The utility of B72.3, carcinoembryonic antigen, and Leu M-1 in cell blocks

An adjunct to fine-needle aspiration diagnosis of bronchioloalveolar carcinoma of the lung




The distinction of bronchioloalveolar carcinoma (BAC) from atypical adenomatous hyperplasia (AAH) or reactive alveolar cell hyperplasia (RAH) can be difficult on aspiration cytology, even when cell block preparations are available. The authors evaluated the usefulness of B72.3, carcinoembryonic antigen (CEA), and Leu M-1 immunostains in differentiating BAC, AAH, and RAH.


Immunostains for B72.3, CEA, and Leu M-1 were performed on cell block sections from 11 lung lesions that were diagnosed cytologically as BAC (6 lesions) and “atypical cells, cannot exclude BAC” (5 lesions). Ten histologic sections of AAH and 8 histologic sections of RAH also were stained.


Among the six lesions that had an unequivocal cytologic diagnosis of BAC, all sections were positive for two of three immunostains. Tissue follow-up confirmed BAC in all six lesions. Among the five lesions that were diagnosed as “atypical cells, cannot exclude BAC,” four lesions were positive for two of three immunostains, and one lesion was negative for all three immunostains. Subsequent tissue follow-up confirmed BAC in four of these lesions. Follow-up histology of the wedge resection on the lesion in the atypical category that was negative for B72.3, CEA, and Leu M-1 showed only AAH. All 10 lesions that had a histologic diagnosis of AAH and 8 lesions that had a histologic diagnosis RAH were negative for B72.3, CEA, and Leu M-1.


Positive staining for at least 2 immunostains among B72.3, CEA, and Leu M-1 provided strong supportive evidence for the diagnosis of BAC, and a negative result for all 3 immunostains was helpful in excluding BAC and in differentiating BAC from AAH and RAH. Cancer (Cancer Cytopathol) 2005. © 2005 American Cancer Society.

Primary lung carcinoma is the most common worldwide cause of cancer incidence and mortality.1 Among the major histologic variants of lung carcinoma, adenocarcinoma is emerging rapidly as a leading type in the United States.2 Adenocarcinoma is the most prevalent form of lung carcinoma in younger males (age < 50 years), in women of all ages, in nonsmokers, and in former smokers.3 Bronchioloalveolar carcinoma (BAC) of the lung is a histologic type of lung adenocarcinoma and is defined in the recent World Health Organization (WHO) histologic classification of lung and pleural tumors as an adenocarcinoma with a pure bronchioloalveolar growth pattern and no evidence of stromal, vascular, or pleural invasion. There are two subtypes of BAC: mucinous and nonmucinous.4 However, the histogenesis of each type is not understood well. Atypical adenomatous hyperplasia (AAH) is a bronchioloalveolar proliferation that resembles but falls short of the criteria for BAC.5 Many studies have suggested that AAH is a probable but not proven precursor lesion of pulmonary adenocarcinoma.6–13 However, in the current WHO lung tumor classification, AAH is categorized as a precursor lesion for the nonmucinous type of BAC.4 No precursor lesion has been identified for the mucinous type of BAC.

The presence of hyperplastic alveolar pneumocytes is not uncommon in cytologic samples. Distinction of BAC from AAH or reactive alveolar hyperplasia (RAH) can be difficult on aspiration cytology. Even when cell block preparations are available, the differential diagnosis remains problematic, because these lesions may appear as a few atypical cells involving rare alveolar spaces. This distinction can be difficult, because there is considerable overlap in the morphologic features between AAH and BAC.4, 9, 11, 14 It is known that the distribution of carcinoembryonic antigen (CEA) immunostain varies with the degree of cellular atypia during the carcinogenic sequence of colorectal and pancreatic adenocarcinomas.15, 16 It also is known that B72.3 and Leu-M1 (CD15) immunostains are positive in the majority of lung adenocarcinomas.17–21 We attempted to study the frequency of CEA, B72.3, and Leu-M1 expression in RAH, AAH, and BAC and to evaluate any differences in expression that may serve as an adjunct in differentiating RAH, AAH, and BAC in cell block material.


Samples were obtained from the archives of the Department of Pathology, Upstate Medical University (Syracuse, NY). An electronic search was made in the cytopathology data base to identify samples that were signed out on fine-needle aspiration (FNA) cytology as “BAC” or “atypical cells, cannot exclude BAC” between January, 1997 and July, 2003. Only patients with adequate cell blocks and histologic follow-up were included, resulting in a total of 11 evaluable patients (6 patients diagnosed with “BAC” and 5 patients diagnosed with “atypical cells, cannot exclude BAC”). In addition, paraffin embedded histologic sections from 10 random patients with AAH and from 8 random patients with RAH also were selected.

The FNA material from each patient was obtained by standard technique, using a 22-gauge needle under fluoroscopic guidance by radiologists. A rapid interpretation was performed by a cytopathologist at the time of aspiration to determine the adequacy of the specimen and to formulate a tentative diagnosis whenever possible. Direct smears were prepared from each pass, and at least one smear was air-dried and stained with Diff-Quik stain; the rest of the smears were fixed in alcohol to be stained with Papanicolaou stain. The needle and syringe were then rinsed in RPMI-1640 solution to be processed for cell block preparation, fixed in 10% formalin, and embedded in paraffin. The 10 samples of AAH were obtained from lobectomies (8 patients) and pneumonectomies (2 patients) for primary lung carcinomas. The histology of the primary tumor was adenocarcinoma in 6 patients and BAC in 4 patients. The eight samples of RAH were obtained from pneumonectomies (2 patients) and lobectomies (six patients). The histology of the associated primary lesion was adenocarcinoma (four patients) and squamous cell carcinoma (four patients). All of the AAH and RAH lesions were separate distinctly from the main tumor. The surgically excised lung specimens also were fixed in 10% formalin and embedded in paraffin.

On histology, AAH was diagnosed according to the following criteria6, 9, 22–24: 1) localized and well defined lesion measuring < 5 mm in greatest dimension comprising of a single layer of atypical cells without central scar or collapse; 2) the atypical cells had abundant, round or dome-shaped cytoplasm resembling type II pneumocytes with hyperchromatic nuclei and occasionally prominent nucleoli, but the nuclear atypia was milder than in adenocarcinoma; 3) the alveolar septa lined by atypical cells were thickened mildly by fibrosis in most instances. We did not subdivide AAH into low-grade and high-grade, because it is not reproducible and, thus, is not recommended in the recent WHO classification.4 RAH also was a small localized lesion characterized by a single row of intermittent or, occasionally, a continuous layer of round, cuboidal, or low columnar proliferating cells along the alveolar wall, with very mild atypia and uniform cytologic features, including small, minimally polymorphic nuclei. Compared with RAH, AAH showed increased cellularity and cellular atypia, such as enlarged hyperchromatic nuclei, an increased nuclear:cytoplasmic ratio, and mildly thickened alveolar septae. Neither AAH nor RAH showed mitotic activity. The recent WHO criteria (adenocarcinoma with a pure bronchioloalveolar growth pattern and no evidence of stromal, vascular, or pleural invasion) was followed for the diagnosis of BAC on excisional specimens.4

For each patient, immunostaining was performed on 5-μm sections of cell block and tissue sections with a monoclonal antibody against CEA (Ventana Medical Systems, Inc., Tucson, AZ), B72.3 (Becton-Dickinson, Mountain View, CA), and Leu M-1 (Ventana Medical Systems, Inc.) on an automatic immunostainer, which incorporated standard capillary-gap and streptavidin-biotin-peroxidase techniques. Positive controls consisted of a histologic block of primary lung adenocarcinoma that was confirmed previously as positive for CEA, B72.3, and Leu M-1. Negative controls included tissue sections from each patient with replacement of the primary antibody by nonimmune mouse serum. For analysis of immunoreactivity, only the cytoplasmic staining of the tumor cells was assessed. Immunoreactivity was considered positive if > 10% of tumor cells were stained and was graded according to the percentage of tumor cells stained as 1+ (10–25% of tumor cells stained), 2+ (26–50% of tumor cells stained), and 3+ (> 50% of tumor cells stained). The immunostained slides were evaluated by three observers (K.K.K., L.D.T., and V.S.C.) who had no prior knowledge of the cytologic or histologic diagnoses.


Cytologic Features

Smears from the six patients who had a cytologic diagnosis of BAC showed three-dimensional clusters; large sheets or occasional papillae of deceptively bland, relatively orderly glandular cells without cilia, with few single cells, in a clean background. The cells were small, and the nuclear:cytoplasmic ratio was increased. The nuclei were fairly uniform, round or oval, with some nuclear overlap, and with irregular nuclear membranes. The nuclear chromatin was finely granular, and some cells had prominent nucleoli. Intranuclear cytoplasmic invaginations were seen in two lesions. Smears from five patients who had a cytologic diagnosis of “atypical cells, cannot exclude BAC” showed few monolayered sheets of glandular cells against a clean background. The cells had well demarcated cytoplasmic borders, and few cells showed a mild-to-moderate increase in the nuclear:cytoplasmic ratio. The nuclei were fairly uniform with inconspicuous nucleoli; however, occasional cells showed nuclear membrane irregularity. Thus, a diagnosis of “atypical cells, cannot exclude BAC” was rendered when all of the cytologic features of BAC were not present to diagnose BAC confidently.


Among the 6 specimens with an unequivocal cytologic diagnosis of BAC (Table 1, Fig. 1), CEA was positive in 5 specimens (83%), B72.3 was positive in 5 specimens (83%), and Leu M-1 was positive in 3 specimens (50%). All six specimens were positive for two of the three immunostains. Lobectomy/wedge resection performed in all six patients confirmed the diagnosis of BAC in each specimen.

Table 1. Expression of B72.3, Carcinoembryonic Antigen, and Leu M-1 in Patients Diagnosed with Bronchioloalveolar Carcinoma and “Atypical Cells, Cannot Exclude Bronchioloalveolar Carcinoma” on Fine-Needle Aspiration Cytology
DiagnosisaB72.3CEALeu M-1
  • BAC: bronchioloalveolar carcinoma, nonmucinous type; CEA: carcinoembryonic antigen.

  • a

    Tissue follow-up was available in all patients.

  • b

    Tissue follow-up from this patient showed only atypical adenomatous hyperplasia.

Atypical cells, cannot exclude BAC03+1+
Atypical cells, cannot exclude BAC3+3+0
Atypical cells, cannot exclude BAC3+2+0
Atypical cells, cannot exclude BAC3+03+
Atypical cells, cannot exclude BACb000
Figure 1.

Examples of bronchioloalveolar carcinoma are shown in a cell block (A) with neoplastic cells lining the alveolar spaces (hematoxylin and eosin stain) and in sections from the same tumor illustrated in A that show cytoplasmic immunostaining for carcinoembryonic antigen (B) and for B72.3 (C). (immunohistochemical stain). Original magnification × 100 (A); × 200 (B,C).

Among the 5 specimens with a cytologic diagnosis of “atypical cells, cannot exclude BAC” (Table 1), immunostaining for CEA was positive in 3 specimens (60%), B72.3 was positive in 3 specimens (60%) and Leu M-1 was positive in 2 specimens (40%). Only 1 specimen was negative for all three markers (B72.3, CEA, and Leu M-1). Tissue follow-up (wedge resection/lobectomy) available from all patients showed BAC in 4 patients and AAH in 1 patient. The latter patients was the one who was negative for all 3 immunostains (B72.3, CEA, and Leu M-1) on cell block sections. All 10 AAH samples (Fig. 2) and 8 RAH samples were negative for CEA, B72.3, and Leu M-1.

Figure 2.

Examples of atypical adenomatous hyperplasia are shown in a paraffin embedded section (A) with atypical cells lining the alveolar spaces (hematoxylin and eosin stain) and in sections from the same tumor illustrated in A with negative staining for carcinoembryonic antigen (B) and for B72.3 (C) (immunohistochemical stain). Original magnification × 100 (A); × 200 (B,C)

Diagnostic Utility of Immunostain

In the current study, all 10 samples of BAC were positive for at least 2 of the 3 immunostains, i.e., CEA, B72.3, and Leu M1. In comparison, all 11 samples of AAH and 8 samples of RAH were negative for B72.3, CEA, and Leu M-1.


FNA biopsy is used widely in evaluating lung masses, and its main indication is a presumptive clinical diagnosis of primary or metastatic malignancy. FNA is the best method for investigating small, peripheral lung lesions or patients who cannot produce adequate sputum.25, 26 With the rising incidence of BAC, it is expected that more BAC will be seen on FNA biopsies.27 Several published studies have described the cytomorphology of BAC on exfoliative or aspiration biopsy specimens.28–31 AAH is a bronchioloalveolar proliferative lesion that resembles but falls short of the criteria for BAC.5 Although the recent WHO classification4 has included AAH in the category of preinvasive lesions, it is clear that AAH is not present in each patient who develops adenocarcinoma, and its frequency varies from 2% in young individuals without lung carcinoma32 to 23.2% in autopsy series of elderly patients with lung carcinoma.33 Alveolar cell hyperplasia also is seen in a variety of nonneoplastic lung diseases, including diffuse alveolar damage, viral or chemical pneumonitis, and postchemotherapy or radiation injury.34

AAH lesions usually are detected during examination of surgically resected lung specimens or at autopsy, because they usually measure < 0.5 cm in greatest dimension. Recent advances in helical computed tomography (CT) have increased the incidence of AAH lesions dramatically as a part of routine health check-ups in Japan.35 With the advent of modern imaging techniques, we speculate that, in the near future, AAH more often will be detected radiologically and subjected to FNA for differentiation from a variety of lung lesions, the most important of which is the nonmucinous subtype of BAC.36 Because there is considerable morphologic overlap between the two lesions, distinguishing between them may be very difficult4, 9, 11, 14 or even impossible and often reflects subjective judgment.10, 22 Both AAH and BAC appear as focal, ground-glass opacities on CT scans; and, although it is believed that AAH is a precursor of BAC, BAC needs to be resected surgically, whereas patients who have suspected AAH can be followed safely by CT.37

To our knowledge, no published study to date specifically has assessed the use of immunostains to help differentiate between RAH, AAH, and BAC in FNA-derived cell block material. Our study showed that BAC, despite the limited material often seen in cell block sections, was immunoreactive for at least 2 of the 3 tested markers, i.e., CEA, B73.2, and Leu M1, in all 10 lesions. In contrast, lack of expression of all three immunostains was helpful in excluding BAC and differentiating BAC from AAH and RAH. However, this immunoprofile did not help to distinguish AAH from RAH.

The results of our current study were similar to those in a study by Rao and Fraire,38 in which immunostains were performed on surgically resected adenocarcinomas of the lung. Those authors studied 23 primary pulmonary adenocarcinomas and 6 associated AAH lesions and found that pulmonary adenocarcinomas reacted positively in 23 of 23 lesions stained with Cam 5.2, in 22 of 23 lesions stained with AE1/AE3, in 17 of 23 lesions stained with Leu M-2, in 22 of 23 lesions stained with CEA, and in 18 of 23 lesions stained with B72.3. AAH reacted positively in 6 of 6 lesions, 6 of 6 lesions, 0 of 6 lesions, 1 of 6 lesions, and 0 of 6 lesions that were stained with Cam 5.2, AE1/AE3, Leu M-1, CEA, and B72.3, respectively. They concluded that primary pulmonary adenocarcinomas almost always would be immunoreactive for at least 4 of the 5 markers, i.e., Cam 5.2, AE1/AE3, Leu M-1, CEA, and B72.3. In contrast, a pattern of positive immunostaining for Cam 5.2 and AE1/AE3 and negative immunostaining for Leu M-1, CEA, and B72.3 may be useful for recognizing AAH and helping to differentiate AAH from pulmonary adenocarcinoma. Shimosato et al.39 also reported that CEA was negative in AAH, moderately positive in very well differentiated adenocarcinoma, and markedly positive in poorly differentiated adenocarcinoma; however, their report did not provide any details regarding the individual patients or immunostaining patterns. Another study by Kitamura et al.7 of surgically resected lung specimens showed that the frequency of CEA expression was significantly lower in AAH than in BAC. Weng et al.40 reported woman age 72 years who had BAC years associated with 14 foci of AAH within a lobectomy specimen, in which CEA was expressed strongly in the BAC but was not seen in any of the AAH lesions.

Recently, Mori et al.8 evaluated the value of CEA expression in the differential diagnosis of 97 lung lesions that included alveolar hyperplasia, AAH, and adenocarcinomas obtained from surgical specimens. Those authors found that CEA was negative in nonneoplastic cells and alveolar hyperplasia, whereas AAH and BAC were positive for CEA. They concluded that there was no difference in the degree of CEA immunostaining between AAH and lung adenocarcinoma. Carey et al.41 studied 10 AAH lesions from pneumonectomy and lobectomy specimens of lung carcinoma and found that many of these lesions, unlike the surrounding lung parenchyma, expressed CEA, albeit not as strongly as the adenocarcinomas. However, the authors did not specify the threshold for positive versus negative results and failed to elaborate on the results. Nakanishi42 studied 27 RAH and AAH lesions in association with primary lung adenocarcinoma and found no significant differences in CEA expression between AAH and adenocarcinoma: AAH more frequently was positive for CEA than RAH. All of the studies discussed above were performed on surgically excised lung specimens, and the differences between their results and ours may be attributable to the differences in classification of lung lesions, the criteria used to evaluate the immunostain results, and the immunostaining technique.

In conclusion, the current findings suggest that B72.3, CEA, and Leu M-1 staining could serve as a useful adjunct in differentiating BAC from AAH and RAH in FNA-derived cell block material. Although our study was limited by the fact that only 1 sample of AAH had FNA-derived cell block material from the 5 samples that were diagnosed as “atypical cells, cannot exclude BAC”, and the other 10 samples of AAH were investigated using histologic sections that were obtained from resected lung specimens, we believe that findings on histologic sections do simulate those on cell block material. Our results indicate that positive staining for two of the three antibodies provides strong supportive evidence for a diagnosis of BAC, and a negative result for all three antibodies is helpful in excluding BAC and differentiating BAC from AAH and RAH. Because more specimens of AAH will be identified with modern imaging techniques and subjected to FNA, differentiating between AAH and BAC may have important diagnostic and therapeutic implications. Further immunohistochemical studies with our suggested panel on larger series of FNA-derived cell block material in patients with confirmed histologic diagnoses of AAH/RAH will be needed to evaluate the validity of this approach in distinguishing BAC from AAH and RAH.


The authors thank QiLui Zhai, M.D., for his help with immunohistochemistry.