Circulating Tumor Cell As a Diagnostic Marker in Primary Lung Cancer

Home , Circulating tumor cell

Published OnlineFirst November 3, 2009; DOI: 10.1158/1078-0432.CCR-09-1095 Published Online First on November 3, 2009 as 10.1158/1078-0432.CCR-09-1095 Imaging, Diagnosis, Prognosis

Circulating Tumor Cell as a Diagnostic Marker in Primary Lung Cancer Fumihiro Tanaka,1 Kazue Yoneda,1 Nobuyuki Kondo,1 Masaki Hashimoto,1 Teruhisa Takuwa,1 Seiji Matsumoto,1 Yoshitomo Okumura,1 Shakibur Rahman,1 Noriaki Tsubota,2 Tohru Tsujimura,3 Kozo Kuribayashi,4 Kazuya Fukuoka,4 Takashi Nakano,4 and Seiki Hasegawa1

Abstract Purpose: To investigate the diagnostic performance of circulating tumor cells (CTC) in discrimination between primary lung cancer and nonmalignant diseases as well as in prediction of distant metastasis. Patients and Methods: We prospectively evaluated CTCs in 7.5-mL samples of peripheral blood sampled from patients with a suspicion or a diagnosis of primary lung cancer. A semiautomated system was used to capture CTCs with an antibody against epithelial cell adhesion molecule. Results: Of 150 eligible patients, 25 were finally diagnosed as having nonmalignant disease, and 125 were diagnosed as having primary lung cancer with (n = 31) or without (n = 94) distant metastasis. CTCs were detected in 30.6% of lung cancer patients and in 12.0% of nonmalignant patients. CTC count was significantly higher in lung cancer patients than in nonmalignant patients, but a receiver operating characteristic (ROC) curve analysis showed an insufficient capability of the CTC test in discrimination between lung cancer and nonmalignant diseases with an area under ROC curve of 0.598 (95% confidence interval, 0.488-0.708; P = 0.122). Among lung cancer patients, CTC count significantly increased along with tumor progression, especially with development of distant metastasis. The area under ROC curve for CTC count in prediction of distant metastasis was 0.783(95% confidence interval, 0.679-0.886; P < 0.001). When patients with one or more CTCs were judged as having metastatic disease, sensitivity and specificity of the CTC test were 71.0% and 83.0%, respectively. Conclusions: CTC is a useful surrogate marker of distant metastasis in primary lung cancer. (Clin Cancer Res 2009;15(22):6980–6)

Primary lung cancer is the leading cause of cancer death in scanning (1). More importantly, even in patients without clin- most industrialized countries, and its high mortality is mainly ically detectable distant metastasis at the time of initial diagno- caused by frequent occurrence of distant metastasis. In fact, sis, distant metastasis may frequently develop during treatment ∼40% of lung cancer patients have distant metastasis detectable or long-time follow-up. Thus, in most lung cancer patients, tu- with current diagnostic modalities such as whole-body comput- mor cells may circulate in the blood with or without apparent ed tomography (CT) and positron emission tomography (PET) distant metastasis, and detection of such circulating tumor cells (CTC) may contribute to improvement in diagnosis and therapy of lung cancer patients (2). Authors' Affiliations: Departments of 1Thoracic Surgery, 2Pathology, and However, in spite of many efforts for development of a sen- 3Thoracic Oncology, and 4Division of Respiratory Medicine, Department sitive detection system of CTCs, clinical significance of CTCs of Internal Medicine, Hyogo College of Medicine, Hyogo, Japan had not been established mainly due to lack of reproducibility Received 5/7/09; revised 6/17/09; accepted 6/18/09; published OnlineFirst and accuracy in detection of CTCs (2, 3). The CellSearch Sys- 11/3/09. Grant support: Grants-in-Aid 20390374 (F. Tanaka) for Scientific Research tem (Veridex LLC), a semiautomated system for quantitative (B) from the Japan Society for the Promotion of Science, Japan. evaluation of CTCs, has been recently developed, in which The costs of publication of this article were defrayed in part by the payment CTCs are immunomagnetically captured with an antibody of page charges. This article must therefore be hereby marked advertisement against epithelial cell adhesion molecule (EpCAM; ref. 3). in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. The most important advantage of the CellSearch system is re- Note: Part of this study was presented at the 44th Annual Meeting of the American Society of Clinical Oncology, May 30 to June 3, 2008, Chicago, IL producibility across different laboratories, which is validated and at the 45th Annual Meeting of the American Society of Clinical Oncol- by a prospective multicenter study in metastatic breast cancer ogy, May 29 to June 2, 2009, Orlando, FL. (4). Based on accumulating data supporting the accuracy and Requests for reprints: Fumihiro Tanaka, Department of Thoracic Surgery, Hyogo precision in evaluating CTCs (5–8), the CTC test has been ap- College of Medicine, Mukogawa 1-1, Nishinomiya-city, Hyogo 663-8501, Japan. Phone: 81-798-45-6885; Fax: 81-798-45-6897; E-mail: [email protected]. proved in the United States of America by the Food and Drug F 2009 American Association for Cancer Research. Administration for monitoring of blood from metastatic doi:10.1158/1078-0432.CCR-09-1095 breast and colon cancer patients. In addition, several clinical

Clin Cancer Res 2009;15(22) November 15, 20096980 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 24, 2021. © 2009 American Association for Cancer Research. Published OnlineFirst November 3, 2009; DOI: 10.1158/1078-0432.CCR-09-1095

Circulating Tumor Cell in Primary Lung Cancer

Patients and Methods Translational Relevance Study design. Patients who presented with a pathologic diagnosis The present study has revealed that circulating tu- of primary lung cancer or with a suspicion of primary lung cancer on mor cells (CTC) can be found and quantitatively eval- chest radiograph and/or CT at the Department of Thoracic Surgery, uated with a semiautomated system (CellSearch) in Hyogo College of Medicine hospital and agreed to the purpose of the patients with primary lung cancer, and that CTC study were eligible. Patients who have apparent or symptomatic count is useful diagnostic marker to predict develop- distant metastasis before enrollment were excluded from the study. In ment of distant metastasis. When the clinical value addition, patients who had concurrent or prior malignancy treated of the CTC test in diagnosis of distant metastasis is within in the previous 5 y were excluded. All patients provided written established in future clinical studies, CTC count will informed consent before enrollment. A 7.5-mL sample of peripheral blood was collected from each become a practical clinical marker, which can be patient, and was served for the CTC test. A complete clinical data noninvasively and repeatedly performed, not only including history, physical examination, and laboratory and radio- in determining clinical stage but also in monitoring graphic studies were also collected. In some primary lung cancer of therapeutic efficacy and predicting prognosis. In patients who underwent thoracotomy, pulmonary venous blood was addition, the present study has revealed that CTCs also sampled during thoracotomy, and was served for the CTC test are detected even in ∼20% of lung cancer patients for another study (12). without clinically detectable distant metastasis, For all patients enrolled, serum levels of carcinoembryonic antigen which suggests that CTC count is a surrogate marker (CEA) were measured using a commercially available electrochemilu- of micrometastasis. Thus, the CTC test may be po- minescence immunoassay system (Roche Diagnostics K.K.) following tentially marker in decision making of adjuvant che- a manufacturer's instruction; the manufacturer-suggested cutoff point of CEA to discriminate between nonmalignant disease and malignant motherapy following complete resection, which tumor was 5 ng/mL (13, 14). For patients with a suspicion of primary should be examined in future studies. lung cancer, bronchoscopic and/or trans-thoracic needle biopsy was ap- plied to obtain pathologic diagnosis; if failed, video-assisted thoracoscopic biopsy was done. For lung cancer patients, whole-body CT, studies have revealed that the CTC test is a potentially useful brain CT or magnetic resonance imaging (MRI), and PET scanning were clinical marker in other malignant tumors such as prostate routinely conducted to evaluate tumor progression. Clinical stage (c- cancer (9–11). In lung cancer, however, little has been re- stage) was determined according to the current tumor-node-metastasis ported about the incidence of CTCs (3), and its clinical signif- classification as revised in 1997; for patents undergoing surgery for lung cancer, pathologic stage (p-stage) was also determined. This study icance remains unknown. Thus, we now conducted a was approved by the Institutional Review Board of Hyogo College of prospective study to examine the presence and incidence of Medicine. CTCs in primary lung cancer patients for evaluating its diag- Evaluation of CTCs (the CTC test). CTCs were isolated from nostic performance in discrimination between primary lung peripheral blood using the CellSearch system (Veridex LLC), and cancer and nonmalignant diseases as well as in prediction of the number of CTCs was determined following a manufacturer's distant metastasis. protocol (3). In brief, epithelial cells, which were captured using

Fig. 1. Flowchart of diagnosis of patients enrolled in the study.

www.aacrjournals.org 6981 Clin Cancer Res 2009;15(22) November 15, 2009 Downloaded from clincancerres.aacrjournals.org on September 24, 2021. © 2009 American Association for Cancer Research. Published OnlineFirst November 3, 2009; DOI: 10.1158/1078-0432.CCR-09-1095

Imaging, Diagnosis, Prognosis

Table 1. Characteristics of patients and prevalence of CTCs count according to final diagnosis

Prevalence of CTC count (per 7.5 mL of peripheral blood) 0 ≥1 ≥2 ≥3 ≥5 ≥10 ≥50

Primary lung cancer All patients (n = 125) 87(69.6%) 38(30.6%) 21(16.8%) 13(10.4%) 9(7.2%) 5(4.0%) 1(0.8%) Histologic cell type Squamous cell carcinoma (n = 22) 13(59.1%) 9(40.9%) 4(18.2%) 2(9.1%) 1(4.5%) 1(4.5%) 0 Adenocarcinoma (n = 85) 63(74.1%) 22(25.9%) 13(15.3%) 8(9.4%) 5(5.9%) 2(2.4%) 0 Other non–small cell carcinoma (n = 9) 8(88.9%) 1(11.1%) 0 0 0 0 0 Small cell carcinoma (n = 9) 3(33.3%) 6(66.7%) 4(44.4%) 3(33.3%) 3(33.3%) 2(22.2%) 1(11.1%) C-stage I(n = 88) 71(80.7%) 17(19.3%) 8(9.1%) 3(3.4%) 1(1.1%) 1(1.1%) 0 II (n = 3) 3(100%) 0 0 0 0 0 0 III (n = 5) 4(80.0%) 1(20.0%) 1(20.0%) 1(20.0%) 1(20.0%) 0 0 IV (n = 29) 9(31.0%) 20(69.0%) 12(41.4%) 9(31.0%) 7(24.1%) 4(13.8%) 1(3.4%) P-stage I(n = 67) 55(82.1%) 12(17.9%) 6(9.0%) 1(1.5%) 0 0 0 II (n = 11) 9(81.8%) 2(18.2%) 1(9.1%) 1(9.1%) 1(9.1%) 1(9.1%) 0 III (n = 15) 14(93.3%) 1(6.7%) 0 0 0 0 0 IV (n = 17) 6(35.3%) 11(64.7%) 6(35.3%) 4(23.5%) 1(5.9%) 0 0 Nonmalignant disease All patients (n = 25) 22(88.0%) 3(12.0%) 1(4.0%) 0 0 0 0

ferroparticles coupled to an anti-EpCAM antibody, were separated in parison among three or more groups) if the sample distribution was a magnetic field, and the enriched samples were then stained with asymmetrical. 4′,6-diamidino-2-phenylindole and anti–cytokeratin-phycoerythrin. Diagnostic performance of CTC count or serum CEA level assessed Contaminated WBC were excluded by negative selection for CD45. by constructing a receiver operating characteristic (ROC) curve, and Stained cells were then analyzed on a florescent microscope using was evaluated by calculating the area under each ROC curve (AUC- the Cell track Analyzer II (Veridex LLC). The criteria for each cell ROC; ref. 15). An AUC-ROC of 1 denotes perfect discrimination of a to be defined as a CTC are as follows: round to oval morphology, test, whereas an AUC-ROC of 0.5 denotes complete lack of discrimina- a visible 4′,6-diamidino-2-phenylindole–positive nucleus, positive tion of a test. P value was calculated for the difference between each cytokeratin staining in the cytoplasm, and negative staining for ACU-ROC and 0.5 (complete useless test). P CD45. All evaluations were done by two authors (K.Y. and F.T., both For each test, two-sided values of <0.05 were considered statistical- completed the “Cell Interpretation Proficiency Assessment” managed ly significant. All statistical manipulations were done using the SPSS for by the Veridex LL for identification of CTCs) independently without Windows software system (SPSS, Inc.). knowledge of clinical characteristics of patients. Statistics. Counts were compared by the χ2 test. Continuous data Results were compared using Student's t test if the distribution of samples was normal, or using nonparametric tests (Mann-Whitney U test Patient characteristics. From October 2007 through Decem- for comparison between two groups and Kruskal-Wallis test for com- ber 2008, 152 consecutive patients were enrolled in the study.

Fig. 2. A, distribution of CTC count in primary lung cancer patients and in patients with nonmalignant disease. B, ROC curves for CTC count and serum CEA level to discriminate primary lung cancer patients from nonmalignant patients (B).

Clin Cancer Res 2009;15(22) November 15, 20096982 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 24, 2021. © 2009 American Association for Cancer Research. Published OnlineFirst November 3, 2009; DOI: 10.1158/1078-0432.CCR-09-1095

Circulating Tumor Cell in Primary Lung Cancer

Table 2. Diagnostic values of CTC count and serum CEA level at selected cutoff points

Biomarker Cutoff point Sensitivity Specificity Accuracy Positive predictive value Negative predictive value

For diagnosis of lung cancer (discrimination between primary lung cancer and nonmalignant disease) CTC count ≥1 30.4% 88.0% 40.0% 92.7% 20.2% ≥2 16.8% 96.0% 30.0% 95.5% 18.8% ≥3 10.4% 100% 25.3% 100% 18.2% Serum CEA level ≥2.5 78.4% 60.0% 75.3% 90.7% 35.7% ≥5.0 45.6% 92.0% 53.3% 96.6% 25.3% ≥10.0 19.2% 96.0% 32.0% 96.0% 19.2%

For prediction of distant metastasis (among primary lung cancer patients) CTC count ≥1 71.0% 83.0% 80.0% 57.9% 89.7% ≥2 41.9% 91.5% 79.2% 61.9% 82.7% ≥3 32.3% 96.8% 80.8% 76.9% 81.3% Serum CEA level ≥2.5 83.9% 23.4% 38.4% 26.5% 81.5% ≥5.0 51.6% 56.4% 55.2% 28.1% 77.9% ≥10.0 29.0% 84.0% 70.4% 37.5% 78.2%

For prediction of distant metastasis (among all patients with a diagnosis or a suspicion of primary lung cancer) CTC count ≥1 71.0% 84.0% 81.3% 53.7% 91.7% ≥2 41.9% 92.4% 82.0% 59.1% 85.9% ≥3 32.3% 97.5% 84.0% 76.9% 84.7% Serum CEA level ≥2.5 83.9% 31.1% 42.0% 24.1% 88.1% ≥5.0 51.6% 63.9% 61.3% 27.1% 83.5% ≥10.0 29.0% 86.6% 74.7% 36.0% 82.4%

After enrollment, whole-body CT or PET scanning incidentally count was significantly higher compared with that in nonmalig- revealed a concurrent malignant tumor in two patients (pancre- nant patients (Table 1; Fig. 2A). atic cancer and bile-duct cancer in one patient each), and these The AUC-ROC for CTC count in discrimination between patients were excluded. Of 150 eligible patients, 17 had been lung cancer patients and nonmalignant patients was 0.598 pathologically diagnosed as having primary lung cancer at en- (P = 0.122; Fig. 2B). At a cutoff point of 1 for lung cancer rollment. Of 133 patients with a suspicion of lung cancer at en- (CTC count, ≥1) or nonmalignant patients (CTC count, 0), rollment, 108 were finally diagnosed pathologically as having sensitivity and specificity were 30.4% and 88.0%, respectively primary lung cancer. Thus, the final diagnosis was primary lung (Table 2). cancer in a total of 125 patients (Fig. 1). The majority of pa- The AUC-ROC for serum CEA level was 0.747 (P < 0.001), tients had c-stage I disease, and 29 patients were diagnosed as suggesting that serum CEA level is a better diagnostic marker having c-stage IV disease. for the diagnosis of lung cancer (Fig. 2B). At a cutoff point of Thoracotomy was done for 110 primary lung cancer patients 5 ng/mL, sensitivity and specificity were 45.6% and 92.0%, consisting of all (n = 94) c-stage I-IIIA, and one of 2 c-stage IIIB, respectively (Table 2). and 15 of 29 c-stage IV patients. Of 96 patients with c-stage I-III CTCs in primary lung cancer patients. Among primary lung disease, 2 (both c-stage IB patients) were diagnosed as having p- cancer patients, small-cell lung cancer patients showed a signif- stage IV disease as pulmonary metastases were found during icantly higher CTC count than non–small cell cancer patients thoracotomy. Finally, distant metastasis was found in a total (Fig. 3A). of 31 primary lung cancer patients (Fig. 1). CTC count significantly increased along with increase in c- The remaining 25 patients were finally diagnosed as having stage (Fig. 3B). The difference in CTC count between c-stage I nonmalignant disease by video-assisted thoracoscopic biopsy disease and c-stage IV disease was statistically significant, but (Fig. 1); there were 10 patients with organizing organized pneu- the difference among c-stage I to III diseases did not reached monia, 7 with nonmalignant tumor including atypical adeno- a statistically significance. matous hyperplasia (in 3 patients), 6 with infectious disease, CTC count significantly increased along with increase in p- and 2 with lymph-proliferative disease. There was no significant stage, and was significantly higher in p-stage IV disease than difference in any patient characteristic between lung cancer pa- any other p-stage disease (Fig. 3C). Of noted, two patients, tients and nonmalignant patients. who migrated from c-stage IB to p-stage IV disease due to CTCs in primary lung cancer patients and nonmalignant pulmonary metastases found during thoracotomy, had one patients. CTCs were identified in peripheral blood of 3 or more CTCs (CTC count, 1 and 3 in one patient each) (12.0%) of 25 nonmalignant patients, and CTC count was in peripheral blood. “1” (n =2)or“2” (n = 1). CTCs were identified in peripheral CTC test for prediction of distant metastasis. Metastatic blood of 38 (30.6%) of 125 lung cancer patients, and CTC lung cancer patients showed a significantly higher CTC count

www.aacrjournals.org 6983 Clin Cancer Res 2009;15(22) November 15, 2009 Downloaded from clincancerres.aacrjournals.org on September 24, 2021. © 2009 American Association for Cancer Research. Published OnlineFirst November 3, 2009; DOI: 10.1158/1078-0432.CCR-09-1095

Imaging, Diagnosis, Prognosis than primary lung cancer patients without distant metastasis were judged as having lung cancer (cutoff point, 1), sensitivity (Fig. 4A). The AUC-ROC for CTC count for prediction of the ab- was only 30.4%, whereas specificity was as high as 88.0%. sence or presence of distant metastasis among primary lung can- And, if the cutoff point was elevated to 3, only 10.4% of lung cer patients was 0.783 (P < 0.001; Fig. 4B), suggesting that CTC cancer patients were correctly judged to be positive, whereas count was a useful surrogate marker of distant metastasis. At a cutoff point of 1 for metastatic lung cancer (CTC count, ≥1) or nonmetastatic lung cancer patients (CTC count, 0), sensitivity and specificity were 71.0% and 83.0%, respectively (Table 2). In contrast, the AUC-ROC for serum CEA for prediction of distant metastasis failed to reach a statistical significance (P = 0.136; Fig. 4B). When the diagnostic performance of the CTC test in prediction of distant metastasis was evaluated among all patients enrolled in the present study (all patients with a suspicion or a diagnosis of lung cancer), similar results were obtained (Table 2; Fig. 4C and D).

Discussion

The present study is the first detailed study evaluating CTCs in patients with a suspicious or a diagnosis of primary lung cancer using the CellSearch system. We first showed that CTC count was significantly higher in lung cancer patients than in nonmalignant patients, but the discriminatory capability might be insufficient with an AUC-ROC of 0.598 (P = 0.122). Second, we showed that CTC count significantly increased along with tumor progression, especially with development of distant metastasis, and the CTC test showed a moderate diagnostic performance in discrimination between nonmetastatic patients and metastatic patients with an AUC-ROC of 0.783 showing a significantly better capability compared with a completely useless test with an AUC-ROC of 0.5 (P < 0.001). Furthermore, these ROC- curves suggested that the optimal cutoff point in diagnosis of lung cancer was 1 (patients with one or more CTCs are judged to be “positive”). Although intense efforts have been continued for development of useful biomarkers, many have not been established for diagnosis or decision-making in therapy of lung cancer (16, 17). Because only a small tumor specimen can be obtained by usual diagnostic modalities such as fine-needle aspiration and transbronchial biopsy, development of biomarkers that are evaluated in peripheral blood samples would be valuable. In addition, blood-based biomarkers have a great advantage of being evaluated easily and repeatedly. Among blood-based biomarkers, serum CEA is most extensively examined, and several studies showed that serum CEA was a useful marker not only to discriminate lung cancer, especially adenocarcinoma of the lung, from benign diseases but also to determine tumor progression and prognosis (13, 14). Despite these promising results, routine use of serum CEA in clinical practice is not recommended mainly due to its insufficient sensitivity and/or specificity as well as lack of reproducibility of the results (16, 17). Thus, we assessed the diagnostic performance of CTC count in comparison with that of serum CEA in the present study. Concerning the diagnostic performance in discrimination between nonmalignant patients and lung cancer patients, serum CEA showed a moderate performance (AUC-ROC, 0.747). However, the CTC test showed an insufficient perfor- Fig. 3. A, distribution of CTC count in patients with primary lung cancer according to histologic cell type. B, distribution of CTC count in patients mance (AUC-ROC, 0.598), which may be mainly due to its with primary lung cancer according to c-stage. C, distribution of CTC count low sensitivity. In fact, when patients with one or more CTCs in patients with primary lung cancer according to p-stage.

Clin Cancer Res 2009;15(22) November 15, 20096984 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 24, 2021. © 2009 American Association for Cancer Research. Published OnlineFirst November 3, 2009; DOI: 10.1158/1078-0432.CCR-09-1095

Circulating Tumor Cell in Primary Lung Cancer

Fig. 4. A, distribution of CTC count in primary lung cancer patients without distant metastasis and in patients with distant metastasis. B, ROC curves for CTC count and serum CEA level to predict the absence or presence of distant metastasis among primary lung cancer patients. C, distribution of CTC count in patients with nonmetastatic disease (nonmalignant disease or primary lung cancer without distant metastasis) and in patients with metastatic disease. D, ROC curves for CTC count and serum CEA level to discriminate metastatic lung cancer patients from nonmetastatic patients. no nonmalignant patients were wrongly judged to be positive of technical issues such as inappropriate blood sampling, (b) (both specificity and positive predictive value, 100%). These false-positive staining of contaminated nonepithelial cells for results indicate that the CTC test is characterized by low sen- cytokeratin/4′,6-diamidino-2-phenylindole during sample pro- sitivity and negative predictive value as well as high specificity cessing, (c) inappropriate judgment in identification of CTCs and positive predictive value. Here, at a cutoff point of 1, by researchers. A false false-positive result can occur, when attention should be paid to the fact that specificity or positive detected CTCs are originated from a clinically undetectable predictive value did not reach 100% (88.0% and 92.7%, malignant tumor. An individual reason for three false-positive respectively). In other words, some patients (7.3%, 3 of 41 cases in the present study remains unclear, and we will contin- patients) were finally diagnosed as having nonmalignant ue to watch for development of malignant tumors. diseases, whereas one or two “tumor cells” were identified in The CTC test showed a significant diagnostic performance peripheral blood. These results are similar as those in a pre- to predict the absence or presence of distant metastasis (AUC- vious study showing that up to three CTCs were detected in ROC, 0.783; P < 0.001), whereas serum CEA showed an a small subset of healthy volunteers or nonmalignant patients. insufficient performance. Thus, when CTCs are identified in Such “false-positive” results may be classified into “true” false- peripheral blood of patients with a suspicion or a diagnosis positive results or “false” false-positive results. A true false- of lung cancer, whole-body PET-scanning and brain magnetic positive result, indicating that patients without any malignant resonance imaging (or CT) for evaluation of possible distant tumor are judged to have a malignant tumor according to the metastasis should be routinely performed even in clinical CTC test, can be brought by several factors as follows: (a) con- T1N0 cases evaluated with thoracic CT scanning. However, tamination of epithelial cells in blood samples due to a variety at a cutoff point of 1, specificity or negative predictive value

www.aacrjournals.org 6985 Clin Cancer Res 2009;15(22) November 15, 2009 Downloaded from clincancerres.aacrjournals.org on September 24, 2021. © 2009 American Association for Cancer Research. Published OnlineFirst November 3, 2009; DOI: 10.1158/1078-0432.CCR-09-1095

Imaging, Diagnosis, Prognosis of the CTC test for prediction of distant metastasis was not sensitivity in identification of CTCs, as a pilot study showed 100% (83.0% and 89.7%, respectively), indicating that that CTCs were detected in most blood samples taken from whole-body PET scanning and brain magnetic resonance im- patients with a variety of malignant tumor including lung, aging (or CT) are still necessary even when the CTC test was prostate, pancreatic, breast, and colon cancer (18). Of noted, negative. The most important advantage of the CTC test may CTCs were detected in all blood samples (n = 55) taken from be its capability of detecting of micrometastasis undetectable non–small cell lung cancer patients, and the mean CTC count with routine diagnostic modalities. In the present study, CTCs was very high (155 cells/mL) compared with that in the pres- were detected in 17 (19.3%) of 88 c-stage I patients, and the ent study. When the CTC chip can be available in future, a presence of distant metastasis was actually confirmed in re- comparative study between the CellSearch system and the sected specimens in 2 of 17 CTC-positive c-stage I patients. CTC chip system may be conducted. In parallel with a longer The most interesting and important question is whether dis- follow-up of patients enrolled in the present study, we are tant metastasis may develop after thoracotomy in CTC- now conducting following studies to reveal biological and positive patients without detectable metastasis. CTCs detected clinical significance of CTCs: (a) in vitro and in vivo studies in clinically “nonmetastatic” patients may represent true CTCs to assess viability and capability to form metastatic foci of or nonmalignant cells that are contaminated or incorrectly CTCs captured from lung cancer patients, (b) molecular stud- identified as CTCs, which may be answered after a longer fol- ies to reveal biological characteristics such as mutations in the low-up of patients. In addition, basic research studies should epidermal growth factor receptor gene of CTCs, (c) transla- be done to reveal whether circulating epithelial cells defined tional studies to evaluate clinical value of the CTC test as a as “CTCs” are true tumor cells and whether CTCs can actually surrogate of therapeutic effect, and (c) clinical studies to as- grow at distant organs to form metastatic foci. More experi- sess the nature of CTCs in pulmonary venous as mentioned mental and clinical studies are needed before the CTC test in the “Patients and Methods” section of the article (12). can be used in clinical practice for decision making of thera- In conclusion, CTC is a surrogate of distant metastasis in py of lung cancer patients. primary lung cancer patients, which can be useful in clinical The CellSearch system, used in the present study, is the on- practice. Further studies to confirm clinical value of the CTC ly commercially available system for detection and identifica- test are warranted. tion of CTCs. However, the most critical issue in the use of the CellSearch system may be its low sensitivity for detecting Disclosure of Potential Conflicts of Interest CTCs, which suggest a need for a more sensitive detection system. Recently, a novel microfluidic platform for detecting No potential conflicts of interest were disclosed. CTCs (“CTC chip”) has been developed (18, 19). This CTC chip consists of an array of 78,000 microposts coated with Acknowledgments anti-EpCAM antibodies, and CTCs are captured by interaction of these cells with the EpCAM-coated microposts under lam- We thank Masami Shiiba and Mayo Yamamoto for helpful assistance for inar flow conditions. The CTC chip may provide a higher preparation of the manuscript.

References 1. Socinski MA, Crowell R, Hensing TE, et al. Treat- ship of circulating tumor cells to tumor re- in non-small cell lung cancer. Ann Thorac Surg ment of non-small cell lung cancer, stage IV. Di- sponse, progression-free survival, and overall 2004;78:1001–10. agnosis and management of lung cancer: survival in patients with metastatic colorectal 14. Matsuguma H, Nakahara R, Igarashi S, et al. evidence-based clinical practice guidelines (2nd cancer. J Clin Oncol 2008;26:3213–21. Pathologic stage I non-small cell lung cancer edition). Chest 2007;132:277–89s. 8. Sastre J, Maestro ML, Puente J, et al. Circulat- with high levels of preoperative serum carci- 2. Alix-Panabières C, Riethdorf S, Pantel K. Circu- ing tumor cells in colorectal cancer: correlation noembryonic antigen: clinicopathologic charac- lating tumor cells and bone marrow microme- with clinical and pathological variables. Ann teristics and prognosis. J Thorac Cardiovasc tastasis. Clin Cancer Res 2008;14:5013–21. Oncol 2008;19:935–8. Surg 2008;135:44–9. 3. Alland WJ, Matera J, Miller MC, et al. Tumor 9. de Bono JS, Scher HI, Montgomery RB, et al. 15. Hanley JA, McNeil BJ. The meaning and the cells circulate in the peripheral blood of all Circulating tumor cells predict survival benefit use of the area under a receiver operating major carcinomas but not in healthy or patients from treatment in metastatic castration-resistant characteristic (ROC) curve. Radiology 1982; with nonmalignant diseases. Clin Cancer Res prostate cancer. Clin Cancer Res 2008;14:6302–9. 143:29–36. 2004;10:6897–904. 10. Naoe M, Ogawa Y, Morita J, et al. Detection of 16. The American Thoracic Society and The Euro- 4. Riethdort S, Fritsche H, Muller V, et al. Detection circulating urothelial cancer cells in the blood pean Respiratory Society. Pretereatment evalua- of circulating tumor cells in peripheral blood of using the CellSearch system. Cancer 2007;109: tion of non-small cell lung cancer. Am J patients with metastatic breast cancer: a valida- 1439–45. Respir Crit Care Med 1997;156:320–2. tion study of the CellSearch System. Clin Cancer 11. Hiraiwa K, Takeuchi H, Hasegawa H, et al. Clin- 17. Molina JR, Yang P, Cassivi SD, Schild SE, Adjei Res 2007;13:920–8. ical significance of circulating tumor cells in AA. Non-small cell lung cancer: epidemiology, 5. Cristofanilli M, Budd GT, Ellis MJ, et al. Circulat- blood from patients with gastrointestinal can- risk factors, treatment, and survivorship. Mayo ing tumor cells, disease progression, and surviv- cers. Ann Surg Oncol 2008;15:3092–100. Clin Proc 2008;83:584–94. al in metastatic breast cancer. N Engl J Med 12. Okumura Y, Tanaka F, Yoneda K, et al. Circu- 18. Nagrath S, Sequist LV, Maheswaran S, et al. 2004;351:781–91. lating tumor cells in pulmonary venous blood of Isolation of rare circulating tumour cells in can- 6. Cristofanilli M, Hayes DF, Budd GT, et al. Circu- primary lung cancer patients. Ann Thorac Surg cer patients by microchip technology. Nature lating tumor cells: a novel prognostic factor for 2009;87:1669–75. 2007;450:1235–9. newly diagnosed metastatic breast cancer. J Clin 13. Okada M, Nishio W, Sakamoto T, et al. Effect of 19. Maheswaran S, Sequist LV, Nagrath S, et al. Oncol 2005;23:1420–30. histologic type and smoking status on interpre- Detection of mutations in EGFR in circulating 7. Cohen SJ, Punt CJ, Iannotti N, et al. Relation- tation of serum carcinoembryonic antigen value tumor cells. N Engl J Med 2008;359:366–77.

Clin Cancer Res 2009;15(22) November 15, 20096986 www.aacrjournals.org Downloaded from clincancerres.aacrjournals.org on September 24, 2021. © 2009 American Association for Cancer Research. Published OnlineFirst November 3, 2009; DOI: 10.1158/1078-0432.CCR-09-1095

Circulating Tumor Cell as a Diagnostic Marker in Primary Lung Cancer

Fumihiro Tanaka, Kazue Yoneda, Nobuyuki Kondo, et al.

Clin Cancer Res Published OnlineFirst November 3, 2009.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-09-1095

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/early/2009/11/02/1078-0432.CCR-09-1095. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.