Study of computed tomography-guided fine needle aspiration cytology of thoracic lesions.

BACKGROUND: Thoracic lesions include a variety of benign and malignant lesions of lung, pleura, chest wall and mediastinum. Transthoracic fine needle aspiration cytology (TFNAC) is a well established technique for work up of thoracic lesions. Computed tomography (CT) has extended the use of FNAC, because it is accurate for localization, needle puncture and above all it permits evaluation of lesions less than 1 cm. This diagnostic modality has a high sensitivity, specificity and is of relatively low cost. AIMS: To assess the role of CT-guided FNAC in the diagnosis of thoracic lesions.
MATERIALS AND METHODS: Eighty three patients with various thoracic lesions were evaluated by CT guided FNAC. The cytologic findings were compared with cell blocks whenever available.
RESULTS: Conclusive opinion was offered on cytology smears in 80 patients. Lesions of the lung were the most common. Neoplastic lesions in our study accounted for 65% of cases. The sensitivity and specificity of the study were 93.33% and 100%, respectively.
CONCLUSION: CT-guided TFNAC is a low cost, safe, minimally invasive and accurate diagnostic procedure with high sensitivity and specificity and when interpreted in conjunction with clinical and radiological data can prevent some of the pitfalls in diagnosis.

Introduction

Thoracic lesions include large variety of benign and malignant lesions of the lung, pleura, mediastinum and vertebrae. Transthoracic fine needle aspiration cytology (TFNAC) is a well established technique for work up of thoracic lesions compared to other imaging modalities available. Computed tomography (CT) is best for more accurate localization and lesions less than 1 cm can be aspirated.[12] In patients with lung cancer which is inoperable owing to local factors or the patient's general condition, FNAC confirms the diagnosis and reveals the tumor type.[3] Hence, TFNAC of small pulmonary lesions helps in earlier diagnosis, improved staging, increased chance of effective intervention and to formulate immediate effective management of thoracic mass lesions.[4] Major surgical procedure like thoracotomy can also be avoided. The aim of the present study was to assess the role of CT-guided FNAC in diagnosis of thoracic lesions. The high sensitivity and specificity of the widely accepted diagnostic method and relatively low cost with minimal complications prompted us to carry out this study.

Materials and Methods

Our study was undertaken at a tertiary care hospital in South India, from September 2007 to August 2009.

The study included 83 patients with various thoracic lesions who underwent CT guided FNAC.

Procedure for CT guided FNAC

The clinical details of all the patients were recorded and after ruling out bleeding diathesis, an informed consent was taken. A non contrast CT scan image of the lesions was obtained and the shortest distance from the lung was chosen for FNAC so that least amount of normal lung parenchyma was traversed.

A 22-gauge lumbar puncture (LP) needle was used for aspiration. The needle with the stylet was inserted into the lesion, the position of the tip of the needle within the lesion was confirmed by CT. Three to five rapid passes were made within the lesion, following which the stylet was removed and a 10 ml syringe was attached to the needle and aspirated. The aspirated material was spread on 3-5 slides and the adequacy of material was confirmed by rapid staining with toluidine blue. In possible cases, the material in the needle hub was used to prepare cell blocks for which the tissues were rapidly fixed in 60% ethyl alcohol and further processed as routine histopathological specimens.

The smears were routinely stained with hematoxylin and eosin, Papanicolaou and May-Grünwald-Giemsa stains. Various special stains like Gram′s, ziehl neelsen stain, Gomori methanamine silver stain and periodic acid Schiff stain were done, whenever needed. The slides were studied and analyzed in terms of radiological, clinical and cytological data to reach a final diagnosis. Smears were correlated with findings of the cell block, wherever available.

Results

The present study comprised 83 patients who underwent CT-guided FNAC for thoracic lesions in the radiology department, at a tertiary care hospital in South India during the study period from September 2007 to August 2009. The age of the patients in the study ranged from 11 years to 87 years. The youngest patient was an 11 years boy who presented with an anterior mediastinal mass, which on cytology was diagnosed as non Hodgkin's lymphoma (NHL). There was only a single patient above the age of 80 years. Maximum number of patients was between the ages of 41-60 years (42.2%). There were 60 males and 23 females with a male to female ratio of 2.6:1 (M:F=2.6:1). Out of 83 cases, the material was sufficient to offer a diagnosis in 80 (96.3%) cases while it was insufficient in 3 (3.6%) cases. The most common biopsied site was lung (72 cases). The lesions were classified as neoplastic (65%) and non neoplastic (35%) and were studied in relation to the site of lesion as lesions of lung, lesions of pleura, lesions of mediastinum and vertebral lesions.

Non-neoplastic lesions

Of the 28 non neoplastic lesions, 12 were granulomatous lesions. The granulomatous lesions showed predominantly epithelioid histiocytes in cohesive clusters and lymphocytes. In addition to these findings, the cases of tuberculosis showed caseous necrosis and a few cases showed positivity for acid fast bacilli (AFB) stain. The case of aspergillosis showed presence of branching septate hyphae on cell block study. There were 13 cases of acute inflammatory lesions and 3 cases of chronic non-specific inflammatory lesions [Table 1]. The cases of acute inflammatory lesions showed sheets of neutrophils, necrotic debris, fibrin and macrophages. The diagnosis in inflammatory lesions were improved by Gram's stain, culture and sensitivity, AFB stains and other special stains.

Table 1

Distribution of non-neoplastic lesions

Neoplastic lesions

Lung lesions

Of the 72 lung lesions, 47 (65.2%) were neoplastic, 23 (32%) were non-neoplastic (inflammatory) and 2 (2.8%) were inconclusive. The most common lesion was 15 cases of squamous cell carcinoma followed by 13 cases of adenocarconima. There were 8 cases of small cell carcinoma and 5 cases of bronchiloalveolar carcinoma. The remaining 4 cases with large pleomorphic, undifferentiated cells were grouped as large cell carcinomas [Table 2].

Table 2

Distribution of neoplastic lesions

The cases of squamous cell carcinoma showed sheets and clusters of pleomorphic squamous cells with abundant cytoplasm and pleomorphic, hyperchromatic nuclei. Few cases showed keratinization. The cases of adenocarcinoma on the other hand showed tumor cells arranged in glandular pattern and the cells had round nuclei with prominent solitary nucleoli. The cases of bronchioloalveolar carcinoma (BAC) were moderately cellular and showed three dimensional cell clusters with mild pleomorphism and intranuclear inclusions. The highly cellular smears with smaller cells having scanty cytoplasm, salt and pepper chromatin and nuclear molding were grouped as small cell carcinomas.

Mediastinal lesions

All the 4 cases that presented as masses in the mediastinum were neoplastic and they included single case of thymoma, NHL, a case of metastatic adenocarcinoma and small cell carcinoma [Table 2]. The case of thymoma showed bland epithelial cells and mature lymphocytes comprising the dual population of cells which are characteristic of lymphoepithelial type of thymoma. The single case of NHL was highly cellular and showed monotonous population of lymphoid cells with speckled nuclear chromatin and small nucleoli.

Lesions of pleura

Of the two lesions in the pleura, one was the benign solitary fibrous tumor and the other was a non-specific inflammatory lesion. The case of benign solitary fibrous tumor was cellular and comprised of poorly cohesive spindle cells with bland nuclei.

Vertebral lesions

Of the four vertebral lesions, three cases were granulomatous lesions and in one case the material was inadequate for a conclusive opinion.

Cell block study

Out of 83 cases, 43 cases (54.3%) were followed by cell block study. The cases included 28 neoplastic lesions and 15 non neoplastic lesions. The sensitivity, specificity and diagnostic accuracy was calculated for only the 43 cases followed by cell block study. Correlating the diagnosis of FNAC and cell block (CB) study in cases of neoplastic lesions was 100% correlation, whereas in non neoplastic lesions two cases which were diagnosed as only inflammatory lesions could be further categorized as aspergilloma and tuberculous lesions. Thus, the sensitivity, specificity of the present study was 93.3% and 100%, respectively.

Discussion

The present study was carried out to categorize the various thoracic lesions by cytomorphology. In 80 cases (96.3%), FNAC yield was diagnostic which is comparable to other studies like Arslan et al. (88%).[5] The most common site biopsied (81.9%) was the lung which is similar to the study of Sonneberg et al.[6] The M:F ratio in the present study was 2.6:1 whereas the age range varied from 11-87 years.

Non neoplastic lesions

Similar to the findings of Fraser et al.,[1] cases of acute inflammatory lesions showed necrotic debris, fibrin, neutrophils and macrophages [Figure 1a]. Conces et al.[7] supports that TFNAC is useful in diagnosis of pulmonary infections. But, Covell et al.[8] states that the above findings are not specific unless the organisms are demonstrated and advocates culture for the same. Silverman et al.[9] claims that one can make a confident diagnosis of a granulomatous process in aspirated material and could identify AFB in 38% cases of granulomatous lesions. Scattered epithelioid cells with abundant necrosis and AFB positivity was observed in the granulomatous lesions in the present study. The specific diagnosis in inflammatory lesions can be improved by integrating with other diagnostic methods such as Gram's stain, culture and sensitivity, special stains and ancillary techniques such as cell block and immunocytochemistry. The single case of aspergilloma was identified in cell block sections that showed septate hyphae which was similar to the finding of Mobeireck.[10]

Figure 1

(a) Acute non specific inflammation: Smear showing numerous polymorphs and necrotic debris in the background. (MGG; ×200), (b) Thymoma - Lymphoepithelial type: Smears show sheets of scattered epithelial cells and lymphocytes (Pap, ×200), (c) Adenocarcinoma lung- tumor cells showing intracytoplasmic PAS positivity (PAS; ×400), (d) SCLC (Cell Block) - Section shows cells with scant cytoplasm, granular chromatin and nuclear molding (H and E; ×400)

Among the neoplastic lesions, the malignant lesions were predominant, accounting for 94.23% of the neoplastic lesions, which is similar to the study by Arslan et al.[5] (88.1%), while the benign lesions accounted for 5.77%. This supports the fact that the usefulness of TFNAC is limited by low yield for specific benign lesions. This has been supported by the findings of Weisbrod et al.[1112] who found it difficult to aspirate benign spindle cell lesions.

Thymoma is the most common primary mediastinal neoplasm in adults and they show dual population of cells consisting of bland epithelial cells and mature lymphocytes as seen in the present study[11] [Figure 1b].

Squamous cell carcinoma was the most common malignant lesion, which was followed by adenoarcinoma and small cell carcinoma. The cytological diagnosis of squamous cell carcinoma (SCC) requires abnormal squamous cells having enlarged, dense hyperchromatic, angular nucleus with irregular chromatin distribution and inconspicuous nucleoli.[9]

The findings of tumor cells arranged in acinar pattern with abundant cytoplasm, round nuclei, solitary nucleoli and intracytoplasmic PAS positivity favors, the findings of mucin secretion in adenocarcinoma as suggested by Secoggs et al.[13] [Figure 1c].

The cellular smears of small cell carcinoma showed scanty cytoplasm, salt and pepper chromatin, nuclear molding, and showed histological correlation in cases followed by cell block[9] [Figure 1d].

Three dimensional cell clusters with mild pleomorphism and intranuclear inclusions are diagnostic indicators of BAC.[9] Similar to the findings of Goulamos et al.[13] and Kim et al.,[14] when the differentiation in the smears of malignant lesions was not apparent, the lesions were grouped as non small cell carcinoma and the smears showing large, highly pleomorphic cells with abundant cytoplasm were grouped as large cell carcinomas.[1314]

The mediastinum is a common site for secondary neoplasms from all sites. Small cell carcinoma of the lung is the most frequent followed by non small cell tumors. Our study showed a case of metastatic small cell carcinoma and an adenocarcinoma. A single case of NHL in 11 years child showed cellular smears of lymphoid cells with speckled nuclear chromatin and small nucleoli. Westcott et al.[15] states that TFNAC accuracy depends on the type of lymphoma.

Cell block study was instrumental in offering a conclusive opinion in cases where the cytologic findings were inconclusive. The sensitivity, specificity and diagnostic accuracy of the present study is comparable to that of Khouri et al.[16] and Hamper et al.[17] Cell block sections helped us in reaching a more conclusive diagnosis, especially in malignant lesions.

CT guided TFNAC is a low cost, safe, minimally invasive and accurate diagnostic procedure and major surgical procedures like thoracotomy can be avoided. The sensitivity and specificity of TFNAC is high for malignant lesions and is the initial procedure of choice. The results of cytology when interpreted in conjunction with clinical and radiological data can prevent some of the pitfalls in diagnosis.