Introduction

Relapse of non-small cell lung cancer (NSCLC) is not uncommon, even when the cancer lesion is completely resected. According to a nationwide registry conducted in Japan, approximately 15 % of pathological stage IA patients suffer from cancer recurrence within 5 years of the initial operation [1]. Such recurrence after complete resection of pulmonary carcinoma is speculated to arise from residual isolated tumor cells (ITCs) in the body. Some ITCs may exist prior to surgery, while others may be derived from circulating tumor cells (CTCs), some of which may be dislodged from the tumor into the peripheral blood during the operation. We herein evaluated the CTC status in patients who underwent pulmonary resection in order to assess the clinical implications.

Methods

Background of patients

In the setting of a referral community healthcare hospital, 23 patients with NSCLC who underwent an anatomical pulmonary resection procedure (1 segmentectomy, 21 lobectomies, and 1 pneumonectomy) (10 males, 13 females; median age 68 years) from June 2014 to October 2014 were enrolled in this clinical study (Table 1). The histological diagnosis was adenocarcinoma in 17 cases, squamous cell carcinoma in 4, and pleomorphic carcinoma in 2; the clinical stage was IA in 15 cases, IB in 3, IIA in 2, IIB in 2, and IIIA in 1; and the pathological stage was IA in 14 cases, IB in 2, IIA in 2, IIB in 2, IIIA in 2, and IV in 1. The patient with clinical IIIA lung cancer underwent induction chemotherapy with 4 cycles of carboplatin and taxotel. Patient characteristics, tumor size [whole size shown by computed tomography (CT), size of the solid component shown by CT, gross size], tumor markers [carcinoembryonic antigen (CEA), cytokeratin fragment A (CYFRA)], maximum standardized uptake value (SUVmax) in the tumor shown by fluorodeoxyglucose positron emission tomography (FDG-PET), type of operation, tumor histology findings, and genetic mutation results are shown in Table 1.

Table 1 Clinical-pathological characteristics
Full size table

Follow-up examinations

The median follow-up period was 12 months (range 5–15 months) and the last follow-up examination was in August 2015. All patients were evaluated at 1-month intervals, which included physical and chest X-ray examinations, and blood tests including tumor markers, with additional thoraco-abdominal CT scans generally performed at 6-month intervals. Mortality and recurrence data were collected by the primary physician (NS).

Detection of CTCs

The detection of CTCs was performed during the operation and postoperative period. Peripheral atrial blood (3 mL) was collected into an EDTA tube before surgery (Before), immediately after resection of the lung (After), and 6 h after the end of surgery (6 h). In addition, pulmonary vein (PV) blood (3 mL) from the resected lung was collected and stored in the same fashion. Tumor cells in the 4 blood samples from each patient were simultaneously extracted using a ScreenCell® CTC selection kit with the size selection method using a micro-pore film that extracts formalin-fixed tumor cells. The extracted cells were stained using a hematoxylin–eosin method and observed with a light microscope, using a previously described method [2]. Tumor cells were classified into 3 categories: no tumor cells detected (N), single cell or less than 4 cells (S), and clustered cells (C).

Statistical analyses

To analyze the relationships between the clinicopathological findings and the CTC status, we examined the following parameters: pathological stage (p-IA vs. greater), whole tumor size (cm) on CT images, solid component size (cm) on CT images, pattern of CT appearance (pure solid vs. glass or ground-glass opacity + solid and solid), serum CEA and CYFRA levels [mean ± standard deviation (SD)], and SUVmax values on fluorodeoxyglucose positron emission tomography (FDG-PET) images (mean ± SD). Patients were classified into 3 categories according to their CTC status: Before-C and After-C, Group I; Before S or N and After C, Group II; and Before S or N and After S or N, Group III. This classification was used because clustered cancer cells in the blood have been reported to be an indicator of a poor prognosis [2, 4]. Statistical analyses were performed using a commercially available statistics software program (StatView 5.0, Abacus Computer, Berkeley, CA). Fisher’s exact test was used to compare two dichotomous variables and the t test to compare mean values. For analyses among multiple groups, a Chi-square or Kruskal–Wallis test was used as appropriate. As for the follow-up data, the relapse-free survival curves were calculated using the Kaplan–Meier method and the log-rank test was utilized to compare two groups, when possible.

Ethical approval

This investigation was approved by the institutional review board of the Hoshigaoka Medical Center following discussions and all patients provided their informed consent to participate in this study.

Results

Detection of CTCs

CTCs were detected in both peripheral and PV samples (Figs. 1, 2). Morphological and quantitative findings of tumor cells detected before surgery, immediately after resection of the lung, at 6 h after the end of surgery, and in the PV blood from the resected lung are shown in Table 2. Seven cases were positive (30.4 %; S, 1; C, 6) among samples collected before surgery, 17 (73.9 %; S, 2; C, 15) were positive among those collected immediately after resection of the lung, 1 (4.3 %; S, 1) was positive among those collected 6 h from the end of surgery, and 19 (82.6 %; S, 3; C, 16) were positive among those collected from PV blood. All of the 17 cases with positive samples obtained immediately after resection of the lung also had positive PV samples, while 10 of those (58.8 %) were negative prior to surgery. According to the CTC status, there were 6 patients in Group I, 9 in Group II, and 8 in Group III. Regarding the status of tumor cells in the PV blood according to our classification (N, S, C), all patients with C status were C in the peripheral arterial blood immediately after resection of the lung (Table 2).

Fig. 1
figure 1

Representative case of invasive non-small cell lung cancer (case no. 11). A tumor located in the right upper lobe (a) showed a high standardized uptake value ((SUV)max = 5.4) in fluorodeoxyglucose positron emission tomography (FDG-PET) findings (b). Clustered tumor cells were observed in the peripheral arterial blood immediately after lung resection (c, hematoxylin–eosin stain) and in the pulmonary vein blood from the resected lung. Stump smear cytology reveals the tumor is adenocarcinoma (d, Papanicolaou stain). Bold arrow tumor cells

Full size image
Fig. 2
figure 2

Representative case of non-invasive adenocarcinoma (case no. 20). A tumor located in the left upper lobe showed a ground-glass opacity with no consolidation and was diagnosed as non-invasive adenocarcinoma (a) with a standardized uptake value [(SUV)max = 0] in fluorodeoxyglucose positron emission tomography (FDG-PET) findings (b). Abnormal alveolar cells were observed in the pulmonary vein blood from the resected lung (c, hematoxylin–eosin stain). Stump smear cytology revealed the tumor to be adenocarcinoma (d, Papanicolaou stain)

Full size image
Table 2 Morphological and quantitative findings of circulating tumor cells
Full size table

Clinical pathological results

The tumor characteristics according to the CTC status are shown in Table 3. The ratio of p-IA was greatest in Group III (100 %), followed by Group II and Group I, which was in contrast to the ratio of pure solid tumors, tumor size, CEA level, and SUVmax value. As shown in Table 4, there were 5 cases of cancer recurrence and 1 cancer death. Survival curves for all groups are shown in Fig. 3. The 1-year relapse-free survival rate was 55.6 % in Group I, 64.8 % in Group II, and 100 % in Group III, indicating no significant difference between Group I and Group II (p = 0.9).

Table 3 Comparison of clinical parameters according to change in circulating tumor cells during surgery
Full size table
Table 4 Relapse and survival
Full size table
Fig. 3
figure 3

Relapse-free survival curves. Patients were classified into the following three categories according to the CTC status: Before-C and After-C, Group I; Before S or N and After C, Group II; and Before S or N and After S or N, Group III. The 1-year relapse free survival rates for the groups were 55.6, 64.8, and 100 %, respectively. The difference between Group I and Group II was not significant (p = 0.9)

Full size image

Discussion

The detection of CTCs during surgery was initially performed using a cytological technique [5], after which polymerase chain reaction (PCR) [6, 7] and flow cytometry [8] methods were introduced to improve the sensitivity and specificity. In recent years, advancements in commercially available equipment and methods focusing on CTC detection have been developed and reports of CTC detection using clinical specimens have also been presented [2, 9].

Among the modern available CTC detection systems, the EpCAM positive selection method using magnified antibodies (CellSearch® system) was the first to show morphological evidence of CTCs during lung cancer surgical procedures [10–12]. However, the sensitivity for the detection of CTCs in peripheral blood samples obtained during lung cancer surgery is not high [11] and clustered cells might be missed by the CellSearch® system. To compensate for the low sensitivity in peripheral blood samples and for clustered cells, the ISET® system was developed [13], although the equipment is costly. In contrast, the CD45 negative selection method (RosetteSep™) does not require a specialized machine and can detect clustered CTCs perioperatively among lung cancer tumor cells [3, 4], although its sensitivity is lower than other CTC detection methods. As a result, tests for perioperative CTCs in lung cancer cases have been performed using pulmonary vein blood from resected lungs with RosetteSep™ [3, 4]. In a recently published report on the CTC status in lung cancer surgery patients that utilized an exhaustive gene analysis, however, cells were collected from the pulmonary vein blood using a flow cytometry method for EpCAM [14].

The ScreenCell® CTC selection kit is a highly sensitive size selection method that collects nearly 100 % of spiked tumor cells in the blood without the use of a specialized machine and can also detect clustered cells [2]. It is considered to be a useful method for revealing CTCs in clinical settings, although its use is limited because the detection method is subjective and overestimation of non-cancer cells that might be defined as cancerous is possible. However, because the method to determine the cell status is the same as that of clinically used cytology techniques, there is a slight possibility of overestimation.

The results of the present investigation showed that patients without clustered CTCs at the time of surgery had a higher rate of p-stage IA, lower tumor size and CEA level, and a higher relapse-free survival rate compared with those with clustered CTCs revealed during surgery. This finding may be attributed to the potential malignancy of the original cancer lesion [15, 16] as well as of clustered CTCs, which are reported to have a high ratio of epithelial–mesenchymal transition markers, such as N-cadherin and vimentin, and a low ratio of Ki-67 antigen [17]. Therefore, clustered CTCs may have the potential to become a lesion associated with recurrence.

In the present cohort, 16 patients did not have CTCs detected in the peripheral blood prior to surgery, of whom 10 (62.5 %) became CTC positive after lung resection. This finding indicates that surgical manipulation for the treatment of lung cancer can dislodge cancer cells thus allowing them to pass into the circulating blood. The presence of CTCs has been reported to be a prognostic indicator in cases of lung cancer [18], with clustered CTCs an indicator of high malignancy [17, 19, 20] and early recurrence [3, 4]. Our study results also showed that the detection of clustered CTCs during surgery for NSCLC was a marker of high malignancy and indicated early recurrence. However, our findings are limited by the small number of patients, short follow-up period, and retrospective setting, and a prospective study with a large number of patients is needed. Nevertheless, if it is confirmed that CTCs caused by surgical manipulation during lung cancer surgery are related to cancer recurrence, such knowledge may contribute to improvements in the outcome of affected patients by greater control of CTC development and dislodgement.

Conclusion

CTCs detected during surgery for NSCLC may become dislodged by surgical manipulation, and clustered CTCs found during the operation may be related to tumor malignancy, thus indicating an unfavorable outcome.