Dominant Papillary Subtype Is a Significant Predictor of the Response to Gefitinib in Adenocarcinoma of the Lung Free

Lung cancer is the leading cause of cancer deaths in many countries, and the global incidence is rising at a rate of 0.5% per annum (1, 2). Platinum-based combination chemotherapy has been shown to improve survival and quality-of-life in patients with advanced non–small-cell lung cancer (NSCLC; ref 3 and 4), which accounts for approximately 80% of all lung cancers. New chemotherapeutic agents, like gemcitabine, vinorelbine, docetaxel, paclitaxel, and irinotecan, were developed in the 1990s. However, chemotherapy for advanced NSCLC has been of limited benefit, with response rates of approximately 30% and a median survival period of about 8 months, and it seems to have reached a plateau (1, 5, 6). It is clear that additional treatment strategies are necessary.

Epidermal growth factor receptor (EGFR) has been shown to play an important role in the growth of many solid tumors and is overexpressed in approximately 40 to 80% of NSCLCs (7, 8, 9). Furthermore, the overexpression of EGFR has been associated with a poor prognosis in several studies on lung cancer (10, 11). EGFR activation occurs when ligands, such as epidermal growth factor, transforming growth factor-α, or amphiregulin, bind to its extracellular domain, resulting in cell proliferation, angiogenesis, metastasis, and antiapoptosis (8, 9). Gefitinib (IRESSA; AstraZeneca, Osaka, Japan) is an orally active, selective EGFR tyrosine kinase inhibitor that blocks downstream of the EGFR signal transduction pathway (12). In this context, gefitinib has a quite different profile from chemotherapeutic agents that have ever been used.

After phase I studies, two multicenter, randomized, double-blind phase II studies (IDEAL1, ref. 13; IDEAL2, ref. 14) were carried out to evaluate the tolerability and efficacy of gefitinib in patients with advanced NSCLC who had been treated previously with platinum-based combination chemotherapy. In total, 426 patients were enrolled in the two studies, and all of the patients had been treated previously with platinum-based combination chemotherapy. In both studies, the administration of two different gefitinib dosages (250 mg/day and 500 mg/day) were compared. No significant differences in efficacy were seen between the two dosages, but the 250 mg/day treatment was better tolerated than the 500 mg/day treatment in both studies. For the 250 mg/day gefitinib arms, the response rates were 18.4% and 12.0% in IDEAL1 and IDEAL2, respectively.

The results of IDEAL1 showed that gefitinib was significantly more effective for the treatment of adenocarcinomas than for other histologies (odds ratio, 3.45) and was also more effective in females than in males (odds ratio, 2.65). These findings were unexpected, given the mechanism of the drug, because squamous cell carcinomas are known to overexpress EGFR, the target of gefitinib, to a greater degree than adenocarcinomas. On the other hand, gefitinib-induced severe acute interstitial pneumonia resulting in a high mortality rate is a serious social problem in Japan (15). Although the precise mechanism of gefitinib’s action is not yet completely understood, clinically it is more important to identify favorable characteristics in the treatment with gefitinib.

Gefitinib has been confirmed to be significantly effective for the treatment of adenocarcinomas, and some investigators have reported that gefitinib is especially effective in adenocarcinomas with bronchiolo-alveolar features (16, 17). To the best of our knowledge, however, these studies only examined biopsy specimens, and no studies examining surgical specimens have been done. Lung cancer is generally characterized by histologic heterogeneity; thus, studies examining histologic features should be done using surgically resected specimens.

In this study, we investigated the clinicopathological and immunohistochemical features of surgically resected adenocarcinoma specimens from patients who were subsequently treated with gefitinib after relapse to identify any characteristics that were associated with a favorable response to gefitinib.

Between June 2002 and October 2003, 253 consecutive patients were treated with a 250 mg daily dosage of gefitinib at our hospital. To select fully treated patients, we defined assessable patients as follows. Response evaluation using chest computed tomography was done after receiving gefitinib at least for > 4 weeks. Among 253 patients, 222 satisfied these criteria. Of these 222 patients, 48 patients (36 patients with adenocarcinomas, 7 patients with squamous cell carcinomas, 3 patients with large cell carcinomas, 1 patient with adenosquamous carcinoma, and 1 patient with pleomorphic carcinoma) had previously undergone a lung resection for primary NSCLC at our hospital. In this study, we analyzed 36 surgically resected adenocarcinoma specimens. At the time of analysis, the mean gefitinib treatment period of these 36 patients was 172 days (range, 29–542 days).

All surgical specimens were fixed with 10% formalin or absolute methanol and embedded in paraffin. The tumors were cut at approximately 5-mm intervals, and serial 4-μm sections were stained with H&E, Alcian blue-periodic acid Schiff method to visualize cytoplasmic mucin production, or Verhoeff van-Gieson method (18) to visualize elastic fibers. Lymphatic permeation and pulmonary metastases were evaluated on sections stained with H&E. Vascular invasion and pleural invasion were evaluated with the Verhoeff van-Gieson method. Three observers (Y. K., G. I., and K. T.) who were unaware of the clinical data independently reviewed all pathologic slides. The histologic diagnoses were based on the revised World Health Organization histologic classification (19). In addition, the histologic subtypes and percentage of each subtype present in the tumor were evaluated with the maximum cut surface of the tumor. The histologic patterns were divided into four distinctive subtypes: bronchioloalveolar carcinoma (BAC), acinar subtype, papillary subtype, and solid adenocarcinoma with mucin. The dominant subtype of each tumor was then documented. Tumor size was measured as the maximal diameter on the cut section of the lung. The pathologic stage was determined according to the classification of the Union Internationale Contre le Cancer (20).

Tissue blocks were cut into 4-μm sections and mounted on silane-coated slides (Matsunami, Tokyo, Japan). The slides were then deparaffinized in xylene, dehydrated in a graded alcohol series, and blocked for endogenous peroxidase with 3% H₂O₂ in absolute methanol. After microwave pretreatment in citrate buffer (pH 6.0) at 95°C for 20 minutes, immunostaining was done at 4°C overnight with a mouse monoclonal antihuman EGFR (Novocastra, Newcastle, United Kingdom; ref. 21) at a dilution of 1:10, a mouse monoclonal antihuman phosphorylated EGFR (provided by Kyowahakko, Tokyo, Japan) that recognizes Try-1173 of the activated EGFR at a dilution of 1:10. As for the use of the antihuman phosphorylated EGFP (p-EGFR), a synthetic peptide (CGSTENAEpYLRVAPQSS), the amino acid sequence of which corresponds to COOH-terminal region of human EGFR, was used as an immunogen to generate a monoclonal antibody specific for the tyrosine-phosphorylated EGFR molecule. Obtained monoclonal antibody (KM2911) was further characterized by ELISA, Western blot assay, and immunohistochemical staining to verify the specificity and sensitivity. Furthermore, we compared the immunostaining of KM2911 with that of another monoclonal antibody against tyrosine-phosphorylated EGFR (MAB3052, Chemicon International, Inc., Temecula, CA) and confirmed the same specificity and sensitivity. The tissues were then exposed to DAKO EnVision+ (DAKO, Glostrup, Denmark) at room temperature for 30 minutes. Staining was visualized by exposure to 3,3′-diaminobenzidine for 3 to 5 minutes. For c-erbB-2, mouse monoclonal antihuman c-erbB-2 (Ventana, Frankfurt, Germany) and the NX/EX automatic stainer (Ventana) were used (22) As positive controls, lung adenocarcinoma specimen, which had been surgically resected at our hospital and had been determined previously to be strongly positive, was used for the EGFR and p-EGFR experiments. Breast cancer specimen, also surgically resected at our hospital and known to be strongly positive, was used for the c-erbB-2 experiment. Negative controls for each antibody were done with nonimmune serum instead of the primary antibodies. The expression of each receptor was scored as follows: − = no discernible staining, or <10% of cells stained; 1+ = >10% of cytoplasmic staining, or plasma membrane staining with weak intensity; 2+ = >10% of plasma membrane staining with moderate intensity; and 3+ = >10% of plasma membrane staining with strong intensity. Staining of 2+ and 3+ were evaluated as positive. As for EGFR and p-EGFR, no universal evaluation criteria exist at present; therefore, we applied the same criteria as c-erbB-2. Although this evaluation criteria basically followed HercepTest (23), we added some modification to evaluate cytoplasmic staining.

All statistical analyses were done with the statistical program StatView, version 5.0 (Abacus Concepts, Berkley, CA). The significance of the relationships between individual clinicopathologic factors; the expression of EGFR, p-EGFR, and c-erbB-2; and a univariate analysis with the Fisher exact probability test was used to evaluate the response to gefitinib. A multivariate regression analysis was conducted according to the Cox proportional hazard model. Kaplan-Meier method was used to calculate survival rates, and a log-rank test was used to evaluate the statistical significance of any differences. A P value of less than 0.05 was considered significant.

The patient characteristics are shown in Table 1. All clinical data were retrieved from medical records. The mean age of the patients was 70 years (range, 43–82 years). Seventeen patients were male and 19 were female. The Eastern Cooperative Oncology Group performance status was 0 for 7 patients, 1 for 25 patients, and 2 for 4 patients. Sixteen patients were current or ex-smokers. Nineteen patients had been treated previously with chemotherapeutic agents for postoperative recurrences. The Response Evaluation Criteria in Solid Tumors (24) was used to evaluate the response of the patients. Seventeen patients experienced a partial response (PR) to gefitinib, 16 patients had a stable disease (SD), and 3 patients had a progressive disease (PD); the overall response (OR) rate was 47%. No clinical differences were observed between the responders and the non-responders (Table 2). Mean duration of response was 258 days; however, 11 of 17 responded patients were continuing gefitinib at the time of analysis.

The details of the pathologic findings for the surgical specimens are summarized in Table 3. The pathologic stage was IA in 3 cases, IB in 6 cases, IIA in 0 cases, IIB in 3 cases, IIIA in 11 cases, IIIB in 11 cases, and IV in 1 case. The stage IV disease was caused by pulmonary metastasis to another lobe. The pathologic stage could not be determined in one case, because only a partial resection had been done. All but 3 specimens were adenocarcinomas of mixed subtype. The dominant histologic subtype was BAC in 7 cases, acinar subtype in 5 cases, papillary subtype in 17 cases, and solid adenocarcinoma with mucin in 7 cases. Both a dominant papillary subtype (P = 0.0021) and the presence of pleural invasion (P = 0.0489) were significantly associated with the response to gefitinib when examined with a univariate analysis (Table 2), but a multivariate analysis revealed that a dominant papillary subtype was the only significant factor (Table 4). In addition, the survival period of the dominant papillary subtype patients was longer than that of the non-papillary subtype patients (Fig. 1). The representative histologic features of the papillary subtype are shown in Fig. 2.

The immunohistochemical evaluation was done according to the scoring system described in Patients and Methods. Immunohistochemistry was not done in one patient, because a tissue block was not available. Nine patients (28%) were positive for EGFR, and 14 (40%) were positive for p-EGFR. None of the patients were positive for c-erbB-2. No substantial association was observed between the immunohistochemical expression of each receptor and the response to gefitinib. The results of immunohistochemistry are summarized in Table 5.

Adenocarcinomas were known to be significantly sensitive to the treatment with gefitinib, despite their lower expression rates of EGFR compared with squamous cell carcinomas (13). Some investigators have reported that gefitinib is particularly effective especially in BAC (16, 17), and clinical trials targeting BAC are now under way (25). In fact, as Hirsch et al. (8) reported, a high expression level of both EGFR and c-erbB-2 was seen in the BAC in a preclinical study. However, the association between the overexpression of EGFR or c-erbB-2 and the sensitivity to gefitinib is still controversial (26, 27, 28), and no data supporting an association has been obtained in clinical studies (29). Moreover, none of the previous studies examined surgical specimens, although BAC cannot be diagnosed with small biopsy specimens (19). In the present study, we investigated the clinicopathological and immunohistochemical features of surgically resected specimens from adenocarcinoma patients who were treated with gefitinib for postoperative recurrences. The surgical specimens were used to determine the dominance of the histologic subtypes according to the revised World Health Organization classification and to precisely evaluate the expressions of EGFR, p-EGFR, and c-erbB-2.

Clinical factors, including age, gender, performance status, smoking history, previous chemotherapy, and recurrence site, were not associated with the response to gefitinib in this study. The immunohistochemical expression profiles of EGFR, p-EGFR, and c-erbB-2 were also not associated with the response. However, both the dominant histologic subtype and the presence of pleural invasion differed significantly between responders and non-responders according to a univariate analysis, whereas a multivariate analysis revealed that only the dominant histologic subtype was a significant factor. In other words, a dominant papillary subtype was the feature that most favored a response to gefitinib, and the survival period of patients with this feature was significantly longer than that of patients with non-papillary subtypes.

The finding that gefitinib is more effective in papillary subtype lesion is of great interest. Drug delivery might be more effective in this histologic subtype, because cancer cells with a papillary structure usually line the fibrovascular stroma. In an in vitro and in vivo study, Hirata et al. (30) showed that the antitumor effect of gefitinib was partly mediated by the inhibition of tumor angiogenesis through direct effects on microvascular endothelial cells that express EGFR. In the papillary subtype, this direct effect on microvascular endothelial cells might be more efficient than in other subtypes.

The results of the present immunohistochemical study suggest that EGFR expression is not a useful predictor of the response to gefitinib. Recently, Paez et al. (31) and Lynch et al. (32) originally showed that EGFR mutations may predict sensitivity to gefitinib. These epoch-making studies arouse an interest about association of EGFR mutations with histologic subtypes.

In the present study, 9 patients (28%) were positive for EGFR and 14 (40%) were positive for p-EGFR. It seems somewhat strange that the positive rate of p-EGFR surpassed that of EGFR; however, we consider that it is simply because the p-EGFR antibody was more sensitive than the EGFR antibody.

The response rate of our study was high even for adenocarcinoma patients; however, patients were not selected at a point of administration of gefitinib for the most likely respond and patient’s selection in the present study strictly followed the definition described in the Patients and Methods section. The relatively high proportion of female (53%) and never-smoker (56%) might lead to this result.

A micropapillary pattern (MPP) of lung adenocarcinoma, which was not included in the revised World Health Organization histologic classification, was first described by Silver and Askin (33). Lung adenocarcinomas characterized by MPP are thought to be more likely to metastasize and have a poor prognosis (34, 35). Most MPP-positive adenocarcinoma cases were included in the papillary subtype in the present study. Miler et al. (16) reported that a never-smoker status was a significant predictor of the response to gefitinib, whereas Wu et al. (36) reported that all of their patients who achieved a complete response with gefitinib had bilateral diffuse small pulmonary metastases. Both a never-smoker status and diffuse pulmonary metastases are frequently observed in MPP-positive adenocarcinoma (35). These reports, combined with the results of the present study, suggest that gefitinib might be effective against MPP-positive adenocarcinoma. In fact, MPP-positive adenocarcinomas (12 cases) were more sensitive to gefitinib than MPP-negative lesions (24 cases) in the present study (P = 0.0328).

In conclusion, the results of the present study indicate that a dominant papillary adenocarcinoma subtype can be an important predictor of the response to gefitinib. Even in patients with pathologic stage IA NSCLC who undergo a complete resection, the 5-year survival rate is about 70% at best (2). Therefore, adenocarcinoma with a dominant papillary subtype might be susceptible to postoperative adjuvant treatments with gefitinib. However, the precise mechanism of how this agent works is still obscure. Additional studies are needed to reveal the relation between the sensitivity to gefitinib and the histology of papillary subtype.