記住我
Abstract
Objectives: Respiratory cytology specimens such as bronchoalveolar lavage (BAL) and bronchial wash (BW) obtained using a fiberoptic bronchoscope are very useful in detecting or ruling out various inflammatory conditions, infections, and neoplastic lesions. A study was carried out to determine the usefulness of respiratory cytology in the diagnosis of pulmonary lesions and the limitations of cytology if any, and correlate the results of cytology with biopsies wherever possible. Methods: All bronchoscopic cytology and biopsy specimens received at the pathology laboratory of this tertiary care institute between June 2014 and May 2017 were analyzed. Cytology smears were stained with Leishman's stain, hematoxylin and eosin (H and E), Papanicolaou (PAP), and Ziehl–Neelsen (ZN) stain for all cases and special stains wherever needed. Slides prepared from biopsy specimens were stained with H and E. Immunohistochemistry was used for confirmation and further typing of malignant lesions and diagnosis rendered was compared with the corresponding cytology diagnosis. Results: A total of 120 specimens of BAL or BW cytology with or without biopsy were analyzed. Thirty-three were diagnosed as non-specific inflammatory lesions. The most common malignancy diagnosed by cytology was adenocarcinoma followed by squamous cell carcinoma. Correlating BAL with biopsy specimens, the sensitivity, specificity, and diagnostic accuracy of BAL were 100%, 88.8%, and 91.6%, respectively. Correlating BW with biopsy specimens, the sensitivity, specificity, and diagnostic accuracy of BW were 85.6%, 85.6%, and 85.6%, respectively. Conclusions: Accurate diagnosis can be made from the examination of bronchoscopic cytology specimens in pulmonary inflammation, tuberculosis, fungal infections, and malignancies. Combining respiratory cytology with biopsy and ancillary techniques can aid in better subtyping of neoplastic lesions.
Keywords: Bronchial wash, bronchoalveolar lavage, bronchoscopy, cytology, pulmonary
How to cite this article: Introduction 
Respiratory diseases are very common and many patients present to respiratory clinics with symptoms, which may attract basic radiologic investigations. Patients presenting with prolonged or severe symptoms and with a strong suspicion of malignancy will require further investigations such as multidirectional computed tomography and magnetic resonance imaging. Many times, a definitive diagnosis may not be possible despite clinical and radiologic evaluation, and routine laboratory tests including pulmonary function tests. Hence, it often becomes necessary to obtain cells or biopsy tissue for pathologic examination to arrive at a definitive diagnosis.
The development of a rigid bronchoscope in the late nineteenth century by Gustav–Kilian formed the foundation of a technology by which the mucosal surfaces of bronchi could be directly visualized and sampled for both tissue and cellular evaluation. This technology further progressed through the introduction of flexible catheters into the bronchi and perhaps even more importantly by the development of a flexible fiberoptic bronchoscope.[1] This opened new avenues for the evaluation of various cytology specimens from the respiratory tract.
Exfoliative respiratory cytology includes sputum cytology, bronchial brushings, bronchial washings, and bronchoalveolar lavage (BAL) cytology. The major role of respiratory cytology is in the diagnosis of primary and metastatic pulmonary tumors and inflammatory and infective conditions.[2]
BAL explores large areas of the alveolar compartment, providing cells as well as non-cellular constituents from the lower respiratory tract.[3] An alteration in the BAL fluid reflects a pathological change in the lung parenchyma. Its usefulness has been accepted worldwide in the diagnostic workup of infectious and non-infectious lung diseases.[4]
Bronchial brushings and washings are complementary to sputum cytology in the diagnosis of pulmonary lesions. Bronchial wash (BW) cytology is a valuable tool and yields almost the same information as a biopsy. It is useful in patients with evidence of obstruction or risk of hemorrhage.[4]
The increasing demand for minimally invasive procedures and the need to extract increasing amounts of diagnostic and predictive information from ever smaller amounts of diagnostic material has placed further demands on the cytopathologist.[5] BAL and BW cytology have several advantages over biopsy. It is a safe, minimally invasive, and well-tolerated procedure, even when performed in high-risk and immunocompromised individuals.[4]
In this study, we will determine the usefulness of bronchoscopic cytology as a diagnostic tool in lung lesions and correlate with the histopathological diagnosis of bronchoscopic biopsy specimens wherever available along with cytochemical and immunohistochemical (IHC) studies.
Materials and Methods This was a cross-sectional, descriptive study carried out in a tertiary care hospital for 3 years, from June 2014 to May 2017, where specimens from 120 patients with pulmonary lesions were analyzed. Approval from ethics committee was obtained. Date of approval: 03/11/2015, Ref ID: SDMIEC: 0417:2015.
Inclusion Criteria: All BW or BAL samples and bronchoscopic biopsies obtained from both male and female patients of all age groups that were received in our laboratory during the study period were included in the study.
Exclusion Criteria: Incorrectly labeled and poorly preserved samples were excluded from the study.
All patients with clinical symptoms suggestive of respiratory disease and/or abnormal imaging studies were carefully examined and selected for bronchoscopy by the clinicians. After obtaining informed consent, a bronchoscopy was performed by the pulmonologists. The aspirated fluid samples (BW or BAL) were sent unfixed to the lab immediately. If the processing was expected to be delayed, the fluids were stored at 4°C for 24–48 h.
After gross features of the fluid such as volume, color, appearance, and presence of any clots were noted, the container of the fluid was shaken to disperse the cells and then placed in a cytospin funnel with a filter paper placed between the slide and funnel and then centrifuged. A minimum of four slides were prepared for each specimen. Two of the smears were then fixed in 95% alcohol and stained with routine hematoxylin and eosin (H and E) and Papanicolaou (PAP) stain and examined under a microscope. Unfixed smears were stained with Leishman's stain and Ziehl–Neelsen (ZN) stain. Mucus or blood clots in the fluid specimens were removed and placed in cassettes for processing as a cell block material.
The endobronchial biopsy specimens were sent fixed in 10% formalin and processed in an automated tissue processor and paraffin blocks were made. From each block, multiple sections were taken at different levels and stained with H and E stain. A slide with special stain PAS was also prepared in selected cases.
All malignant lesions identified in H and E sections of biopsies or cell blocks were subjected to IHC staining using TTF1, p63, chromogranin, synaptophysin, and PanCK as per the case requirement.
Parameters studied:
Adequacy of specimenIndicators of inflammation such as an abundance of neutrophils/lymphocytes/eosinophils.Infectious agents such as bacteria/fungi/viral inclusions.Reactive cellular changesDysplastic or malignant cellular features and categorizing them into various malignancies as per the new 2015 World Health Organization (WHO) classification.Gram's stain/KOH preparation/culture sensitivity reports of BW or BAL fluids sent to the microbiology laboratory were also collected.
Statistical analysis: The data collected were tabulated and analyzed by proportions and percentages, and diagnostic accuracy of the test was obtained using SPSS software package version 22.
Results Of the 120 patients, 74 (61.67%) were males and 46 (38%) were females. The majority (28.33%) of the patients were between the age group of 51 and 60 years. The most common indication for FOB was lung mass in 41 (34.2%) patients followed by non-resolving pneumonia in 38 (31.7) and tuberculosis in 26 (21.7%) patients. Metastasis to the lungs was suspected in eight (6.7%) patients. Other indications were pleural effusion, lung abscess, pulmonary alveolar proteinosis, and foreign body aspiration. Among 120 bronchoscopic cytology specimens, 86 were BW and 34 were BAL. Bronchoscopic biopsy specimens were received along with cytology specimens (either BW or BAL) for 47 patients.
Bronchial wash: Eighty-six samples of BW were analyzed. Twenty-nine (33.72%) were neoplastic lesions, out of which 11 (12.8%) were reported as suspicious for malignancy, 12 (13.9%) as adenocarcinomas [Figure 1]a and [Figure 1]b, 4 (4.65%) as squamous cell carcinomas [Figure 2]a and [Figure 2]b, 1 (1.16%) as carcinoid tumor [Figure 3]a and [Figure 3]b, and 1 (1.16%) as metastatic carcinoma. Twenty-nine (33.72%) were non-neoplastic lesions, out of which 21 (24.4%) were non-specific inflammation, 5 (5.81%) were tuberculosis [Figure 4], 2 (2.32%) were fungal infections, and 1 (1.16%) was nocardiosis [Figure 5]. Twenty-seven (31.4%) BW fluids were reported as negative for any pathological lesions (neither malignancy nor inflammation) and 1 (1.16%) was unsatisfactory.
Figure 1: (a) Bronchial wash showing malignant cells in acinar arrangement (H and E, 40×); (b) Bronchoscopic biopsy showing adenocarcinoma (H and E, 40x); (c) TTF1-positive adenocarcinoma cells (IHC, 10×)
Figure 2: (a) Group of malignant squamous cells in the bronchial wash (H and E, 40×); (b) Bronchoscopic biopsy showing squamous cell carcinoma (H and E, 40×)
Figure 3: (a) Carcinoid tumor cells in the bronchial wash (H and E, 40×); (b) Biopsy of carcinoid tumor (H and E, 40×); (c) Carcinoid tumor cells negative for TTF1 (IHC, 40×); (d) Carcinoid tumor cells positive for synaptophysin (IHC, 40×)Bronchoalveolar lavage: In the present study, 34 BAL fluids were analyzed. Fourteen (41.17%) among 34 were non-neoplastic, out of which 1 (2.9%) was diagnosed as tuberculosis, 1 (2.9%) as pulmonary alveolar proteinosis, and 12 (35.3%) as non-specific inflammation [Figure 6]. Seven (20.6%) out of 34 patients had neoplastic lesions, of which 4 (11.8%) were squamous cell carcinomas, 2 (5.8%) were adenocarcinomas, and 1 (2.9%) were reported as suspicious for malignancy. Twelve (35.3%) BAL fluids were reported as negative for any pathological lesions and 1 (2.9%) was reported unsatisfactory due to excessive hemorrhage.
Taking both BW and BAL into consideration among neoplastic lesions, the most common malignancy diagnosed by cytology (BW/BAL) was adenocarcinoma in 14 (11.6%) patients, followed by squamous cell carcinoma in 8 (6.66%) and carcinoid tumor in 1 (0.83%) patients. Twelve (10%) patients were diagnosed as suspicious for malignancy and 1 (0.83%) as metastatic lung carcinoma.
Fluid specimens of all 120 patients were tested for tubercular bacilli by staining one slide of each specimen with ZN stain. Six (5%) were positive for acid-fast bacilli. Reports of sputum AFB were correlated for the above six Acid Fast Bacilli (AFB)-positive bronchoscopic cytology specimens (BW/BAL), which showed that only one sputum smear out of the total six was positive for AFB. The rest five were sputum-negative cases.
The FOB fluids BW/BAL obtained from 120 patients were also sent for microbiologic studies such as Gram's stain/KOH preparation/culture and sensitivity for identification and isolation of bacteria or fungi. On reviewing these reports, 42 fluid specimens showed positive results. Among 42 cases where bacteria were isolated, the most common organism isolated was Klebsiella species in 15 (35.7%) patients, followed by Pseudomonas species in 9 (21.42%), Staphylococcus species and Streptococci in 6 patients each (14.3%), Enterobacter species in 5 (11.9%), Escherichia coli in 4 (9.5%), and Citrobacter freundii in 1 (2.4%). More than one organism was isolated in four cases. In two cases, Gram's stain showed fungal hyphae but was not isolated in culture.
Out of 120 cases studied, a bronchoscopic biopsy was performed in 47 patients and the following diagnosis was made [Table 1]. All (14) bronchoscopic biopsies diagnosed as adenocarcinoma on morphology showed positivity for special stain PAS and IHC marker TTF1 [Figure 1]c. IHC analysis was performed for three out of five biopsies diagnosed as squamous cell carcinoma and all were negative for TTF1 and positive for P63. Two tumors diagnosed as carcinoid tumors showed positivity for synaptophysin [Figure 3]c, chromogranin, and Pan-CK and were negative for TTF1 [Figure 3]d. One of the two cases diagnosed as metastatic deposits was positive for estrogen receptors.
Correlation between cytopathology diagnosis and histopathology (HP) diagnosis for neoplastic and non-neoplastic lesions
The BAL cytology showed 100% sensitivity and 88.8% specificity with a positive predictive value of 75%, a negative predictive value of 100%, and a diagnostic accuracy of 91.6% [Table 2].
The BW cytology showed 85.6% sensitivity and specificity with a positive predictive value of 90%, a negative predictive value of 80%, and diagnostic accuracy of 85.7% [Table 3].
In this study, 95.8% of the tumors could be morphologically subtyped through biopsy. Among cytology specimens, 85.7% and 62% of tumors could be subtyped morphologically by BAL and BW, respectively.
Limitations
When the cellularity is low in malignant conditions, a diagnosis of suspicious malignancy may be made, which further leads to a repeat procedure or the need for a biopsy. In this study, 10% of the cases were reported as suspicious for malignancy on cytology specimens (BW/BAL).Reactive cellular changes may sometimes be misinterpreted as malignant lesions, or a malignant lesion may be completely missed due to excess obscuring elements such as inflammatory cells or hemorrhage.Accurate tumor subtyping may not always be possible on cytology specimens due to a lack of architectural patterns or low cellularity. Discussion The use of exfoliative cytology in the diagnosis of various pulmonary lesions has been revolutionized using fiberoptic bronchoscopy. Various combinations of cytologic techniques such as BAL, BW, bronchial brushings (BB), and forceps biopsy (FB) are required for the high diagnostic yield in the management of neoplastic and non-neoplastic pulmonary lesions. As these techniques became easier and more accessible, their popularity increased, and have now become the first-line diagnostic tools for malignancies. Pulmonary cytology is now well-established throughout the world as a diagnostic procedure in the evaluation of patient with suspected lung malignancy and other pulmonary lesions.[6],[7]
The present study was carried out for 3 years from July 2014 to May 2017 (1-year retrospective and 2 years prospective). Cytological examination of 120 fluid specimens (BAL/BW) received was carried out with histopathological correlation in 46 cases.
In the present study, the majority of the patients affected were males with a male-to-female ratio of 1.6:1. Other similar studies such as Anantham et al. and Ghildiyal et al. also showed that males were more commonly affected than females with a male-to-female ratio of 1.6:1 and 1.9:1, respectively.[8],[9]
Diagnosis of neoplastic lesions on BAL cytology
In the present study, out of 34 BAL fluids analyzed, 7 were categorized as neoplastic, 26 were non-neoplastic, and 1 was inadequate. Out of 34 cases, a biopsy was available only for 12 cases. On comparing these, we found that three were correctly classified as malignant and one case, which was diagnosed malignant on BAL, was negative for malignancy on biopsy. Thus, three were truly positive (75%) and one was false positive (25%) for the neoplastic lesion. Among eight cases, which were diagnosed as non-neoplastic, all were correctly diagnosed as non-neoplastic on biopsy. Thus, the true negative percentage was 100% and there were no false-negative cases. The histologic subtype was correctly diagnosed as squamous cell carcinoma in one case on BAL.
In a similar study by Bhat et al. (n = 902), 301 patients were categorized as neoplastic and 601 as non-neoplastic on BAL. Biopsy-confirmed neoplastic cases were 760. On correlating with biopsy, 270 were truly positive, 31 were falsely positive, 490 were falsely negative, and 111 were truly negative on BAL.[10]
In another study by Gaur et al. (n = 196), 41 BAL specimens were diagnosed as neoplastic and 155 were non-neoplastic. Seventy-one were proven to be neoplastic on biopsy. On correlating BAL with biopsy, 28 were truly positive, 13 were falsely positive, 43 were falsely negative, and 112 were truly negative on BAL[11][Table 4].
Table 4: Comparing sensitivity, specificity, and diagnostic accuracy of neoplastic and non-neoplastic lesions in BAL among other studiesDiagnosis of neoplastic lesions on bronchial wash cytology
Among 86 BW fluids analyzed, 29 (33.72%) were diagnosed as neoplastic lesions, 56 as non-neoplastic, and 1 was inadequate. Out of these 86 cases, a biopsy was performed for only 35 cases. On correlating these 35 cases with biopsy diagnosis, it was seen that 18 were correctly diagnosed and 2 were falsely diagnosed as a neoplastic lesion on BW. Thus, the true-positive percentage was 90% and the false-positive percentage was 10%. Among the non-neoplastic cases, 12 cases were correctly diagnosed but three cases were diagnosed as neoplastic on biopsy. Hence, the false-negative percentage was 20% and the true-negative percentage was 80% in this study.
In a study by Ahmad et al. (n = 73), 34 were diagnosed as neoplastic on BW, and 39 as non-neoplastic. Out of 34 neoplastic cases, 33 were accurately diagnosed and only 1 was falsely reported as neoplastic. Thus, the true-positive percentage was 97% and the false-positive percentage was 2.94%. Among 39 non-neoplastic cases, 31 were true negatives (79.5%) and 8 were false negatives (20.5%) as they were found to be malignant on biopsy.[4]
A study by Bodh et al. (n = 200) consisted of 155 biopsy-proven neoplastic cases. Sixty-two cases were diagnosed as neoplastic on BW and 138 as non-neoplastic. Among these 62 cases, 57 were truly positive and 5 were falsely diagnosed as neoplastic on BW. Among 138 cases, which were diagnosed as non-neoplastic, 40 were true-negative cases and 98 were false-negative[12][Table 5].
Table 5: Comparing sensitivity, specificity, and diagnostic accuracy of neoplastic and non-neoplastic lesions in BW among other studiesMost common neoplastic lesion
In our study, out of 120 fluid specimens analyzed (BW/BAL), 36 were diagnosed as neoplastic. Among these, the most common malignancy was adenocarcinoma followed by squamous cell carcinoma. This was comparable with studies by Gaude et al. and Ghildiyal et al., which also found that the most common malignancy was adenocarcinoma, followed by squamous cell carcinoma.[3],[9] However, in a study by Raiza et al., squamous cell carcinoma was the most common malignancy followed by adenocarcinoma.[13]
Accuracy of morphologic subtyping for neoplastic lesions on cytology
In the present study, accurate morphologic subtyping of neoplastic lesions could be performed on cytology specimens (BAL/BW) of 12 patients (33.3%). This was much lower compared to a study by Ahmad et al., where correct tumor subtyping on cytology could be performed in 61% of cases.[4]
Diagnosis of tuberculosis
In the present study, one BAL fluid and five BW were diagnosed as tuberculosis based on the presence of AFB on smears. The sputum smears of these patients, which were examined before the FOB showed that only one out of six sputum smears was positive for AFB. None of the FOB fluids or sputum specimens were subjected to Mycobacterium tuberculosis on culture in this study.
In a study by Gaude et al. (n = 175), 16 patients were diagnosed with tuberculosis on BAL based on smear positivity for AFB (11 cases) and culture positive (16 cases) for M. tuberculosis. Among these, only seven patients showed sputum positivity for culture.[3]
In another study by Bachh et al. (n = 75), all sputum smear-negative cases of suspected tuberculosis were studied. Out of these, 20 sputum specimens were positive on culture. On evaluating BW specimens of these patients, 39 were culture positive and 21 out of these were smear positive.[14]
Hence, BAL/BW analysis by smear examination along with culture is useful in diagnosing tuberculosis in sputum-negative cases.
Diagnosis of other pulmonary infections
In this study, the reports of fluid specimens sent for microbiologic analysis were studied. On reviewing the reports of 120 fluid specimens, 42 showed positive results. It was seen that the most common organism isolated was Klebsiella in 15 patients (35.7%), followed by Pseudomonas in 9 (21.42%). Staphylococcus and Streptococci were isolated from six patients each (14.3%). Enterobacter was isolated from five (11.9%) patients, E. coli from four (9.5%) patients and Citrobacter freundii from 1 (2.4%) patient. More than one organism was isolated from four patients. In two cases, Gram's stain showed fungal hyphae but was not isolated in culture. No opportunistic infections by Cytomegalovirus and Pneumocystis jirovecii were reported in our study.
In a study by Radha et al. where 91 BAL specimens were analyzed, 38 were positive for bacterial infections. The organisms isolated were Klebsiella in 15 (39%) patients, Pseudomonas in 9 (23%), Acinetobacter in 5 (13%), Escherichia coli in 6 (15%), and methicillin-resistant Staphylococcus aureus in 3 (7.8%) patients. Seven cases of fungal infection were diagnosed, out of which 4 were caused by Candida, 2 were aspergillosis, and 1 was caused by Pneumocystis.[15]
Kshatriya et al. (n = 100) studied various cytology specimens obtained by FOB (BW/BAL/Brushings) and found that 24 out of 100 patients had infective etiology. Among these, in 20 cases (83%), bacteria were isolated, with Pseudomonas being the most common organism and four (17%) were fungal organisms.[16]
A study by Gaude et al. evaluated the diagnostic utility of 175 BAL specimens and found that in 32 cases, various organisms were isolated. S. pneumoniae in 10 (31.25%) cases, Haemophilus influenzae in 8 (25%), S. aureus in 6 (18.75%), and Klebsiella in 4 (12.5%). In four (12.5%) cases, hyphae of Aspergillus species were isolated.[3]
Hence, we can infer that BW/BAL can be very useful in identifying various microorganisms causing LRTI in all age groups of patients and different clinical settings.
Conclusions From this study, we can conclude that BW and BAL cytology is very useful in diagnosing various pulmonary lesions in all age groups. These cytology specimens can yield excellent diagnoses when correlated with good clinical history and imaging studies. Various infectious agents can also be easily identified by subjecting the respiratory cytology specimens to microbiologic studies. Both centrally and peripherally located tumors can be diagnosed and subtyped accurately on cytology specimens with proper sampling techniques. Awareness of the limitations and potential pitfalls is essential in avoiding false-negative or false-positive diagnoses. Further, combining respiratory cytology with bronchoscopic biopsies or ancillary techniques such as immunocytochemistry aid in accurate diagnosis and subtyping of neoplastic pulmonary lesions, which can lead to individualized treatment modalities.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
References 
Source of Support: None, Conflict of Interest: None
CheckDOI: 10.4103/joc.joc_138_22
留言 (0)