Context.—In 1999, the World Health Organization redefined bronchioloalveolar carcinomas (BACs) as those neoplasms with only a pure lepidic growth pattern and no invasion.

Objectives.—The present study examined 45 lung cancers with a BAC component (1) to determine whether these tumors would be classified as BACs by current World Health Organization standards, (2) to quantitate the BAC component within these tumors, and (3) to see if phenotypic differences exist between the so-called invasive and noninvasive regions of these tumors.

Design.—Retrospective review of hematoxylin-eosin–stained slides and classification of histologic grade, tumor subtype, and percentage of pure BAC pattern, with further characterization by immunohistochemical staining for thyroid transcription factor 1, cytokeratin 7, cytokeratin 20, and Ki-67 antibodies.

Results.—Only 7 (15.6%) of the 45 tumors examined could be classified as BAC by current strict World Health Organization criteria. Those tumors, classified as nonmucinous and mixed, showed similar immunohistochemical staining for cytokeratin 7, cytokeratin 20, and thyroid transcription factor 1; mucinous tumors showed disparate staining. Significant differences in immunohistochemical staining and tumor cell proliferation were seen for the regions of tumors designated as lepidic, infiltrative, and leading edge and for the regions of tumors with different histologic grades (ie, well, moderately, and poorly differentiated).

Conclusions.—Nonmucinous and mixed BACs are phenotypically similar and show identical immunohistochemical staining patterns; mucinous tumors, on the other hand, show disparate immunohistochemical staining. Pulmonary neoplasms designated as adenocarcinomas with a BAC component represent a heterogenous group with a range of cell types, differentiation, growth, and immunophenotypes. Within an individual neoplasm, there are regional differences in these parameters as well.

In 1999, the World Health Organization (WHO) redefined bronchioloalveolar carcinoma (BAC) to include only those lesions with a so-called pure lepidic (classic) growth pattern and no evidence of invasion by a tumor. If stromal, vascular, or pleural invasion was seen, the neoplasm was classified as an adenocarcinoma, mixed type, with bronchioloalveolar features.1,2 This narrow definition was applied because lesions with these features that measure less than 2 cm are often curable, whereas those with an invasive component are much more likely to recur and metastasize.3–5 BAC is the only subtype of adenocarcinoma with a substantially better prognosis than other forms of lung cancer.6,7 Although BAC has traditionally been regarded as an uncommon type of pulmonary adenocarcinoma, the new WHO classification will likely make the diagnosis of BAC more uncommon.

Invasion in BAC typically involves epithelial elements infiltrating a central fibrous scar.4 In fact, studies of BACs have shown that the prognosis is directly related to the size, percentage, and pattern of scar tissue present in the lesion.3–5,7 Since the presence of an infiltrative component is so closely related to the prognosis, it seems reasonable to hypothesize that the infiltrative portion of the neoplasm will be phenotypically different from the noninfiltrating neoplastic cells. The goals of this study were (1) to determine what percentage of neoplasms with a BAC component were considered pure BAC by current WHO criteria, (2) to quantitate the BAC component in neoplasms with some BAC growth pattern, and (3) to determine whether there were phenotypic differences between noninvasive and invasive neoplastic cells in neoplasms with BAC features. For goal 3, we chose to study distributions of cytokeratin 7 (CK7), cytokeratin 20 (CK20), thyroid transcription factor 1 (TTF-1), and Ki-67 (MIB-1) immunostaining in the lepidic and infiltrating portions of the mucinous and nonmucinous subtypes of adenocarcinomas with BAC features. Since an inverse relationship between TTF-1 and Ki-67 expression has been reported for pulmonary adenocarcinoma,8 Ki-67 immunostaining was also used to help assess tumor cell proliferation within the lepidic, infiltrative, and leading-edge portions of tumors.

Forty-five lung cancers previously diagnosed as having a BAC component were selected from the surgical pathology files of the University of California (Los Angeles) (UCLA) Department of Pathology and Laboratory Medicine, UCLA Center for Health Sciences.

The original microscopic glass slides were reviewed and reevaluated for histologic grade, tumor subtype (mucinous, nonmucinous, or mixed), and percentage of tumors having a pure BAC pattern within each neoplasm (designated BAC component).

Histology Grade

The histology grade (ie, degree of cellular differentiation) was defined as well differentiated, moderately differentiated, or poorly differentiated on the basis of the cytologic features and architectural characteristics of the invasive component when present. For correlation and regression analyses between particular subtypes of tumors, we used the following scoring system: 1 indicates well differentiated; 2, moderately differentiated; and 3, poorly differentiated.

BAC Subtype

A tumor was classified as mucinous if it contained a predominance of well- to moderately differentiated goblet or mucin-producing columnar tumor cells with clear, gray, or foamy material within its cytoplasm. In addition, these tumors often showed copious amounts of extracellular viscous mucus that distended adjacent alveolar spaces. Neoplasms with characteristic Clara cell or type II pneumocyte differentiation (without mucin production) were labeled nonmucinous. Specimens showing features of both mucinous and nonmucinous tumors in at least 10% of the neoplasm were labeled mixed. For correlation and regression analyses, we used the following grading system: 1 indicates mucinous; 2, nonmucinous; and 3, mixed.

BAC Component

A tumor was called pure if it was composed of neoplastic tumor cells (of the nonmucinous or mucinous BAC subtype)—seen lining mildly to moderately thickened alveolar septae without an associated central sclerotic region.

A tumor was called infiltrative if it showed dense fibroelastic and desmoplastic connective tissue infiltrated by neoplastic cells—seen in glandular configurations, as nests of cells with no glandular differentiation, or individually.

For the present study, a system was devised to help categorize the tumor specimens by the percentage, or component, of lepidic (ie, pure) BAC versus infiltrative tumor. The BAC component was determined by comparing the amount of lepidic BAC to the amount of infiltrative BAC. Each tumor was then assigned a BAC component score. The BAC component was numbered as follows: 5 indicates 100% lepidic growth and 0% infiltration (ie, pure); 4, 75% to 99% lepidic growth and 1% to 25% infiltration; 3, 50% to 74% lepidic growth and 26% to 50% infiltration; 2, 25% to 49% lepidic growth and 51% to 75% infiltration; and 1, 0% to 24% lepidic growth and 76% to 100% infiltration (Table 1).

Table 1.

Bronchioloalveolar Carcinoma Component (BAC)

Bronchioloalveolar Carcinoma Component (BAC)
Bronchioloalveolar Carcinoma Component (BAC)

Immunohistochemistry

Following the initial examination, the paraffin tissue block was recut, and serial sections were taken for immunohistochemical staining with antibodies to TTF-1, CK7, CK20, and Ki-67 (MIB-1). The paraffin sections were cut at 3 to 4 μm and baked overnight at 60°C. Slides were then deparaffinized in xylene and graded ethyl alcohols and brought to water. Endogenous peroxidase activity was quenched by treating the slides with 3% hydrogen peroxide in methyl alcohol for 10 minutes. Heat-induced epitope retrieval was performed on the slides using 0.01M citrate buffer, pH 6.0, in a vegetable steamer (Black and Decker, Towson, Md) for CK7 and CK20 and using 1mM EDTA at pH 8 in a pressure cooker (Biocare Medical, Walnut Creek, Calif) for TTF-1 and Ki-67. After heating for 25 minutes in the vegetable steamer and 3 minutes in the pressure cooker, the slides were cooled and then washed in 0.01M phosphate-buffered saline. Slides were then immunostained on a Dako Corporation (Carpinteria, Calif) immunostainer with a primary antibody overnight for 30 minutes, which was followed sequentially with rabbit anti-mouse immunoglobulin (Dako) for 15 minutes and by Envision+ (rabbit, peroxidase) (Dako) for 30 minutes. Diaminobenzidine and hydrogen peroxide were used as the substrates for the peroxidase enzyme. Slides were counterstained with hematoxylin and coverslipped. Primary antibodies included mouse monoclonal antibodies against CK7 (1:50 dilution), CK20 (1:50 dilution), Ki-67 (1:100 dilution), and TTF-1 (1:300 dilution) (all from Dako). For the negative control, normal mouse serum (Dako) was used in place of the primary antibodies.

The TTF-1, CK7, and CK20 immunohistochemical stains were read and assigned a numeric score depending on the degree of staining for both the lepidic and infiltrative regions (0 = no staining, 1 = mild staining, 2 = moderate staining, and 3 = marked staining). The Ki-67 stains were read and assigned a numeric score indicating the number of positive tumor cells per quartile of tissue within the lepidic, infiltrative, or leading-edge regions of tumors (1 = 0%–25% tumor cells staining, 2 = 26%–50% tumor cells staining, 3 = 51%–75% tumor cells staining, and 4 = 76%–100% tumor cells staining).

Statistics

Group means were compared by the Student t test. The relationships between variables were described by the correlation and multivariate regression analyses. A P value of <.05 was considered statistically significant. Statistical analyses were performed using SPSS software (SPSS Inc, Chicago, Ill).

Patients and Tumors Studied

Forty-five lung cancers were selected from the surgical pathology files of the UCLA Department of Pathology and Laboratory Medicine. The specimens were taken from female patients (n = 30) with an average age of 64.9 years (range, 44–80 years) and male patients (n = 15) with an average age of 69.3 years (range, 57–77 years). Tumors in the female patients were predominantly peripheral (28 of 30, 93%) and exhibited an average size of 2.8 cm (range, 0.6–7.0 cm). Tumors in the male patients were less predictable as to location (6 central tumors and 9 peripheral tumors) and exhibited a larger average size of 4.1 cm (range, 0.8–12.5 cm).

BAC Subtype

In this series of 45 neoplasms with a BAC component, only 7 (15.6%) could be classified as BAC by current WHO criteria (Figure 1, A through C). All cases of nonmucinous BAC (n = 34) showed moderate-to-strong TTF-1 reactivity (scoring an average of 2.9 and 2.2 within the lepidic and infiltrative regions across all BAC component scores, respectively), strong positive CK7 reactivity (scoring an average of 3.0 and 2.6 within the lepidic and infiltrative regions across all BAC component scores, respectively), and little-to-no staining for CK20 (scoring an average of 0.00 and 0.10 within the lepidic and infiltrative regions across all BAC component scores, respectively). The lepidic areas of tumor growth showed statistically significant greater CK7 and TTF-1 reactivity when compared to the infiltrative regions of tumors (Table 2; Figure 1, D); however, both regions exhibited prominent staining.

Figure 1.

A, Low-power view of a nonmucinous bronchioloalveolar carcinoma (BAC) with a pure lepidic growth pattern (grade 5, hematoxylin-eosin, original magnification ×12.5). B, Low-power view of an adenocarcinoma with a central scar (S) that contains an infiltrating neoplasm, with a lepidic growth pattern at the periphery (hematoxylin-eosin, original magnification ×12.5). C, High-power view of a portion of (A) demonstrating the lepidic growth of a nonmucinous neoplasm (hematoxylin-eosin, original magnification ×200). D, Thyroid transcription factor 1 (TTF-1) immunostaining of the neoplasm shown in (C) demonstrating strong positivity in the nuclei (TTF-1, avidin-biotin complex [ABC] peroxidase technique). E, High-power view of a mucinous neoplasm (hematoxylin-eosin, original magnification ×200). F, TTF-1 staining of the neoplasm shown in (E) that demonstrates no positive staining (TTF-1, ABC peroxidase technique)

Figure 1.

A, Low-power view of a nonmucinous bronchioloalveolar carcinoma (BAC) with a pure lepidic growth pattern (grade 5, hematoxylin-eosin, original magnification ×12.5). B, Low-power view of an adenocarcinoma with a central scar (S) that contains an infiltrating neoplasm, with a lepidic growth pattern at the periphery (hematoxylin-eosin, original magnification ×12.5). C, High-power view of a portion of (A) demonstrating the lepidic growth of a nonmucinous neoplasm (hematoxylin-eosin, original magnification ×200). D, Thyroid transcription factor 1 (TTF-1) immunostaining of the neoplasm shown in (C) demonstrating strong positivity in the nuclei (TTF-1, avidin-biotin complex [ABC] peroxidase technique). E, High-power view of a mucinous neoplasm (hematoxylin-eosin, original magnification ×200). F, TTF-1 staining of the neoplasm shown in (E) that demonstrates no positive staining (TTF-1, ABC peroxidase technique)

Close modal
Table 2.

Immunostaining Within the Lepidic and Infiltrative Regions of Tumors Versus the Percentage of Bronchioloalveolar Carcinoma (BAC) Within Nonmucinous Tumors (n = 34)*

Immunostaining Within the Lepidic and Infiltrative Regions of Tumors Versus the Percentage of Bronchioloalveolar Carcinoma (BAC) Within Nonmucinous Tumors (n = 34)*
Immunostaining Within the Lepidic and Infiltrative Regions of Tumors Versus the Percentage of Bronchioloalveolar Carcinoma (BAC) Within Nonmucinous Tumors (n = 34)*

Similarly, the cases of mixed BAC (n = 6) showed moderate-to-strong positive TTF-1 reactivity (scoring an average of 2.33 and 1.83 within the lepidic and infiltrative regions across all BAC component scores, respectively), strong positive CK7 staining (scoring an average of 2.83 and 2.67 within the lepidic and infiltrative regions across all BAC component scores, respectively), and no detectable CK20 reactivity. However, no statistically significant difference was found between the lepidic and infiltrative immunoreactivity of the mixed tumors (see Table 3).

Table 3.

Immunostaining Within the Lepidic and Infiltrative Regions of Tumors Versus the Percentage of Bronchioloalveolar Carcinoma (BAC) Within Mixed Tumors (n = 6)*†

Immunostaining Within the Lepidic and Infiltrative Regions of Tumors Versus the Percentage of Bronchioloalveolar Carcinoma (BAC) Within Mixed Tumors (n = 6)*†
Immunostaining Within the Lepidic and Infiltrative Regions of Tumors Versus the Percentage of Bronchioloalveolar Carcinoma (BAC) Within Mixed Tumors (n = 6)*†

The 5 mucinous tumors showed a different immunohistochemical profile: weak TTF-1 reactivity (scoring an average of 1.00 and 0.6 within the lepidic and infiltrative regions across all BAC component scores, respectively), strong CK7 reactivity (scoring an average of 3 and 2.80 within the lepidic and infiltrative regions across all BAC component scores, respectively), and mild CK20 reactivity (scoring an average of 1.00 within the lepidic and infiltrative regions across all BAC component scores, respectively). However, a comparison of the lepidic regions to the infiltrative regions of mucinous tumors revealed no statistically significant differences in immunoreactivity (Table 4; Figure 1, E and F).

Table 4.

Immunostaining Within the Lepidic and Infiltrative Regions of Tumors Versus the Percentage of Bronchioloalveolar Carcinoma (BAC) Within Mucinous Tumors (n = 5)*†

Immunostaining Within the Lepidic and Infiltrative Regions of Tumors Versus the Percentage of Bronchioloalveolar Carcinoma (BAC) Within Mucinous Tumors (n = 5)*†
Immunostaining Within the Lepidic and Infiltrative Regions of Tumors Versus the Percentage of Bronchioloalveolar Carcinoma (BAC) Within Mucinous Tumors (n = 5)*†

Finally, through the use of multivariate analysis, we examined TTF-1 reactivity within the regions of lepidic tumor growth of all BAC subtypes versus the BAC component and showed a statistically significant inverse correlation between the degree of lepidic TTF-1 reactivity and the percentage of pure BAC (r = −0.37; P = .01; see Table 5). In addition, statistically significant differences were observed when comparing lepidic TTF-1 reactivity to the BAC subtype, showing an increase in TTF-1 reactivity from mucinous to nonmucinous to mixed tumors (r = 0.36, P = .01; see Table 5). Although statistically significant, these correlations were relatively weak (Table 5). Evaluation of TTF-1 reactivity within the infiltrative regions showed no significant correlations.

Table 5.

Correlation Coefficients for Thyroid Transcription Factor 1 (TTF-1), Cytokeratin (CK) 7, and CK20 Reactivity Versus the Percentage of Bronchioloalveolar Carcinoma (BAC), Histology Grade, and BAC Subtype*

Correlation Coefficients for Thyroid Transcription Factor 1 (TTF-1), Cytokeratin (CK) 7, and CK20 Reactivity Versus the Percentage of Bronchioloalveolar Carcinoma (BAC), Histology Grade, and BAC Subtype*
Correlation Coefficients for Thyroid Transcription Factor 1 (TTF-1), Cytokeratin (CK) 7, and CK20 Reactivity Versus the Percentage of Bronchioloalveolar Carcinoma (BAC), Histology Grade, and BAC Subtype*

Statistically significant differences were also observed when comparing lepidic CK20 reactivity to the BAC subtype, showing a progressive increase in CK20 reactivity from mixed to nonmucinous to mucinous tumors (Table 5; Figure 2, A through D). Similarly, statistically significant differences were observed when comparing infiltrative CK20 reactivity to the BAC subtype, indicating a progressive increase in CK20 reactivity from mixed to nonmucinous to mucinous tumors (Table 5). Although statistically significant, these correlations were again relatively weak (Table 5). Evaluation of TTF-1 reactivity within the infiltrative regions showed no significant correlations.

Figure 2.

A and B, A nonmucinous bronchioloalveolar carcinoma (BAC) that demonstrates positive staining for cytokeratin 7 (CK7) (A) and negative staining for cytokeratin 20 (CK20) (B) (original magnification ×100). C and D, A mucinous BAC that is positive for both CK7 staining (A) and CK20 staining (B) (original magnification ×100). E, Ki-67 staining of a nonmucinous neoplasm showing much less positivity in the center (C) of the neoplasm than at the periphery (P) of the neoplasm (Ki-67, avidin-biotin complex immunoperoxidase technique, original magnification ×100)

Figure 2.

A and B, A nonmucinous bronchioloalveolar carcinoma (BAC) that demonstrates positive staining for cytokeratin 7 (CK7) (A) and negative staining for cytokeratin 20 (CK20) (B) (original magnification ×100). C and D, A mucinous BAC that is positive for both CK7 staining (A) and CK20 staining (B) (original magnification ×100). E, Ki-67 staining of a nonmucinous neoplasm showing much less positivity in the center (C) of the neoplasm than at the periphery (P) of the neoplasm (Ki-67, avidin-biotin complex immunoperoxidase technique, original magnification ×100)

Close modal

No significant correlations were observed when comparing CK7 reactivity within the lepidic or infiltrative portions of a tumor to the percentage of BAC, histology grade, or BAC subtype (Table 5).

Cellular Differentiation

The tumors examined were segregated according to the degree of cellular differentiation (ie, histology grade), and most (27 of 45) were called moderately differentiated. Most well-differentiated tumors were given high BAC component scores (ie, had small infiltrative components), and most tumors that were called poorly differentiated had relatively low BAC component scores (ie, had a predominantly infiltrative component; see Table 6).

Table 6.

Bronchioloalveolar Carcinoma (BAC) Component Versus Histology Grade

Bronchioloalveolar Carcinoma (BAC) Component Versus Histology Grade
Bronchioloalveolar Carcinoma (BAC) Component Versus Histology Grade

Of the 45 cases evaluated, 5 were considered well differentiated, 27 were considered moderately differentiated, and 13 were considered poorly differentiated. All tumors stained similarly for TTF-1 within their respective BAC subtype. Regression analysis detected no significant correlation between the histology grade and the degree of immunohistochemical staining for CK7, CK20, or TTF-1 (Table 5). Statistically significant differences were seen, however, when comparing the histology grade to CK7 and TTF-1 staining within the lepidic or infiltrative region of the nonmucinous tumors (Table 7). Statistically significant differences were also seen when comparing the histology grade to all tumors as a whole (ie, across all BAC subtypes; see Table 8).

Table 7.

Histology Grade Versus Cytokeratin (CK) and Thyroid Transcription Factor 1 (TTF-1) Immunostaining in Nonmucinous Tumors*

Histology Grade Versus Cytokeratin (CK) and Thyroid Transcription Factor 1 (TTF-1) Immunostaining in Nonmucinous Tumors*
Histology Grade Versus Cytokeratin (CK) and Thyroid Transcription Factor 1 (TTF-1) Immunostaining in Nonmucinous Tumors*
Table 8.

Histology Grade Versus Cytokeratin (CK) and Thyroid Transcription Factor 1 (TTF-1) Immunostaining Across All Bronchioloalveolar Carcinoma Subtypes*

Histology Grade Versus Cytokeratin (CK) and Thyroid Transcription Factor 1 (TTF-1) Immunostaining Across All Bronchioloalveolar Carcinoma Subtypes*
Histology Grade Versus Cytokeratin (CK) and Thyroid Transcription Factor 1 (TTF-1) Immunostaining Across All Bronchioloalveolar Carcinoma Subtypes*

Cellular Proliferation

Each quartile of tumor tissue was assigned a numeric score that correlated with the lepidic, leading-edge, and infiltrative regions of a tumor. These regions were then segregated by BAC component scores and assessed.

The lepidic and leading-edge regions of tumors consistently showed increased tumor cell proliferation when compared to the infiltrative region for all BAC component scores (Figure 2, E; Table 9). Tumor cell proliferation within the lepidic region was graded as 25% to 50% for all BAC component scores; however, BAC component-1 tumors (those tumors with a predominantly infiltrative component) contained the highest number of proliferating tumor cells within their lepidic, leading-edge, and infiltrative regions. In addition, a mild progressive increase in tumor cell proliferation was observed as the BAC component score decreased (ie, when examining BAC component-1 → BAC component-5 tumors, or as the infiltrative component increased; P = .01; see Tables 9 and 10).

Table 9.

Tumor Cell Proliferation Versus Broncioloalveolar Carcinoma Grade

Tumor Cell Proliferation Versus Broncioloalveolar Carcinoma Grade
Tumor Cell Proliferation Versus Broncioloalveolar Carcinoma Grade
Table 10.

Correlation Coefficients for Tumor Cell Proliferation (Ki-67 Reactivity) Versus the Percentage of Bronchioloalveolar Carcinoma (BAC), Histology Grade, and BAC Subtype*

Correlation Coefficients for Tumor Cell Proliferation (Ki-67 Reactivity) Versus the Percentage of Bronchioloalveolar Carcinoma (BAC), Histology Grade, and BAC Subtype*
Correlation Coefficients for Tumor Cell Proliferation (Ki-67 Reactivity) Versus the Percentage of Bronchioloalveolar Carcinoma (BAC), Histology Grade, and BAC Subtype*

Similarly, when comparing Ki-67 immunoreactivity to the histology grade, statistically significant differences were seen within the lepidic and leading-edge regions of a tumor versus the more infiltrative areas (Table 11).

Table 11.

Tumor Cell Proliferation (Ki-67 Reactivity) Versus Histology Grade for All Tumors*

Tumor Cell Proliferation (Ki-67 Reactivity) Versus Histology Grade for All Tumors*
Tumor Cell Proliferation (Ki-67 Reactivity) Versus Histology Grade for All Tumors*

The logistic regression model was used to determine how each immunostain (within either the lepidic or infiltrative region) related to the BAC component, histology grade, and BAC subtype (Table 12). Only those comparisons (ie, lepidic or infiltrative tumor vs TTF-1, CK20, and CK7 immunoreactivity) that are statistically significant are shown.

Table 12.

Logistic Regression Analysis Comparing the Lepidic or Infiltrative Regions of Tumors to the Bronchioloalveolar Carcinoma (BAC) Component, Histology Grade, and BAC Subtype*

Logistic Regression Analysis Comparing the Lepidic or Infiltrative Regions of Tumors to the Bronchioloalveolar Carcinoma (BAC) Component, Histology Grade, and BAC Subtype*
Logistic Regression Analysis Comparing the Lepidic or Infiltrative Regions of Tumors to the Bronchioloalveolar Carcinoma (BAC) Component, Histology Grade, and BAC Subtype*

A similar approach was used for Ki-67, but a linear regression model was used instead of logistic regression. Statistically significant results were found for Ki-67 expression within the leading-edge regions (Table 13).

Table 13.

Linear Regression Analysis for Ki-67 Immunostaining Within the Leading-Edge Portions of Tumors*

Linear Regression Analysis for Ki-67 Immunostaining Within the Leading-Edge Portions of Tumors*
Linear Regression Analysis for Ki-67 Immunostaining Within the Leading-Edge Portions of Tumors*

The WHO's recent reclassification of BAC has renewed interest in this subtype of pulmonary adenocarcinoma. In our series, however, only 7 (15.6%) of 45 pulmonary adenocarcinomas with a BAC component fulfilled current WHO criteria for the diagnosis of BAC.1 Thus, unless more sensitive imaging techniques (eg, high-resolution computed tomography, fluorodeoxyglucose-positron emission tomography) are implemented to detect early lung cancers, this neoplasm will be very unusual (or rare) in most centers by WHO criteria.9 The present study documents that adenocarcinoma of the lung with a BAC component is a heterogenous group of neoplasms that differ in cell type, degree of differentiation, growth pattern, and immunohistochemical staining characteristics; furthermore, within any given neoplasm, there are regional differences.

For the present study, we devised a grading system that placed individual bronchioloalveolar cell tumors in distinct categories on the basis of the percentage, or component, of lepidic versus infiltrative growth (Table 1). To our knowledge, no other study has examined bronchioloalveolar cell carcinoma in such a manner.

The present series (total tumors = 45) represents the different subsets of bronchioloalveolar cell carcinoma encountered at the UCLA Medical Center from 1996 to 2001 and, indeed, mimics the proportions of the different subtypes of BAC found in general—a preponderance of nonmucinous BAC (n = 34) and smaller fractions of mucinous BAC (n = 5) and mixed BAC (ie, nonmucinous + mucinous BAC) (n = 6).

In addition to the distinct morphologic characteristics of BAC, the use of specific immunohistochemical markers may help the pathologist further classify adenocarcinoma with BAC features. For this purpose, CK7 and CK20 are often employed to help differentiate primary lung neoplasms from pulmonary metastases.10,11 Most recently, immunostaining with TTF-1 has added specificity to the immunohistochemical characteristics of lung neoplasms.8,10–13 TTF-1 marks both neoplastic and normal pulmonary bronchioloalveolar cells and is currently applied worldwide to determine, with a high degree of certainty, whether a lesion is primary to the lung or metastatic.

Recently, Pelosi et al8 comprehensively described the role of TTF-1 immunoreactivity in stage I nonsmall cell carcinomas of the lung to include bronchioloalveolar cell carcinoma and further reinforced the value of this marker as a means of diagnosing pulmonary adenocarcinoma. Their study describes TTF-1 immunoreactivity in the lung predominantly as a product of the glandular component of normal and neoplastic lung cells. Indeed, the most common type of BAC, the nonmucinous variant, is thought to be derived from the type II pneumocyte/Clara cell. This cell is said to secrete lung surfactant and other proteins vital to normal lung function.8 The results of the present study show strong TTF-1 immunoreactivity in 100% of the nonmucinous and mixed tumors examined. In fact, these tumors showed similar staining patterns for CK7 and CK20 as well (Tables 2 and 3). Although classified as distinct subgroups of BAC, the homologous immunostaining of nonmucinous and mixed tumors in this study suggests some form of commonality—perhaps even a common cell (or cells) of origin.

When the lepidic and infiltrative regions of neoplasms were compared, the regions of lepidic growth exhibited a somewhat higher degree of TTF-1 immunoreactivity. This suggests that more well-differentiated tumor cells (found within the lepidic regions of tumor growth) are more likely to retain the ability to express common antigens that more infiltrative or high-grade cells may lose (Tables 2, 7, and 8). This correlates with our finding that tumors with high BAC component scores (ie, tumors with a relatively higher percentage of lepidic or pure BAC; see Table 2) were the more well-differentiated tumors (Table 6).

In addition, multivariate analysis of TTF-1 immunostaining for all tumors (across all BAC subtypes) showed statistically significant greater TTF-1 immunoreactivity within the lepidic portions of more infiltrative tumors (ie, those tumors with a relatively higher percentage of infiltrative vs lepidic growth and relatively lower BAC component scores; see Table 1). The reasons for this finding are unclear but may indicate an increased expression of antigen by the more infiltrative and, therefore, more aggressive tumors within their respective regions of lepidic growth. However, although statistically significant, this inverse correlation was relatively weak.

Our findings show that all of the nonmucinous tumors, the majority of the mixed tumors, and only a minority of the mucinous tumors showed TTF-1 reactivity. Unlike the nonmucinous variant of BAC, which is often uniformly positive for CK7 and TTF-1 (further confirmed within the present study, see Table 2), the immunophenotype of mucinous BAC may be TTF-1, CK7, or CK20 positive or negative (and in any combination; see Table 4).10,13,14 In fact, immunostaining for the mucinous variant of BAC has often caused pathologists a great deal of consternation. The lack of TTF-1 positivity and different cytokeratin expression has prompted some researchers to suggest that these tumors do not arise from the lung at all. This idea, coupled with the notion that mucinous BAC heralds a much graver prognosis for the patient, has generated much interest in this subtype.5–7 

Indeed, a recent paper by Goldstein and Thomas10 showed that it is often difficult to distinguish a mucinous BAC from a mucinous tumor metastatic from the colon or other site. Their findings show that TTF-1 staining within these tumors was often patchy, if seen at all. Also, CK20 staining was a common feature seen within these mucinous tumors. More recently, Lau et al13 further confirmed that merely staining mucinous bronchioloalveolar cell carcinomas for CK7, CK20, and TTF-1 is not a reliable means of differentiating primary from metastatic mucinous lung tumors. Because of this, the use of villin staining has been proposed to differentiate primary pulmonary neoplasms from those that are metastatic from the colon.10,14 Although the mucinous tumors in the present study were relatively few in number (5 of 45), our findings further illustrate the variable staining found within these tumors for the common pulmonary immunostains CK7, CK20, and TTF-1.

TTF-1 regulates the transcription of genes encoding lung-specific proteins. TTF-1 immunoreactivity is observed within the nonneoplastic alveolar cells and bronchiolar Clara cells, ciliated cells, and basal cells but not within the bronchial epithelial cells.15 The cells of mucinous BAC are morphologically similar to the columnar cells within the bronchi. While some researchers suggest that pure mucinous tumors of the lung are metastatic lesions from the colon, we feel that it is much more likely that many of the TTF-1–negative mucinous BACs identified are not metastatic lesions at all but that they are merely derived from a different pulmonary epithelial cell, namely, the bronchial columnar epithelial cell, rather than bronchiolar or alveolar cells, which give rise to nonmucinous BAC and other adenocarcinomas of the lung. If this is the case, then the term bronchioloalveolar carcinoma is a misnomer for the mucinous neoplasms with a BAC growth pattern.

To date, several research groups have also examined cell proliferation by means of Ki-67 staining in lung carcinoma. Kitamura et al16 recently reported significant differences in well-differentiated versus poorly differentiated lung adenocarcinomas. The study by Saleh et al17 showed no significant difference in cell proliferation between BAC and conventional lung adenocarcinoma. To our knowledge, however, no group has examined tumor cell proliferation in different regions within the neoplasm.

Our findings demonstrate increased tumor proliferation at the leading-edge and lepidic regions of tumors. The central infiltrating component, which is often thought to be more aggressive, actually demonstrated less Ki-67 expression. In addition, when examining all tumors as a whole and comparing Ki-67 reactivity to the BAC component (ie, the percentage of pure or lepidic growth), a progressively higher degree of staining was observed within the leading-edge and lepidic regions for those tumors with a larger infiltrative component (ie, those tumors with low BAC component scores; see Table 9). Again, this was most evident at the leading edge of the tumor (Table 13). Thus, the more aggressive, infiltrating cells exhibited less proliferative activity than the more well-differentiated tumor cells at the periphery of the neoplasm (presumably the site of growth for these lesions) (Figure 2, E). However, tumors with a more infiltrative component (ie, more aggressive tumors) showed the highest tumor cell proliferation within their respective leading-edge regions. This stands to reason if tumors that exhibit greater infiltration are thought to be the more aggressive tumors.

In summary, nonmucinous and mixed BACs show identical staining patterns for CK7 and TTF-1 and thus express a common phenotype. The mucinous tumors examined in this study show disparate staining for CK7, CK20, and TTF-1, as documented in other studies.10,13 Because the staining pattern of mucinous BAC differs from that of all other lung neoplasms, immunohistochemistry is less helpful in distinguishing primary tumors from metastatic lung tumors. That the immunohistochemical phenotype of mucinous BAC mimics that of normal bronchial epithelium suggests to us that the TTF-1–negative mucinous tumors are pulmonary neoplasms derived from a more central airway epithelial cell that is different from that of the nonmucinous or mixed BACs or other pulmonary adenocarcinomas. That cell proliferation is greatest at the advancing edge of the tumor indicates that studies of rates of neoplastic cell proliferation should focus specifically on these regions of a neoplasm.

This project was funded in part by NIH SPORE grant P50 CA90338 and the Piansky Family Trust Endowment (Dr Fishbein).

Brambilla
,
E.
,
W. D.
Travis
, and
T. V.
Colby
.
et al
.
The new World Health Organization classification of lung tumours.
Eur Respir J
2001
.
18
:
1059
1068
.
Franklin
,
W. A.
Diagnosis of lung cancer: pathology of invasive and preinvasive neoplasia.
Chest
2000
.
117
:(
suppl 1
).
80S
89S
.
Higashiyama
,
M.
,
K.
Kodama
, and
H.
Yokouchi
.
et al
.
Prognostic value of bronchiolo-alveolar carcinoma component of small lung adenocarcinoma.
Ann Thorac Surg
1999
.
68
:
2069
2073
.
Yamashiro
,
K.
,
S.
Yasuda
, and
A.
Nagase
.
et al
.
Prognostic significance of an interface pattern of central fibrosis and tumor cells in peripheral adenocarcinoma of the lung.
Hum Pathol
1995
.
26
:
67
73
.
Clayton
,
F.
Bronchioloalveolar carcinomas: cell types, patterns of growth, and prognostic correlates.
Cancer
1986
.
57
:
1555
1564
.
Hsu
,
C. P.
,
C. Y.
Chen
, and
N. Y.
Hsu
.
Bronchioloalveolar carcinoma.
J Thorac Cardiovasc Surg
1995
.
110
:
374
381
.
Barkley
,
J. E.
and
M. R.
Green
.
Bronchioloalveolar carcinoma.
J Clin Oncol
1996
.
14
:
2377
2386
.
Pelosi
,
G.
,
F.
Fraggetta
, and
F.
Pasini
.
et al
.
Immunoreactivity for thyroid transcription factor-1 in stage I non-small cell carcinomas of the lung.
Am J Surg Pathol
2001
.
25
:
363
372
.
Yap
,
C. S.
,
C.
Schiepers
, and
M. C.
Fishbein
.
et al
.
FDG-PET imaging in lung cancer: how sensitive is it for bronchioloalveolar carcinoma?
Eur J Nucl Med Mol Imaging
2002
.
29
:
1166
1173
.
Goldstein
,
N. S.
and
M.
Thomas
.
Mucinous and nonmucinous bronchioloalveolar adenocarcinomas have distinct staining patterns with thyroid transcription factor and cytokeratin 20 antibodies.
Am J Clin Pathol
2001
.
116
:
319
325
.
Ordonez
,
N. G.
Value of thyroid transcription factor-1 immunostaining in distinguishing small cell lung carcinomas from other small cell carcinomas.
Am J Surg Pathol
2000
.
24
:
1217
1223
.
Fabbro
,
D.
,
C.
Di Loreto
, and
O.
Stamerra
.
et al
.
TTF-1 gene expression in human lung tumours [abstract].
Eur J Cancer
1996
.
32
:
512
517
.
Lau
,
S. K.
,
M. J.
Desrochers
, and
D. J.
Luthringer
.
Expression of thyroid transcription factor-1, cytokeratin 7, and cytokeratin 20 in bronchioloalveolar carcinomas: an immunohistochemical evaluation of 67 cases.
Mod Pathol
2002
.
15
:
538
542
.
Tan
,
J.
,
G.
Sidhu
, and
M. A.
Greco
.
et al
.
Villin, cytokeratin 7, and cytokeratin 20 expression in pulmonary adenocarcinoma with ultrastructural evidence of microvilli with rootlets.
Hum Pathol
1998
.
29
:
390
396
.
Nakamura
,
N.
,
E.
Miyagi
, and
S.
Murata
.
et al
.
Expression of thyroid transcription factor-1 in normal and neoplastic lung tissues.
Mod Pathol
2002
.
15
:
1058
1067
.
Kitamura
,
H.
,
Y.
Kameda
, and
N.
Nakamura
.
et al
.
Proliferative potential and p53 overexpression in precursor and early stage lesions of bronchioloalveolar lung carcinoma.
Am J Pathol
1995
.
146
:
876
887
.
Saleh
,
H. A.
,
J.
Haapaniemi
, and
G.
Khatib
.
et al
.
Bronchioloalveolar carcinoma: diagnostic pitfalls and immunocytochemical contribution.
Diagn Cytopathol
1998
.
18
:
301
306
.

The authors have no relevant financial interest in the products or companies described in this article.

Author notes

Reprints: Michael C. Fishbein, MD, Department of Pathology and Laboratory Medicine, David Geffen School of Medicine at UCLA, 10833 Le Conte Ave, Los Angeles, CA 90095 ([email protected])