The recurrence of GC is reportedly frequent but not uniform [42]. Up to 46.4% of GC patients relapse after curative resection in our center. Among the GC patients whose tumors re-emerged, the cumulative rates of recurrence within 1, 2, 3, and 5 years were 38.8%, 69.4%, 80.1%, and 92.1%, respectively. Recently, immunotherapies such as anti-PD-1/PD-L1 have received considerable attention. Its combined use with conventional therapies has increased the survival rate of patients and highlighted the importance of immune factors surrounding the tumors [17, 18]. Although PD-1/PD-L1 has become a potential immune marker for the prognosis and prediction of a series of malignant tumors, only 14.3–29.6% of GC patients express PD-L1, and the positive rate of PD-L1 in our center is less than 10% [43–45]. Therefore, in addition to PD-L1, there are other immunosuppressive molecules that, alone or together with PD-L1, downregulate antitumor immunity. They promote immune suppression by reducing the infiltration of TILs, inhibiting T cell activation, recruiting inhibitory cells such as Tregs or MDSCs, or increasing the expression of inhibitory receptors [35–38, 46–48].
The suppression of the immune microenvironment is considered to be one of the signs of GC progression and plays an important role in tumor re-appearance [46, 49]. Therefore, indicators of immunosuppression may have predictive value for GC recurrence. However, different immunosuppressive indicators may be expressed in different tumors; it is therefore necessary to identify immunosuppressive indicators that play a role in GC. Based on our previous research, we identified seven immunosuppressive indicators that are significant in the prognosis of GC [41]. Lasso Cox regression analysis was used to fit recurrence data, eliminate the correlation between the indicators, and, finally, obtain the IRS based on six immunosuppressive indicators. The IRS reflects different risks of recurrence by classifying different immunosuppressive states.
Distinguishing the infiltration of TILs is critical for deconstructing the immunosuppressive microenvironment. Accumulating evidence indicates that TILs, which represent the local immune response [50–52], in many malignant tumors are related to the recurrence of several cancers [53–55]. Our research also confirmed that several immune biomarkers (CD3+, CD8+, CD45RO+) were significantly associated with recurrence (S14 Fig). However, this has a strong relationship with the location of TILs, and TILs at the IM seem to show more significant differences. Our data revealed that different IRS groups have different immune infiltration characteristics, and the level of IRS seems to be related to the effects of CD8+ TILs in the IM. Increased infiltration of CD8+ TILs enhances the local antitumor activity. In the low- and middle-risk groups, with the increase in IRS, the infiltration of CD8+ TILs in the IM tended to decrease. However, in the high-risk group, this trend was reversed. Although no reduction in the infiltration of CD8+ TILs was observed in the high-risk group, the CD8+ TILs in the high-risk group displayed noticeable exhaustion patterns, which seems to imply the loss of CD8+ TILs function. The lack of CD8+ TILs in the IM may be linked to local recurrence [56]. We hypothesized that this may be due to the suppression of the local immune response, causing micrometastases to occur despite treatment. In our study, patients with local recurrence, especially in the perigastric lymph node group, had a higher IRS. This suggests that the IRS can effectively reflect the degree of local immune suppression. In summary, we termed the low-, middle-, and high-risk subgroups as immune-complete, immune-deficient, and immune-suppressive, respectively, which may contribute to selecting appropriate postoperative adjuvant treatment options.
Adjuvant chemotherapy is the standard adjuvant treatment for advanced resectable GC. Although chemotherapy can partially control the primary tumor, it provides little benefit for inhibiting the metastasis of cancer cells owing to the heterogeneity of tumors and the effect of suppressing immune surveillance [57, 58]. A growing body of evidence has shown the efficacy of chemotherapy in addition to its direct cytostatic/cytotoxic effects but that it is also affected by the (re)activation of antitumor immune responses [59, 60]. Chemotherapy can increase the immunogenicity of tumor cells and inhibit the immunosuppressive pathway of tumor formation [59]. It can also stimulate the anti-cancer immune response to kill residual tumor cells or residual micrometastases in the dormant stage [61]. In our study, IRS was able to predict the sensitivity of GC patients receiving adjuvant chemotherapy. In the middle-risk group, wherein the immune function was not completely suppressed, chemotherapy significantly reduced the recurrence rate, but in the high-risk group with immunosuppression, chemotherapy was ineffective. For these patients, new therapies need to be developed. Immunotherapy has advantages for a variety of cancers, and immunotherapy can improve the outcome of some patients who are resistant to chemotherapy [62–65]. At present, most patients who respond to immunotherapy have so-called “hot” tumors [66]. Hot tumors possess a high level of immune cell infiltration, but inhibitory receptors (such as PD-1, TIM-3) inhibit the antitumor effect [20, 22, 66]. Our data revealed that GC patients in the high-risk group were compatible with this feature; thus, we hypothesized that the prognoses of GC patients in the high-risk group might be improved through immunotherapy. Unfortunately, the current follow-up time of immunotherapy patients in our center is only 1 year, and there are no corresponding data to verify our hypothesis, which is the limitation of our study.
In this study, we defined a signature constructed using six immunosuppressive indicators to assess the risk of recurrence in GC patients. IRS can be treated as a classifier of local immune infiltration, which contributes to the selection of adjuvant treatment options for GC patients after resection. For patients in the low-risk group, the benefit of chemotherapy is not significant, and the decision to undergo chemotherapy should be determined according to individual conditions. Chemotherapy is recommended for patients in the middle-risk group. GC patients in the high-risk group who receive no benefit from chemotherapy alone may potentially benefit from immunotherapy. Our IRS and IRS-based nomogram showed significant prognostic value for recurrence and were verified in three independent cohorts. Our data also showed that GC patients with higher IRSs might be more likely to relapse in a short period of time after surgery and tend to relapse locoregionally, especially in the perigastric lymph nodes. Taken together, our results showed that the IRS can assist surgeons in the individualized selection of postoperative follow-ups, examinations, and treatment plans for patients. However, our research was based on retrospective data, and it is essential to verify it with prospective, multi-center clinical trials.