Patient characteristics. A total of 98 NSCLC patients were enrolled, including 65 males and 33 females, and the age distribution was between 18 and 82 years old (average age was 52 ± 9.3). There were 60 cases of lung adenocarcinoma, 33 cases of lung squamous cell carcinoma, and 5 cases of other NSCLCs. According to IASLC2009 (TNM staging standard for lung cancer, 2009, 7th edition), TNM staging was performed on the enrolled patients. Among them, 48 patients were stage I, 13 patients were stage II, 29 patients were stage III, and 8 patients were stage IV. There were 38 patients with benign lung diseases including 18 males and 20 females with the age distribution from 18 to 70 years (average age was 46 ± 11.7) (Table 2).
Table 2. Patients Characteristics and prevalence of circulating tumor cells.
Characteristics
|
No.
|
CTCs(CTC Units/5ml)
|
|
|
|
|
M
|
P25-P75
|
P
|
M
|
P25-P75
|
P
|
M
|
P25-P75
|
P
|
M
|
P25-P75
|
P
|
Benign lung diseases
|
38
|
0
|
0-0
|
< 0.01
|
0
|
0-0
|
< 0.01
|
0
|
0-0
|
0.013
|
0
|
0-0
|
< 0.01
|
NSCLC
|
98
|
1
|
0-2
|
|
1
|
0-3
|
|
0
|
0-1
|
|
3
|
1-6
|
|
Pathological type
|
|
|
|
0.845
|
|
|
0.528
|
|
|
0.904
|
|
|
0.579
|
AC
|
60
|
1
|
0-2.75
|
|
1.5
|
0-3
|
|
0
|
0-1
|
|
3
|
1-6
|
|
SC
|
33
|
1
|
0-2
|
|
1
|
0-2
|
|
2
|
0-0.5
|
|
2
|
1-5
|
|
Others
|
5
|
1
|
0.5-3
|
|
2
|
0-3
|
|
1
|
0.5-1
|
|
2
|
1-7
|
|
TNM stage
|
|
|
|
0.850
|
|
|
0.954
|
|
|
0.505
|
|
|
0.926
|
Ⅰ
|
48
|
1
|
0-2
|
|
1
|
0-3
|
|
0
|
0-1
|
|
3
|
1-6
|
|
Ⅱ
|
13
|
1
|
0.5-3
|
|
1
|
0-4
|
|
0
|
0-1
|
|
3
|
1-11
|
|
Ⅲ
|
29
|
1
|
0-2.5
|
|
1
|
0-3
|
|
0
|
0-0.5
|
|
3
|
1-6
|
|
Ⅳ
|
8
|
1.5
|
0-4.5
|
|
1
|
0.25-5
|
|
0.5
|
0-1.75
|
|
3
|
0.25-12.75
|
|
Age
|
|
|
|
|
|
|
|
|
|
|
|
|
|
≤ 60y
|
74
|
1
|
0-2
|
0.446
|
1
|
0-3
|
0.470
|
0
|
0-1
|
0.353
|
3
|
1-6
|
1
|
> 60y
|
24
|
1
|
0-2
|
|
0
|
0-3.75
|
|
0
|
0-0
|
|
2.5
|
0-6
|
|
Abbreviations: NSCLC, non-small cell lung cancer; AC, Adenocarcinoma; SC, Squamous carcinoma; CTCs, circulating tumor cells; M, median; P25-P75, inter-quartile range.
Comparison of the number of CTCs between groups. The number of all subtypes of CTCs and the total number of CTCs in NSCLC were higher than those in the benign lung disease group (Mann-Whitney U test: The U value of epithelial CTCs group was 822.5, P < 0.01; the U value of epithelial-mesenchymal CTCs group was 859, P < 0.01; the U value of mesenchymal CTCs group was 1487, P = 0.013; and the U value of total CTCs was 605.5, P < 0.01). There was no statistically significant difference in the number of CTCs between lung adenocarcinoma, lung squamous cell carcinoma, and other NSCLC. According to the Kruskal-Wallis test, there was no statistically significant difference in the number of CTCs between TNM stages. Also, there was no significant difference in the number of CTCs between NSCLC patients at different ages (≦ 60 years or > 60 years) (Table 2). The detection rates of CTCs in stage I, II, III, and IV lung adenocarcinoma were 81%, 80%, 89%, and 67%, respectively, while lung squamous cell carcinoma was 71%, 100%, 80%, and 100%, respectively (supplementary table 3).
ROC curve analysis to determine the cut-off value and assess the diagnostic performance. Taking the pathological results as standard, the ROC curve of the total number of CTCs in the NSCLC group was plotted to compare with those in the benign lung disease group (Figure 2). The area under the curve (AUC) was 0.837, 95% CI was 0.76-0.914. The critical value corresponding to the maximum value of the Youden index was 0.5 CTC/5mL. That was when the number of CTCs ≥ 0.5 was considered positive, the sensitivity was 81.6% and the specificity was 86.8%. Among them, the diagnostic sensitivity of stage I, II, III, and IV NSCLC was 79.2%, 84.6%, 86.2% and 75.0%, and the false-negative rate was 20.8%, 15.4%, 13.8%, and 25.0%, respectively (supplementary table S4).
COX proportional hazard regression analysis. A total of 63 of the 98 NSCLC patients were effectively followed up for three years. COX proportional hazard regression analysis revealed that the tumor stage was a risk factor for recurrence and metastasis in NSCLC patients (P = 0.006), while gender, age, and smoking were not risking factors for recurrence and metastasis (P > 0.05) (Table 3). The Exp(B) of tumor staging was 1.813, and the 95.0% CI was 1.186-2.772, indicating that for each upgrade of tumor stage, the risk of recurrence and metastasis was increased by 1.813times.
Table 3. COX proportional hazard regression analysis of follow-up information for 63 NSCLC patients.
|
|
|
95.0% CI for Exp (B)
|
|
P
|
Exp(B)
|
Lower
|
Upper
|
Stage
|
0.006
|
1.813
|
1.186
|
2.772
|
Smoking
|
0.843
|
0.895
|
0.299
|
2.680
|
Gender
|
0.745
|
0.820
|
0.248
|
2.709
|
Age
|
0.517
|
0.985
|
0.941
|
1.031
|
The progress prediction ability of CTCs. The 63 followed-up patients were grouped according to the TNM stage, chemotherapy, pathological type, smoking, gender, and age. For each prognostic factor, the progress of the CTC ≥ 0.5 group has no difference from that of all patients (P > 0.05): TNM stage (P = 0.952), chemotherapy (P = 0.877), pathological type (P = 0.649), smoking (P = 0.968), gender (P = 0.61), age (P = 0.877), as shown in Supplementary Table S5.
Kaplan-Meier survival analysis. Due to the close relationship between PFS and TNM staging as well as whether chemotherapy is performed, finally 14 stage IIIA patients of the followed-up 63 NSCLC patients met the same TNM staging and the same treatment conditions. The 14 patients who underwent radical surgery and subsequent four rounds of adjuvant chemotherapy were divided into two groups according to the total number of CTCs (CTCs ≥ 0.5, 10 cases and CTCs < 0.5, 4 cases). Kaplan-Meier survival analysis results showed that the DFS (progression-free survival) of patients with the total number of CTCs ≥ 0.5 was significantly lower than that of patients with the total number of CTCs < 0.5 (P = 0.022) (Figure 3).