Endoscopic ultrasonography (EUS) was initially. Diagnosing Sarcoidosis Using Endosonography-Guided Fine-Needle Aspiration*

Transcription

1 Diagnosing Sarcoidosis Using Endosonography-Guided Fine-Needle Aspiration* Annette Fritscher-Ravens, MD; Parupudi V. J. Sriram, MD; Theodoros Topalidis, MD; Hans P. Hauber, MD; Andreas Meyer, MD; Nib Soehendra, MD; and Almuth Pforte, MD Study objectives: The ability to diagnose sarcoidosis cytologically has been reported previously, but the method is rarely used. Endoscopic ultrasonography (EUS) is a sensitive technique for detecting mediastinal lymph nodes, which in addition provides an opportunity to carry out guided fine-needle aspiration (FNA) cytology. We report herein on the use of EUS-FNA in the diagnosis of sarcoidosis. Patients and methods: Nineteen patients with suspected sarcoidosis were investigated using EUS-FNA with a linear echoendoscope and a 22-gauge Hancke-Vilman needle. Measurements and results: In all 19 patients, EUS revealed enlarged mediastinal lymph nodes (mean size, 2.4 cm), located subcarinally (n 15), in the aortopulmonary window (n 12), or in the lower posterior mediastinum (n 5). The nodes had an isoechoic or hypoechoic appearance, with atypical vessels in five cases. The amount of aspirate obtained using EUS-FNA was adequate in all patients, and contained blood in excess of normal in some, indicating a high degree of vascularity. Cytology demonstrated epithelioid cell granuloma formation, suggesting sarcoidosis. Mycobacterial cultures were negative in all of the patients except one, in whom the final diagnosis was tuberculosis. The specificity and sensitivity of EUS-FNA in the diagnosis of sarcoidosis were 94% and 100%, respectively. Conclusions: EUS of mediastinal lymph nodes in sarcoidosis reveals certain characteristic features. However, it is not capable of differentiating the lesions from tuberculosis or malignancy. EUS-FNA is a safe and sensitive method of aspirating material for cytology and mycobacterial cultures. We believe it will provide a useful alternative in the diagnosis of sarcoidosis. (CHEST 2000; 118: ) Key words: biopsy; cytology; endosonography; mediastinum; sarcoidosis Abbreviations: ACE angiotensin-converting enzyme; ATS American Thoracic Society; EUS endoscopic ultrasonography; FNA fine-needle aspiration; TBLB transbronchial lung biopsy Endoscopic ultrasonography (EUS) was initially used as a method of staging GI malignancies, 1 and more recently in the diagnosis and staging of lung cancer as well. 2 4 EUS has definite advantages compared to CT imaging of the mediastinum, as it allows assessment of the echo structure in lesions 1 cm in diameter, 5 and provides an opportunity to obtain material from mediastinal lymph nodes using *From the Departments of Interdisciplinary Endoscopy (Drs. Fritscher-Ravens, Sriram, and Soehendra), and Internal Medicine, Pulmonology (Drs. Hauber, Meyer, and Pforte), University Hospital Eppendorf, Hamburg; and the Cytopathological Institute Atay (Dr. Topalidis), Hannover, Germany. Manuscript received November 11, 1999; revision accepted May 2, Correspondence to: Annette Fritscher-Ravens, MD, Department of Interdisciplinary Endoscopy, University Hospital Eppendorf, Martinistrasse 52, Hamburg, Germany; FRI- RAV@t-online.de EUS-guided fine-needle aspiration (FNA), for tissue confirmation. Structures within a radius of 5 cm from the esophagus are accessible. The entire posterior mediastinum, the aortopulmonary window, and masses along the great vessels can be visualized, but the anterior mediastinum (and the pretracheal region in particular) cannot be imaged due to the air-filled trachea. With a high-resolution ultrasound transducer, it is possible to detect even very small lesions (3 to 4 mm). Studies of EUS-FNA have reported a specificity and positive predictive value of 100% in the diagnosis of mediastinal malignancies. 6 8 We report herein on the use of EUS-guided FNA to diagnose sarcoidosis, a granulomatous disease of unknown etiology with pulmonary and mediastinal lymph node involvement in 80% of patients Clinical Investigations

2 Table 1 Clinical, Laboratory, and Imaging Characteristics in Patients With a Prior History of Cancer and a Suspicion of Sarcoidosis* Patient No./ Age/Sex Clinical Presentation/History ACE, U/L TCT Bronchoscopy EUS in Mediastinum Follow-up, mo Cytology Final Disease 1/48/F Routine CXR, right hilar prominence 1986 breast cancer 1990 renal transplant (end-stage renal disease) 1996 endometrial cancer 2/58/M Routine CXR, paratracheal opacity 1995 malignant melanoma 3/47/M On TCT for staging colon cancer LN in mediastinum, asymptomatic 4/68/M 1998 malignant melanoma Routine CXR asymptomatic 5/64/M 1997 non-small cell lung cancer, right upper lobe curative resection; restaging 38 Polycyclic mass in right hilum 48 Pretracheal, precarinal, infracarinal, aortopulmonary window Multiple LNs up to 3 cm 59 2-cm mass in left hilum Multiple LNs up to 1.6 cm in the middle and lower mediastinum 71 Mass in right lower lobe 5, 1.6-cm LN; right lower lobe 7, 2.3 cm and multiple 1 cm Normal mucosa; histology, NECG; BAL, normal cell differentiation; CD4/CD8, 3.4; TB culture negative Normal mucosa; histology, normal; BAL, normal cell differentiation; CD4/CD8, 3.4; TB culture negative 5, 7: three 2-cm LNs, isoechoic, atypical vessels 5, 7: multiple LNs up to 3 cm, isoechoic, atypical vessels Not done 5, 7: cm mass; 8: multiple LNs up to 1.2 cm, hypoechoic Increased vascularity; BAL, lymphocytosis, atypical cells; CD4/ CD8, 2.5; TB culture negative 27 Mass in left hilum Normal mucosa; cytology, normal cell differentiation; CD4/CD8, 1.2; TB culture negative 5: cm mass isoechoic, atypical vessels; 7: multiple LNs up to 1 cm 5, 7: 2.8-cm isoechoic, atypical vessels 19 NECG Sarcoidosis lung 16 NECG Sarcoidosis 24 NECG NECG Sarcoidosis surgery 18 NECG Sarcoidosis 10 NECG NECG Sarcoidosis mediastinoscopy *CXR chest radiography; TCT thoracic CT; LN lymph node; TB tuberculosis; M male; F female; 5, 7, 8 numbers based on the ATS mediastinal lymph node mapping. 13 Normal range, 8 to 52 U/L. Based on a combination of noncaseating epithelioid cell granuloma (NECG), tuberculosis (TB) cultures, and stable clinical follow-up. Histology was obtained by microbiopsy specimen taken using EUS-FNA. CHEST / 118 / 4/ OCTOBER,

3 Table 2 Clinical, Laboratory, and Imaging Features of Patients Without a Prior History of Cancer and a Suspicion of Sarcoidosis* Patient No./Age/Sex Clinical Presentation/History ACE, U/L CT Bronchoscopy EUS in Mediastinum Follow-up, mo Cytology Final Disease 1/34/F Dyspnea; routine CXR: mediastinal mass 2/54/M Asymptomatic; routine CXR: hilar mass 40 Mediastinal mass suspect for NHL; multiple foci in both lungs 77 Bihilar mass, 2, 4, 5: multiple LNs up to 1.6 cm 3/58/M Hypercalcemia cm nodules in upper mediastinum; 2.5- cm mass near left atrium 4/59/M Asymptomatic 48 Mass in right middle lobe 5, 7: multiple mediastinal LNs up to 1.2 cm 5/49/F Asymptomatic; routine CXR: hilar mass 6/67/M Hypercalcemia 80 4L, 5: multiple 2 cm LNs and left pleural LN 7/70/M Dyspnea; CXR: hilar mass 8/72/F Suspicion of malignant lymphoma on TCT: multiple LNs mediastinum Granulomatous mucosal inflammation; histology, inflammation; BAL: lymphocytosis; CD4/CD8: 9.7; TB culture negative Mucosal inflammation; histology, chronic inflammation; BAL, lymphocytosis; CD4/CD8, 5.4; TB culture negative Normal mucosa; histology, normal; BAL, normal cell differentiation; CD4/CD8, 2.6; TB culture negative Chronic bronchitis; BAL, granulocytosis; CD4/CD8, 1.7; TB culture negative 7: mediastinal mass 3 cm Normal mucosa; BAL, histiocytosis; CD4/CD8: 8.6; TB culture negative 42 Mass in left hilum cm 5: multiple LNs up to 1.3 cm 42 Para-aortal and paracaval multiple LNs up to 2 cm Acute bronchitis; BAL, lymphocytosis; CD4/ CD8, 3.0; TB, culture negative Acute inflammation; BAL, granulocytosis; CD4/CD8, 2.6; TB culture positive Discrete chronic bronchitis; BAL, lymphocytosis; CD4/CD8, 1.46; TB culture negative 9/22/F Retinal vasculitis 43 Bihilar mass Normal mucosa; BAL, lymphocytosis; CD4/ CD8, 0.7; TB culture negative 10/54/M Dyspnea, weight 59 Small 1-cm LN in Normal mucosa; BAL, extreme lymphocytosis; loss aortopulmonary window CD4/CD8, 3.7; TB culture negative 11/38/M Poor appetite, 49 Multiple LNs up to 2 Hypervascularity; BAL, lymphocytosis; CD4/ weakness CD8, 7.5; TB culture negative 12/40/F Weight loss, weakness cm subcarinally and in aortopulmonary window 45 Multiple LNs up to 2 cm in left hilum Normal mucosa; BAL, lymphocytosis; CD4/ CD8, 5.2; TB culture negative 13/26/F Parotiditis 35 Bihilar LNs Acute bronchitis; BAL, material not representative; TB culture negative 14/31/M Erythema nodosum 47 Granulomatous disease in both lungs; bihilar LNs Mucosal inflammation; BAL, lymphocytosis; CD4/CD8, 5.0; TB culture negative 7: multiple LNs up to 3 cm, isoechoic 19 NECG Sarcoidosis 7, 8: multiple cm LNs, hypoechoic 5: 2.4-cm mass, isoechoic, partly hypoechoic 5: mass 2.1 cm; 7, 8: multiple LNs up to 1.8 cm, hypoechoic 4R, 5: multiple LNs 2 cm; 7: 3.2 cm, isoechoic, atypical vessels 21 NECG Sarcoidosis 16 NECG Sarcoidosis 15 NECG Sarcoidosis lung 14 NECG Sarcoidosis mediastinoscopy 5: 2 cm, hypoechoic 15 NECG Sarcoidosis 5: 2.2-cm LN, hypoechoic with calcifications; 7: 1.3-cm isoechoic TB culture positive 13 NECG TB 8: up to 1.4 cm, isoechoic 12 NECG Sarcoidosis 7: 1.5-cm LN, hypoechoic 12 NECG Sarcoidosis 5: 1.5-cm LN, hypoechoic 7 NECG Sarcoidosis 5: 4-cm mass, hypoechoic, with vessels coursing through; 7: multiple LNs up to2cm 7: multiple LNs up to 1.8 cm; 5: 2.6- cm LN, isoechoic with atypical vessels 6 NECG Sarcoidosis 4 NECG Sarcoidosis lung 7: up to 2-cm LNs, hypoechoic 3 NECG Sarcoidosis parotis 5, 7: up to 3.2-cm LNs, hypoechoic 3 NECG Sarcoidosis muscle biopsy *NHL non-hodgkin s lymphoma; see Table 1 for abbreviations. Normal range, 8 to 52 U/L. Based on a combination of NECG at cytology/histology, negative TB cultures, and a stable clinical follow-up. 930 Clinical Investigations

4 Table 3 Differential Diagnosis of Sarcoidosis and Tuberculosis Using Lymph Node FNA Cytology* Criteria Since spontaneous remission is observed in 70% of cases, the need for tissue diagnosis is under critical discussion, especially in asymptomatic stage I patients. 10,11 However, there is an absolute need for tissue confirmation in symptomatic patients in whom treatment decisions are necessary, or if alternative diagnoses such as malignant lymphoma, mediastinal involvement of solid tumors, or tuberculosis need to be excluded. This generally requires transbronchial biopsy or mediastinoscopy. We evaluated the feasibility of using EUS-guided FNA for cytodiagnosis of sarcoidosis. Materials and Methods Sarcoidosis Tuberculosis Epithelioid cells, epithelioid cell granuloma ( fish-in-the-net appearance) Histiocytic-type giant cells Langhans giant cells Eosinophils Macrophages Endothelial cells Granulocytes Cell necrosis, debris (caseating necrosis), protein precipitates ( dirty background ) * sometimes positive; often positive; strongly positive; usually negative. Nineteen patients (12 men, 7 women; mean age, 55 years; range, 22 to 72 years) with mediastinal lymph node enlargement were included in the study. There was a prior history of malignancy in five patients (group 1 in Table 1), in whom recurrent malignancy had to be excluded. In the absence of features suggesting local or regional recurrence, a differential diagnosis of sarcoidosis was entertained in these patients in view of raised angiotensin-converting enzyme (ACE) levels (patients 3 and 4 in Table 1), lymphocytosis, an elevated CD4/CD8 ratio in BAL (patients 2 and 4), or bronchoscopic biopsy specimen showing noncaseating granulomas (patient 1). Patients with undiagnosed mediastinal lymph nodes and a compatible clinical history of sarcoidosis, as defined elsewhere, comprised group 2 (Table 2). 12 Of these, three patients had respiratory symptoms, three others had only constitutional symptoms, and two had hypercalcemia. Three patients had extrapulmonary complaints in the form of erythema nodosum, parotiditis, or retinal vasculitis. Eight patients had no symptoms, and had been found to have mediastinal lymph nodes on chest radiographs carried out in the course of a routine clinical examination. Chest radiographs of sarcoidosis are categorized into three groups: the radiographic appearance of stage I disease is bilateral lymphadenopathy, with or without enlarged right paratracheal nodes but with normal lung fields; in stage II, there is additional evidence of pulmonary parenchymal involvement; and patients in stage III have diffuse parenchymal abnormality without hilar adenopathy. Radiography frequently suggests fibrosis with small lung volumes, elevated diaphragms, and honeycombing. 12 The changes observed in the study group were consistent with stage I disease in 12 patients, stage II in 5 patients, and stage III in 2 patients. CT scanning demonstrated enlarged mediastinal lymph nodes in all patients, and the locations were described according to the American Thoracic Society (ATS) mapping of mediastinal lymphadenopathy. 13 Bronchoscopy with BAL was carried out in 18 patients. In seven patients, BAL showed lymphocytosis combined with a CD4/CD8 ratio 3.5. Three others exhibited lymphocytosis with a normal CD4/CD8 ratio. Associated hypercalcemia was observed in two patients, and raised ACE levels were seen in five patients. Mycobacterial culture of BAL fluid was negative for tuberculosis in all patients. To diagnose the etiology of the mediastinal lymph nodes, all of these patients would have had to undergo a blind transbronchial biopsy, CT-guided FNA, or mediastinoscopy. We have reported earlier on the feasibility and safety of using EUS-guided FNA for cytodiagnosis of mediastinal lesions, 4,14 and we used this technique to establish the diagnosis in the present study. EUS evaluation of the mediastinum was performed using an echoendoscope with an electronic multielement curved lineararray ultrasound transducer (Pentax 34 UX; Pentax Ltd; Hamburg, Germany; or Olympus GF UC 30 P; Olympus Optical; Hamburg, Germany). A 22-gauge Hancke-Vilmann needle (GIP Medizintechnik; Grassau, Germany) was advanced through the wall of the esophagus into the target lesions, and material was aspirated using a 10-mL syringe. Two or three punctures were needed to obtain material for cytology and mycobacterial cultures. The aspirated material was stained using the May-Grünwald-Giemsa technique, and interpreted by an experienced and independent cytopathologist. The cytologic criteria used in the diagnosis of sarcoidosis are given in Table Results Enlarged lymph nodes were easily demonstrated using EUS and were documented according to the ATS mapping of mediastinal lymphadenopathy. 13 Lymphadenopathy was demonstrated in at least one location in all patients, and in two or more regions in 12 patients. On EUS, lymph nodes with an echo texture similar to that of the adjacent tissue were considered to be isoechoic (n 10), while those with comparatively poor echo morphology were labeled as hypoechoic (n 11). Two patients had lymph nodes with either type of morphology. The size of the lymph nodes varied from 1.0 to 4.1 cm (mean, 2.4 cm). In seven patients, some small atypical vascular structures were seen coursing through the lymph nodes, as demonstrated on color Doppler ultrasound examinations (Fig 1). As in our experience, such findings were not observed in other lymph nodes, except one metastasis of renal cell cancer; we called these vessels atypical. When the EUS and CT findings were compared, the location of the lymph node groups differed considerably (Tables 1, 2). In nine patients, EUS showed additional lymph node groups in the right paratracheal region (level 4R), in the aortopulmonary window (level 5), subcarinally (level 7), and in the para-aortal region (level 8), which were not seen on CT. However, in five CHEST / 118 / 4/ OCTOBER,

5 Figure 1. There is a large lymph node near the pulmonary artery (with blue and red color flow) containing a small vessel (blue color). Directly beneath, there is another without color flow in this position. patients, nodes located on CT pretracheally (level 2), in the left paratracheal area (level 4L), and paracavally could not be demonstrated on EUS, either due to interference from tracheal air (regions 2 and 4L) or because the site is too distant for imaging by EUS via the esophagus. EUS-Guided FNA Two or three EUS-guided FNAs were carried out in a target lesion (Fig 2). The number of punctures was based on the macroscopic appearance of cellularity in the smears. The main purpose was to obtain adequate material. The first author was trained in Figure 2. An echo-poor lymph node 2.4 cm in diameter, close to the root of the aorta (A), with the aortic valve seen at the bottom of the Figure (bend arrow). The needle is within the lymph node. The arrow is pointing to the needle tip. 932 this type of assessment, as a cytopathologist was not present. Since the cytologic interpretation depends not on the number of granulomas obtained, but on specific criteria observed in the granulomas (Table 3), it was not the number of smears that was the focus of interest, but the quality of the cell material obtained If more than one node was detected, the best assessable hypoechoic nodes without calcifications and those with vessels inside were selected for puncture.3 Since there are no predictive features regarding the malignancy of lymph nodes, the decision on which one to carry out a puncture depends largely on the operator s judgment. This procedure is similar to the approach used in lymph node sampling for histology in mediastinoscopy or thoracoscopy, where (as in EUS-FNA) small micrometastases may be missed, as lymphadenectomy is not intended.2,3 It is also based on the decision of the individual surgeon and the level of suspicion in the node. Cytology demonstrated noncaseating epithelioid cell granulomas in all of the patients (Fig 3, 4). Apart from sarcoidosis, the differential diagnosis in noncaseating granulomas includes tuberculosis, histoplasmosis, and berylliosis, etc. While tuberculosis is a possibility, histoplasmosis is not endemic in Germany, and there was no occupational exposure to beryllium in any of the patients. The cytologic distinction between sarcoidosis and tuberculosis is well delineated using fine-needle aspirates (Table 3) The mycobacterial cultures in the EUS-FNA were negative in all cases except one. In the patient concerned, the smears showed an excess amount of Langhans giant cells in a dirty background along with epithelioid cell granulomas, suggestive of tuberculosis (Fig 4). Since the Ziehl-Neelsen stain did not demonstrate acid-fast bacilli, the patient was initially thought to have sarcoidosis clinically, but subsequent Figure 3. Cytology, showing an epithelioid cell granuloma in sarcoidosis (May-Gru nwald-giemsa, original 900). Clinical Investigations

6 mycobacterial cultures were positive. Since the final diagnosis was tuberculosis in this patient, the case was regarded as a false-positive result. Four-drug antituberculosis therapy was begun, and the patient is receiving follow-up. During the clinical follow-up period (median of 12 months for patients with no history of cancer, and 18 months for five patients with prior cancer; range, 3 to 24 months), there was no clinical evidence of malignancy in any of the patients. Additional histologic confirmation was available from other sites in six patients, which supported the diagnosis. In the seventh patient, who had a suspected lung cancer relapse, specimens from an additional biopsy via mediastinoscopy and thoracoscopy confirmed sarcoidosis. Corticosteroid therapy was begun in five patients with pulmonary symptoms. Thus, the final diagnosis was considered as sarcoidosis in 18 patients, based on cytologic evidence of noncaseating granulomatous inflammation along with clinical follow-up and negative mycobacterial cultures. The overall diagnostic accuracy and sensitivity of EUS-FNA in the diagnosis of sarcoidosis were 94% and 100%, respectively. An unusual feature observed in the EUS-FNA of the lymph nodes in these patients, compared to our general experience with this procedure in other mediastinal lesions, was an excess amount of blood in the aspirates, particularly in those in whom atypical vessels were detected. Discussion Figure 4. Cytology in patient 7 (Table 2), showing a Langhans giant cell on a dirty background in tuberculosis (May-Grünwald-Giemsa, original 900). The differential diagnosis of enlarged hilar lymph nodes includes several infectious, inflammatory, immunologic, and neoplastic disorders. The most common cause of bihilar lymphadenopathy in the Western world is sarcoidosis. 19 Since sarcoidosis is known to resolve spontaneously in 70% of patients, the need for tissue diagnosis has been the subject of some debate. The arguments of the opponents of tissue diagnosis are based on risk/benefit and cost/ benefit considerations, while those of its proponents are based on the individual patient s perspective. 10 Although many of the patients may be asymptomatic (stage I sarcoidosis), alternative diagnoses such as tuberculosis and other granulomatous disorders, Hodgkin s lymphoma, etc, cannot be ruled out with certainty. There is therefore an absolute requirement for tissue confirmation in symptomatic patients. Treatment decisions have to be made in these patients, as well as in those with a history of cancer, when alternative diagnoses such as malignant lymphoma, mediastinal involvement of solid tumors, or any kind of metastasis need to be excluded. To avoid missing an opportunity when there is an alternative diagnosis of a disease that might be at a treatable stage, some authors recommend carrying out investigations to establish the diagnosis. 20 Although the value of cytology in the diagnosis of sarcoidosis was established long ago, only a few centers prefer to use it routinely. Pisircriler et al 21 diagnosed intrathoracic sarcoidosis using epithelioid cell granulomas in 441 patients, based on cytologic examinations of transbronchial biopsy specimens from mediastinal lymph nodes. The results were confirmed histologically using transbronchial lung biopsy (TBLB) or thoracoscopy. The sensitivity of cytology was 92%. Cytologic evaluation of material obtained from transbronchial biopsies led to a diagnosis of intrathoracic sarcoidosis in 143 of 158 patients. The diagnosis was confirmed in 102 cases by histology via mediastinoscopy, thoracoscopy, or open biopsy. 22 Tambouret et al 23 analyzed the use of FNA in the clinical examination of 32 patients with noncaseating granulomas, detected in all and proven by histology in half of the patients. Lohela et al 24 diagnosed granulomatous inflammation using ultrasound-guided FNA cytology of lymph nodes. The results were consistent with sarcoidosis in 88% of the patients, and the diagnosis was confirmed on histology and during the clinical follow-up. At present, bronchoscopy with BAL, TBLB, and/or mediastinoscopy are the standard diagnostic procedures for sarcoidosis. BAL is helpful in the diagnosis, as it can determine the lymphocyte subpopulation or when lymphocytosis with a CD4/CD8 ratio 3.5 is found. Although specificity rates of up to 90% have been reported, the sensitivity of BAL for diagnosing sarcoidosis is 60%. 25 The diagnostic yield of TBLB ranges from 40 to 90%, 11,26 but the CHEST / 118 / 4/ OCTOBER,

7 rate increases logarithmically with the number of biopsies performed per patient. 27 Using TBLB in 42 patients, Mitchell et al 26 demonstrated epithelioid cell granulomas in the lung in 88% of the patients, establishing the diagnosis of sarcoidosis. However, this procedure is associated with a significant rate of major complications, such as pulmonary hemorrhage (0.6 to 5.4%) or pneumothorax in about 1 to 10%. 28,29 Mediastinoscopy has a sensitivity of up to 88% in patients in whom there is a suspicion of sarcoidosis In a meta-analysis of 16,895 procedures, the reported major complication rates at experienced centers varied between 1.4% and 2.3%. 10 Apart from the potential complications, further disadvantages of the examination include the need for hospitalization and general anesthesia. By contrast, no complications have been observed in several series of EUS-FNA examinations in the mediastinum, and the procedure can be performed on an outpatient basis. 2 4,6 8,14 The sensitivity and specificity rates reported in studies of EUS-FNA in the lymph nodes in combination with lesions in the lung have been 89 to 96% and 100%, respectively. 2 4,6,14 The present study of EUS-FNA for diagnosing sarcoidosis in 19 patients shows that the technique may be able to serve as a powerful tool for simultaneous imaging and cytodiagnosis of the mediastinal lymph nodes. However, further studies including larger patient numbers will be required before this can be confirmed. The combination of EUS with simultaneous cytologic sampling of cell material using FNA in mediastinal lymphadenopathy is well established, 2 8 and a few cases of sarcoidosis have been observed in some studies. Mishra et al 33 recently reported seven cases of sarcoidosis in 108 EUS-guided mediastinal FNA procedures. Williams et al 8 diagnosed sarcoidosis in five men by aspirating granulomatous material from bulky posterior nodes. The patients represented a subgroup from an overall series of 120 patients undergoing EUS-FNA of mediastinal lymph nodes. All of the patients showed a clinical response to steroid treatment. In our own experience with EUS- FNA of mediastinal lymphadenopathy, we have also diagnosed a few cases of sarcoidosis and/or tuberculosis. 4,14 In cooperation with a cytopathologist who is well trained in the cytodiagnosis of sarcoidosis, we extended this method to confirm the diagnosis of suspected sarcoidosis. The typical EUS morphology observed in sarcoidosis represents either isoechoic or hypoechoic lymph nodes with atypical vascular structures within them, seen on color Doppler ultrasound imaging. When punctured, the nodes with atypical vessels show excess blood in the aspirates, confirming the highly vascular nature of these lesions. This first prospective study of the diagnostic value of EUS-FNA in sarcoidosis shows specificity and sensitivity rates of 94% and 100%, respectively. Sarcoid-like lesions were observed in association with lung, breast, colon, and other cancers, which may occur in association with the malignancy or may follow it. As all the patients who had no previous history of cancer were followed up for a median of 12 months, malignancy associated with the sarcoidosis is less likely. In those with a history of cancer, however, it remains a theoretical possibility, although such lesions have not been described previously in association with cancer of the urinary bladder, endometrium, or cervix, or with melanoma. To conclude, EUS-FNA is a safe alternative procedure for establishing a cytologic diagnosis of sarcoidosis. Before it can be recommended as a routine diagnostic procedure in suspected sarcoidosis, however, a study comparing it with standard procedures such as TBLB and/or mediastinoscopy will be needed to further define the potential role of this highly sensitive technique. References 1 Grimm H, Binmöller KF, Hamper K, et al. Endosonography for preoperative locoregional staging of esophageal and gastric cancer. Endoscopy 1993; 25: Silvestri GA, Hoffmann BJ, Bhutani MS, et al. Endoscopic ultrasound with fine needle aspiration in the diagnosis and staging of lung cancer. Ann Thorac Surg 1996; 61: Gress F, Savides T, Sandler A, et al. Endoscopic ultrasonography, fine-needle aspiration biopsy guided by endoscopic ultrasonography, and computed tomography in the preoperative staging of non-small-cell lung cancer: a comparison study. Ann Intern Med 1997; 127: Fritscher-Ravens A, Soehendra N, Schirrow L, et al. Role of transesophageal endosonography-guided fine-needle aspiration in the diagnosis of lung cancer. Chest 2000; 117: Vilman P. Endoscopic ultrasonography guided fine needle aspiration biopsy of lymph nodes. Gastrointest Endosc 1996; 43: Serna DL, Aryan HE, Chang KJ, et al. An early comparison between endoscopic ultrasound-guided fine-needle aspiration and mediastinoscopy for diagnosis of mediastinal malignancy. Am Surg 1998; 64: Giovannini M, Seitz JF, Monges G, et al. Fine-needle aspiration cytology guided by endoscopic ultrasonography: results in 141 patients. Endoscopy 1995; 27: Williams DB, Sahai AV, Aabakken L, et al. Endoscopic ultrasound guided fine-needle aspiration biopsy: a large single center experience. Gut 1999; 44: Crystal RG. Sarcoidosis. In: Fauci AS, Braunwald E, Isselbacher KJ, eds. Harrison s principles of internal medicine. 14th ed. New York, NY: McGraw Hill; 1998: Reich JM, Brouns MC, Connor EA, et al. Mediastinoscopy in patients with presumptive stage I sarcoidosis: a risk/benefit, cost/benefit analysis. Chest 1998; 113: Pauli G, Pelletier A, Bohner C, et al. Transbronchial needle aspiration in the diagnosis of sarcoidosis. Chest 1984; 85: Winterbauer RH. Sarcoidosis. In: Petersdorf RG, Adams RD, Braunwald E, et al, eds. Harrison s principles of internal 934 Clinical Investigations

8 medicine. 10th ed. New York, NY: McGraw Hill, 1983: Mountain CF. A new international staging system for lung cancer. Chest 1986; 89(suppl 4):225S 233S 14 Fritscher-Ravens A, Petrasch S, Reinacher-Schick A, et al. Diagnostic value of endoscopic ultrasonography guided fineneedle aspiration cytology of mediastinal masses in patients with intrapulmonary lesions and nondiagnostic bronchoscopy. Respiration 1999; 515: Takahashi M. Color atlas of cancer cytology. 2nd ed. Stuttgart, Germany: Thieme, 1981; 54 57,293, Cardozo PL. Atlas of clinical cytology. Weinheim, Germany: Edition Medizin, 1976; Kline TS. Handbook of fine-needle aspiration biopsy cytology. St Louis, MO: Mosby, 1981; Leonhard P. Atlas der klinischen Zytodiagnostik bei Erkrankungen im Thorax bereich. Leipzig, Germany: Barth, 1977; 50: Winterbauer RH, Betic N, Moores KD. A clinical interpretation of bilateral hilar adenopathy. Ann Intern Med 1973; 78: Newman LS, Rose CS, Maier LA. Sarcoidosis. N Engl J Med 1997; 336: Pisircriler R, Atay Z, Lang W. Cytological diagnosis of intrathoracic epithelioid cellular inflammatory process. Pneumologie 1990; 44: Atay Z. Cytodiagnosis of infrathoracal sarcoidosis. Verh Dtsch Ges Pathol 1971; 55: Tambouret R, Geisinger K, Silverman J, et al. The clinical application of FNA in the diagnosis and management of sarcoidosis [abstract]. Acta Cytol 1997; 41: Lohela P, Tikkakoski L, Strengell L, et al. Ultrasound guided fine-needle aspiration cytology of non-palpable supraclavicular lymph nodes in sarcoidosis. Acta Radiol 1996; 37: Costabel U, Zaiss A, Wagner DJ, et al. Value of bronchoalveolar lavage lymphocyte subpopulations for the diagnosis of sarcoidosis. In: Grassi C, Rizzatu G, Pozzi E, eds. Sarcoidosis and other granulomatous disorders. Amsterdam, The Netherlands: Elsevier, 1988: Mitchell DM, Mitchell DN, Collins JV, et al. Transbronchial lung biopsy through fiberoptic bronchoscope in diagnosis of sarcoidosis. BMJ 1980; 280: Gilman MJ, Wang KP. Transbronchial lung biopsy in sarcoidosis: an approach to determine the optimal number of biopsies. Am Rev Respir Dis 1980; 122: Harrow EM, Oldenburg FA, Lingenfelter MS, et al. Transbronchial needle aspiration in clinical practice: a five-year experience. Chest 1989; 96: McDougall JC, Cortese DA. Bronchoscopic lung biopsy. In: Prakash UBS, ed. Bronchoscopy. New York, NY: Raven Press, 1994; Mikhail JR, Shepherd M, Mitchell DN. Mediastinal lymph node biopsy in sarcoidosis. Endoscopy 1979; 1: Van Schil PE, Van Hee RH, Schoofs EL. The value of mediastinoscopy in preoperative staging of bronchogenic carcinoma. J Thorac Cardiovasc Surg 1989; 97: Cybulsky IJ, Bennett WF. Mediastinoscopy as a routine outpatient procedure. Ann Thorac Surg 1994; 58: Mishra G, Sahai AV, Penman ID, et al. Endoscopic ultrasonography with fine-needle aspiration: an accurate and simple diagnostic modality for sarcoidosis. Endoscopy 1999; 31: CHEST / 118 / 4/ OCTOBER,