Can p40 (Polyclonal) Replace p63 (Clone 4A4) in the Cytologic Diagnosis of Pulmonary Non–Small Cell Carcinoma?

Abstract

Objectives:

Differentiating squamous cell carcinoma from adenocarcinoma (ACA) in cytology specimens can be challenging. Recent literature showed p40 had higher specificity than p63 for this purpose.

Methods:

We identified 190 cytology cases with p40 (polyclonal) and p63 (monoclonal clone 4A4) immunohistochemistry, including specimens from fine-needle aspirations (FNAs) and effusions.

Results:

ACAs of lung origin stained for p40 and p63 in 21% and 20% of cases, respectively, regardless of specimen site. Among lung FNAs of primary pulmonary ACAs (n = 42), 14% were positive for p40 and 24% were positive for p63. Of the 20 pulmonary ACAs in effusions, more cases showed p40 positivity (40%) compared with FNAs, whereas p63 were positive in 15%. Among metastatic ACAs from other sites (n = 14), more cases were positive for p40 than p63.

Conclusions:

Polyclonal p40 yields a level of false positivity in ACAs similar to p63, which is highest in effusions and is not limited to lung origin.

Materials and Methods

With the evolution of targeted therapies, accurate identification of squamous differentiation in cytology specimens has become critical. This issue is best exemplified in lung specimens, in which the distinction between squamous cell carcinoma (SCC) and adenocarcinoma (ACA) is crucial in guiding therapy. Cytology specimens are not uncommonly the only diagnostic material on which treatment is based. Often this biopsy specimen is from fine-needle aspiration (FNA), which has the advantage of being a minimally invasive tool that can sample sites not amenable to core biopsy or surgical excision. These samples are often of modest cellularity and must be used judiciously for special stains to conserve material for possible subsequent molecular studies. Other cytology specimens such as effusions are also common and sometimes the only material available for diagnostic testing in metastatic cancer. To conserve tissue for molecular studies, a minimalist approach for immunohistochemistry (IHC) panels performed on cytology specimens is of the utmost importance.^1,2

Antibodies against p63 have been the most common tool used to identify squamous differentiation in both cytology and histology specimens.³ p63 is a p53 homologue that is involved in stem cell commitment toward squamous differentiation. It is often used for lung specimens in a panel with markers of ACA differentiation, TTF-1 and napsin A. An N-terminally truncated isoform of p63 (p40) has recently been described.⁴ Antibodies to both p40 and p63 have consistently been shown to have high sensitivity for SCC in histology specimens.^5,6 The superiority of p40 IHC has been attributed to its better specificity for lung SCC vs ACA, with p63 showing a higher rate of false positivity (range, 18%-31%) compared with p40 (0%-3%).^3,5,6 Thus, recommendations have been made that p40 should replace p63 for routine use in differentiating SCC from ACA.^4,5 We have been using p63 (clone 4A4), and for the past 2 years, we added p40 (polyclonal) to a panel we implemented to distinguish SCC from ACA.

Few studies have evaluated the clinical performance of p40 IHC in cytology specimens. The existing studies are focused on image-guided FNA specimens of primary lung masses,^7-9 have variable sample sizes (ranging from 30-144), and use different antibody clones. These studies report high sensitivities for detecting lung SCC using both p63 and p40 on IHC-stained cell block sections similar to those seen in histologic studies.^7-9 They also report greater specificity using p40 than p63, with false-positive rates for p63 of up to 63% in adenocarcinomas compared with 0% to 6% for p40.^7-9 Other important and common cytology specimens that have not been well studied with regard to p40 staining include effusions and metastases as well as specimens from sites other than lung. Our study is the largest to date to evaluate p40 and p63 staining in cytology specimens and includes a variety of types of specimens from multiple primary sites.

Approval was obtained from the Institutional Review Board of the New York University (NYU) Langone Medical Center, New York. Cases were identified through a search of our pathology database (Powerpath; Sunquest, Tucson, AZ) using search terms p40, p63, and cytology from January 2014 through March 2016. Inclusion criteria included the following: (1) cytology specimen with cell block, (2) both p40 and p63 IHC performed on cell block sections, and (3) H&E-, p40-, and p63-stained cell block sections available for review. Glass slides were retrieved for each case from the archives of the NYU Langone Medical Center Department of Pathology.

Results

Cytology Specimen Collection

All FNAs were performed by interventional radiologists and/or pulmonologists by computed tomography or ultrasound guidance in the presence of a cytotechnologist for on-site assessment. The air-dried smears were stained with Diff-Quik for on-site adequacy assessment. Additional smears were fixed in 95% alcohol and stained with Papanicolaou stain in the cytopreparatory laboratory. Cell blocks were prepared from RPMI needle rinses of each pass and an additional one or two passes dedicated to cell block.

Cell Block Preparation

Effusion specimens were collected by clinicians and transported to the laboratory within 24 hours. All effusion specimens were collected fresh with no fixative added. Each effusion specimen was stained with Diff-Quik (one cytospin), Papanicolaou (one ThinPrep; Hologic, Marlborough, MA), and H&E (one cell block).

For FNAs, the cellular material was retrieved by centrifuging a 50-μL cell block tube containing the aspirate in RPMI solution at 3,000 revolutions per minute (rpm) for 8 minutes (Beckman Coulter Allegra 6 centrifuge; Beckman Coulter, Jersey City, NJ). The supernatant was discarded and the cell pellet was clotted using three drops of thrombin and four drops of plasma. A wood applicator was used to slide the clot onto a lens paper, after which the lens paper was folded and gently pressed with the clot inside to form a compact cell aggregate. It was then placed in a cassette and fixed in 10% buffered formalin for at least 6 hours. Then the clot was embedded in paraffin wax and processed in the histology laboratory. Sections were cut at a 4-μm thickness for H&E and immunohistochemical staining.

IHC

Discussion

Data Collection

We identified 190 cytology specimens with IHC results of both p40 and p63, including 31 lung SCCs, 80 lung ACAs, and five lung adenosquamous carcinomas ❚ Table 1❚. Concurrent, concordant histologic diagnoses were present for 72% and 73% of SCC and ACA cytologic diagnoses, respectively. Among cases of ACA diagnosed on cytology without concurrent tissue biopsy specimens, 31% had a molecular study that showed ALK, EGFR, or KRAS mutation, and 91% showed immunoreactivity for TTF-1 and napsin A ❚Table 2❚. All three ACAs negative for TTF-1 and/or napsin A were morphologically classic ACAs with gland/acinar formations.

Of the 31 cases of SCCs, 100% were positive ❚Table 3❚ and ❚Image 2❚ for both p40 and p63, with most cases showing strong intensity (score 3+). Sixty-one percent of these were primary lung FNA specimens; 16% were from lymph nodes, 16% were from effusions, and 7% were from other metastatic foci. In 97% of cases, the staining extent was diffuse (≥2+) for both markers with overall stain scores (intensity + extent) of 4 or higher. The staining pattern did not differ between primary site and metastasis, and 100% of effusion specimens showed equivalent positive staining. TTF-1 was negative in 100% of cases (data not shown). All adenosquamous carcinomas of the lung (n = 5) also showed intense and diffuse staining pattern (overall stain scores of ≥4) for both stains.

Of the 80 cases diagnosed as lung ACA, 17 (21%) showed positive staining (overall score ≥2) for p40, 16 (20%) showed positive staining (overall score ≥2) for p63, and five (6%) were positive for both markers (Table 2 and ❚Image 3❚). For p40-positive ACA cases, effusion specimens showed the highest rate of positive staining (47% vs 35% in lung masses and 18% in other FNAs; ❚Table 4❚). In contrast, for p63-positive ACA cases, the greatest number of positive cases was seen in lung masses (Table 4). Seventy-five (94%) cases showed positive staining for TTF-1, and 73 (91%) cases were positive for napsin A. Of the lung adenocarcinomas that were p40 or p63 positive, more than 50% showed strong and diffuse staining with overall stain scores of 4 or higher (Table 2).

Sensitivities and specificities for distinguishing pulmonary SCC from ACA were calculated first using a score of 2 or higher as positive and then using an overall score of 4 or higher as positive ❚Table 5❚. Using a score of 2 or higher as positive, the sensitivity and specificity were 100% and 79% for p40 and 100% and 80% for p63. Setting the cutoff to an overall score of 4 or higher improved the specificities of both p40 and p63 for differentiating SCC from ACA to a similar extent; the sensitivity and specificity were 100% and 89% for p40 and 100% and 90% for p63, respectively. With the change in criteria, the change in specificity was significant for both p40 (79% to 89%; McNemar test P = .000) and p63 (80% to 90%; McNemar test P = .000).

Molecular study results were available for 94% of the p40-positive lung ACA cases and 21% of the p63-positive cases (Table 2). Sixty-five percent of p40-positive cases and 19% of p63-positive cases were from patients with ALK, EGFR, or KRAS alterations. One case positive for both p40 and p63 had an associated specimen with a BRAF mutation; another case positive for both markers had an associated specimen with a nonsense p53 mutation, and one p63-positive case had a BRAF V600E mutation (data not shown).

Twenty-five cases of lung SCC (n = 5) or ACA (n = 20) in effusion specimens were identified. All five (100%) cases of SCC were positive for both p40 and p63. Of the 20 effusion ACA cases, eight were positive for p40 and three were positive for p63. p63 demonstrated a significantly higher specificity for SCC than p40 in effusion specimens ❚Table 6❚.

We identified nine cases of lung small cell carcinoma (five primary lung and four metastases), with eight cases negative for both markers and only one demonstrating focal positive staining for p40 and p63 (Table 1). The remaining lung cases included eight poorly differentiated carcinomas (all cases negative for both markers), as well as two mixed tumors with adenosquamous differentiation that showed positive staining for both markers.

There were 14 cases of nonpulmonary ACAs (Table 1). Positive p40 staining was observed for ACA from breast (n = 1), ovary (n = 2), pancreaticobiliary tract (n = 2), and endometrium (n = 1), and positive p63 staining was seen in one pancreatic, one ovarian, and one endometrial case. Head and neck SCC and urothelial cancers were positive for both p40 and p63 in all cases. A variety of other types of neoplasms were also examined, and the detailed results are reported in Table 1.

The importance of identifying squamous differentiation in cytology specimens is growing with the increasing development of new targeted therapies. Patients with specific genetic alterations occurring primarily in adenocarcinomas, such as EGFR mutations, are now eligible for small-molecule inhibitor therapies such as erlotinib and gefitinib.¹⁰ Conversely, it is important to identify squamous cell carcinomas so as to avoid treatment with vascular endothelial growth factor inhibitors, which have a known risk of hemorrhage in patients with SCC.¹¹ Since many patients with lung cancer are diagnosed at a high stage, cytology specimens may be the only diagnostic material available to guide treatment; thus, judicious use of the material for diagnosis is critical. The objective of this study was to evaluate polyclonal p40 and p63 clone 4A4 in a variety of cytology specimens from lung as well as other sites to help guide a minimalist approach to IHC.

Our results show that p40 and p63 both show intense and diffuse staining for lung SCC, as has been reported previously in histology and cytology specimens.^5,7,9 Whereas prior studies have reported small improvements in sensitivity by p40 compared with p63 (ranging from a 2%-3% increase) in cytology specimens, we show equivalent sensitivity (100%) for both markers.^7,8 Our findings hold true regardless of the site (primary or metastasis).

In this study, we have found equivalent specificities for p40 and p63 (79% and 80%, respectively) in lung cytology specimens for differentiating SCC vs ACA. This is in contrast to several smaller prior studies reporting that p40 has a higher specificity than p63. For example, one study of cell block sections from 46 FNAs of lung adenocarcinomas reported zero cases staining for p40 and two cases staining for p63 with overall specificities of 100% for p40 and 96% for p63.⁸ Similarly, another study looking at FNA specimens of 30 SCCs and 30 ACAs reported a 100% specificity for p40 vs 80% for p63.⁷ Differences in preanalytic and analytic variables, including antibody type, may contribute to the difference in results between our study and Vogt et al.⁷ While Collins et al⁸ used monoclonal antibodies to both markers, Vogt et al⁷ used a rabbit polyclonal p40 antibody and a monoclonal p63 antibody (4A4), similar to those used in our study.

References

While the three aforementioned studies were done on FNA specimens, our study also includes effusions, which is another cytology specimen type that is commonly used in cancer diagnosis. We identified five cases of lung SCC and 20 cases of lung ACA in effusions. All five cases of SCC were positive for both p40 and p63. In ACA effusion specimens, the frequency of false-positive staining was higher for p40 than for p63 (40% vs 15%). Our data suggest that p63 is more specific for SCC than p40 in effusion specimens. IHC performance statistics for FNAs cannot necessarily be applied to effusions, and IHC results should be interpreted with caution.

We also sought to evaluate p40 and p63 staining for neoplasms other than lung primaries in cytology specimens. In studies of surgical specimens, p40 has demonstrated superior specificity as well. For example, in a study that included 152 large cell lymphomas, p63 showed positive staining for 54% of cases compared with zero by p40.⁵ In our study, we observed that both p40 and p63 stain several other types of adenocarcinomas and that p40 gives more frequent false-positive results than p63. This emphasizes that false-positive staining is observed in both p40 and p63 and is not limited to lung adenocarcinomas.

Studies have shown superior specificity of p40 over p63 for the differential diagnosis of lung SCC vs small cell carcinoma (SmCC) in cytology specimens.¹² In a study of 50 SmCC and 25 SCC FNA specimens, specificities of p40 and p63 were 96% and 66%, respectively.^5,12 However, by using a higher threshold for positive, they were able to “rescue” p63’s specificity, increasing it from 66% to 94%. Using similar antibodies and a score of 2+ or higher as a cutoff for positive, we find that the sensitivity for both p40 and p63 is 100%, and the specificity for both p40 and p63 is 89%. Although limited by low sample size, we do not observe any advantage of polyclonal p40 for this purpose at our institution.

In summary, we report equivalent performance by p40 and p63 in FNAs of lung primaries and metastases for the differential diagnosis of SCC and ACA using a polyclonal rabbit antibody for p40 and a monoclonal antibody for p63. Both markers show a similar rate of false-positive staining in lung ACA, although p63 is a superior stain in effusion specimens. Both markers also stain adenocarcinomas from sites other than lung. It is important to recognize the shortcomings of each of these immunohistochemical markers and interpret results in the context of cytomorphology with the understanding that one is not superior to the other. In terms of maximizing cost savings and tissue preservation in cell blocks, we found that the diagnostic performance of polyclonal p40 IHC was inadequate for it to supplant p63 IHC in our laboratory.

Author notes