institutions 2 to 14 days before transesophageal echography. Twenty-one patients underwent transesophageal echography 2 days prior to CT. The interpretation of transesophageal echography was done by one author (B.V.) in all cases. The CT interpretation was adopted from the contributing radiologists according to the medical records. Tumors were graded preoperatively by their invasion status: T4 aortic invasion, T2 no invasion, TX indistinct invasion. Statistical analyses were performed on a personal computer (SPSS for Windows, Version 10.0; SPSS; Chicago, IL). All tests ( 2, Fisher exact test) were two tailed, and p 0.05 was considered significant; 95% confidence intervals were used. Preoperative Ultrasonographic Criteria Passing the aortic lumen, the ultrasound generates a welldefined borderline representing the aortic wall and the covering pleura. Sometimes these anatomic layers are represented by two lines, depending on the amount of fat between the aorta and the mediastinal pleura (Fig 1, top right, B). Invasion of the aortic wall is represented by disappearance of both lines (Fig 1, bottom right, D) and lack of synchronous movement of the consolidation during respiration. Invasion of the mediastinal pleura but not the aortic wall is recognizable by disappearance of the outer line and the pulsating movement of the vessel along the invaded pleura. Respiratory movement of the consolidated lung along the borderline proves that the mediastinal pleura is not involved. A tissue border can be detected by ultrasound unequivocally if the ultrasound hits the investigated surface vertically as in the greater part of the aortic arch and the descending aorta. Near the hilum behind the left main bronchus and at the subaortic nodes, the border between a tumor and the aortic wall runs parallel to the sound direction; therefore, the echogenic borderline may be incomplete, simulating an invasion despite an intact aortic wall. Verifying Transesophageal Echography Results by Pathology Eighty patients, particularly those with T4 tumor, had N3 or M1 disease as well and were therefore excluded from surgery. Eighteen patients had severe respiratory dysfunction, and 16 patients refused the operation; thus, only 96 patients underwent thoracotomy. All operations were performed within 10 days after ultrasound examination. The transesophageal echography estimations and the CT reports were compared with the intrathoracic findings and histologic results. One case was examined at autopsy. Results Preoperative Staging By transesophageal echography, 61 of the 201 patients showed typical signs of invasion of the aorta, 129 were declared not invasive, and in 11 cases the results were inconclusive (Table 1). In comparison, the CT scans of these patients were judged invasive in 14 cases, noninvasive in 12 cases, and inconclusive in 66 cases ( invasion possible or not excluded ). In Figure 1. CT scan (top left, A) showing tumor contact with the aorta approximately 5 cm in length and no detectable aortic wall structure. In contrast, transesophageal echography (top right, B) shows the aortic wall (solid arrow) and the mediastinal pleura (solid arrowhead) completely preserved (T2). Bottom left, C: CT scan showing left upper lobe consolidation with minor contact to the aortic arch and the descendent aorta. Bottom right, D: Transesophageal echography showing subtotal abolishment of the borderline between the tumor and the aortic lumen (open arrowheads). Surgery revealed a broad invasion of the aortic wall; no resection was performed. CHEST / 127 / 2/ FEBRUARY,

3 Variables Table 1 Comparison of Tumor Invasion by Transesophageal Echography and CT* Transesophageal Echography T T TX *Transesophageal echography showed in 61 of the 201 patients typical signs of invasion of the aorta; 129 were declared not invasive. In 11 cases, the results were inconclusive. CT scans of these 201 patients were judged invasive in 14 cases, and 12 cases were noninvasive. The majority of CT scans (175 cases) were inconclusive or not commented on by means of aortic invasion. 109 patients, the question of invasion was not commented on at all in the CT report, although unequivocal invasion existed by transesophageal echo interpretation in some of these patients. Conversely, of the 14 cases estimated to have aortic invasion by CT, 5 cases were clearly noninvasive by means of transesophageal echography. Comparison of Pathology Results With Preoperative Staging Twelve patients had histologically proven aortic invasion (pathologic T4 [pt4]) correctly staged by transesophageal echography in 10 cases (83.3%; confidence interval [CI], 51.6 to 97.9%). One case had been understaged as T2, and one was inconclusive (TX). In contrast, by CT only two cases were correctly staged as T4 (16.7%; CI, 2.1 to 48.4%). One case was understaged as T2, and the remaining nine cases were labeled inconclusive. Transesophageal echography has significantly higher sensitivity (83%) than CT interpretation (17%) in the evaluation of T4 tumors (p 0.05) [Table 2, top left, A]. The remaining 85 patients had histologically CT proven T2 disease (pathologic T2 [pt2]) disease. Transesophageal echography findings in these patients were correct in 79 cases (92.9%; CI, 85.6 to 97.4%), inconclusive in 4 cases, and overstaged as T4 in two cases. In contrast, by CT only four cases (4.7%) were correctly staged as T2. Six cases were overstaged as T4 (Fig 2), and the majority (75 cases) were inconclusive. Of these 75 inconclusive cases, the question of invasion was not certain in 33 of them; in 42 cases, the question of invasion was not discussed at all. Transesophageal echography has significantly higher sensitivity (93%) than CT (5%) in the evaluation of T2 tumors (p 0.01) [Table 2, top right, B]. In 89 of 97 patients, transesophageal echography estimation was correct, yielding an accuracy of 91.8%. Discussion An exact clinical staging of bronchial carcinoma is necessary for planning surgical intervention, especially in view of combined multimodality therapy. With regard to the invasion of mediastinal structures, especially the thoracic aorta, the predominantly used imaging procedure CT has not met expectations. 4 Neither the amount of contact ( 3 cm and/or 30% circumference) nor the presence of mass effect on adjacent structures are reliable signs of either invasion or unresectability. 5,6,8 MRI has the same limitations, with a slightly better but not significant accuracy. 7,8 The critical limitations of CT in respect to detecting tumor invasion are biological and technical. 8 First, the distinction of tissue borders by CT requires an appreciable difference in radiograph absorption of the contacting tissues and a certain quantity to be accepted as an altered absorption by the computer. Second, the pleura itself is too thin to be detected between the aortic wall and a consolidated lung if Table 2 Comparison of Pathologic Stage With Transesophageal Echography and CT Findings* T4 T2 TX Sensitivity, % 95% CI pt4 (n 12) Transesophageal echography p % CT % pt2 (n 85) Transesophageal echography p % CT % *Data are presented as No. unless otherwise indicated. Twelve patients had histologic aortic invasion (pt4), correctly staged by transesophageal echography in 10 cases (83.3%; CI, 51.6 to 97.9%). In contrast, by CT only two cases were correctly staged as T4 (16.7%; CI, 2.1 to 48.4%), and nine were labeled inconclusive. Sensitivity of detecting T4 tumors by transesophageal echography was significantly higher (83%) than by CT (17%; p 0.05). Eighty-five patients had histologic pt2 disease, correctly staged by transesophageal echography in 79 cases (92.9%: CI, 85.6 to 97.4%). In contrast, by CT only four cases were correctly staged as T2 (4.7%; CI, %); the majority (75 cases) were inconclusive. Transesophageal echography has significantly higher sensitivity (93%) than CT (5%) in the evaluation of T2 tumors (p 0.01). 440 Clinical Investigations

4 Figure 2. CT scan (top left, A) and MRI (top right, B) show suspicion of aortic invasion, while tranversal (bottom left, C) and longitudinal (bottom right, D) transesophageal echography shows preserved borderlines for aortic wall (solid arrows) and mediastinal pleura (solid arrowheads). Calcified intima plaques (open arrow) guide determination of the aortic vessel architecture. An extrapleural lobectomy was performed because of inflammatory adhesion. Histology disclosed the tumor restricted to the lung without invasion of the visceral pleura, and only inflammatory adhesions of the parietal pleura were found. there is not enough fat between them acting as a biological contrast medium. In addition, the acquisition of data is disarranged by the pulsating movement of the aorta. Even a faster acquisition by means of electron-beam CT does not show a better distinction between a mass and the vessel wall, demonstrated on the study of the left pulmonary artery. 11 The results of CT can be enhanced by artificial pneumothorax, 12,13 and by dynamic CT 14 or cine CT, 15 but even these techniques are not able to differentiate between tumor invasion and inflammatory adhesion. These technical limitations seem to be the reason for the high number of inconclusive CT assessments. Since this study was not designed to evaluate CT imaging, we only adopted the CT assessment from the written radiology reports, reflecting the usual practice. Reports have showed the ability of intravascular ultrasound to exclude tumor invasion of the aorta in the case of inflammatory adhesion 16 or confirm invasion, 17 an interesting additional sonographic technique to this field. Of course, there are some limitations: if the sound beam hits the border obliquely, the borderline can disappear and invasion may be suspected. But often this disadvantage is corrected by the real-time mode: pulsating movement of the aorta and respiratory movement of the lung, while disturbing the CT-imaging, are very helpful in the ultrasound study because visible motion between the two structures demonstrates that there is no invasion. If there is adhesion present, the ultrasound has the capability to distinguish between invasion and postinflammatory scaring. Preserved borderlines for the aortic wall and the mediastinal pleura rule out invasion. As the scarring becomes thicker, it actually becomes easier to show the intact layers. Only in freshly developed inflammation, where the edema fluid changes the echogenicity of the pleura, is an invasion difficult to rule out. This study, the results of which are not unexpected, 18,19 shows that the accuracy of ultrasound in detecting or excluding malignant invasion of the aorta can reach 90%, far more than ever reported from CT or MRI studies. 20 The advantage of ultrasound is its detection of borderlines on the basis of a different sound conduction, widely independent of the thickness of the structure. This quality of resolution is well documented at many sites of the body, eg, by imaging the cardiac valves. Because of this, it is possible to diagnose the aortic wall as intact or invaded by the cancer. 21,22 Based on these results, we believe that a transesophageal ultrasound study should be performed if chest radiography or CT demonstrate a tumor abut- CHEST / 127 / 2/ FEBRUARY,

5 ting against the aorta, and the therapeutic approach (direct surgery, neoadjuvant therapy, or exclusion of surgery) should be modified according to the transesophageal echography evaluation. More invasive procedures such as intravascular ultrasound may be helpful in the rare cases in which transesophageal echography is limited. References 1 White PG, Adams H, Crane MD, et al. Preoperative staging of carcinoma of the bronchus: can computed tomography scanning reliably identify stage III tumors? Thorax 1994; 49: McLoud T, Swenson SJ. Imaging techniques for diagnosis and staging of lung cancer. Clin Chest Med 2002; 23: Colice GL. Chest CT for known or suspected lung cancer. Chest 1994; 106: Verschaekelen JA, Bogaert J, De Wever W. Computed tomography in staging for lung cancer. Eur Respir J 2002; 19(Suppl):40s 48s 5 Glazer HS, Kaiser LR, Anderson DJ, et al. Indeterminate mediastinal invasion in bronchogenic carcinoma: CT evaluation. Radiology 1989; 173: Herman SJ, Winton TL, Weisbrod GL, et al. Mediastinal invasion by bronchogenic carcinoma: CT signs. Radiology 1994; 190: Kameda K, Adachi S, Kono M. Detection of T-factor in lung cancer using magnetic resonance imaging and computed tomography. J Thorac Imaging 1988; 3: Webb WR, Gatsonis C, Zerhouni EA, et al. CT and MR imaging in staging non-small cell carcinoma: report of the Radiologic Diagnostic Oncology Group. Radiology 1991; 178: Sakio H, Yamaguchi Y. Transesophageal endoscopic echocardiography in lung cancer involving mediastinal organs in the assessment of resectability. Nippon Kyobu Geka Gakkai Zasshi 1989; 37: Pothoff G, Curtius JM, Wassermann K, et al. Transösophageale Echographie im Staging von Bronchialkarzinomen. Pneumologie 1992; 46: Takahashi M, Shimoyama K, Murata K, et al. Hilar and mediastinal invasion of bronchogenic carcinoma: evaluation by thin-section electron-beam computed tomography. J Thorac Imaging 1997; 12: Watanabe A, Shimokata K, Saka H, et al. Chest CT combined with artificial pneumothorax: value in determining origin and extent of tumor. AJR Am J Roentgenol 1991; 156: Yokoi K, Mori K, Miyazawa N, et al. Tumor invasion of the chest wall and mediastinum in lung cancer: evaluation with pneumothorax CT. Radiology 1991; 181: Murata K, Takahashi M, Mori M, et al. Chest wall and mediastinal invasion by lung cancer: evaluation with multisection expiratory dynamic CT. Radiology 1994; 191: Ohtsuka T, Minami M, Nakaji J, et al. Cine computed tomography for evaluation of tumors invasive to the aorta: seven clinical experiences. J Thorac Cardiovasc Surg 1996; 112: Tabata T, Tanita T, Ono S, et al. Preoperative endovascular echocardiography in lung cancer patient for detecting aortic wall invasion: a case report. Kyobu Geka 1998; 51: Yasui K, Kanazawa S, Mimura H, et al. Assessment of aortic invasion by pulmonary carcinoma with the use of intra-aortic endovascular sonography: a case report. J Thorac Cardiovasc Surg 2001; 122: Varadarajulu S, Schmulewitz N, Wildi SF, et al. Accuracy of EUS in staging of T4 lung cancer. Gastrointest Endosc 2004; 59: Wallace MB, Ravenel J, Block MI, et al. Endoscopic ultrasound in lung cancer patients with a normal mediastinum on computed tomography. Ann Thorac Surg 2004; 77: Rankovic BS. Transesophageal ultrasound for identification of a lung cancer infiltration into great vessels. Chest 2001; 120(Suppl):318S 21 Engberding R, Daniel WG, Erbel R, et al. Diagnosis of heart tumors by transoesophageal echocardiography: a multicentre study in 154 patients. Eur Heart J 1993; 14: Wang KY, Lin CY, Kuo-Tai J, et al. Use of transesophageal echocardiography for evaluation of resectability of lung cancer. Acta Anaesthesiol Sin 1994; 32: Clinical Investigations