Cytologic Findings and Ancillary Tests Results of Sclerosing Pneumocytoma: Our Institutional Experience

ABSTRACT

Introduction

Sclerosing pneumocytoma (SP) is a rare benign tumor and a potential diagnostic pitfall. Our aim was to review the cytologic features of our surgically diagnosed SP cases including the clinical, immunohistochemical and available molecular findings.

Materials and Methods

A computerized search from 2013 to 2020 for surgical cases with corresponding cytology specimens diagnosed as SP was performed. The clinical data, cytology, and surgical specimens were collated for analysis.

Results

Six cytology specimens were collected. All were female (mean age = 35). Three have incidental lung nodules and three with cough. Cytologic findings showed variable architectural pattern (papillary, solid, singly scattered, acinar/rosette‐like) and cellular heterogeneity (surface, stromal, epithelioid, plasmacytoid cells). Atypia was identified in 4/6 cases. The original cytology diagnoses were negative = 1, SP = 2 and adenocarcinoma = 3. The latter diagnoses were amended to SP after review of the surgical specimens. The three false positive cases on review have cytologic features mimicking adenocarcinoma. Immunohistochemical stains showed tumor cells (surface and stromal) were positive for TTF‐1, and EMA with only the surface cells positive for pancytokeratin and Napsin A. Though two cases sent for molecular testing were negative for AKT1 or CTNNB1 exon 3 mutation, our panel did not evaluate AKT1 exon 4.

Conclusions

SP is a diagnostic pitfall with 50% initially misdiagnosed as adenocarcinoma. Integrating the clinical/radiologic findings, cytologic features, and performance of immunohistochemistry on cell block are helpful in avoiding misdiagnosis. Molecular testing for recurrent mutations, if present, could be helpful for diagnosis and possible therapy options. However, routinely used molecular testing may not always capture relevant molecular markers for SP.

1. Introduction

Sclerosing pneumocytomas (SP) is rare benign tumors of the lung that can histologically and cytomorphologically resemble adenocarcinomas. With the extensive utility of noninvasive procedures, such as endobronchial ultrasound‐guided fine needle aspiration for a radiologically identified lung mass or nodule, our task is to render an accurate diagnosis. Given their rarity, differentiating these tumors from adenocarcinomas pre‐operatively is important for management purposes.

Currently, these tumors are classified under the category of adenomas based on the recent fifth edition of the World Health Organization Classification of Thoracic tumors [1]. SP is defined as tumor of pneumocytic origin composed of dual population of surface cells resembling type 2 pneumocytes and rounds cells [1]. The tumor can show variable architectural patterns ranging from solid, papillary, sclerotic, and hemorrhagic. Occasionally, they can be misdiagnosed malignant tumors on histology [2, 3]. There are very few publications on the cytologic findings of SP [4, 5, 6, 7]. Due to its rarity, the heterogenous cell population, variability of cytologic features, and range of architectural pattern which is best appreciated on histologic sections and may not always be present on the cell block, it can be a diagnostic pitfall in cytology.

Herein, we describe our institutional experience on the clinical, radiologic, cytologic, immunohistochemical, and available molecular findings of sclerosing pneumocytoma as well as a brief review of literature.

2. Materials and Methods

This study was performed after approval from the Cleveland Clinic's institutional review board was obtained. One case was previously published as a case report [7].

A computerized search using Pathtools [8] was performed for a period of eight years (2013–2020). The search criteria was limited to surgical specimens diagnosed as SP (biopsy or resection) with concurrent or prior fine needle aspiration (FNA). The clinical data (age, sex, clinical presentation, smoking history, clinical follow‐up) and radiologic findings were retrieved from the electronic medical records.

The available Papanicolaou and Diff‐Quik stained cytology slides (ThinPrep slides and smears) as well as corresponding hematoxylin and eosin‐stained cell blocks were re‐evaluated. The tumor cellularity was assessed as follows: low (few lesional cells in one slide), moderate (lesional cells in more than one slide), high (lesional cells present on all slides). The cytologic features were evaluated for the presence of two cell types: surface cells (polygonal to cuboidal cells with moderate amounts of vacuolated cytoplasm), stromal cells (round, plasmacytoid, to spindle cells with more dense cytoplasm), tumor cell architecture (solid sheets, papillary, single cells, clusters/acini and rosette‐like), nuclear features (atypia, nuclear membrane irregularity/grooves, nucleoli, intranuclear inclusions), mitotic figures, presence or absence of stroma, necrosis, and bloody background (fresh blood, hemosiderin‐laden macrophages, hemosiderin pigment). The presence and type of inflammatory cells in the background were also noted. Immunohistochemical stains performed on the cell block were also reviewed. The molecular test results which were performed on two cases initially diagnosed as adenocarcinoma were also searched for recurrent mutations recently described in sclerosing pneumocytoma such as AKT1 internal tandem duplications and point mutations [9].

As part of a routine clinical work‐up for lung adenocarcinoma, molecular DNA hotspot testing was performed on two of the cases. Following extraction of tumor DNA from microdissected formalin fixed paraffin embedded tissue (FFPET), liquid based cytology, cell block (CB) specimens and library construction utilizing the custom Cancer Hotspot Panel v.1 (LifeTechnologies, Grand Island, NY). DNA sequencing of gene mutation hotspot regions was performed on the MiSeq instrument (Illumina, San Diego, CA). A combination of NextGENe software (Softgenetics, State College, PA) and customized bioinformatics pipeline was used to analyze FASTQ files to identify hotspot mutations in 50 genes.

3. Results

A total of six cases were identified (four in‐house, one referral consultation and one previously published case report). All patients were female with an age range of 22–58 years (mean age = 35 years.). Four were non‐smokers and one had a 34‐pack‐year smoking history. No smoking history was provided in the consult case. Three patients presented with an incidental well‐delineated lung nodule and three patients complained of cough. The nodules were located in the right upper lobe (2), left lower lobe (2), right middle lobe (1) and right hilar lymph node (1). The tumor size ranged from 1.1–3.0 cm (mean = 2.0). Clinical follow‐up was available in five patients. Two patients had no recurrent lesion at 24 and one patient at 48 months. One patient showed a decrease in size of the nodule at 60 months. One patient was diagnosed with lung adenocarcinoma on her right upper lobe one year after the initial diagnosis of SP on a wedge resection on the same lobe with negative margins. She was treated with lobectomy, chemotherapy, and radiation. Per the most recent follow‐up, the patient clinically remains without evidence of disease (25 months from initial diagnosis of SP and 13 months from initial diagnosis of adenocarcinoma). The patient with lymph node involvement by SP was a consult case, so no clinical follow‐up is available.

All specimens were obtained through endobronchial ultrasound‐guided fine needle aspiration by interventional pulmonologists. Rapid on‐site evaluation (ROSE) was performed in three cases and the following preliminary diagnoses were as follows: non‐small cell carcinoma (1), neoplastic cells present (1) and atypical cells (1). The cellularity of the six cases ranged from low (1/6, 17%), moderate (2/6, 33%) to high (3/6, 50%). Cytologic findings showed the presence of surface cells (6/6, 100%) and stromal cells (6/6, 100%). The surface cells were composed of cuboidal epithelium admixed with occasional histiocytoid cells. The stromal cells were either round, epithelioid, plasmacytoid, spindle‐shaped to histiocytoid. The architectural pattern showed the following results: papillary (5/6, 83%) (Figure 1), solid (5/6, 83%) (Figure 2), and single cells (5/6, 83%) (Figure 3). A combination of all three architectural pattern (papillary, solid, and single cells) was present in 5/6 (83%) of cases. The single cells were plasmacytoid and/or epithelioid to histiocytoid in appearance (Figure 3). Rarely, clusters/acini and rosette‐like architecture were present (1/6, 17%) (Figure 4). The nuclear features evaluated showed the following findings: nuclear atypia of varying degrees (4/6, 67%) (Figure 5), nuclear irregularity /grooves (2/6, 33%), small, minute nucleoli (3/6, 50%), and intranuclear pseudoinclusions (2/6, 33%) (Figure 3). No mitotic figures or necrosis were identified. A bloody background was identified in 5/6 (83%) whereas hyalinized stroma was noted in 4/6 (67%) of cases. Macrophages were the predominant inflammatory cells in 5/6 (83%) of cases. The original cytology diagnoses were as follows: Negative for malignancy. Bronchial epithelial cells present (1), SP (2), and adenocarcinoma (3). The three cases initially diagnosed as adenocarcinoma were amended after surgical resection of the specimen showed SP.

FIGURE 1.

Tumor cells arranged in papillary architecture with some histiocytoid cells admixed with the surface cells (Diff‐quik, ×60). [Color figure can be viewed at wileyonlinelibrary.com]

FIGURE 2.

Tumor cells arranged in solid sheets and admixed with some histiocytoid cells (Diff‐quik, ×60). [Color figure can be viewed at wileyonlinelibrary.com]

FIGURE 3.

Singly scattered plasmacytoid tumor cells with rare intranuclear pseudoinclusion (Diff‐quik, ×60). [Color figure can be viewed at wileyonlinelibrary.com]

FIGURE 4.

Tumor cells arranged in glandular/acinar/rosette‐like pattern (Diff‐quik, ×60). [Color figure can be viewed at wileyonlinelibrary.com]

FIGURE 5.

Tumor cells showing atypia (nuclear enlargement, irregular nuclear membrane, occasional nuclear grooves, and minute to inconspicuous nucleoli) (Papanicolaou, ×60). [Color figure can be viewed at wileyonlinelibrary.com]

Five cases have available cell blocks which showed papillary architecture, recapitulating the architecture on cytology (Figure 6A,B). Bloody background (4/5, 80%), hyalinized stroma (2/5, 40%), and solid architecture (1/5, 20%) were also present.

FIGURE 6.

Sclerosing pneumocytoma showing papillary architecture with surface/cuboidal epithelial cells (arrow) and underlying stromal/round cells (A, thin prep, Papanicolaou, ×40). Corresponding cell block (B, H&E ×40). [Color figure can be viewed at wileyonlinelibrary.com]

Immunohistochemical (IHC) stains were performed either on the cell block or surgical specimens only or on both cell block and surgical specimens. Immunohistochemical stains showed both surface and stromal cells positive for TTF‐1 and EMA whereas pancytokeratin AE1/AE3 and CK7 were only positive only on the surface cells (Figure 7). Napsin A was focally positive on the surface cells. CK20, chromogranin, synaptophysin, CD56, SMA and CAMTA‐1 were all negative.

FIGURE 7.

Tumor cells arranged in papillary architecture (cell block, H&E), with pancytokeratin showing positive staining only of the surface cells/cuboidal cells whereas TTF‐1 and EMA stain both the surface and stromal cells. [Color figure can be viewed at wileyonlinelibrary.com]

Molecular tests performed on the two cases initially diagnosed as adenocarcinoma were negative for BRAF, EGFR, HER2, KRAS, and MET mutation. Upon manual review of the sequencing data, no AKT1 or CTNNB1 exon 3 hotspot alterations were detected.

4. Discussion

SPis a rare benign lung tumor first described as Sclerosing hemangioma (histiocytoma, xanthoma) of the lung by Liewbow and Hubbel in 1956 because it had similar features to the skin lesion called “sclerosing hemangioma”, “histiocytoma” or “xanthoma” [10].

SPhas a female predominance. Though it affects a wide age range, it is more common in middle age group. Most patients are asymptomatic, and tumors are discovered incidentally on imaging studies [7]. However, when symptomatic they can present with cough and hemoptysis. Our patients had similar case presentations. Only one patient had a prior history of smoking (34‐pack‐year).

Radiographically, they are typically described as a solitary, peripherally located, circumscribed mass which can occasionally be cystic or calcified. However, bilateral, and multiple masses have also been reported [11, 12]. On contrast computerized tomography, they show enhancement and are FDG‐avid, raising the concern for malignancy. Tumor size can be variable with an average size of 2.0 cm; findings which corroborated with our results and prior studies. Four of our cases showed peripherally located solid nodules with well‐defined borders and no preference for specific lobe or laterality. One case was initially described on imaging as right upper lobe mass. However, during bronchoscopy it was an endobronchial lesion, which is a rare location for SP [13]. While lymph node metastasis is rare, it does occur as seen in one of our cases [14]. Hence, an endobronchial location and presence of lymph node metastasis does not negate the possibility of sclerosing pneumocytoma in the right clinical setting in collaboration with the cytologic and immunohistochemical findings.

One of our cases had both SP and adenocarcinoma on the same lobe approximately a year apart. While a similar occurrence is uncommon, the coexistence of pulmonary sclerosing hemangioma and primary adenocarcinoma in the same nodule has been reported in a study by Liu et al. [15].

Our cytologic findings of SP showed moderate to high tumor cellularity except for one which was from an endobronchial lesion diagnosed as negative for malignancy with bronchial epithelial cells present. On slide review, there were few small vaguely papillary clusters showing both surface with underlying stromal cells and scant stroma. This finding is comparable to the paper by Kaw and Nayak, wherein the surface cells were initially interpreted as bronchial cells with reactive changes [16].

Like other studies, the dual population of surface and stromal cells was present in all our cases with papillary, solid sheets and single cells as the most common architectural patterns [4]. Of note, while most of the surface cells in our study were composed of cuboidal cells with scant to moderate dense cytoplasm, some of them appeared histiocytoid based on their vacuolated cytoplasm (Figure 1). Likewise, singly dispersed histiocytoid cells were present. While at first glance they appear to resemble macrophages, their intimate admixture with both the surface and round cells in the solid sheets as well as similarity in nuclear features ascertain the tumoral nature of these vacuolated cells. The vacuolated cytoplasm is not uncommon and has been previously described on largest series on the histologic and immunohistochemical features of SP by Devouassoux‐Shisheboran et al. [17]. The variable morphology of the stromal cells (epithelioid, plasmacytoid, and spindle‐shaped) was also noted in our cases; which could briefly raise the differential diagnosis of carcinoma, neuroendocrine tumor, and sarcomatoid neoplasm. As expected, a bloody background and macrophages were present in 83% of our cases. Only 67% of our cases showed the presence of a hyalinized stroma compared to 100% in the study by Maleki et al. [4]. The degree of nuclear atypia promoted misdiagnosis of adenocarcinoma in 50% of our cases one case with atypia was appropriately called SP only because there was a corresponding biopsy, which was extremely helpful. A misdiagnosis of malignancy in cases of SP is common and our results are similar to several case reports and a case series [4, 6, 7]. The presence of minute nucleoli was noted in 3/6 of our cases best appreciated on Papanicolaou stained slides. Nuclear grooves were identified in 2/6 cases similar to the findings in the series by Maleki et al. [4]. However in contrast to their studies where they found rare intranuclear inclusions in all their cases, we only saw them in two of our cases.

Due to the tumor heterogeneity of SP, the importance of having a cell block cannot be overemphasized especially for evaluating architectural pattern and for interpretation of immunohistochemical stains. In SP, the cores of the papillae are composed of round/stromal cells and not true fibrovascular cores [18] whereas papillary lung adenocarcinoma shows neoplastic glandular cells growing along the surface of fibrovascular cores [19]. When papillary architecture is present, the surface cells of SP are positive for pancytokeratin, CAM5.2 CK7, EMA, TTF‐1, Napsin A and negative for PAX8 and neuroendocrine markers (Table 1). Whereas the stromal cells are positive for EMA, TTF‐1 but negative pancytokeratin, CAM5.2 CK7, Napsin, PAX8 and neuroendocrine markers. In cases when papillary architecture is not identified, a useful clue on immunohistochemistry is the observation that TTF1‐positive cells seem to prominently outnumber keratin positive cells [18]. Due to the rarity of SP, other tumors come into the differential diagnosis which includes: (a) well‐differentiated papillary lung adenocarcinoma due to the presence of nuclear atypia, small nucleoli and papillary architecture, (b) papillary thyroid carcinoma which shows nuclear grooves, small nucleoli and intranuclear inclusions and (c) neuroendocrine tumor based on the presence of monotonous population of singly scattered bland appearing epithelioid and plasmacytoid cells. Immunohistochemical stains useful in the aforementioned differential diagnosis of SP are listed in Table 1.

TABLE 1.

Immunohistochemical stains useful in the differential diagnosis of sclerosing pneumocytoma.

	Pan cytokeratin CAM5.2 CK7	Epithelial membrane antigen	TTF‐1	Napsin A	PAX‐8	Neuroendocrine markers
Sclerosing pneumocytoma
Surface/cuboidal	Positive	Positive	Positive	Positive	Negative	Negative
Stromal/round cells	Negative	Positive	Positive	Negative	Negative	Negative
Lung adenocarcinoma	Positive	Positive	Positive	Positive	Negative	Negative
Carcinoid	Positive	Often Negative	Often Negative	Negative	Negative	Positive
Metastatic papillary thyroid carcinoma	Positive	Variable Positive/Negative	Positive	Negative	Positive	Negative

Molecular studies performed on our two cases which were initially diagnosed as adenocarcinoma did not have BRAF, EGFR, ERBB2 (HER2), KRAS or MET mutation. Diagnostic molecular work‐up of lung adenocarcinoma at our institution is performed mainly on limited cytology specimens, usually for non‐small cell lung carcinoma (NSCLC) and pan‐cancer actionable markers. A recent study by Boland et al showed recurrent genetic abnormalities in the AKT1 gene a serine–threonine kinase in the mTOR pathway, which is an oncogene with a critical role in cell differentiation, proliferation, and survival [18]. Their study showed AKT1 mutation was present in ~80% of their SP. Based on this finding, we retrospectively reviewed the molecular tests performed in our two cases. Surprisingly, neither were positive for an exon 3 hotspot alteration. However, this negative result may be due to the limited clinical testing which was performed focusing only on AKT1 exon 3. Yeh et al. showed that approximately 50% of the AKT1 mutations were internal tandem duplications involving exon 4, which is not evaluated by our clinical hotspot panel and may account for the negative results [9]. Our clinical panel does assess CTNNB1 exon 3 mutations, which were also observed in a small proportion of cases in the study by Yeh et al.; however, no CTNNB1 alterations were observed. Hotspot testing for actionable NSCLC markers (i.e., BRAF, EGFR, ERBB2, KRAS, MET) can help rule in a diagnosis of adenocarcinoma and rule out a diagnosis of SP if positive; however, this testing is less definitive if negative. Most molecular tests are not routinely evaluating AKT1 in a comprehensive manner, due to the scant available diagnostic tissue and assay limitations. Hence, pathologists and clinicians should be cognizant of potential molecular testing pitfalls when evaluating for SP.

Clinical follow‐up or limited resection will suffice for SP. With the era of precision medicine, one patient in the study by Boland et al. with AKT1 mutation was treated with everolimus (mTOR pathway inhibitor) [20]. However, their patient showed no objective radiographic evidence of treatment response after 4 months of treatment. Nevertheless, in the future, the development of other effective mTOR pathway inhibitors may prove to be beneficial in patients with AKT1 mutation.

Overall, the difficulty with diagnosing SP on cytology aside from its rarity is the random cell sampling which occurs during the FNA procedure resulting in heterogenous cell population and the orientation may not always show the typical architecture seen on histology [6]. Furthermore, the bland appearing surface cells may be interpreted as background pneumocytes or bronchial epithelial cells. On the other hand, papillary fragments lined by the atypical surface cells could be interpreted as adenocarcinoma.

In conclusion, SP is a diagnostic pitfall on cytology with 50% of our cases initially misdiagnosed as adenocarcinoma. Familiarity with the cytologic features, integration of clinical/radiologic findings and procurement of additional material for cell block which not only helps with the architectural pattern but also with the ancillary tests will be key in avoiding misdiagnosis. Molecular testing for recurrent mutations, if present, could be helpful for diagnosis and possible therapy options.

Author Contributions

Maria Luisa C. Policarpio‐Nicolas: conceptualization, slide review, project supervision, formal analysis, writing of the original draft, and final reviewing and editing of the manuscript. Sydnee Webb: data collection and writing – review and editing of the final manuscript. Elizabeth M. Azzato: data analysis, writing and editing of the original draft. Rema Rao Chaari: data collection, final reviewing and editing of the manuscript. Erika Hissong: data collection, final reviewing and editing of the manuscript. Jennifer A. Brainard: slide review, formal analysis, final reviewing and editing of the manuscript.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.