The Presence of Small-Size Circulating Tumor Cells Predicts Worse Prognosis in Non–Small Cell Lung Cancer Patients

A total of 80 lung cancer patients were recruited from Beijing Shijitan Hospital of Capital Medical University (Beijing, China) between February 2021 and July 2023. The inclusion criteria were age 18–80 years, newly diagnosed, never previously treated with any treatment, and had NSCLC. Forty-eight patients were eligible and information on their clinical characteristics is presented in Table 1. There were 27 patients with early-stage lung cancer, and 21 with late-stage lung cancer. The staging was based on tumor node metastasis (TNM) version 8 with I–IIIA as early stage and IIIB–IV as late stage. Pathologic types included 42 of 48 adenocarcinomas (87.5%, with 1 mucinous adenocarcinoma), 5 of 48 squamous carcinomas (10.4%) and 1 of 48 large cell carcinoma (2.1%). To detect CTCs and CTECs, 6 mL of peripheral blood was drawn from each of the 48 NSCLC patients. These 48 NSCLC patients had not undergone any treatment at the time of blood sampling and they were followed up afterward. The design of the study is shown in Figure 1.

This study involved human participants and was approved by the Ethical Committee of Beijing Shijitan Hospital, affiliated with Capital Medical University. The experiments were conducted after collecting the informed consent of each subject, and the study conformed with the Code of Ethics of the World Medical Association¹⁴  (Declaration of Helsinki) in the British Medical Journal (18 July 1964).

Subtraction enrichment (SE) was performed according to the manufacturer's updated instructions with minor modifications (Cytelligen, San Diego, California). In its simplest terms, 6 mL of blood were collected into a tube containing an anticoagulant solution and centrifuged to separate plasma. Sedimented blood cells were resuspended with 3 mL human CTCs buffer and loaded on the nonhematologic cell separation matrix in a 50-mL tube. Samples were centrifuged, followed by the collection of the solution containing white blood cells (WBCs) and tumor cells above red blood cells. The solution containing WBCs was incubated with magnetic beads conjugated to a cocktail of anti-leukocyte monoclonal antibodiess for 30 minutes. WBCs bound to immuno-beads were subsequently removed using a 50-mL magnetic separator (Cytelligen). The remaining nonhematologic cells were mixed with cell fixative, then smeared on formatted and coated CTC slides and dried for subsequent immunostaining–fluorescence in situ hybridization (iFISH) processing.

Regarding iFISH, dried monolayer cells on the coated CTC slides were hybridized with centromere probe 8 (CEP8) SpectrumOrange (Vysis, Abbott Laboratories, Chicago, Illinois), which has been approved by the US Food and Drug Administration to identify aneuploid solid tumor cells. Samples were subsequently incubated with the indicated monoclonal antibodies, including Alexa Fluor (AF)594–anti-CD45 (clone 9.4) and Cy5-anti-CD31(clone WM59). Conjugation of diverse antibodies to each specific fluorescent dye was performed at Cytelligen. After washing, samples were mounted with mounting media containing DAPI (Vector Laboratories, Burlingame, California) and subjected to the automated iFISH CTC 6-channel 3D scanning and image analyzing system codeveloped by Carl Zeiss (Oberkochen, Germany), MetaSystems (Altlussheim, Germany), and Cytelligen. CTC slides were automatically loaded on a Zeiss fluorescence microscope (AXIO Imager Z2) and afterward subjected to automated X-Y scanning with cross Z-sectioning of all cells performed at 1-μm steps of depth. X-Y-Z 3D scanning was performed in each of the 6 fluorescence color channels. Positive target cells were defined as DAPI⁺/CD45⁻/CD31⁺ or DAPI⁺/CD45⁻/CD31⁻ with diploid or aneuploid chromosome 8.

Following high-throughput scanning and acquiring and processing cell images, subsequent comprehensive characterization and classification of CD31⁻ CTCs and CD31⁺ CTECs as well as statistical analyses were performed upon phenotypic, karyotypic, and cell morphological characterization of the trielement in the intracellular bio-chain, with particular focus on cell size and cell cluster. In this technique, leukocytes were used as negative control cells while the definition of large- and small-size cells¹⁵  was generally based on the size of leukocytes (5 μm) as a threshold, with cells larger than 5 μm being called large-size cells and those smaller than 5 μm designated as small size cells (Figure 2, A through X).

Statistical analysis and graphical plots were performed using SPSS 26 (IBM) and GraphPad Prism 9 (La Jolla, California). The differences of categorical variables in distribution among groups were analyzed using the χ² test or the Fisher exact test. Differences of continuous variables with normal distribution among 2 groups were compared by t test. Receiver operating characteristic (ROC) curve analysis was used to evaluate diagnostic accuracy and to determine cutoff points. Kaplan-Meier survival curves and log-rank tests were used to compare the differences in PFS rates between the 2 groups. The Cox proportional hazards regression model was used to determine hazard ratios for PFS. P < .05 was considered a statistically significant difference.

First, we analyzed the proportional distribution of the number of CTCs and CTECs in patients (Table 1). We defined CTCs or CTECs ≥1 as positive; otherwise as negative. CTCs were present in all patients. The CTEC positivity rate was 81.3% (39 of 48) and it was higher in late-stage patients (95.2%, 20 of 21) than in early-stage patients (70.4%, 19 of 27), and statistically significant (P = .03). Meanwhile, we compared the differences in CTC and CTEC counts between lung cancer stages. However, there was no significant difference between the early and late stages for either the CTC number (Figure 3, A; P = .67) or the CTEC number (Figure 3, C; P = .50). We further stratified CTCs and CTECs (Figure 3, B and D), then determined that the number of early-stage patients with either CTCs or CTECs was not less than that of late-stage patients.

CTCs and CTECs had diverse cell subtypes and differed in their distribution in the blood (Figure 4, A). The size of CTCs and CTECs in peripheral blood differed. Therefore, we categorized CTCs and CTECs. Classification was based on the size of leukocytes (5 μm): when a cell was larger than that it was called large size and it was small size when less than 5 μm. Among the 48 newly diagnosed NSCLC patients, 840 CTCs (74.3%, 840 of 1130) and 290 CTECs (25.7%, 290 of 1130) were detected (Figure 4, A). CTCs comprised the majority of the blood in NSCLC patients, and large-size CTCs dominated among CTCs. Small-size CTCs accounted for 19.0% (160 of 840) of CTCs, while large-size CTCs were in the majority (81.0%, 680 of 840); the situation was similar for CTECs, with small-size CTECs being 13.1% (38 of 290) and large-size CTECs being 86.9% (252 of 290; Figure 4, B and C).

Using small- and large-size cells as categories, we grouped all cells as negative or positive (Table 2). Among CTCs, the positivity rates of small- and large-size CTCs were 66.7% (32 of 48) and 93.8% (45 of 48), respectively; within CTECs, the positivity rates of small- and large-size CTECs were 27.1% (13 of 48) and 81.2% (39 of 48), correspondingly. By analyzing the relationship between different clinical features (Figure 4, D through F), we discovered that the positivity rate of small-size CTCs was higher in patients with lymph node metastasis than in those without lymph node metastasis (P = .04); although small-size CTCs were more often positive in patients with late-stage compared to early-stage lung cancer, it was not statistically significant (P = .06); nevertheless, this was not the case for large-size CTCs. For various CTECs, large-size CTECs were more often positive in patients with late-stage compared to early-stage lung cancer (P = .03).

To investigate whether CTCs and CTECs of different cell sizes had an impact on the prognosis of patients with NSCLC, we followed all patients until July 2023 and used death or relapse as a marker for termination of follow-up. The median follow-up time was 11.3 months (0.5–29.3 months). During the follow-up period, 22 patients had a recurrence, 23 patients were progression-free, and death occurred in 3 patients. First of all, we figured out that small-size CTCs and large-size CTECs had the highest diagnostic value by plotting ROC curves, and then calculated the optimal cutoff values for each subtype of cells (Figure 4, G). In view of the above results, we separately plotted the PFS curves for all patients as well as only late-stage patients (Figure 5, A through H). Presence of ≥1 small-size CTC was a significant risk factor for reducing the PFS of NSCLC patients (P = .007; Figure 5, A). In other sizes of CTCs, CTECs had no significant influence on PFS. We continued to analyze the survival of patients with late-stage lung cancer and realized that presence of ≥1 small-size CTC was equally clinically significant for PFS (P = .04; Figure 5, B). Our results suggested that the number of small-size CTCs was predictive of PFS in all patients with NSCLC, and in advanced patients, the median PFS was significantly shorter in patients with ≥1 small-size CTC (2.9 months) than in patients with <1 small-size CTC (17.7 months). On top of that, survival univariate analysis (Table 3) showed that the following factors influenced prognosis: size, stage, lymph node metastasis, distant metastasis, and small-size CTCs. Cox multivariate survival analysis was performed revealing that small-size CTCs and stage were independent prognostic factors (Figure 6). The risk of disease progression in patients with small-size CTCs positive was 3.446 times higher than in patients with small-size CTCs negative (P = .03, 95% CI: 1.159–10.249). To rule out the influence of treatment on prognosis, we subgrouped patients (see Supplemental Table in the supplemental digital content at https://meridian.allenpress.com/aplm in the January 2025 table of contents.) and performed a multivariate analysis, which discovered that treatment was not a prognostically independent factor (P = .39).

Lung cancer incidence and mortality rates in China have been on the rise in the past few years. According to 2020 global cancer statistics,¹⁶  approximately 820 000 people were diagnosed with lung cancer in China, and 715 000 people died from lung cancer. The 5-year survival rate for lung cancer patients is only 19%.³  Difficulty in early diagnosis and detection of metastases is an important factor in the low 5-year survival rate of lung cancer. Accordingly, it is important to find markers for disease diagnosis and disease progression monitoring. Low-dose computed tomography has been investigated in several randomized trials in the United States and Europe as a tool for early lung cancer detection and screening.¹⁷  In 2011, the American National Lung Screening Trial first reported a statistically significant reduction of mortality from lung cancer of about 20% in low-dose computed tomography compared to chest X-ray screening.¹⁸  At this stage, a relatively mature method for early screening of lung cancer existed. However, it was still tough predicting the prognosis and progression of lung cancer and in consequence, it was important to find indicators to predict the development of lung cancer as early as possible.

In this study, liquid biopsy was a noninvasive technique that allowed efficient, timely, and dynamic observation of changes in a patient's condition, with CTCs and CTECs being the cell types identified. It has been demonstrated that CTCs and CTECs are useful for the early screening of lung cancer and could determine the efficacy of treatment and prognosis. CTCs are formed after cells entered the bloodstream from the primary or metastatic tumor, and eventually, they could end up forming new metastatic foci through blood circulation, leading to tumor recurrence and metastasis. Tumor tissues contain abundant large blood vessels, and the cells lining the walls of these vessels are known as tumor-derived endothelial cells (TECs). Endothelialization of tumor cells and tumorization of vascular endothelial cells are the main mechanisms of TEC formation.¹⁹  When they enter the blood or lymph fluid, they form CTECs, which are involved in blood and lymph node metastasis of tumors. Together, CTCs and CTECs constitute cell-based circulating tumor biomarkers, while they act as nonhematopoietic aneuploid circulating rare cells (apCRCs) and are distinguished from hematopoietic apCRCs by differences in CD45 expression and centromere probe (CEP).¹⁹  CEP is an important indicator for the diagnosis of malignant cells and aneuploid chromosome 8 is observed in neoplastic cells of almost all solid tumors, including lung cancer.²⁰  CD45 is expressed on all leukocytes. With the presence of CD45⁻/CEP8⁺, a diagnosis of nonhematopoietic apCRCs can be made. To further differentiate between CTCs and CTECs, attention was given to CD31 expression. CD31 is one of the most representative endothelial cell markers. The presence of DAPI⁺/CD45⁻/CD31⁻/CEP8⁺ indicated CTCs, and when DAPI⁺/CD45⁻/CD31⁺/CEP8⁺ was present, it was CTECs. SE-iFISH technology has been proven to be highly sensitive and specific for the detection of various CTCs and CTECs,²¹  and not only that, it takes into account the 3 elements of the intracellular biocontainment chain (nucleic acids, proteins, and cellular morphology) in a holistic manner, allowing for the 1-step counting of CTCs, karyotyping, phenotyping, and molecular typing, as well as detecting the expression of specific tumor markers, and identifying a variety of CTCs and CTECs of varying clinical significance. A rise in the number of CTCs after treatment compared with before treatment has been associated with a poorer prognosis in patients with advanced adenocarcinoma, and pretreatment vimentin⁺ CTECs were more likely to recur and had a shorter PFS.²²  In patients with early-stage NSCLC, triploid circulating aneuploid cells and small-size circulating aneuploid cells were more likely to relapse and had shorter disease-free survival.²³  In advanced lung cancer, ≥2 small-size CTCs was associated with poorer PFS.²⁴  In advanced NSCLC patients undergoing anti-programmed death receptor-1 (anti–PD-1) therapy, the prognosis was worse with programmed cell death ligand-1⁺ CTECs and might be resistant to anti–PD-1 therapy.²⁵  Not only that, CTCs and CTECs also had an early diagnostic role: when small tetraploid CTCs and CTECs were present, early adenocarcinoma was more likely compared to patients with benign nodules, while when large polyploid CTCs and CTECs existed, they could distinguish between early and late stages of adenocarcinoma.²⁶  However, in spite of these previous studies, there was still a lack of reports on the prognosis related to different sizes of CTCs and CTECs in patients with all stages of NSCLC.

Many types of CTC and CTEC detection methods have been reported, which can be roughly categorized into 3 main groups.²⁷  The first one, the positive capture method based on epithelial antigen and polypeptide expression, led to a low detection rate caused by its dependence on epithelial antigen expression. The second one, the negative enrichment method by CD45 antibody, not only couldn't effectively remove most of the leukocytes in the blood but also caused damages to CTCs and CTECs because of the hypotonic cleavage method. The third one, a filtration method, was based on the size of the cells; although it was more convenient, small CTCs and CTECs with clinical significance were also filtered. Recently, a novel SE-iFISH technique has been reported.²⁸  This technique includes 2 steps: isolation and identification. The SE isolation method removes most hematologic leukocytes by using special immuno-magnetic beads that are conjugated to antibodies specific for multiple leukocytes’ surface antigens. In the iFISH identification method, positive immunostaining of tumor markers combines with fluorescence in situ hybridization of chromosome 8 aneuploidy^20,29  to subtype CTCs and CTECs. SE-iFISH considers the 3 elements of the intracellular biological chain (nucleic acids, proteins, and cell morphology) as a whole, and detects chromosomal karyotypes, tumor marker protein phenotypes, and cell morphology on cells synchronously and in situ. Compared with other techniques, it has 4 major advantages: it does not rely on the expression of tumor markers to isolate CTCs and CTECs from a variety of solid tumors, it can isolate small CTCs and CTECs, the cells can avoid damage, and cells are complete enough to be used for single-cell analysis. Compared to the traditional enrichment and identification method, SE-iFISH demonstrated higher sensitivity for CTC detection, showing a 92% positive rate in the identical population of lung cancer patients.³⁰  Similar high CTC positivity detected by SE-iFISH was also observed in gastric (90.5%) and esophageal (87%) carcinoma patients.^30,31  Moreover, in contrast to a conventional lengthy FISH protocol, which takes more than 20 hours, the time required for an iFISH experiment including antibody staining is as short as 3–4 hours.²⁰ 

In this study, we used the SE-iFISH technique to detect CTCs, CTECs, and their different sizes of cell subtypes to find predictive indicators related to prognosis. We found that CTCs were present in all patients and that the CTEC positivity rate was higher in late-stage patients than in early-stage patients; this might be due to the fact that the increase in the number of CTECs allowed for more angiogenesis, which provided nutrients to the tumor, which in turn caused the tumor to be less susceptible to eradication and development of an advanced stage. Although there was no statistically significant difference in the number of CTCs and CTECs between early- and late-stage patients, the distribution of the number of CTCs and CTECs in early-stage patients was more in the 0–10 range, and the amount of patients with the number of CTCs or CTECs in the range greater than 10 increased in late-stage patients; this change may be correlated with the size of the tumor burden, which also suggested that even in the early stages of the disease, the escape of tumor cells into the circulation system already existed. Therefore, we believed that the analysis of CTCs and CTECs will be helpful for the prognosis of patients. We speculated that there was a correlation with patient survival. Next, we explored small- or large-size CTCs and CTECs. CTCs accounted for most of the total number of all cells, and large-size CTCs were more numerous among CTCs. The positivity rate of small-size CTCs was higher in patients with advanced-stage lymph node metastasis, while this was not the case for large-size CTCs. Large-size CTECs were more frequently positive in patients with late-stage compared to early-stage lung cancer. These results might suggest that patients with NSCLC positive for small-size CTCs and large-size CTECs might be more likely to develop lymph node metastases or advanced stages. To confirm this conclusion, Kaplan-Meier survival curves were performed. Ultimately, presence of ≥1 small-size CTC was found to be a risk factor for shorter PFS not only in all patients but also in patients with advanced disease. Furthermore, small-size CTCs were an independent prognostic factor. This also supported the previous statement that patients were more likely to relapse when they had a specific cell type—small-size CTCs, which led to disease progression. In terms of the findings of this study, we thought that early-stage patients might also show a similar trend, namely, the median PFS of early-stage patients with ≥1 small-size CTC was shorter than the patients with <1 small-size CTC, and we will then follow up further with early-stage patients to determine the significance of small-size CTCs as a predictor of PFS in early-stage patients. Not only that, some studies have reported that when patients with advanced NSCLC had vimentin⁺ small-size CTCs they were more likely to have liver metastasis and poor prognosis,²⁴  which might be because tumor cells with epithelial-mesenchymal transition properties were smaller in size than those without such properties,³²  which made tumor cells invasive and metastatic, thus affecting patient prognosis. Given the above results, we believe that small-size CTCs may be a risk factor for NSCLC and have some value in predicting the early and late stages and prognosis of NSCLC in patients. This finding would provide important information for clinical work as a way to individualize treatment. If combined with the detection of the expression of specific tumor markers, such as programmed cell death ligand-1⁺, epithelial cell adhesion molecule, etc, it might be possible to accurately determine the efficacy of a patient for a specific drug, or combined with single-cell gene sequencing, and it would be expected to explore new drug therapy targets. All of these tests could be done by nearly noninvasive SE-iFISH technology, which is believed to be greatly helpful in the treatment of tumor patients.

There are a few limitations to this study. First, the findings are limited because of the lack of healthy population controls. Second, the sample size is relatively small. So, a prospective study with a larger cohort size is necessary to validate the finding. Finally, the pathologic types are predominantly adenocarcinomas, with squamous carcinomas and other types remaining low; more pathologic types need to be included in the future.

We utilized the SE-iFISH technique to detect CTCs, CTECs, and their subtypes in the peripheral blood of NSCLC patients. Presence of ≥1 small-size CTC was associated with poor PFS and an independent prognostic indicator; the presence of small-size CTCs was associated with advanced lung cancer and lymph node metastasis and was considered to be a predictive risk factor.

The authors thank the thoracic surgery team of Beijing Shijitan Hospital, Capital Medical University.