Association of Diffuse, Random Pulmonary Metastases, Including Miliary Metastases, With Epidermal Growth Factor Receptor Mutations in Lung Adenocarcinoma Yosuke Togashi, MD 1 ; Katsuhiro Masago, MD, PhD 1 ; Takeshi Kubo, MD 2 ; Yuichi Sakamori, MD 1 ; Young Hak Kim, MD 1 ; Yukimasa Hatachi, MD 1 ; Akiko Fukuhara, MD 1 ; Tadashi Mio, MD, PhD 1 ; Kaori Togashi, MD, PhD 2 ; and Michiaki Mishima, MD, PhD 1 BACKGROUND: Although the association of multiple pulmonary metastases, and particularly miliary metastases, with response to gefitinib treatment in patients with nonsmall cell lung cancer has been reported, the association of miliary pulmonary metastases with epidermal growth factor receptor gene () mutations remains unclear. METH- ODS: The authors retrospectively investigated the association of diffuse, random pulmonary metastases in patients with lung adenocarcinoma. The study included 163 Japanese patients who had unresectable, advanced lung adenocarcinoma diagnosed between April 2003 and March 2010. Computed tomography scans that were obtained at the time of diagnosis were analyzed by 2 investigators. For the purposes of this study, diffuse, random pulmonary metastases were defined as multiple nodules (n ¼ 50; 3 cm in greatest dimension) distributed diffusely and randomly throughout the lungs. RESULTS: Of 163 patients, 55 had pulmonary metastases, and mutations were detected in 22 of those 55 patients. The mutations were identified preferentially among women (P ¼.15) and were identified significantly among patients who had a smoking history of <10 pack-years (P ¼.0057). Diffuse, random pulmonary metastases were identified in 11 of 22 patients who had mutations and in 4 of 33 patients who had the wildtype (P ¼.0043). On the basis of multivariate analyses, mutations were associated independently with a smoking history of <10 pack-years (P ¼.026) and with diffuse, random pulmonary metastases (P ¼.012). CONCLU- SIONS: When patients with lung adenocarcinomas who had mutations developed pulmonary metastases, they tended to be diffuse and random, including military metastases. However, such metastases were much less common in patients who had lung adenocarcinomas with wild-type. Cancer 2011;117:819 25. VC 2010 American Cancer Society. KEYWORDS: pulmonary metastases, miliary metastases, epidermal growth factor receptor gene mutation, tyrosine kinase inhibitor, lung cancer. The epidermal growth factor receptor () is recognized as an important molecular target in cancer therapy. 1 Because patients with nonsmall cell lung cancer (NSCLC) who have somatic activating mutations of the gene generally respond to tyrosine kinase inhibitors (TKIs) (gefitinib or erlotinib), such mutations are very important markers and are useful in deciding the course of treatment for patients with NSCLC. 2-4 Although miliary pulmonary metastases are observed most often in patients with thyroid carcinoma, 5 these metastases sometimes are observed in patients with lung cancer; an association between multiple pulmonary metastases, particularly miliary metastases, and response to gefitinib therapy in patients with NSCLC has been reported. 6 Moreover, we recently reported that 1 patient with thyroid carcinoma who had an mutation 7 had miliary pulmonary metastases and responded to gefitinib therapy. From this finding, we hypothesized that, for cancers in which the patient has accompanying mutation(s), including NSCLC and thyroid carcinoma, miliary pulmonary metastases often develop. In this Corresponding author: Katsuhiro Masago, MD, PhD, Department of Respiratory Medicine, Graduate School of Medicine, Kyoto University, 54 Shogoin-Kawaracho, Sakyo-Ku, Kyoto 606-8507 Japan; Fax: (011) 81-75-751-4643; masago@kuhp.kyoto-u.ac.jp. 1 Department of Respiratory Medicine, Graduate School of Medicine, Kyoto University, Kyoto, Japan; 2 Department of Diagnostic Imaging and Nuclear Medicine, Graduate School of Medicine, Kyoto University, Kyoto, Japan DOI: 10.1002/cncr.25618, Received: March 31, 2010; Revised: June 27, 2010; Accepted: August 3, 2010, Published online September 30, 2010 in Wiley Online Library (wileyonlinelibrary.com) Cancer February 15, 2011 819
Table 1. Patient Characteristics (n¼55) Characteristic No. of Patients (%) Median/mean age [range], y 68/67.7 [35-86] Sex Women 31 (56.4) Men 24 (43.6) Smoking status Nonsmoker 25 (45.5) Smoker 30 (54.5) Former 23 (41.8) Current 7 (12.7) Performance status a 0-1 47 (85.5) 2-4 8 (14.5) Administration of -TKIs Yes 31 (56.4) No 24 (43.6) First or second-line therapy with -TKIs Yes 24 (43.6) No 31 (56.4) Mutated 22 (40) Wild type 33 (60) indicates epidermal growth factor receptor; TKIs, tyrosine kinase inhibitors. a Performance status was evaluated before the administration of - TKIs. article, we retrospectively investigated the association of diffuse, random pulmonary metastases, including miliary metastases, with mutations in patients with lung adenocarcinoma. MATERIALS AND METHODS Patients Japanese patients (n ¼ 163) with unresectable, advanced lung adenocarcinoma who were diagnosed between April 2003 and March 2010 and who had their tumors analyzed for gene mutation at Kyoto University Hospital were enrolled in this study. Patient characteristics are listed in Table 1. We obtained written informed consent from all patients. Mutational Analysis Formalin-fixed, paraffin-embedded tissue blocks or cytologic samples were used for DNA analysis. We adopted the peptic nucleic acid-locked nucleic acid (PNA-LNA) polymerase chain reaction (PCR) clamp method, according to previously described protocols. 8 Briefly, PNA Figure 1. These chest computed tomography images from a human aged 35 years who was a current smoker with an epithelial growth factor receptor gene mutation (L858R) reveal miliary pulmonary metastases. (A) Profuse, tiny, discrete, rounded, pulmonary opacities that generally are uniform in size are distributed diffusely throughout the lungs. A primary lesion was identified in the right upper lobe (arrow). (B) Four weeks after the initiation of gefitinib, the patient achieved a partial response. clamp primers inhibit amplification of the wild-type sequence, and LNA probes are used to detect specific mutant sequences in the presence of wild-type sequences. A synergistic effect of these primers causes specific PCR amplification of mutant sequences. Specific PNA-LNA probe sets for each mutation were developed to detect >95% of the mutations reported previously in Japan. 9 820 Cancer February 15, 2011
Radiologic Findings in Lung Cancer/Togashi et al Figure 2. These chest computed tomography images from a women aged 67 years who was a nonsmoker with an epithelial growth factor receptor gene mutation (L858R) reveal diffuse, random pulmonary (nonmiliary) metastases. (A) Multiple nodules are distributed diffusely throughout the lungs and a primary lesion in the right upper lobe (arrow). The nodules are not uniform in size. (B) Four weeks after the initiation of erlotinib, the patient achieved a partial response. Figure 3. These chest computed tomography images from a women aged 79 years who was an exsmoker with the wildtype epithelial growth factor receptor gene reveal miliary pulmonary metastases. (A) Bilateral pleural effusions and profuse, tiny, discrete, rounded pulmonary opacities that generally are uniform in size are distributed diffusely throughout the lungs. (B) Four weeks after the initiation of gefitinib, the patient achieved a partial response. Imaging Studies Two physicians who were blinded to results from the mutational analyses reviewed the chest computed tomography (CT) scans, including high-resolution CT scans, that were obtained at the time of lung cancer diagnosis. For the purposes of this study, diffuse, random pulmonary metastases, including miliary metastases, were defined as multiple nodules (a count 50 nodules and 3 cm in greatest dimension) that were distributed diffusely and randomly throughout the lungs. We used a previous definition of miliary pulmonary metastases, 10 that is, profuse, tiny, discrete, rounded pulmonary opacities (3 mm in greatest dimension) generally uniform in size that are distributed diffusely and randomly throughout the lungs (Figs. 1-3). To avoid confusion, we distinguished a bronchioloalveolar carcinoma (BAC) pattern, which appears as centrolobular distribution (Fig. 4). Cancer February 15, 2011 821
Statistical Analysis The univariate correlation between each independent variable was examined using either the chi-square test or the Fisher exact test. A multivariate logistic regression model was applied to estimate odds ratios (ORs) and 95% confidence intervals (CIs). Only those variables with P values <.1 in univariate analysis were included in the multivariate analysis. All tests were 2-tailed, and P values <.05 were considered statistically significant. All preceding statistical analyses were performed using JMP 8 software (SAS Institute, Cary, NC). Figure 4. These chest computed tomography images from a women aged 36 years who was an exsmoker with the wildtype epithelial growth factor receptor gene reveal a bronchioloalveolar carcinoma pattern. (A) Patchy consolidation, ground-glass opacity, and centrolobular nodules are distributed bilaterally in the lower lungs. (B) Four weeks after the initiation of gefitinib, she developed progressive disease. RESULTS mutations were detected in 77 of 163 patients (47%) and were not detected in 86 of 163 patients (53%). Twenty-two of 77 patients who had mutations had lung metastases (29%), and 33 of 86 patients with wildtype had pulmonary metastases (38%; P ¼.19; chi-square test). The clinical characteristics of the 55 patients who had pulmonary metastases are summarized in Table 1. The profiles of patients with mutations were similar to those reported previously. 11,12 The mutations were identified preferentially (but not significantly) among women (15 of 31 patients; P ¼.15; chisquare test) but were identified significantly in patients who had a smoking history of <10 pack-years (17 of 30 patients; P ¼.0057; chi-square test). Diffuse, random pulmonary metastases and miliary metastases were identified in 11 patients (50%) and 7 patients (32%), respectively, of the 22 patients who had mutations and in 4 patients (12%) and 1 patient (3%), respectively, of the 33 patients who had wild-type. mutations were associated significantly with both diffuse, random pulmonary metastases and miliary metastases (P ¼.0043 and P ¼.0049, respectively; Fisher exact test) (Table 2). In contrast, mutations were not associated with a BAC pattern (P ¼.72; Fisher exact test). -TKIs were administered to 31 of 55 patients who had pulmonary metastases, and 19 of those 31 patients (61%) achieved a partial response (Table 3). -TKIs were more administered frequently to patients who had mutations (18 of 22 patients; 82%) than to patients who had the wild-type gene (13 of 33 patients; 39%; P ¼.0024; Fisher exact test). The response rate to -TKIs was higher preferentially (but not significantly) among women (15 of 22 patients; P ¼.25; Fisher exact test) and among patients who had a smoking history of <10 pack-years (15 of 21 patients; P ¼.13; Fisher exact test). In contrast, the response rate was significantly higher among patients who had mutations (16 of 18 patients; 89%) than among patients who had the wild-type gene (3 of 13 patients; 23%; P ¼.0005; Fisher exact test). In addition, the response rate was significantly higher in patients who had diffuse, random pulmonary metastases (11 of 13 patients; 85%) than in patients who had other pulmonary metastases (8 of 18 patients; 44%; P ¼.032; Fisher exact 822 Cancer February 15, 2011
Radiologic Findings in Lung Cancer/Togashi et al Table 2. Clinical Characteristics of Patients With Pulmonary Metastases (n¼55) Characteristic No. of Patients P Mutated Wild-Type No./total no. of patients 22/77 33/86.19 Age, y <70 13 19.91 70 9 14 Sex Men 7 17 Women 15 16.15 Smoking status, pack-years <10 17 13 10 5 20.0057 a Performance status 0-1 19 28 2-4 3 5.88 Pulmonary metastasis pattern Diffuse random (miliary) 11 (7) 4 (1) Others (BAC) 11 (3) 29 (7).0043 a /.0049 a indicates epidermal growth factor receptor; TKIs, tyrosine kinase inhibitors; BAC, bronchioloalveolar pattern. a Statistically significant in Chi square test or Fisher exact test (P <.05). test). Four of the 15 patients who had diffuse, random pulmonary metastases had the wild-type gene. It is noteworthy that 2 of those 4 patients achieved a partial response to gefitinib (Fig. 3). We constructed a multivariate logistic regression model to identify the factors associated significantly with mutations and the response rate in this cohort. mutations were associated independently with a smoking history of <10 pack-years (OR, 4.40; 95%CI, 1.25-17.61; P ¼.026) and with diffuse, random pulmonary metastases (OR, 6.12; 95%CI, 1.58-27.97; P ¼.012) (Table 4). In contrast, the response rate was associated independently only with mutations (OR, 44.38; 95%CI, 4.28-1602.28; P ¼.0075) (Table 5). Miliary metastases were not analyzed, because the number of patients with such metastases was too small, and these were included in the diffuse, random pulmonary metastases category. Table 3. Clinical Characteristics of Patients With Pulmonary Metastases Who Received Epidermal Growth Factor Receptor-Tyrosine Kinase Inhibitor Treatment (n¼31) Characteristic -TKI Response DISCUSSION Although this was a small, retrospective study, 11 of the 15 patients who had diffuse, random pulmonary metastases had mutations (73%), and mutations were associated independently with diffuse, random pulmonary metastases. That is, patients who had lung adenocarcinoma with mutations tended to develop diffuse, random pulmonary metastases, including miliary metastases. However, diffuse, random pulmonary metastases were much less common in lung adenocarcinomas with wild-type. The reason for this association is unclear. A random pulmonary metastatic pattern means that hematogenous metastasis is associated with angiogenesis. 13,14 It is long established that and its ligands play important roles in tumorigenesis. 15 However, signaling also may play an important role in cancer progression. Because signaling regulates the synthesis and secretion of several different angiogenic growth factors in tumor cells (ie, vascular endothelial growth factor, interleukin-8, and PR SD or PD No. of Patients 19 12 Age, y <70 9 9 70 10 3.16 Sex Men 4 5 Women 15 7.25 Smoking status, pack-years <10 4 6 10 15 6.13 Performance status 0-1 17 11 2-4 2 1.84 No. of previous chemotherapy regimens 0-1 18 6 2 1 6.007 a Pulmonary metastasis pattern Diffuse random (miliary) 11 (6) 2 (0) Others (BAC) 8 (1) 10 (3).032/.059 a Mutated 16 10 Wild type 3 2.0005 a indicates epidermal growth factor receptor; TKI, tyrosine kinase inhibitor; PR, partial response; SD, stable disease; PD, progressive disease; BAC, bronchioloalveolar pattern. a Statistically significant in Chi square test or Fisher exact test (P <.05). P Cancer February 15, 2011 823
Table 4. Association Between Patient Characteristics and Epidermal Growth Factor Receptor Mutations Variable OR (95% CI) P Univariate analysis Age (<70 y/ 70 y) 1.06 (0.36-3.24).91 Sex (women/men) 2.28 (0.75-7.34).15 Smoking history (<10 pack-years/ 10 pack-years) 5.23 (1.63-19.23).0077 a Performance status (0 or 1/2) 1.13 (0.25-6.04).88 Diffuse random pulmonary metastases (yes/no) 7.25 (2.03-30.90).0037 a BAC pattern pulmonary metastases (no/yes) 1.69 (0.41-9.09).48 Multivariate analysis Smoking history (<10 pack-years/ 10 pack-years) 4.40 (1.25-17.61).026 a Diffuse random pulmonary metastases (yes/no) 6.12 (1.58-27.97).012 a OR indicates odds ratio; CI, confidence interval; BAC, bronchioloalveolar pattern. a Statistically significant in logistic regression model analysis (P <.05). Table 5. Association Between Patient Characteristics and Response to Epidermal Growth Factor Receptor-Tyrosine Kinase Inhibitor Treatment Variable OR (95% CI) P Univariate analysis Age ( 70 y/<70 y) 3.33 (0.73-18.87).14 Sex (women/men) 2.68 (0.55-14.08).23 Smoking history (<10 pack-years/ 10 pack-years) 3.75 (0.78-19.82).10 Performance status ( 2/0 or 1) 1.30 (0.11-29.41).84 No. of previous chemotherapy regimens (0 or 1/ 2) 18.00 (2.43-378.61) 014 a Diffuse random pulmonary metastases (yes/no) 6.88 (1.35-53.43).033 a BAC pattern pulmonary metastases (no/yes) 5.88 (0.66-131.57).14 (mutated/wild type) 26.67 (4.51-250.91).0010 a Multivariate analysis No. of previous chemotherapy regimens (0 or 1/ 2) 13.06 (0.94-516.25).087 Diffuse random pulmonary metastases (yes/no) 9.61 (0.75-305.60).12 (mutated/wild type) 44.38 (4.28-1602.28).0075 a OR indicates odds ratio; CI, confidence interval; BAC, bronchioloalveolar pattern;, epidermal growth factor receptor. a Statistically significant in logistic regression model analysis (P <.05). basic fibroblast growth factor), mutations that accompany cancer, including NSCLC and thyroid carcinoma, tend to develop angiogenic metastases, such as diffuse, random pulmonary metastases. 16 On the basis of our multivariate analysis, the response rate to -TKIs was associated independently with mutations but not with diffuse, random pulmonary metastases. This finding may have been because most patients with diffuse, random pulmonary metastases also had mutations and had achieved a partial response to the -TKIs. That is, diffuse, random pulmonary metastases were associated strongly with mutations. It is noteworthy that 2 of the 4 patients who had diffuse, random pulmonary metastases and the wild-type achieved a partial response to gefitinib therapy. The PNA-LNA PCR clamp method that we adopted can detect >95% of the mutations reported previously in Japan, 9 and its sensitivity is 97%. 17 However, this method does not allow the detection of all mutations. We speculate that the 2 patients with the wild-type who achieved a partial response to gefitinib in fact had an mutation. -TKIs should be used in patients with diffuse, random pulmonary metastases even when the wild-type is present, because there is a possibility that such patients in fact have an mutation. Although, according to some previous reports, BAC pathologic subtype was a predictor of response to gefitinib, and mutations were observed preferentially in tumors that had BAC features, 18-20 others reported that there was no association between BAC subtype and mutations. 21,22 In our study population, there was no association between a BAC pattern and mutations or response to gefitinib. However, a BAC pattern based 824 Cancer February 15, 2011
Radiologic Findings in Lung Cancer/Togashi et al on imaging appearance may include pathologic BAC histology. Although this was a small, retrospective study, we observed an association between mutations and diffuse, random pulmonary metastases, including miliary metastases. The method that we used to detect mutations cannot detect all mutations. Therefore, to avoid missing mutations, larger studies to identify the features of mutations associated with cancer should be performed. We speculate that NSCLC-associated mutations have behavior similar to that of thyroid carcinoma-associated mutations, such as the development of diffuse, random pulmonary metastases, including miliary metastases. In the future, it may be possible to establish a new category for mutation-associated cancers that would include NSCLC and thyroid carcinoma. CONFLICT OF INTEREST DISCLOSURES The authors made no disclosures. REFERENCES 1. Baselga J, Arteaga CL. Critical update and emerging trends in epidermal growth factor receptor targeting in cancer. J Clin Oncol. 2005;23:2445-2459. 2. Jackman DM, Yeap BY, Sequist LV, et al. Exon 19 deletion mutations of epidermal growth factor receptor are associated with prolonged survival in nonsmall cell lung cancer patients treated with gefitinib or erlotinib. Clin Cancer Res. 2006;12:3908-3914. 3. Riely GJ, Pao W, Pham D, et al. Clinical course of patients with nonsmall cell lung cancer and epidermal growth factor receptor exon 19 and exon 21 mutations treated with gefitinib or erlotinib. Clin Cancer Res. 2006;12:839-844. 4. Yamamoto H, Toyooka S, Mitsudomi T. Impact of mutation analysis in nonsmall cell lung cancer. Lung Cancer. 2009;63:315-321. 5. Herold CJ, Bankier AA, Fleischmann D. Lung metastases. Eur Radiol. 1996;6:596-606. 6. Park K, Goto K. A review of the benefit-risk profile of gefitinib in Asian patients with advanced nonsmall-cell lung cancer. Curr Med Res Opin. 2006;22:561-573. 7. Masago K, Asato R, Fujita S, et al. Epidermal growth factor receptor gene mutations in papillary thyroid carcinoma. Int J Cancer. 2009;124:2744-2749. 8. Nagai Y, Miyazawa H, Huqun, et al. Genetic heterogeneity of the epidermal growth factor receptor in nonsmall cell lung cancer cell lines revealed by a rapid and sensitive detection system, the peptide nucleic acid-locked nucleic acid PCR clamp. Cancer Res. 2005;65:7276-7282. 9. Kosaka T, Yatabe Y, Endoh H, Kuwano H, Takahashi T, Mitsudomi T. Mutations of the epidermal growth factor receptor gene in lung cancer: biological and clinical implications. Cancer Res. 2004;64:8919-8923. 10. Hansell DM, Bankier AA, MacMahon H, McLoud TC, Muller NL, Remy J. Fleischner Society: glossary of terms for thoracic imaging. Radiology. 2008;246:697-722. 11. Mitsudomi T, Yatabe Y. Mutations of the epidermal growth factor receptor gene and related genes as determinants of epidermal growth factor receptor tyrosine kinase inhibitors sensitivity in lung cancer. Cancer Sci. 2007;98:1817-1824. 12. Toyooka S, Matsuo K, Shigematsu H, et al. The impact of sex and smoking status on the mutational spectrum of epidermal growth factor receptor gene in non small cell lung cancer. Clin Cancer Res. 2007;13:5763-5768. 13. Lee KS, Kim TS, Han J, et al. Diffuse micronodular lung disease: HRCT and pathologic findings. J Comput Assist Tomogr. 1999;23:99-106. 14. Xu Y, Liu YJ, Yu Q. Angiopoietin-3 inhibits pulmonary metastasis by inhibiting tumor angiogenesis. Cancer Res. 2004;64:6119-6126. 15. Mitsudomi T, Yatabe Y. Epidermal growth factor receptor in relation to tumor development: gene and cancer. FEBS J. 2010;277:301-308. 16. De Luca A, Carotenuto A, Rachiglio A, et al. The role of the signaling in tumor microenvironment. J Cell Physiol. 2008;214:559-567. 17. Tanaka T, Nagai Y, Miyazawa H, et al. Reliability of the peptide nucleic acid-locked nucleic acid polymerase chain reaction clamp-based test for epidermal growth factor receptor mutations integrated into the clinical practice for nonsmall cell lung cancers. Cancer Sci. 2007;98:246-252. 18. Miller VA, Kris MG, Shah N, et al. Bronchioloalveolar pathologic subtype and smoking history predict sensitivity to gefitinib in advanced nonsmall-cell lung cancer. J Clin Oncol. 2004;22:1103-1109. 19. Lynch TJ, Bell DW, Sordella R, et al. Activating mutations in the epidermal growth factor receptor underlying responsiveness of nonsmall-cell lung cancer to gefitinib. N Engl J Med. 2004;350:2129-2139. 20. Pao W, Miller V, Zakowski M, et al. EGF receptor gene mutations are common in lung cancers from never smokers and are associated with sensitivity of tumors to gefitinib and erlotinib. Proc Natl Acad Sci U S A. 2004;101:13306-13311. 21. Shigematsu H, Lin L, Takahashi T, et al. Clinical and biological features associated with epidermal growth factor receptor gene mutations in lung cancers. J Natl Cancer Inst. 2005;97:339-346. 22. Motoi N, Szoke J, Riely GJ, et al. Lung adenocarcinoma: modification of the 2004 WHO mixed subtype to include the major histologic subtype suggests correlations between papillary and micropapillary adenocarcinoma subtypes, mutations and gene expression analysis. Am J Surg Pathol. 2008;32:810-827. Cancer February 15, 2011 825