Circulating tumor cells as a predictive biomarker in patients with small cell lung cancer undergoing chemotherapy

Accepted Manuscript Title: Circulating tumor cells as a predictive biomarker in patients with small cell lung cancer undergoing chemotherapy Authors: Charu Aggarwal, Xingmei Wang, Anjana Ranganathan, Drew Torigian, Andrea Troxel, Tracey Evans, Roger B. Cohen, Bhavesh Vaidya, Chandra Rao, Mark Connelly, Anil Vachani, Corey Langer, Steven Albelda PII: DOI: Reference:

S0169-5002(17)30447-6 http://dx.doi.org/doi:10.1016/j.lungcan.2017.08.008 LUNG 5445

To appear in:

Lung Cancer

Received date: Revised date: Accepted date:

18-4-2017 1-8-2017 7-8-2017

Please cite this article as: Aggarwal Charu, Wang Xingmei, Ranganathan Anjana, Torigian Drew, Troxel Andrea, Evans Tracey, Cohen Roger B, Vaidya Bhavesh, Rao Chandra, Connelly Mark, Vachani Anil, Langer Corey, Albelda Steven.Circulating tumor cells as a predictive biomarker in patients with small cell lung cancer undergoing chemotherapy.Lung Cancer http://dx.doi.org/10.1016/j.lungcan.2017.08.008 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Circulating tumor cells as a predictive biomarker in patients with small cell lung cancer undergoing chemotherapy Running title: CTC in Small Cell Lung Cancer

Charu Aggarwal1, Xingmei Wang2, Anjana Ranganathan1, Drew Torigian3, Andrea Troxel4, Tracey Evans1, Roger B. Cohen1, Bhavesh Vaidya5, Chandra Rao4, Mark Connelly4, Anil Vachani6, Corey Langer1 and Steven Albelda5 1 - Division of Hematology and Oncology. University of Pennsylvania, Philadelphia, PA 2 - Division of Biostatistics and Epidemiology. University of Pennsylvania, Philadelphia, PA 3– Division of Radiology. University of Pennsylvania, Philadelphia, PA 4- Division of Biostatistics Department of Population Health , New York University School of Medicine, NY 5 – Janssen Diagnostics, LLC, Huntingdon Valley, PA 6 - Division of Pulmonary and Critical Care, Thoracic Oncology Group. University of Pennsylvania, Philadelphia, PA

Text Word Count: 3730 Abstract Word Count: 253 Tables: 3 Figures: 3 Supplementary Tables: 4 Supplementary Figure: 4 Corresponding Author: Charu Aggarwal, MD, MPH 1

Assistant Professor University of Pennsylvania, Department of Medicine Hematology-Oncology Division 10-137 South Pavilion, Perelman Center for Advanced Medicine 3400 Civic Center Boulevard Philadelphia, PA 19104 Tel: 215-662-6318 | Fax: 215-349-5326 Supported by Grant from American Cancer Society, IRG-78-002-31

2

Highlights 

CTCs are detectable at baseline in SCLC and correlate with stage and outcomes



CTCs were measured at baseline, prior to each chemotherapy cycle, and at relapse



CTCs added independent prognostic information for patients undergoing chemotherapy



The first trial to prospectively evaluate markers of both DNA damage and apoptosis on CTC



We show that assessment of these biomarkers on CTCs was feasible

Abstract Background: There are no biomarkers for assessment of disease burden or activity of therapy in SCLC.

Patients and Methods: We conducted a prospective study enumerating serial CTCs in patients with newly diagnosed limited disease (LD) and extensive stage (ED) SCLC. CTCs demonstrating DNA damage and apoptosis based on γH2AX and M30 staining were also assessed. We correlated CTC number with disease stage, survival outcomes and tumor burden by RECIST.

Results: Between 03/2011-10/2013, 50 evaluable patients were enrolled (20 LD). Baseline CTC number was higher for ED (median CTC 71 vs. 1.5 for LD; p 0.0004). Patients with < 5 CTC had longer PFS but not OS (11 vs. 6.7 months, p 0.0259 and 15.5 vs. 12.9 months, p 0.4357). A higher cutoff (CTC <50 or CTC ≥ 50) was significantly correlated with both OS (20.2 vs. 11.8 months, p 0.0116) and PFS (10 vs. 4.8 months, p 0.0002). Patients with < 5 CTC on day 1 of cycle 2 had longer PFS (10 vs. 3.17 months, p <0.001) and OS (18 vs. 9 months, p 0.0001). Patients with an increase in 2HAX-positive CTCs after chemotherapy had longer OS compared to patients without an increase (25.3 vs. 9 months, p 0.15).

3

Conclusions: This study demonstrates that CTCs at baseline and Cycle 2 of chemotherapy correlate with disease stage and survival in patients with SCLC, suggesting that CTCs may be used as a surrogate biomarker for clinical response. Confirmatory prospective clinical trials are needed before we can incorporate routine evaluation of CTCs into clinical practice. Key words: Small Cell Lung Cancer, CTC, biomarker

Introduction Small Cell Lung Cancer (SCLC) accounts for approximately 13% of all lung cancer cases 1,2 Most SCLC patients present with distant metastases and have a poor prognosis. All previous drug development strategies in SCLC have been typified by therapeutic empiricism without a sophisticated strategy for patient selection. Better biomarkers to identify patients destined to do well or poorly as early as possible in the treatment course could be very useful to accelerate the development of new agents. Given the difficulty in obtaining enough analyzable tissue from SCLC patients, there is an unmet need for a noninvasive biomarker that is prognostic and/or predictive of benefit with new therapeutic agents.

One simple user-friendly way to non-invasively obtain cancer cells for some cancer types is through analysis of circulating tumor cells (CTCs). CTCs can be detected by “CellSearch®”, that enriches and enumerates CTCs utilizing an EpCAM-immunoferrofluid magnetic separation and differential staining of circulating white blood cells vs. cells of epithelial origin by selecting cells that are negative for CD45 (a pan leukocyte antibody) and positive with pan cytokeratin (CK) and nuclear material (DAPI)3-8.

CTCs in SCLC Although it has been very difficult to identify CTCs in non-small cell lung cancer9, this does not appear to be the case for SCLC. Hou et al. reported detection of CTCs in 85% of patients with SCLC using CellSearch®; CTC number fell following chemotherapy10. In another study, Naito et al. also detected ≥ 2 CTCs/ 7.5 ml in 68% of patients with SCLC, and found that patients with <8 CTC lived longer than those with ≥8 CTC at baseline, after treatment, and at the time of 4

relapse11. Normanno et al and Hilermann et al., each individually reported detection of at least one CTC in 90% and 84% of patients with SCLC at baseline12-14. Change in CTC number following chemotherapy provided additional clinical information in the Normanno study (a significant decline in CTC count, after one cycle of chemotherapy was associated with a lower risk of death (HR 0.24, 95% CI 0.09–0.61))12. Similarly, Hiltermann and colleagues showed that CTC count after one cycle of chemotherapy served as the strongest predictor of overall survival (HR 5.7; 95% CI 1.7–18.9; p = 0.004). Together these studies demonstrate that capture of CTCs by CellSearch® at baseline is feasible, reproducible, and predictive of response to chemotherapy. However, practical application of CTCs in SCLC remains limited as the previous studies have been restricted to certain subsets of SCLC (for example, the Normanno et al. study evaluated patients with extensive stage disease alone) and variable CTC cutoffs have been utilized for prognostic determination.

CTCs as Biomarkers In addition to being able to enumerate CTCs, CellSearch® also allows one to obtain additional pharmacodynamic information using immunohistochemical markers. The key criteria for such markers are that an antibody exists that can be used to stain cells on the CellSearch® platform and that the marker should be a validated measure of chemotherapy-induced DNA damage and/or cell death/apoptosis. Based on previous literature using, CellSearch® , two such makers that fit these criteria are the DNA damage marker γH2AX15,16 and apoptosis marker M3017. 18.

Phosphorylated H2AX (γH2AX) is a marker of DNA damage and appears in the nucleus within minutes in a dose-dependent manner in cells treated with cytotoxic chemotherapy15. Wang et al demonstrated that the percentage of γH2AX-positive cells increased in epithelial tumor cell lines [MCF7 (human breast adenocarcinoma), PC-3 (human prostate adenocarcinoma), HT-29 (human colorectal adenocarcinoma), and SKOV-3 (human ovarian adenocarcinoma)] treated with therapeutic concentrations of topotecan ex vivo. In vivo, the percent of γH2AX-positive CTCs increased post-treatment from a mean of 2% at baseline to 38% after a single day of chemotherapy in patients with a variety of advanced malignancies enrolled in phase I clinical trials16. 5

M30 antibody recognizes a caspase-cleaved neoepitope of Cytokeratin 18 that is only revealed during apoptosis. M30 positive CTCs can be detected in patients with various malignancies. Rossi et al sequentially assessed CTCs and M30-positive CTCs in breast cancer patients. Overall the number of total and M30-positive CTC decreased during treatment in six and increased in two of eight patients. They suggested that changes in the number of M30 positive CTCs may predict response to therapy17,18.

We designed this study to prospectively assess the relationship of CTCs to disease stage, and survival, at baseline and prior to each chemotherapy cycle. In addition, our study sought to evaluate the applicability of CTCs as a potential, non-invasive, pharmacodynamic biomarker.

Patients and Methods Study Design We conducted a prospective trial at Abramson Cancer Center of the University of Pennsylvania, Philadelphia. Adult patients with a diagnosis of small cell lung cancer with measurable disease, receiving first line chemotherapy or chemoradiation, were included. All patients signed informed consent. The study was reviewed and approved by the Institutional Review Board at the University of Pennsylvania.

Peripheral blood was collected for CTC evaluation before the initiation of therapy (baseline) and Days 2 and 3 of chemotherapy for cycles 1, 2 and Day 1 of chemotherapy for Cycles 3 and 4. Another blood sample was drawn at the time of relapse, before initiation of salvage therapy. Blood samples were drawn into 10-mL evacuated tubes (CellSave, Janssen Diagnostics, LLC, Raritan, NJ). All CTC evaluations were performed on the CellSearch System (Janssen Diagnostics LLC, Raritan, NJ) without knowledge of patient clinical status. Computed tomography scans of the chest, abdomen and pelvis or positron emission technology (PET) scans were performed at baseline and every 6 to 12 weeks after initiating treatment, at the discretion of the treating clinician. Detection of 2HAX and M30 6

The samples were processed on the AUTOPREP using CellSearch® Epithelial Cell Kit (Janssen Diagnostics, Raritan, NJ) and then analyzed on the CellTracks® Analyzer II®. The CTCs were stained with the anti-2HAX-FITC antibody (Millipore (catalog # 16-202A, clone JBW301) at a final concentration of 1ug/ml. Monoclonal antibody M30 (Peviva, Stockholm, Sweden) was used at a final concentration of 0.05ug/ml to enumerate CTCs from patient samples and stained with anti M30-PE.

Radiographic Assessments Tumor measurements, and response assessments, all using RECIST 1.1, were performed by a dedicated thoracic radiologist (D.T.) on the baseline and follow-up scans. The radiologist was aware that these patients were undergoing CTC collection but this individual was blinded to the results.

Statistical Analysis The objective of this analysis was to assess the prognostic role of CTCs at baseline, prior to each chemotherapy cycle, with the additional exploratory goal to evaluate the role of CTCs as pharmacodynamic biomarkers. CTC numbers were characterized using mean, standard deviation (SD), median, and range. We correlated CTC number with disease stage and number of metastatic sites (mets), using non-parametric rank-based tests for association. CTC numbers were also correlated with response to therapy, progression-free survival (PFS) and overall survival (OS). PFS was calculated from the date of diagnosis until the earlier of date of progression or death. OS was calculated from the date of diagnosis until date of death or the date of last follow-up. Patients were censored at last follow-up if death had not occurred. Survival curves were compared using log-rank test. Cox proportional hazards regression was used to determine univariate and multivariate hazards ratios for OS. OS and PFS were analyzed for the overall population, and also stratified by stage alone, by dichotomized baseline CTC value alone (cutoffs set at 5, 10 and 50 CTCs respectively). These exploratory cutoffs were based on previously published results of CTC and analyses from various other disease groups including, but not limited to breast cancer, colorectal cancer, and small cell lung cancer7,8,13,14.

7

Percentages of total CTCs expressing markers of 2HAX or M30 were analyzed at baseline, and during consecutive chemotherapy days in a given cycle. PFS and OS were stratified by an increase of 2HAX or M30 during consecutive chemotherapy days in a given cycle.

Results Patient Characteristics Between March 2011 and October 2013, 50 patients were enrolled. Patient characteristics are displayed in Table 1. At the time of this analysis, 37 patients had disease progression (76%) and 33 patients (66%) had died.

SCLC and overall survival The average length of follow up for all patients on the study was 15.8 months. Median PFS of the entire population was 7.3 months, and median OS was 14.9 months (Figure 1A and 1B). Median PFS was significantly higher for LD compared with ED (12.7 months vs. 5.8 months, p value <0.0001). OS for LD patients (27 months) was also significantly higher than in patients with ED (14.1 months, p value 0.0048), (Figure 1C and 1D).

CTC numbers at baseline CTCs were present in 47 patients (94%) at baseline before chemotherapy. Median CTC number was 7 cells/7.5 ml. The number of patients with detectable CTCs was successively lower with each day of therapy (n=47 with CTC on cycle 1 day 1, n=43 on cycle 1 day 2 and n=37 on cycle 1 day 3). CTC number at baseline, (cycle 1 day 1) strongly correlated with disease stage. Whereas the median CTC number in ED patients was 91, the median CTC number was only 1.5 in LD patients (p value 0.0004). Twenty-six patients had ≥ 5 CTCs at baseline (7 with LD and 19 with ED). Amongst the 27 patients with ED, median baseline CTC count was higher in patients with ≥3 (n=10) compared to 1-2 sites of metastatic disease (n=17), although this difference did not quite achieve statistical significance (1267.5 vs. 33, p-value 0.06).

Prognostic Significance In addition to correlation with stage, there was a statistically significant relationship to PFS. For patients with < 5 CTCs at baseline, median PFS was 11 months; for patients with ≥ 5 CTCs, it 8

was 6.7 months (p = 0.0259) (Figure 2A). There was no significant difference in OS based on baseline CTC number. For patients with < 5 CTCs at baseline, median OS was 15.5 months whereas for patients with ≥ 5 CTCs, OS was 12.9 months (p=0.4357) (Figure 2B).

We conducted similar analyses using different CTC cutoffs (5, 10 and 50 CTCs, Table 2). For patients with < 50 CTCs, PFS and OS were both significantly longer compared to CTCs ≥50 (PFS 10 vs. 4.7 months, p value= 0.0002, OS 20.2 vs. 11.8 months, p value= 0.0116).

CTC numbers on Day1, Cycle 2 of chemotherapy CTC numbers dropped after chemotherapy. Forty-one patients had detectable CTCs on day 1 of cycle 2, and fewer patients had detectable CTCs on day 2 and day 3 of cycle 2 (n= 36 and n=34 respectively). On day 1 of cycle 2, thirty-three patients had CTCs < 5, and 8 patients had CTCs ≥ 5. Median PFS and OS were significantly better for patients with <5 CTCs at this time point (p<0.0001 and p=0.0004 respectively), (Figures 2C and 2D). Using higher cut-offs, (i.e., CTC cutoffs of 10 and 50), we observed a significant correlation of cycle 2 CTCs to OS and PFS (Table 2).

Correlation of CTCs within disease sub groups (LD and ED) CTCs were correlated within individual stage sub-groups (LD; 20 patients, ED; 30 patients) to OS and PFS. Baseline CTCs did not correlate with disease stage (Supplemental Table 1). Within the ED group, patients who had < 5 CTCs on day 1 of cycle 2 had longer OS (16 vs. 9 months, p=0.0008), and PFS (7 vs. 3 months, p<0.0001), Supplemental Table 1.

Correlation of tumor burden and CTCs Tumor burden was assessed through independent radiology review using RECIST 1.1. We correlated CTCs with RECIST measurements at baseline, stage of disease and response to therapy. For 44 patients, tumor measurements were inversely related to response to therapy (Kruskal-Wallis Test p=0.036), and directly related to stage of disease at presentation (KruskalWallis Test p=0.01). CTCs at baseline also correlated directly with tumor measurements at baseline; larger tumors or those patients with greater tumor burden had a greater number of CTCs at baseline (Spearman correlation coefficient, 0.45, p=0.0021). On the other hand, CTCs 9

at baseline were not related to response (Kruskal-Wallis Test p=0.61), (Supplemental Figures 1 A, B, C).

Serial CTCs while on chemotherapy Serial CTC data were available on 36 pts. There was a consistent decline in the number of CTCs detectable during the course of therapy. The average decrease from cycle 1 to cycle 2 was 43% (median 98% decrease). Patients with any decline in CTC number were included in the analysis of survival outcomes. Patients with a decrease in CTCs during successive days on C1 (i.e., days 1, 2 and 3) had numerically longer OS compared to patients without a decrease, but this was not statistically significant (17.6 months vs. 10.7 months, p=0.72); PFS was comparable in patients with and without a decrease in CTCs (11 vs. 7.2 months, p=0.45), (Figures 3A and 3B). Decrease in CTCs was not significantly correlated with response (Fisher’s exact test p=0.26). Kinetics of CTC decrease is shown in Supplemental Figure 2 (2A and 2B).

Univariate and Multivariate Cox Regression Multivariate Cox regression analyses were conducted, adjusting for age, ECOG performance status, extent of disease (ED vs. LD), number of metastatic sites, and baseline CTCs. Older age (p=0.032) and higher baseline CTCs (p=0.043) were associated with a higher hazard of death. Similar multivariate analyses for PFS showed that ED and higher baseline CTCs were predictive of a higher hazard of progression (p=0.005 and p=0.015 respectively), Table 3. Changes in H2AX and M30 expression. To validate 2HAX staining as a measure of chemotherapy-induced DNA damage in small cell lung cancer cells, we first treated human H69 lung cancer cells with etoposide to induce DNA damage and stained the cells with an anti- 2HAX antibody. As shown in Supplemental Figure 3, the low basal level of expression of 2HAX (Supplemental Figure 3A) was markedly increased and easily detectable after 24 (Supplemental Figure 3B) and 48 (Supplemental Figure 3C) hours of etoposide treatment. We next took control H69 cells or etoposide-treated H69 cells and “spiked” these into 7.5 ml whole human blood, which was then put into a CellSave® tube. The sample was processed using the standard clinical CellSearch® instrument. The unstained channel was stained with the FITC- anti- 2HAX antibody. 10

As shown in

Supplemental Figure 3D, CellSearch® readily identified the small cell lung cancer cells (DAPI+/cytokeratin+/CD45-). Whereas none of the control cells stained positively for 2HAX (Supplemental Figure 3D, upper panels), more than 90% of the positive cells in the etoposidetreated H69 cells stained positively for 2HAX (Supplemental Figure 3D, lower panels). These studies demonstrate that CellSearch® was able to identify small cell lung cancer cells with chemotherapy-induced DNA damage.

We also validated the staining of M30. Apoptotic MCF7 cells were spiked into whole blood, which was processed as above, but stained with anti-M30 antibody. The apoptotic MCF7 cells (CK+/Dapi+/CD45-) were uniformly positively for M30 (Supplemental Figure 3E upper panel), while the non-apoptotic contaminating leukocytes (CD-/Dapi+/CD45+) were negative for M30 staining (Supplemental Figure 3E lower panel). We were able to count CTCs expressing 2HAX in 19 patients during Cycle 1 (Supplemental Table 2).

At baseline, the percentage of 2HAX staining cells was generally low

(Supplemental Figure 4); the percent of the CTCs that expressed 2HAX averaged 5%, with a median value of 3%. For unclear reasons, two patients (Patient 2 and Patient 19) showed very high (20% and 22%) basal expression levels. After treatment during Cycle 1 (Day 1 or Day 2), 9 of the 18 patients showed an increase in the percent of 2HAX-positive cells (Supplemental Figure 4 and Supplemental Table 2). The group of patients with an increase in expression of 2HAX-positive cells had a numerically longer OS compared to patients without a significant increase (25.3 vs. 9 months, p=0.15). No difference was noted in PFS (4.8 vs. 4.6, p=0.5) We were able to assess M30 levels at baseline in 11 patients (Supplemental Table 3). At baseline, the expression of M30-positive cells averaged 9% with a median of 7% (Supplemental Figure 4). Three patients had baseline values greater than 10%.

After treatment, 6 of the 11

patients showed an increase in the percent M30 positive CTCs (Supplemental Figure 4). The group of patients with an increase in expression of -positive cells had a similar OS and PFS compared to patients without a significant increase in M30 (OS, 9 vs. 11.83 months, p=0.9; PFS 4.4 vs. 4.87 months, p= 0.8). We compared the patients who had significant changes in 2HAX-

11

positive cells versus those with changes in M30 (Supplemental Table 4) and saw no clear correlation.

Discussion Our single center prospective study confirms that CTCs can be detected and followed serially in most patients with SCLC using the CellSearch® technology. CellSearch® is the most widely used method for CTC detection. Despite widespread availability and common use in other malignancies like breast cancer and prostate cancer, its utility in patients with SCLC is not well established. CTCs have been evaluated as a biomarker in SCLC in several other studies10,11,13,18. Similar to these reports, we were able to detect CTCs in 80% of patients at baseline. Using a cut off of 5 CTCs, we demonstrated a correlation with PFS. Thus, our data corroborate earlier observations that CTCs correlate with disease stage, and PFS. Additionally, we show that CTC number is directly related to tumor measurements and number of metastatic sites. Patients with higher CTCs at baseline and at cycle 2 of chemotherapy had inferior survival outcomes. Unlike other studies that evaluated CTCs before and after a single cycle of chemotherapy, our study measured CTCs serially during different days of each cycle of chemotherapy (through 4 cycles). We found that in a majority of the patients the number of CTCs declined rapidly after the initial cycle of chemotherapy, often with CTCs disappearing as early as Day 3 of chemotherapy. This suggests that CTCs could be used as an early indicator of treatment efficacy when testing an experimental agent in this disease without having to wait weeks or months for scan results. Additionally, we used prospective, independent radiographic review and RECIST measurements to calculate date of progression and estimate tumor burden. Compared to earlier studies that used limited cutoffs (cutoffs of 8 or 50 for e.g.), we explored additional cutoff values, including CTCs greater than or equal to 10, and greater than or equal to 50. The latter cutoff appeared to be the most prognostic with respect to both PFS and OS. All patients in our study received standard chemotherapy regimens (platinum and etoposide) with median OS and PFS similar to historical data and to those observed in the real world setting, thereby reducing the possibility of selection bias.

12

Our study also prospectively evaluated the role of CTCs as pharmacodynamic biomarkers. To our knowledge, this is the first trial in SCLC to evaluate markers of DNA damage and apoptosis in response to chemotherapy, and correlate these with survival. We had hypothesized that the presence of CTCs expressing increased levels of 2HAX or M30 immediately after chemotherapy would reflect a good response to the administered chemotherapy and correlate with increased PFS or OS. We were not able to confirm this hypothesis. We saw no significant correlations with M30 expression levels and could only detect a trend (p=0.015) towards improved overall survival for patients with an increase in percentage of 2HAX expressing CTCs. This may reflect our small sample size, the fact that CTCs might not accurately reflect intratumoral changes in cell viability at these early time points, or that these markers may not ideal. Although negative, our study did show the feasibility of assessing CTCs for potential biomarkers. Additional studies will hopefully identify ways that CTCs could be utilized as a novel, non-invasive way to assess drug effect from chemo- and other targeted therapies, and assess biologic proof of concept in early drug development.

Despite significant improvements in treatment of non-small cell lung cancer over the past 2 decades, the treatment of SCLC sadly has not changed. Given the difficulty in obtaining tissue from SCLC patients, noninvasive biomarkers that are both prognostic and predictive of benefit could accelerate the development of new therapeutic agents. Our study adds to the available literature suggesting that CTCs are detectable at baseline, and reinforces the following observations from previous studies10-13: 1) a higher CTC count is seen in patients with ED SCLC, 2) CTC count correlates with disease burden and stage, 3) a reduction CTC count occurs following chemotherapy, 4) CTC count at baseline and cycle 2 correlates with response to chemotherapy, and 5) CTCs can be interrogated for biomarkers of treatment effect. These observations validate the role of CTCs as surrogate markers of response to therapy, and as tools to predict outcomes post treatment, and may also serve as a predictor of recurrence.

Our study raises a number of important questions for future study. Should CTC measurements guide treatment decision-making in the absence of radiographic documentation? At the current time, the answer to this question must be no. However, clinical trials to assess the potential utility of an early change in therapy based upon poor response by CTC analysis are certainly 13

warranted given the dismal outcomes in patients who fail to respond to their initial platinum based chemotherapy. Further, we need to develop technologies to increase isolation and accurate characterization of CTCs that may enable us to identify patients with evidence of metastatic spread before radiographic or clinical evidence of progressive disease. This would be especially helpful to select treatment regimens and set realistic goals of therapy.

This study has several limitations that must be acknowledged. For one, this single center study included a relatively small number of patients for analysis, which likely limited our statistical power. Our study also had slightly more female patients compared to other studies although the overall PFS and OS that we observed was very similar to reports in the literature for SCLC. Second, we used a CTC cutoff value (<5 or >=5) that has been used on other CTC trials in breast and colorectal cancer, but this cutoff may not be optimal for SCLC. Finally, many patients on our study did not have detectable CTCs by Day 3 or with subsequent chemotherapy administration. Only 36 patients had CTC data available for serial interpretation making it difficult to perform and interpret serial assays for DNA damage beyond day 2.

In conclusion, our study demonstrates that CTCs add prognostic information for patients undergoing chemotherapy for SCLC. The clinical applicability of CTCs to inform treatment decision-making requires further prospective study. Consideration should be given to incorporation of CTC molecular analysis, including CTC DNA and cell free DNA, in future prospective randomized trials. Conflicts of Interest Disclosure In the interest of full transparency, on behalf of my co-authors and myself we would like to disclose that there are no potential conflicts of interest and all authors have read and approved the manuscript.

14

References:

1.

Puglisi M, Dolly S, Faria A, et al: Treatment options for small cell lung cancer -

do we have more choice? Br J Cancer 102:629-38, 2010 2.

Jemal A, Bray F, Center MM, et al: Global cancer statistics. CA Cancer J Clin

61:69-90, 2011 3.

Riethdorf S, Fritsche H, Muller V, et al: Detection of circulating tumor cells in

peripheral blood of patients with metastatic breast cancer: a validation study of the CellSearch system. Clin Cancer Res 13:920-8, 2007 4.

Racila E, Euhus D, Weiss AJ, et al: Detection and characterization of carcinoma

cells in the blood. Proc Natl Acad Sci U S A 95:4589-94, 1998 5.

Gupta GP, Massague J: Cancer metastasis: building a framework. Cell 127:679-

95, 2006 6.

Yilmaz M, Christofori G: Mechanisms of motility in metastasizing cells. Mol

Cancer Res 8:629-42, 2010 7.

Cohen SJ, Alpaugh RK, Gross S, et al: Isolation and characterization of

circulating tumor cells in patients with metastatic colorectal cancer. Clin Colorectal Cancer 6:125-32, 2006 8.

Cristofanilli M, Budd GT, Ellis MJ, et al: Circulating tumor cells, disease

progression, and survival in metastatic breast cancer. N Engl J Med 351:781-91, 2004 9.

Allard WJ, Matera J, Miller MC, et al: Tumor cells circulate in the peripheral

blood of all major carcinomas but not in healthy subjects or patients with nonmalignant diseases. Clin Cancer Res 10:6897-904, 2004 10.

Hou JM, Krebs MG, Lancashire L, et al: Clinical Significance and Molecular

Characteristics of Circulating Tumor Cells and Circulating Tumor Microemboli in Patients With Small-Cell Lung Cancer. J Clin Oncol, 2012 11.

Naito T, Tanaka F, Ono A, et al: Prognostic Impact of Circulating Tumor Cells in

Patients with Small Cell Lung Cancer. J Thorac Oncol, 2012 12.

Normanno N, Rossi A, Morabito A, et al: Prognostic value of circulating tumor

cells' reduction in patients with extensive small-cell lung cancer. Lung Cancer 85:314-9, 2014

15

13.

Hiltermann TJ, Pore MM, van den Berg A, et al: Circulating tumor cells in small-

cell lung cancer: a predictive and prognostic factor. Ann Oncol 23:2937-42, 2012 14.

Hiltermann TJ, van der Wekken AJ, Groen HJ: Moving forward with circulating

tumor cells and lung cancer. J Thorac Dis 4:440-1, 2012 15.

Chowdhury D, Keogh MC, Ishii H, et al: gamma-H2AX dephosphorylation by

protein phosphatase 2A facilitates DNA double-strand break repair. Mol Cell 20:801-9, 2005 16.

Wang LH, Pfister TD, Parchment RE, et al: Monitoring drug-induced

gammaH2AX as a pharmacodynamic biomarker in individual circulating tumor cells. Clin Cancer Res 16:1073-84, 2010 17.

Rossi E, Basso U, Celadin R, et al: M30 neoepitope expression in epithelial

cancer: quantification of apoptosis in circulating tumor cells by CellSearch analysis. Clin Cancer Res 16:5233-43, 2010 18.

Smerage JB, Budd GT, Doyle GV, et al: Monitoring apoptosis and Bcl-2 on

circulating tumor cells in patients with metastatic breast cancer. Mol Oncol 7:680-92, 2013

Figure 1

Figure 2 16

Figure 3

17

18

TABLE 1. Patient Characteristics

N

Limited Stage

Extensive Stage

20

30

65 (45-79)

64 (43-80)

Age in years, Median (Range) Gender Male

7

35%

13

43%

Female

13

65%

17

57%

0

11

55%

8

27%

1

6

30%

13

43%

2

2

10%

7

23%

3

1

5%

2

7%

Current Smoker

10

50%

12

40%

Past Smoker

10

50%

15

50%

Never Smoker

0

0%

2

7%

Unknown

0

0%

1

3%

ECOG Performance Status

Smoking Status

Pack Years, Median (Range)

40 (10-70)

40 (0-145)

Number of Metastases 1

0

0%

9

30%

2

0

0%

10

33%

3

0

0%

6

20%

>3

0

0%

5

17%

Cisplatin and Etoposide

15

75%

7

23%

Carboplatin and Etoposide

5

25%

23

77%

Chemotherapy Received

19

20

Table 2. Survival Outcomes using different CTC cutoffs Outcome

Time Point

CTC Value

Months

P value

OS

Baseline

CTC < 5

15.47

0.2962

CTC ≥ 5

12.90

CTC < 10

15.47

CTC ≥ 10

14.10

CTC < 50

20.27

CTC ≥ 50

11.83

CTC < 5

17.97

CTC ≥ 5

9.03

CTC < 10

17.97

CTC ≥ 10

6.32

CTC < 50

17.57

CTC ≥ 50

6.13

CTC < 5

10.97

CTC ≥ 5

6.70

CTC < 10

10.40

CTC ≥ 10

6.70

CTC < 50

10.03

CTC ≥ 50

4.77

CTC < 5

10.03

CTC ≥ 5

3.17

Cycle 2

PFS

Baseline

Cycle 2

21

0.5692

0.0116

0.0004

<.0001

0.0002

0.0067

0.0280

0.0002

<.0001

CTC < 10

9.80

CTC ≥ 10

3.17

CTC < 50

9.57

CTC ≥ 50

2.60

<.0001

<.0001

Table 3. Multivariate Analysis Sub Outcome

Predictor

category of

HR (95_CI)

p-value

0.80(0.27 , 2.36)

0.68

PS 2

1.53(0.36 , 6.55)

0.56

PS 3

7.45(1.11 , 50.0)

0.03

predictor OS

ECOG PS

Number of Mets

PFS

global p-value

PS 1

0.11

0.77

0.83(0.22 , 3.05)

Stage (ED vs LD)

0.07

3.15(0.88 , 11.3)

Age at Diagnosis

0.03

1.06(1.00 , 1.11)

Baseline CTC

0.04

1.00(1.00 , 1.00)

0.54

0.84(0.35 , 2.03)

0.69

PS 2

1.06(0.30 , 3.76)

0.92

PS 3

4.32(0.51 , 36.4)

0.17

ECOG PS

Number of Mets

< 3 vs. ≥3

PS 1

< 3 vs. ≥3

0.86

0.91(0.29 , 2.82)

Stage (ED vs. LD)

0.005

4.01(1.51 , 10.7)

Baseline CTC

0.015

1.00(1.00 , 1.00)

22