Circulating tumor cells are a good predictor of tumor recurrence in clinical patients with gastric cancer.

Circulating tumor cells (CTCs) as a liquid biopsy have great potential in clinical applications and basic cancer research, but their clinical use in gastric cancer remains unclear. This study investigated whether CTCs could be used as a potential prognosis predictor in patients with gastric cancer. A total of 120 patients with pathologically confirmed gastric cancer were enrolled from January 1, 2015, to December 1, 2019. All patients were initially diagnosed without previous treatment, and then the number of CTCs was detected using the NEimFISH method before radical surgical resection. Regular follow-up was performed in all patients, and the correlations between the number of CTCs and clinical endpoints, such as disease-free survival (DFS) and overall survival (OS), were evaluated. The univariate and multivariate hazard ratios were calculated using the Cox proportional hazard model. Based on the number of CTCs, we defined CTCs ≥ 2 per 7.5 mL of whole blood as the positive group and CTCs < 2 as the negative group. Among the 120 patients who underwent CTC detection before surgery, the rate of CTC-positive patients was 64.17% (77/120) of which stage I and II patients accounted for 22.50% and stage III patients accounted for 41.67% (P = 0.014). By detecting CTCs before surgery and at the time of recurrence, the number of CTCs tends to increase concomitantly with disease progression (median: 2 VS 5 per 7.5 mL). Multivariate analysis showed that age (HR, 0.259; 95% CI, 0.101-0.662; P = 0.005), D-dimer (HR, 3.146; 95% CI, 1.169-8.461; P = 0.023), and lymph node metastasis (HR, 0.207; 95% CI, 0.0071-0.603; P = 0.004) were factors correlated with CTCs. In addition, the median follow-up of all the patients was 38.0 months (range of 28-80 months); the DFS in CTC-positive patients was significantly shorter than that of the CTC-negative patients, and a significant difference was found based on the Cox proportional hazard regression model analysis (44.52 ± 2.83 m vs. 74.99 ± 2.78 m, HR = 4.550, P = 0.018). The OS was shorter in the CTC-positive group than in the CTC-negative group before the operation, but the result was not significant based on the Cox proportional hazard regression model analysis (47.58 ± 2.46 m vs. 70.68 ± 3.53 m, HR = 2.261, P = 0.083). The number of CTCs tends to increase concomitantly with disease progression. In addition, the detection of CTCs was an independent predictor of shorter DFS in gastric cancer. However, the relationship between CTCs and OS needs to be determined in future studies.