Adequacy of Cytologic Samples by Ultrasound-Guided Percutaneous Transthoracic Fine-Needle Aspiration Cytology of Peripheral Pulmonary Nodules for Morphologic Diagnosis and Molecular Evaluations: Comparison With Computed Tomography–Guided Percutaneous Transthoracic Fine-Needle Aspiration Cytology

This study was conducted in close collaboration with the pathology unit and the thoracic surgery unit of “Università della Campania Luigi Vanvitelli” University (Naples, Italy), from November 2016 to March 2018. The research was approved by the Institutional Review Board. The study included 40 patients with a newly detected pulmonary nodule who had been scheduled to undergo transthoracic FNAC. All patients included in the study presented with peripheral pulmonary nodules with at least 1 of the following features: subpleural location, peripheral location with extensions to the visceral pleura, or paravertebral location. The patients were thoroughly informed of risks, complications, and possible inadequacy of the rate of the FNAC procedures and gave their written informed consent to undergo the procedure. In addition, we reevaluated CT-guided transthoracic FNAC procedures carried out during the same period, selecting 40 consecutive patients who met the same inclusion criteria (subpleural location, peripheral location with extensions to the visceral pleura, and paravertebral location).

Ultrasonography-guided percutaneous transthoracic FNAC was performed by a cytopathologist with the collaboration of a thoracic surgeon. In our series, we conducted US with an intercostal approach, using a convex 3.5- to 5-MHz probe in cases of relatively deeper pulmonary nodules, or a linear 7- to 10-MHz probe in cases of superficial nodules. A thoracic surgeon performed transthoracic US evaluation using an Esaote 25 Lab Gold (Esaote, Genova, Italy). A 23-G, 150-mm needle connected to a syringe mounted on the holder was used in all cases. Thereafter, a radiologist, with the collaboration of a cytopathologist, performed a CT-guided percutaneous transthoracic FNAC using a 64-slice scanner (SOMATOM Sensation 64, Siemens AG, Erlangen, Germany). ROSE was always performed to provide indications on the number of passes and the choice of vials in both US-guided and CT-guided cases. A total of 1 to 4 passes were performed for each patient. A total of 1 to 2 smears were air-dried and DiffQuik–stained for ROSE; additional smears were alcohol fixed and Papanicolaou stained. Residual material in the hub of the needle from the first pass, as well as material from the additional passes, was suspended in formalin for cell block (CB) preparation. Based on the ROSE results, in 2 cases, an additional pass was performed, and the samples were suspended in phosphate-buffered saline for flow cytometry evaluation. The time required to perform the procedure and the number of passes for each patient were recorded. All of the patients were clinically monitored in the recovery room, and 2 hours after the procedure a chest radiography was performed to diagnose possible complications.

The cellularity of each sample was assessed by counting the neoplastic cells in both smears and CBs. For statistical purposes, we defined 4 classes of cellularity: acellular samples (lack of representative cells in the sample), slightly cellular samples (fewer than 100 representative cells in the sample), moderately cellular samples (between 100 and 500 representative cells in the sample), and highly cellular samples (more than 500 representative cells in the sample).¹²  Immunocytochemistry (ICC) was carried out on CB sections and included the following antibodies: TTF1 (thyroid transcription factor 1), p63, Napsin, p40, cytokeratin 7 (CK7), CK20, CDX2 (caudal type homeobox 2), CD3, CD20, MUM1 (multiple myeloma 1), CD10, bcl6 (B-cell lymphoma 6), bcl2 (B-cell lymphoma 2), CD5, CD30, Ki-67, ALK (anaplastic lymphoma kinase), CK, PD-L1 (programmed death ligand-1), STAT6 (signal transducer and activator of transcription 6), CD34, CD99, actin, S100, Mart-1, WT1 (Wilms tumor), calretinin, and estrogen receptor. ICC was performed on the Ventana platform (VENTANA BenchMark Ultra system; Ventana, Tucson, Arizona). The ICC markers used for diagnostic purposes to classify pulmonary non–small cell carcinomas were TTF1, Napsin, and CK7 for adenocarcinoma; and p63, p40, and CK 5/6 for squamous cell carcinoma. Generally, a basic ICC panel, including TTF1 and p40 in many cases, was performed, with an additional ICC evaluation only in selected cases. Flow cytometry was performed using BD FACSCanto II System and included the following fluorescein antibodies: CD3, CD5, CD19, CD20, CD22, CD23, CD56, and κ and λ light chains.

The smears were used for the extraction of DNA for next-generation sequencing, whereas CB sections were used for ICC (ALK, PD-L1) and fluorescence in situ hybridization. Next-generation sequencing was performed to analyze epidermal growth factor receptor (EGFR) and BRAF mutational status in 11 cases of lung adenocarcinoma and 2 cases of pulmonary melanoma metastases, respectively. Therascreen EGFR RGQ PCR Qiagen analysis (Quigen, Hilden, Germany) was conducted to evaluate EGFR mutational status in 3 additional cases, which could not be evaluated by next-generation sequencing because of the scanty amount of DNA. In all cases, smears were used for the extraction of DNA by Qiamp DNA Micro Kit (Qiagen), according to the manufacturer's instructions. Extracted DNA was eluted in 20 μL of elution buffer and subsequently quantified by a Qubit 2.0 Fluorometer (Life Technologies, Carlsbad, California) using the Qubit dsDNA HS Assay kit, according to the manufacturer's recommendations. The extracted DNA was stored at −20°C.

In 1 case, fluorescence in situ hybridization was performed on CB sections using the commercially available ALK (ZytoLight SPEC ALK Dual Color Break Apart Probe; ZytoVision, Bremerhaven, Germany) and ROS1 (ROS Proto-Oncogene 1; ZytoLight SPEC ROS1 Dual Color Break Apart Probe, ZytoVision). PD-L1 immunocytochemical test was performed on CB slides and was considered positive for a percentage of positive neoplastic cells 1% or more.

Pearson χ² test was used to identify a possible relationship between the frequency of adequacy for diagnosis, cellularity, occurrence of postoperative complications, and time of execution in the 2 series. Differences were considered to be statistically significant for values of P < .05. All of the statistical analyses were carried out using IBM SPSS statistics software (version 20, Armonk, New York).

The clinical and pathologic features of the series are summarized in Table 1. The results are summarized in Table 2.

Our US-guided transthoracic pulmonary FNAC series included 40 patients, 28 male (70%) and 12 female (30%). Their ages ranged from 47 to 84 years, with a mean age of 67.96 and a median age of 70. The diameter of the pulmonary nodules ranged from 21 to 100 mm, with a mean diameter of 46.5. In 25 of 40 cases (62.5%), the lesions were in the right lung, and in the remaining 15 cases (37.5%), they were in the left lung. The pulmonary nodules were peripheral, subpleural, and paravertebral in 21 (52.5%), 14 (35%), and 5 (12.5%) cases, respectively. The CT-guided FNAC series included 40 patients, 28 male (70%) and 12 female (30%), ranging in age from 38 to 83 years, with a mean age of 66.65. The diameter of the pulmonary nodules ranged from 12 to 62 mm, with a median diameter of 38. In 20 of 40 cases (50%), the lesions involved the right lung, and in the remaining 20 cases (50%) they involved the left lung. Of the 40 pulmonary nodules, 22 cases (55%) were peripheral, 17 (42.5%) were subpleural, and 1 (2.5%) paravertebral. No significant difference was observed between the 2 groups in terms of clinical features.

Of the 40 US-guided FNAC cases, a final diagnosis was made in 36 cases (90%), with 30 neoplastic diagnoses (75%) and 6 inflammatory diagnoses (15%). The diagnoses of neoplasia included 11 pulmonary adenocarcinomas, 9 pulmonary squamous cell carcinomas, 3 non–small cell lung carcinomas not otherwise specified (NSCLCs-NOS), 2 diffuse large B-cell lymphomas, 2 mesenchymal neoplasms (a solitary fibrous tumor and a leiomyosarcoma), and 3 metastases (2 intestinal adenocarcinomas and 1 melanoma). In 26 of 30 neoplastic cases (86.67%), ICC was required for diagnosis, whereas in the remaining 4 cases (13.33%) the diagnosis (1 pulmonary adenocarcinoma and 3 NSCLCs-NOS) was based exclusively on morphologic features because the CB had an acellular result. A thoracoscopic biopsy was performed on a patient with a cytologic diagnosis of diffuse large B-cell lymphoma, which was confirmed by the histologic diagnosis. Five patients underwent surgical resection, and the histologic examination confirmed the previous cytologic diagnosis (3 pulmonary adenocarcinomas, 2 squamous cell carcinoma, and 1 solitary fibrous tumor). Of 40 cases, 2 (5%) were not diagnostic, and another 2 (5%) were indeterminate for malignancies, being suspicious but not diagnostic for neoplastic disease. Following a nondiagnostic cytologic examination, 1 patient underwent a CT-guided FNAC, which led to a diagnosis of squamous cell carcinoma. The second patient with a nondiagnostic cytology and the 2 patients with indeterminate/suspicious cytology were lost to follow-up.

Of the 40 CT-guided FNAC cases, 31 cases (77.5%) were diagnosed with neoplastic lesions, 11 of which had a diagnosis of pulmonary adenocarcinoma, 11 pulmonary squamous cell carcinoma, 5 NSCLCs-NOS, 1 soft tissue neoplasm (pulmonary hamartoma), and 3 metastases (1 gross bowel adenocarcinoma, 1 breast carcinoma, and 1 melanoma). Of 40 cases, 6 (15%) were negative for neoplastic lesions and were referable to inflammation. ICC was performed in 24 of 31 neoplastic cases (77.42%). Thirteen patients underwent surgical excision, and in all cases the histology confirmed the diagnosis of malignancy. Specifically, the diagnosis of the histotype was confirmed in 8 cases (5 pulmonary squamous cell carcinomas and 3 pulmonary adenocarcinomas). In 4 patients with a cytologic diagnosis of NSCLC-NOS, the histologic diagnosis identified an adenocarcinoma in 3 cases and squamous cell carcinoma in 1 case. Of the 40 cases, 2 (5%) were suspected/indeterminate for malignancies, and 1 (2.5%) was not diagnostic. In 1 patient with a suspected/undetermined cytologic diagnosis of malignant tumor, the histology diagnosed a large cell neuroendocrine carcinoma. The remaining 2 patients with suspected/indeterminate cytology and no diagnostic cytology were lost to follow-up. Pearson exact test showed no significant difference in adequacy for diagnosis between the 2 series (P = .58).

Of the 30 US-guided transthoracic pulmonary FNAC neoplastic cases, there were 7 slightly cellular smears (23.3%), 9 moderately cellular smears (30%), and 14 highly cellular smears (46.7%). In addition, 4 CBs (13.3%) were acellular, 3 CBs (10%) slightly cellular, 10 CBs (33.3%) moderately cellular, and 13 CBs (43.3%) highly cellular.

Of the 30 CT-guided transthoracic pulmonary FNAC neoplastic cases, 5 smears (16.13%) were slightly cellular, 14 (45.16%) moderately cellular, and 12 (38.71%) highly cellular. CBs were acellular in 6 cases (19.35%), slightly cellular in 4 (12.9%), moderately cellular in 13 (41.94%), and highly cellular in 8 (25.81%). Pearson exact test showed no significant differences in cellularity between the 2 series (Table 2).

Pearson exact test showed no significant differences in adequacy for ICC, nor did the molecular test for predictive evaluations find differences between the 2 series (Table 2).

Ultrasonography-guided FNAC procedures lasted from 7 to 18 minutes, with a mean time of 13.1. Computed tomography–guided FNAC procedures lasted from 14 to 39 minutes, with a mean time of 23.6.

The statistical analyses were based on the median of the execution time in both series. Pearson exact test showed a significant difference in execution time between the 2 series (Table 2).

One to three passes were performed for each individual patient in the US-guided FNAC series, with a mean number of passes of 2.2. One to four passes were performed for each individual patient in the CT-guided FNAC series, with a mean number of 2.17.

The statistical analysis was based on the median of the number of passes in both series. Pearson exact test failed to show a significant difference in the number of passes between the 2 series (Table 2).

In the US-guided FNAC series, only 1 patient experienced a pneumothorax (1 of 40 cases; 2.5%), whereas a slight hemorrhagic suffusion occurred in 2 (5%). No case of intraparenchymal hemorrhage was observed. Conversely, of the 40 CT-guided FNAC cases, a pneumothorax occurred in 9 (22.5%), intraparenchymal hemorrhagic suffusion in 3 (7.5%), and intraparenchymal hemorrhage in 1 (2.5%).

Pearson exact test showed a significant difference in the occurrence of complications between the 2 series (Table 2).

In the framework of the Italian health system, CT-guided FNAC had a total cost of about $220, whereas the total cost of US-guided FNAC was about $125.

Computed tomography is currently considered the gold standard imaging technique for the assessment of pulmonary lesions, not only for diagnostic evaluation but also to guide transthoracic cytologic sampling. Nevertheless, CT-guided transthoracic FNAC presents some disadvantages, such as the use of ionizing radiation, the need to move the patient to the radiology service, the long execution time, and the high rate of pneumothorax reported in some series.⁷  In this setting, US-guided FNAC represents a promising alternative and could be taken into consideration in selected patients. Proper patient selection is of paramount importance to optimize the results because the ultrasound beam does not easily penetrate the aerated pulmonary parenchyma. Therefore, we performed US-guided FNAC only in patients presenting with peripheral pulmonary nodules without aerated pulmonary parenchyma between the lesion and the pleura. Based on the experience at our institution, about 20% of all peripheral pulmonary masses have these features and may be sampled by US-guided FNAC. The operator's thorough skill with the procedure is essential to obtain representative cytologic samples. The cytopathologists in our institution have more than 10 years' experience in US-guided FNAC of superficial lesions and are able to perform a US scan autonomously. Furthermore, they participate in the execution of cytologic sampling by CT-guided FNAC of pulmonary and abdominal masses, in close collaboration with the radiologists and the endoscopists. Together with the thoracic surgeon, the cytopathologists have developed specific experience in pulmonary US scan because they performed US in all of the patients referred to the thoracic surgery unit of our institution for about 2 months before beginning the US-guided FNAC. Thus, they evaluated the US features of pulmonary and pleural masses and the normal anatomic structures; furthermore, they compared each patient's US and CT features. Indeed, US evaluation and US-guided FNAC are operator-sensitive techniques, and therefore require the operator's thorough knowledge of the US features of the chest region, to be able to recognize the anatomic points of reference. Because ribs block the US beam, intercostal spaces represent the only acoustic window for the evaluation of pulmonary nodules. The preventive study of the patient's chest X-ray is essential for the choice of the correct acoustic window. The anatomic structures of the chest can be differentiated by their distinct echogenicity. A US scan recognizes the soft tissue of the chest wall as echogenic layers and the pleura as 1 or 2 thin hyperechogenic lines. The pulmonary parenchyma is generally hypoechogenic, but the peripheral parenchyma is usually characterized by the presence of horizontal echogenic lines, due to the change in acoustic impedance in the pleura-lung interface.¹³  The US features of malignant pulmonary nodules include hypoechoic appearance, presence of solid or colliquated areas, irregular borders, pleural infiltration, and failure of the nodule to move during breathing.^8,14  Ultrasonography features of pulmonary nodules and thoracic structures are shown in Figure 1. A US was performed with an intercostal approach using a standard 3.75-MHz curvilinear probe in cases of deeper pulmonary nodules, whereas a 7-MHz linear probe was occasionally preferred in selected cases with superficial nodules. A thoracic surgeon first conducted a transthoracic US evaluation, whereas the FNAC procedure was performed by a cytopathologist with the collaboration of the thoracic surgeon.

The position of the patient during the procedure depended on the anatomic location of the nodule. In cases of posterior lesions, the patient lies in a prone position; in cases of anterior lesions, the patient lies in a supine position with his or her hands under his or her head; in cases of dorsolateral lesions, the patient lies on the contralateral side with the corresponding arm raised above his or her head. During the US scan, all blood vessels were identified before performing the FNAC procedure to avoid injury. The needle was inserted relative to the center of the US probe. The trajectory of the needle was visible on the transducer line, to modulate the direction and depth of the needle insertion during the procedure (Figure 2). Some authors have suggested the use of a probe with a central hole to better follow the needle trajectory and identify the needle tip in the lesion.¹⁵  Because the diameter of the nodule may be overestimated by the US scan because of the presence of peripheral atelectasis, we generally prefer to insert the needle into the center of the nodule. Only in cases with evident central necrosis was the needle inserted into the periphery of the lesion. The high diagnostic accuracy of US-guided pulmonary FNAC in selected cases is at least partly due to the possibility of monitoring the trajectory of the needle in real time, allowing for the sampling of large lesions at different points and the evaluation of the possible heterogeneity of the neoplasia. The patient's collaboration is especially important in this setting. Indeed, in cases of small lesions located behind the ribs, the patient is asked to hold his or her breath in order to better localize the lesion with US and sample it accurately (Figure 3). Because the US-guided FNAC is faster than the CT-guided FNAC, the patient feels more comfortable during the procedure and is generally more collaborative.

This study demonstrates that lung US-guided FNAC, in association with ROSE, provides adequate and highly cellular cytologic samples, allowing for an accurate cytologic diagnosis in both benign and malignant pulmonary lesions. Because in recent years several predictive markers have been defined in pulmonary carcinoma, the FNAC of a pulmonary nodule must ensure not only the possibility of a diagnosis, but also a sufficient amount of biologic material for ancillary techniques. In this context, the cellularity of the sample is most relevant. Furthermore, because ancillary techniques require different vials, it is of considerable importance that there be not only an adequate total cellularity of the sample, but also an adequate cellularity in each individual vial. Our study demonstrates that transthoracic pulmonary US-guided FNAC provides samples similar to CT-guided FNAC in terms of cellularity, with no significant difference between the 2 series as to smear cellularity, CB cellularity, and number of passes. Overall, US-guided FNAC samples had adequate and representative results in 95% of cases, without any significant differences compared with the CT-guided series. Ultrasonography-guided FNAC proved to be a versatile technique: in our series, it allowed for the diagnosis of a wide spectrum of malignant lesions, including primary pulmonary carcinoma (squamous cell carcinoma and adenocarcinoma), pleural neoplasm (solitary fibrous tumor), secondary pulmonary lesions (metastasis from intestinal adenocarcinoma and melanoma), pulmonary mesenchymal neoplasm (leiomyosarcoma), and pulmonary localization of B-cell non-Hodgkin lymphoma (Figure 4). Obviously, ROSE requires an excellent management of the biologic sample and allows for the proper distribution into the different vials. Furthermore, US-guided FNAC is a safe procedure with a low complications rate.¹⁶  In our series, a pneumothorax occurred in only 1 case and minimal hemorrhagic suffusions in 2 cases. This study demonstrates that US-guided FNAC is a safer procedure than CT-guided FNAC in selected patients. Indeed, in the CT-guided FNAC series, 9 cases of pneumothorax, 3 cases of intraparenchymal hemorrhagic suffusion, and 1 case of intraparenchymal hemorrhage were recorded. The lower complication rate of US-guided FNAC may be due to, in accordance with Sconfienza et al,⁷  the lower number of punctures made possible by the real-time monitoring of the needle. A lower rate of pneumothorax was also reported by Trovato et al¹⁵  (<1% in a series of 453 transthoracic US-guided FNACs of subpleural nodules). Ultrasonography-guided FNAC requires less time than CT-guided FNAC; it is a radiation-free technique and is more feasible with bedridden or debilitated patients because there is no need to take them to the radiology service. Finally, US is a less expensive procedure than CT. If we consider the minimal expenditure reimbursed to hospitals within the framework of the Italian health system, CT-guided FNAC has a total cost of about $220, whereas for US-guided FNAC the amount is approximately $125. Therefore, for our hospital, for instance, the US-guided FNAC series had a minimum total cost of about $5000, compared with the CT-guided FNAC series, which had a minimum cost of about $8800.

In conclusion, transthoracic pulmonary US-guided FNAC is a radiation-free procedure with low risk of complications and high clinical applicability and diagnostic utility in selected patients. In our experience, US-guided FNAC and CT-guided FNAC are comparable in terms of quality of the cytologic sample, although the former proved to be safer, faster, and less expensive. Caution is essential in the selection of the lesions because deep-seated nodules represent the greatest limitation of this procedure.

We thank Mary Claire Barber and Rosanna Scala for linguistic revision of the manuscript.