This single arm, single-center, phase II trial evaluated the efficacy and safety of a combination therapy of a PD-1 inhibitor with albumin paclitaxel and apatinib as second-line treatment in patients with mGC, providing an option for patients who were previously treated with first-line platinum- or oxaliplatin-based chemotherapy.

This was the first time that the PD-1 inhibitor and small molecule anti-angiogenic agents were combined with chemotherapy in an attempt to improve the current dilemma of second-line therapy. Nivolumab and sintilimab have been approved as immune checkpoint inhibitors for the first-line treatment of advanced GC/EGJC [4]. But till now, there has been no evidence that PD-1 inhibitor monotherapy is comparable to single-agent chemotherapy [10]. Apatinib was approved as third-line therapy for advanced GC in 2014 [14]. The combination of PD-1 inhibitor and anti-angiogenesis has a coordinated function, which has been demonstrated in several works [12]. However, the efficacy of the addition of chemotherapy to the above combination in the second-line therapy is still questionable.

In our study, it was observed that the median PFS was 6.2 months (95% CI, 3.9–9.3), and OS was 10.1 months (95% CI, 7.5–14.1). Compared with the results of previous studies, a significantly increased survival benefit was shown. These results suggested that the outcomes of the three-drug combination were more favorable than those of double drugs in previous studies. In the RAINBOW study [6], the median PFS with ramucirumab plus paclitaxel was 4.4 months (95% CI, 4.2–5.3) and the median OS was 9.6 months (95% CI, 8.5–10.8). The outcomes were even worse in the REGARD study, with the median OS of 5.2 months and the median PFS of 2.1 months [7]. In the similar population-based study of nivolumab combined with paclitaxel plus ramucirumab as second-line treatment for advanced GC [15], the 6-month PFS rate was 46.5% (80% CI, 36.4%–55.8%), and median survival time was 13.1 months (95% CI, 8.0–16.6) as 60.5% of patients had PD-L1 CPS ≥ 1. These outcomes were better than those of the present study, possibly due to a higher proportion of patient with PD-L1 positive status than that of our study. Overall, the addition of a PD-1 inhibitor to chemotherapy and anti-angiogenic agents in the second-line therapy showed significant superiority over chemotherapy alone or combined with anti-angiogenic agents [16].

With regard to tumor response, the ORR and DCR of a PD-1 inhibitor plus apatinib were 20.9% and 88.3%, which were better than chemotherapy alone [8, 9] or combined with anti-angiogenic agents [17]. The ORR and DCR of apatinib combined with chemotherapy were 18.52% and 92.59%, respectively [18]. By contrast, an objective response was observed in 20 (69%, 95% CI, 49–85) of 29 patients, including one patient with MMR-deficient tumor in the EPOC study [19].

The safety profile in this study was generally consistent with that of monotherapy with a PD1 inhibitor, apatinib or albumin paclitaxel, or any combination of two or three drugs for mGC (e.g., nivolumab plus paclitaxel and ramucirumab [15] or lenvatinib plus pembrolizumab [19]), with the notable exception of proteinuria, a frequent toxic effect of apatinib. The addition of a PD-1 inhibitor to apatinib did not increase the known toxic effects associated with apatinib. Common toxicities reported in this study, such as bone marrow suppression, hypertension, hand-foot reaction, and hypothyroidism, were mild to moderate and were manageable with appropriate dose modifications and supportive care. And the incidence of these AEs was similar to that reported in the EPOC1706 trial [19] and in trials involving patients with mGC [12, 15, 18]. No rare serious toxic effects occurred and no patients reported treatment-related deaths.

There were some limitations in our study. This was a single-armed perspective study from a single institution with limited samples resulting in some bias. As check649 and orential16 studies had published the outcome of first-line therapy with immunotherapy, therefore, a further study would be needed to investigate the efficacy of this combination for patients with prior immunotherapy history. Finally, objective response was not assessed by an independent central review, which might lead to the overestimation of anti-tumor activity results in this study.

In summary, the combination of a PD-1 inhibitor with apatinib and albumin paclitaxel showed promising efficacy and acceptable safety profile in patients with mGC who failed first-line chemotherapy. These results warranted further investigation. We look forward to a phase III clinical study to further confirm our findings.