Granulomatous inflammation appears in the histology of various pulmonary settings, including infectious diseases such as pulmonary tuberculosis (PTB) and non-infectious lung diseases such as sarcoidosis [1,2]. Epithelioid Cell Granulomas (ECG), the hallmark of tuberculous disease representing the host-defense mechanism against the inciting agent, consists of blood-derived macrophages, epithelioid cells (differentiated from macrophages/histiocytes by their elongated, distorted nuclei and indistinct cytoplasm) and multinucleated giant cells (Langhans giant cells), surrounded by T-lymphocytes [3]. Histological ECG detection in bronchoscopic biopsy samples may be used for early clinical diagnosis of PTB [4,5]. Recent findings have shown that a cytological approach is useful for non-neoplastic lung diseases [6], and cytologic detection of ECG obtained by fine-needle aspiration is useful for diagnosing unusual types of extrapulmonary tuberculosis such as tuberculous lymphadenitis [7], cutaneous tuberculosis [8], and bone tuberculosis [9]. In contrast, the existence of ECG in sputum cytology is quite rare; there are case reports of ECG detection in the sputum of PTB patients [10,11]. The use of noninvasive sputum cytology for detecting ECG in PTB has not been discussed. The aim of this study was to determine the diagnostic value of ECG in the sputum cytology of patients with PTB and other granulomatous lung diseases.

The Center for Pulmonary Diseases at Tokyo National Hospital has the largest number of beds for tuberculosis in Tokyo Metropolitan City, with 300 beds in the respiratory disease division (including 100 beds in the tuberculosis ward and 200 beds in the ward for other respiratory diseases). If sputum cytology and culture examinations are ordered, the fresh sputum samples in the laboratory are divided into bacteriological and pathological specimens. For pathology, the sputum materials are prepared as smears on glass slides, immediately fixed in 95% alcohol, and Papanicolaou-stained [12] with hematoxylin, OG-6, and EA-50 (eosin-Y and light green). Cytology specimens were reviewed and diagnosed by agreement between a cytologist and a pathologist; the presence and absence of malignant findings were evaluated, and the presence of non-malignant findings such as epithelioid cells, ECG, necrosis, fungus, and foreign bodies was routinely noted. Spontaneous sputum samples were generally used in the laboratory; however, when difficulties were encountered, induced sputum sampling was usually used.

From 2003 to 2011, 11,782 admitted patients underwent sputum cytology examinations in our department. The whole disease population characterized by granulomatous lung diseases (>10 cases) was as follows: PTB (3716 patients, 32%), pulmonary non-tuberculous mycobacteriosis (PNTM, 1117 patients, 9%), chronic pulmonary aspergillosis (629 cases, often encountered as sequelae of PTB, 5%), sarcoidosis (135 patients, 1%), hypersensitivity pneumonitis (77 patients, 0.7%), pneumoconiosis (72 patients, 0.6%), pulmonary cryptococcosis (12 patients, 0.1%), and Wegener’s granulomatosis (12 patients, 0.1%). In Japan, approximately 1-2% of active pulmonary mycobacteriosis cases are complications of lung cancer [13]; in this study, 35 (0.9%) of 3716 PTB patients and 10 (0.9%) PNTM patients also had lung cancer. Lung cancer without pulmonary mycobacteriosis (1345 patients, 11%), a non-granulomatous lung disease, was also listed, because tumor-associated sarcoid reactions are sometimes present in the lymph nodes or in the parenchyma around the tumor [14,15].

Sputum cytology identified 32 ECG-positive cases (0.3%). The background factors, examination findings, bacterial findings, radiographic findings, and pathophysiology were analyzed in these selected cases.

Chi-square test or Fisher’s exact test was used for statistical analyses of various parameters. Differences were considered significant at P<0.05.

This study was approved by the institutional review board and required no informed consent.

Thirty-two ECG-positive cases (Figure 1) identified by sputum cytology were diagnosed with pulmonary mycobacterioses, including PTB (31 cases) and PNTM (one case). ECG was not detected in other granulomatous or non-granulomatous lung diseases such as chronic pulmonary aspergillosis, sarcoidosis, hypersensitivity pneumonitis, pneumoconiosis, and lung cancer without complicating pulmonary mycobacteriosis. The frequency of ECG was significantly higher in PTB (0.8%) than in PNTM (0.09%) (P<0.05). In all 32 cases, ECG were present in a background of inflammatory cells; Langhans giant cells were detected in seven cases, two of which exhibited necrosis (PTB only; Figure 2A and 2B). Of the 35 cases with PTB and lung cancer, only one (2.9%) was ECG-positive according to sputum cytology results.

Figure 1: Epithelioid cell granuloma (a cluster of epithelioid cells showing elongated and distorted nuclei and indistinct cytoplasm) without necrosis in patients with sputum smear-positive pulmonary tuberculosis (sputum cytology, Papanicolaou-stain, original magnification 40×).

Figure 2: (A) Giant Langhans cells with peripherally oriented and overlapped nuclei in patients with sputum smear-positive pulmonary tuberculosis (sputum cytology, Papanicolaou-stain, original magnification 40×). (B) Necrosis and scattered epithelioid cells in patients with sputum smear-positive pulmonary tuberculosis (sputum cytology, Papanicolaou-stain, original magnification 40×).

Clinicopathological features of the ECG positive PTB cases (n=31) are shown in Table 1. The PTB cases included 23 men with a median age of 70 years. Sixteen patients (approximately 50% of the subjects) were aged over 70 years. Underlying diseases included diabetes in four cases, malignant tumors after surgery in two cases, and human immunodeficiency virus positivity in one case; no other significant findings were observed. Radiography revealed 21 cases with a cavitary shadow (68%) and extensive spreading of the disease (exceeding hemithorax) in nine cases (29%). In 2 of 31 cases, endobronchial tuberculosis was suspected on radiography and computed tomography.

Table 1: Clinicopathological features of the ECG-positive cases in PTB (n=31)

Twenty-nine cases (94%) showed positive sputum culture for M. tuberculosis consistent with sputum cytology, and the remaining cases were also culture-positive with other sputum samples. Twenty-eight cases (90%) were sputum smear-positive for acid-fast bacilli, and 11 (35%) also had a large (3+) amount of mycobacteria on the sputum smear. The frequency of ECG positivity in smear-positive PTB cases was 1.1% (28/2521 cases).

There was one case of ECG positivity in PNTM (Mycobacterium avium complex lung disease (MACLD). Radiography did not show nodular bronchiectasis, a typical radiographic pattern of MACLD [16], but fibrocavitary MACLD, with extensive spreading of the disease and upper lobe cavitary shadow on radiography, similar to PTB in this study. Bacteriological examination of the smear and culture of PNTM sputum were also positive, like most TB cases in this study. Thus, no notable differences in clinical features between PTB and PNTM were observed.

The significance of sputum cytology in patients with respiratory diseases, especially PTB, has been discussed by Nasiell et al. [17] in 1972; 0.2% of epithelioid cells and/or multinucleated giant cells were observed in smear cytological specimens of sputum and/or bronchial secretions. Of the 66 patients with epithelioid cells and/or multinucleated giant cells, a majority had PTB (40 patients), followed by lung cancer (11 patients), and sarcoidosis (three patients). Roger et al. [18] reported sputum cytology in PTB patients revealed an extremely high incidence of epithelioid cells (28%) and Langhans giant cells (5%). Further, Tani et al. [19] reported a high frequency of epithelioid cells (56%) in PTB patients using cell-block preparations. These reports indicate that the presence of epithelioid cells and Langhans giant cells in sputum and/or bronchial secretions implies a strong indication of PTB; however, they do not describe the clinical characteristics of these PTB patients.

Recently, Bhatia and Singh [6] reviewed cytological diagnoses of granulomatous inflammation and noted ECG has been observed in tuberculosis, sarcoidosis, occupational lung diseases, and fungal infections such as aspergillosis if polymorphous exudates are also present. They also indicated that the presence of necrosis with ECG is suggestive of tuberculosis.

In this study, ECG positivity on sputum cytology was 0.3% among patients with respiratory diseases and, unlike Nasiell’s report [17], ECG positivity was restricted to pulmonary mycobacteriosis, especially PTB (0.8% in PTB and 0.09% in PNTM), whereas in cases with other granulomatous respiratory diseases such as fungal diseases, sarcoidosis, and cases with lung cancer, ECG were absent. Concomitant findings of Langhans giant cells with or without necrosis were specific findings of PTB.

Previous studies have described histological ECG-positive rates of 61% in sputum-smear negative PTB [4] and 66% in sputum-smear negative PNTM [20] by transbronchial biopsy. Cytological ECGpositive rates obtained by fine-needle aspiration in patients with extrapulmonary tuberculosis were higher; 20% showing ECG without necrosis and 62% showing ECG with necrosis/abscesses in tuberculous lymphadenitis [7]; 89% of lupus vulgaris showing cohesive ECG in cutaneous tuberculosis [8]; 85% with ECG in tuberculosis of the bone [9]. Although the ECG-positive rate was extremely low in comparison to these invasive and expensive examinations, sputum cytology is noninvasive and low-cost, and is routinely performed to rule out lung cancer in patients admitted with respiratory diseases. Through this analysis of a large sputum cytology database, we believe that although sputum cytology is not a primary procedure, it could be a useful aid for the diagnosis of pulmonary mycobacteriosis, especially PTB.

This study revealed the clinical features of ECG-positive cases; patients had PTB with cavity (68%) on radiographs and were positive for acid-fast bacilli in sputum smear (90%). Furthermore, the ECGpositive rate in sputum smear-positive PTB (1.1%) was greater than the rate in culture-positive PTB (0.8%). These findings suggest ECG in sputum cytology is associated with PTB severity.

Although there was only a single case of PNTM (MACLD) with ECG-positive sputum, the clinical portrait was atypical and severe, similar to the ECG-positive PTB cases, in its radiographic and bacteriological characteristics. ECG in sputum cytology may be associated with the clinical situation rather than the species of Mycobacterium.

Since detecting M. tuberculosis is the gold standard for the diagnosis of tuberculosis, ECG-positivity in sputum cytology is of course no more important than bacteriological examination for PTB diagnosis; however, sputum cytology is a non-invasive, cost-effective, and routine examination, so ECG-detection in sputum cytology may be a useful aid to cope with unsuspected, but highly infective PTB patients. Furthermore, the coexistence of necrosis and Langhans giant cells are highly indicative of PTB diagnosis.

This work was supported in part by a grant from Research in the Policy Based Medical Service Network for Chest Diseases in Japan. The authors thank Mr. Kenshi Urata, Pathology Division, National Hospital Organization Tokyo National Hospital, for technical assistance.