Many studies have been conducted to investigate the correlation between peripheral blood-derived inflammation markers and tumor prognosis[17, 18]. Liu showed that CRP/Alb was an independent prognostic marker for patients with ovarian cancer. Also, the NLR and PLR are prognostic factors in patients with non-small cell lung cancer after stereotactic radiation therapy. The independent risk factors for poor GC prognosis include NLR, PLR, fibrinogen, PNI, GPS, CRP/Alb, among others. Also, some studies have combined these systemic inflammatory markers with or without TNM stage to provide new prognostic tools[21–23]. However, most of these studies were retrospective, and reported inconsistent results, particularly on the significance of each inflammatory index and the threshold value.
Therefore, in addition to the peripheral blood-derived inflammation markers reported in previous studies, this study further explored the prognostic value of some conventional systemic inflammatory marker in patients with GC. We prospectively analyzed 598 GC patients after radical surgery and found that independent risk factors for poor prognosis of GC included age, T stage, N stage, number of lymph nodes retrieved, and PLR. Currently, TNM staging is the standard prognostic tool for GC. Given the excellent prognosis of early gastric cancer, our focus was to analyze the prognosis of patients with AGC (stage II/III). Many studies have demonstrated that the number of lymph node dissection has an impact on the prognosis of GC; therefore, we further excluded patients with an inadequate amount of lymph node dissection (less than 16)[24–26].
Independent risk factors for AGC included age, PLR, TNM stages, and chemotherapy. Previous studies have investigated the effect of age on GC prognosis; however, most of these studies found no significant association between age and GC prognosis[27–29]. Takatsu analyzed 5000 GC cases and found that early-onset GC (age < 40 years) was likely to present lymph node metastases. But the survival rate of young GC patients was equivalent to that of older GC patients (age ≥ 60 years). In the present study, older GC patients (age > 60 years) had a worse prognosis, which was closely related to tumor recurrence.
Nutritional status is associated with survival in patients with malignant tumors, including GC. Preoperative underweight and low PNI are considered poor prognostic factors. Park suggested that careful nutritional intervention after surgery could improve the survival rate. Besides, a meta-analysis concluded that a low PNI is significantly associated with poor overall survival except for stage IV GC. However, consistent with the results of Li et al., we observed that PNI was not associated with prognosis. Also, our results did not show a correlation between fibrinogen and the prognosis of GC. Recent studies have shown that fibrinogen is one of the risk factors for poor prognosis in upper gastric cancer. Fibrinogen is the primary acute-phase protein, and as a critical component of the hemostatic system, it regulates the systemic inflammatory state and cancer progression. However, its clinical significance in the prognosis of GC has not been elucidated.
The NLR and PLR are the most extensively studied markers of peripheral blood-derived inflammation, which are associated with the prognosis of GC. Accumulating evidence has shown that NLR and PLR are associated with distant metastases during GC progression[34, 35]. Kim reported that although both the PLR and NLR could reflect the prognosis, the NLR was more predictive of overall survival than the PLR in GC. Also, they suggested that NLR and PLR might be associated with lymph node metastasis in early gastric cancer. On the contrary, Zhu et al. indicated that NLR and PLR could not predict lymph node metastasis and prognosis in early gastric cancer. In the present study, PLR was significantly correlated with the prognosis of GC, but there was no statistical difference between NLR and prognosis of GC. This observation seems to be inconsistent with previous studies, but the exact mechanism is still unclear. However, the inconsistencies could be because most previous studies focused on overall survival as the primary outcome, whereas herein, we used tumor recurrence and tumor-related mortality as observational indicators, which seem to be more accurate. In addition, the clinicopathological characteristics were similar between the PLR elevating group (PLR < 154) and the PLR decreasing group (PLR ≥ 154), except for tumor size, which further suggested that PLR might influence the prognosis of tumor through other mechanisms. A recent meta-analysis has revealed that PLR is associated with prognosis of GC.
Abnormal levels of CRP and Alb have been related to poor prognosis of tumor patients. It is noteworthy that the combinations of these two indicators can enhance the accuracy to predict the recurrence of multiple tumors. Among them, the most common evaluation criteria after combination include GPS and CRP/Alb. Besides, many studies have used GPS to predict the prognosis of various tumors, including GC. Hsueh recently recommended the use of GPS as a predictive and prognostic factor in patients with GC. A significant correlation was observed between the GPS, short-term postoperative complications, and long-term survival outcomes in patients with GC undergoing D2 gastrectomy. Some studies have indicated that GPS and mGPS, used either alone or in combination, represent an independent prognostic factor for long-term outcome in resected GC[41, 42]. However, Walsh’s results did not show a correlation between prognosis of patients and mGPS levels, although mGPS was associated with advanced GC stage. Liu et al. retrospectively analyzed 455 patients with resectable GC and showed that CRP/Alb, rather than GPS and mGPS, was associated with overall survival. Similarly, Xu and Lu et al. also observed that CRP/Alb and CRP/prealbumin were associated with recurrence of GC based on the data from a phase III randomized clinical trial. On the contrary, our results showed that although CRP/Alb and GPS were associated with prognosis in the univariate analysis, the multivariate analysis showed that both were not related to long-term survival.
Consistent with the results of previous clinical trials, our findings showed that chemotherapy could significantly improve the prognosis of AGC, especially for patients with stage III GC[45, 46]. Moreover, many studies have investigated the correlation between peripheral blood-derived inflammation markers and the effects of chemotherapy, to guide the selection of chemotherapy-sensitive patients. A study suggested that the sensitivity of chemotherapy (oxaliplatin/5-fluorouracil combination) might be closely related to NLR, PLR, and their changes in metastatic gastric cancer. Hirahara believed that the combination of NLR and PLR might be more effective in predicting the chemotherapy response in patients with metastatic gastric cancer. Besides, Tang and Chen concluded that PLR could predict the efficacy of neoadjuvant chemotherapy of GC patients treated with oxaliplatin and capecitabine regimens. In the current study, PLR effectively predicted adjuvant chemotherapy (oxaliplatin/5-fluorouracil combination) response in patients with AGC after surgery. For patients with PLR ≥ 154, chemotherapy significantly improved long-term survival, including DFS and DSS; however, patients with PLR < 154 did not benefit from adjuvant chemotherapy. Thus, we recommend that AGC patients with PLR ≥ 154 should actively receive adjuvant chemotherapy (oxaliplatin/5-fluorouracil combination) after surgery, whereas patients with PLR < 154 need to be cautious when choosing adjuvant chemotherapy.
This study systematically evaluated the relationship between peripheral blood-derived inflammation markers and the prognosis of GC. Notably, the potential ability of inflammatory markers to predict the effects of chemotherapy was further demonstrated. However, this study had a few limitations. Importantly, this was an observational study and was therefore influenced by other confounding factors. Also, the sample size was relatively small, and this may have reduced the reliability of the findings. Thus, these results need to be further validated by large multicenter randomized clinical trials.