Bronchoalveolar Lavage in the Diagnosis of Peripheral, Primary Lung Cancer

Bronchoalveolar Lavage in the Diagnosis of Peripheral, Primary Lung Cancer* Alichael Pirozynski, 1\1. D., F. C. C. p.

Numerous case reports have shown the advantage of using bronchoalveolar lavage (HAL) in cytologic diagnosis of primary and secondary malignant neoplasms of the respiratory system. The aim of this study was to determine the usefulness of HAL in the diagnosis of peripheral, primary lung cancer. Of 1,864 patients referred to the Hronchological Department for endoscopic examination, 145 patients were studied: six with large cell lung cancer, 22 with adenocarcinoma, 15 with alveolar cell lung cancer, 40 with small cell lung cancer, and 62 with squamous cell lung cancer. In 94 patients (64.8 percent), HAL was diagnostic, revealing malignant cells. In 52 (35.9 percent) of these patients, the cytologic diagnosis agreed with the 6nal pathologic diagnosis of the resected tumor. The result of HAL was affected by

the type of cancer and size of the tumor. Highest yields were seen in adenocarcinoma (59.2 percent) and alveolar cell lung cancer (80 percent). The average size of the tumor in the group with correct cell typing was 4.9± 1.8 cm; in patients with nondiagnostic HAL, the average size was 2.6± 1.2 cm. HAL provided the highest (statistically significant, p<0.05) diagnostic yield (64.8 percent) in comparison with other sampling techniques: brush biopsy (29.8 percent), catheter biopsy (26.8 percent), and forceps biopsy (32.7 percent). The diagnostic yield of HAL and transbronchial 6ne needle aspiration biopsy (58.3 percent) did not signi6cantly differ. HAL proved to be a valuable diagnostic tool in detecting peripheral, primary, pulmonary malignant neoplasms. (Chest 1992; 102:372-74)

lavage (BAL) \vas demonstrated by B ronchoalveolar Sackner et all to be a safe and reliable Inethod

were earned out aecording to widely at'Cepted rnethods!l Fluoroscopic guidance was always used to loeate the hlnlor. Cancer cell types were identified ac(.'()rding to the criteria listed hy Johnson and Frahle" and Lindtor and Rennard.~ Statistical analysis was carried out using the x2 test.

for sampling of material from distal airways and alveoli. Numerous studies in the 1970s demonstrated usefulness of this method in assessing interstitial lung diseases,2.3 but nlost have not appreciated the use of BAL in the diagnosis of pulmonary malignant neo-

For editorial comment see page 331 plasms, although the criteria of cytologic diagnosis of lung cancer have been established in the past. 4 .5 The aim of this study was to assess if BAL can be useful in the diagnosis of peripheral pulmonary malignant neoplasms.

RESULTS

One hundred forty-five patients with peripheral, prilnary lung cancer of deterlnined histologic type \vere included in the study (Table 1). There were six subjects with large cell lung cancer, 22 with adenocarcinoma, 15 \\rith alveolar cell lung cancer, 40 with small cell lung cancer, and 62 \\rith squamous cell lung cancer. In all, HAL \\'as carried out immediately follo\\ring other sampling methods (fluoroscopically guided transbronchial fine needle biopsy, brush, forTable 1- General Characteristics of the Studied Population

MATERIALS ANI) METH()DS

A retrospective analysis of BAL data was carried out in 1,864 patients. Only patients \\'ho had primary, peripheral pulnlonary mali~nant neoplasnls, nornlal endos(.'opic findings meeting the criteria of the Japanese Oncolo~cal Society, Ii and had BA L perfi)rmed durin~ the initial bronchoscopy \\'ere selected for the study. One hundred fi)rty-five patients fulfilled the selection criteria (Table

Sex

Age, yr

No. of Patients

~I

F

~Iean

Range

145

76

69

52,1 ± 10,6

22-78

1).

Bronchoalveolar lavage was carried out according to the recornmendations of the European BAL Task Group.7 A total volume of 200 ml of 0.9 percent NaCI. \\'anned to 37°C, given in ten equal alliquots were used. The returned samples \\'ere pooled and processed. The fluid \\'as not filtered, only centrifuged. Smears were made from the cell pellet (three slides). Paraffin blocks were made from the rernaining nlc.lterial. Three slides were reviewed. Hernatoxylin-eosin staining was used. Other endoscopic sarnpling nlethods (forceps biopsy, catheter biopsy, transbronchial fine needle aspiration biopsy) were perfi)rmed prior the lavage pro(.·edure. All *Frorn the Departrnent of Bronchology, Institute of Tuberculosis and Lung Diseases, \Varsaw, Poland. Manuscript re<:eived April 15: revision aceepted Novernher 12.

372

Table 2-Results ofBronchoalveolar lAvage (BAL) in 145 Patients with Peripheral Lung Cancer BAL Group Large cell lung cancer Adeno(.·. ard He )Ina Aln,"olar cell lung cancer Squanlous cell lung cancer Snlall cell lung cancer 1()tal, No. (%)

No. of Patients 6 22 15 62 40

145 (HX»)

Positive

Negative

3

3 4

lR

12 37 24 94 (64.83)

3 25

16 51 (35.17)

SAL in Diagnosis of Primary Lung Cancer (Michael Pirozynski)

ceps, and catheter biopsies). In 94 (64.8 percent of 145 patients) BAL sampled cells that were malignant (Table 2). The cytologic examination did not reveal any neoplastic cells despite the presence of cancer in only 35.2 percent of the cases. The correct type of cancer was found in 52 patients (38.9 percent). The highest diagnostic yield was found in patients with adenocarcinoma (59.1 percent) and alveolar cell lung cancer (80 percent). In the remaining, the correct type was identified in less than 50 percent of the patients. Comparison of BAL results with other sampling techniques demonstrated BAL's higher diagnostic yield (Table 3). The differences were all statistically significant (p<0.05) except between BAL and aspiration transbronchial fine needle biopsy. The diagnostic efficacy of BAL was also affected by the tumor size on the chest roentgenogram. In patients with the correct cell typing, the average size of the tumor was 4.9 ± 1.8 cm. In patients with nondiagnostic BAL, the average size of the tumor was smaller, 2.6± 1.2 cm (Table 4). DISCUSSION

In the past, the diagnosis of lung cancer was rarely achieved by BAL.6,8 In visible tumors-those seen during a routine bronchoscopic examination - Lundgren et al8 and Kvale and coworkers9 demonstrated that tissue biopsy specimens yielded superior results in comparison with cytologic examination of bronchial washings and secretions. However, when comparing results of similar findings in peripheral lung tumors, the superiority of cytologic examination of bronchial washings and secretions was evident. 9 Unfortunately, the available data did not include analysis of the effect of BAL on diagnostic yield. A few studies have been published on the usefulness of BAL in the diagnosis of primary and secondary malignant neoplasms of the respiratory tract. Levy et allO reported a high efficacy of BAL in diagnosing cancer but studied only five patients. Baglin et alii demonstrated a diagnostic yield of 40 percent in ten patients with histologically confirmed peripheral pulmonary malignant neoplasms. The same authors ll had shown a considerably better yield in diffuse malignant Table 3-Comparison of Various Sampling Techniques in 145 Patients with Peripheral, Primary Lung Cancer Biopsy Method

No. of Patients

Diagnosis Achieved, No. (%)

Brush biopsy Catheter biopsy Transbronchial aspiration fine needle biopsy Forceps biopsy Bronchoalveolar lavage

124 112 24

37 (29.8) 30 (26.8) 14 (58.3)

107 145

35 (32.7) 94 (64.8)

Table 4-Effect of Tumor Size (in Centimeters) on Bronchoalveolar lAvage (BAL) Results BAL*

No. of patients Mean diameter of tumor Range

(+)

(+/-)

(-)

52 4.9 ± 1.8 2-9

42 2.6 ± 1.3 1-6.5

51 2.6 ± 1.2 1-6

*( + ) = cell type correctly identified; (+ / - ) = only malignant cells present, cell type could not be correctly identified; and ( - ) = malignant cells not identified.

processes of the lungs (diagnosis was made in 9 of 11 patients). Linder and coworkers 12 found BAL to be diagnostic in 24 of 35 patients with proven pulmonary malignant neoplasms. Using this method, Wiesner et al 13 diagnosed adenocarcinoma in two patients, while others have shown either the usefulness of BAL in diffuse pulmonary changes of neoplastic etiology-in bronchoalveolar cancer,14,15 metastatic spread of adenocarcinoma to the lung,16 lymphangitic carcinomatosislO-or in the course of hematologic malignant neoplasms, as demonstrated by Costabel et al. 17 Case reports have also shown its value in establishing the diagnosis of pulmonary changes in cases of lymphomas l8 . 19 and Waldenstrom's macroglobulinemia. 20 The results of this study demonstrate the usefulness of BAL in establishing the diagnosis of pulmonary malignant neoplasms. As in other studies,5,8.12 the diagnostic yield was affected by the cell type of the tumor. It was higher for alveolar cell lung cancer and adenocarcinoma. The sensitivity of BAL was 64.8 percent. The correct cell type was demonstrated in 35.9 percent of the studied patients, but one must remember that cytologic diagnosis of a malignant neoplasm is much easier than correct cell typing. 5.12 Although BAL proved to be a valuable tool in diagnosis of lung cancer, one must keep in mind the limitations of cytologic diagnosis - mainly the similarity ofseverely dysplastic and malignant cells in patients with coexisting pneumonias or inflammatory changes and neoplasia; thus, care should be taken in interpreting results in such patients. 4,5,21 Unfortunately, not all patients underwent other sampling techniques-forceps, brush, and transbronchial fine needle biopsy-but still the comparison of the results is favorable for BAL (Table 3) (differences statistically significant p<0.05). Only fine needle biopsy achieved the results comparable with BAL (58.3 percent vs 64.8 percent; differences statistically not significant). Brush biopsy specimens showed a lower diagnostic yield (28.3 percent) which could only be explained by the fact that one samples only one bronchus at the same time - hopefully the one invaded CHEST I 102 I 2 I AUGUS'T, 1992

373

by the neoplastic process. Bronchoalveolar lavage, on the other hand, samples more than one bronchus at the same time, increasing the probability of reaching the malignant lesion. The diagnostic yield of BAL can also be influenced by lavaging the area after all other sampling methods (brush, catheter biopsy). The results of BAL were also in8uenced, in this study, by the size of the lesion. Correct cell typing was seen in tumors exceeding the diameter of 3 em. Nondiagnostic BAL was seen in smaller lesions, either because the tumor itself was so small that it could not be reached sufficiently by the lavage 8uid or the shedding of the malignant cells in such cases is small (statistical significance, p
374

8 Lundgren R, Bergman F, Angstrom T. Comparison of transbronchial fine needle aspiration biopsy, bronchial washings, brush biopsy and forceps biopsy in the diagnosis of lung cancer. Eur J Respir Dis 1983; 64:378-85 9 Kvale PA, Bode FR, Kini S. Diagnostic accuracy in lung cancer: comparison of techniques used in association with flexible fiberoptic bronchoscopy. Chest 1976; 69:752-57 10 Levy H, Horak DA, Lewis MI. The value of bronchial washings and bronchoalveolar lavage in the diagnosis of lymphangitic carcinomatosis. Chest 1988; 94:1028-30 11 Baglin JY, Danel C, Carnot F, Lacronique J, Jaubert F, Chretien J. Interest of bronchoalveolar lavage (BAL) in the diagnosis of lung tumors with normal fiberoptic bronchoscopic examination. Bethesda, Md: International Conference on Bronchoalveolar Lavage, 1984:26 12 Linder J, Radio SJ, Robbins RA, Chafouri M, Rennard SI. Bronchoalveolar lavage in the cytologic diagnosis of carcinoma of the lung. Acta Cytoll987; 31:796-801 13 Wiesner B, Knoll ~ Jager J, Kessler G. Der Wert der bronchoalveolaren Lavage (BAL) fur die Diagnostik des Adenokarcinoms der Lunge. Z Erkr Atmungsorgane 1986; 167:158-62 14 Springmeyer SC, Hackman R, Carlson JJ, McClellan JE. Bronchoalveolar cell carcinoma diagnosed by bronchoalveolar lavage. Chest 1983; 83:278-79 15 Sestini ~ Rottoli L, Cotli C, Miracco C, Luzi E Bronchoalveolar lavage diagnosis of bronchio-alveolar carcinoma. Eur J Respir Dis 1985; 66:55-8 16 Fedullo AJ, Ettensohn DB. Bronchoalveolar lavage in the lymphangitic spread ofadenocarcinoma to the lung. Chest 1985; 87:129-31 17 Costabel U, Bross KJ, Matthys H. Diagnosis by bronchoalveolar lavage of cause of pulmonary infiltrates in hematologic malignancies. Br Med J 1985; 290:1041 18 Davis WB, Cadek JE. Detection of pulmonary lymphoma by bronchoalveolar lavage. Chest 1987; 91:787-90 19 Myers JL, Fulmer JD. Bronchoalveolar lavage in the diagnosis of pulmonary lymphomas. Chest 1987; 91:642-43 20 Kabayashi H, Ii K, Hizawa K, Maeda T. Two cases of pulmonary Waldenstrom's macroglobulinemia. Chest 1985; 88:297-99 21 Biyoudi-Vouenze R, Tazi A, Hance AJ, Chastre J, Basset F, Soler E Abnormal epithelial cells recovered by bronchoalveolar lavage: are they malignant? Am Rev Respir Dis 1990; 142:686-90 22 Radke JR, Conway WA, Eyler WR, Kvale PA. Diagnostic accuracy in peripheral lung lesions: factors predicting success with 8exible fiberoptic bronchoscopy. Chest 1979; 76:176-79 23 Nathan MH. Management of solitary pulmonary nodules. JAMA 1974; 227:1141-44

BAL in Diagnosis of Primary Lung Cancer (Michael Pirozynski)