Comparative Impact of PD-1 and PD-L1 Inhibitors on Advanced Esophageal or Gastric/Gastroesophageal Junction Cancer Treatment: A Systematic Review and Meta-Analysis

Programmed death 1 (PD-1) and PD ligand 1 (PD-L1) inhibitors have demonstrated varying effectiveness in treating esophageal or gastric/gastroesophageal junction (G/GEJ) cancer. Hence, this systematic review and meta-analysis evaluated the efficacy and safety of anti-PD-1/PD-L1 treatment in patients with esophageal or G/GEJ cancer by analyzing the types of medications. Randomized controlled trials comparing anti-PD-1/PD-L1 to control therapy were identified by searching PubMed, EMBASE, and ClinicalTrials.gov. The outcomes included overall survival (OS), progression-free survival (PFS) rates, and serious adverse events (SAEs), evaluating the differences in therapy types, including a comparison between PD-1 and PD-L1 inhibitors. Eight studies were included in the analysis. PD-1/PD-L1 inhibitors affected the overall OS rate increment without influencing the PFS rate (HR, 0.837; 95% CI, 0.753–0.929; p = 0.001; HR 0.991; 95% CI, 0.778–1.263; p = 0.942, respectively). Anti-PD-1 was significantly more beneficial for increasing OS and PFS than PD-L1 inhibitors. Anti-PD-1 and PD-L1 use was not significantly associated with SAE development in esophageal or G/GEJ cancer patients. PD-1/PD-L1 inhibitor use was associated with improved OS and PFS rate increase among PD-1 and PD-L1 inhibitors. Considering response variations to anti-PD-1/PD-L1 usage, more individualized treatments should be introduced in clinical practice.


Introduction
Esophageal or gastric/gastroesophageal junction (G/GEJ) cancer is one of the most fatal cancers [1,2]. Although these cancer types could not be together considered, controversies remain to differentiate esophageal or G/GEJ in terms of histological distinctions and clinical outcomes [3]. In addition, recently, a study was also conducted that examined esophageal and gastroesophageal junction cancer together [4]. Typically, esophageal or G/GEJ cancer is often asymptomatic in the early stages, resulting in a loss of opportunity for appropriate treatments in advanced cases. Still, recommended therapies primarily include surgical or systemic treatment such as platinum-based chemotherapy, the five-year survival rate remains around 15-25%, and high recurrence or metastasis rates have been reported [5]. Pathophysiologically, gastrointestinal tumors are traditionally considered nonimmune-related malignancies. However, current reports have provided a new direction in the treatment by demonstrating the effectiveness of blocking specific immunosuppressive substances, such as programmed death receiver 1/programmed death ligand 1 (PD-1/PD-L1) [6]. Current studies have indicated that the number of tumor-infiltrating lymphocytes is related to tumor prognosis [7] and the correlation between malignant tumors and immune cells such as T cells. Thus, in order to enhance immune activity against cancer cells, PD-1/PD-L1 inhibitors have been developed. With regard to pathophysiological  Table 1 presents the basic characteristics of the eight studies selected for inclusion [24][25][26][27][28][29][30][31]. In total, 4206 patients were included. Pembrolizumab (PEM) was administered at 200 mg every 3 weeks to 1239 patients across four studies [25,28,29,31]. Nivolumab (NIV) was administered to 540 patients at 240 mg every 2 weeks and 3 mg/kg every 2 weeks, respectively, in two studies [24,27]. Similarly, 185 patients in one study were treated with 10 mg/kg avelumab (AVE) every 2 weeks [26]. In yet another study, a total of 228 patients received 200 mg of camrelizumab (CAM) every 2 weeks [30].

SAE
For safety, four studies were included [24,27,28,30], and a population size of 2109 was used in the analysis. As observed in the forest plot analysis, the intervention group had 1.033 times upper odds for SAE than the control ( Figure 4). Moreover, the results showed that SAEs were not significantly different between the intervention and control groups. (OR 1.033, 95% CI: 0.717-1.488, I 2 = 55.776%, p = 0.861).

PFS
For secondary efficacy outcome, a total of seven studies with a population of 3700 were included [24][25][26][27][28][29][30] in the analysis of the HR of PFS. As shown in Figure 3a

Risk of Bias and Strength of Evidence
Deviations from intended intervention bias items were a major concern for all eight studies [24][25][26][27][28][29][30][31]. Regarding the bias in the outcome measurement, two studies were at high risk, while the remaining were of some concern. The details of the risk of bias for each study are shown in Figure 5. Table 2 illustrates the quality of evidence using the GRADEpro method for the effects of PD-1/PD-L1 inhibitors compared to the control in the outcome. Egger's regression test suggested no evidence of publication bias (p = 0.819) ( Figure 6).   Table 2. Summary of findings for outcome comparing PD-1/PD-L1 inhibitors to control group based on the GRADE approach.

Discussion
The present systematic review and meta-analysis was conducted to evaluate the efficacy and safety of PD-1 and PD-L1 inhibitors in patients with esophageal or G/GEJ cancer patients. Our findings suggested that both PD-1 and PD-L1 inhibitor use was associated with increased OS in these patients, as compared to the PFS. As surrogate endpoints for detecting the benefits of anticancer therapy, OS and PFS are considered to be of prime importance for drug approval. One of the first immunotherapies approved by the FDA in 2011, ipilimumab, showed extended survival among patients with advanced melanoma. To guarantee the approval of drugs and provide one of the most direct measures of true clinical benefit, prolonging the efficacy of survival rates prove most useful in identifying promising agents for cancer treatment [32]. Despite taking into consideration various confounders resulting from biases, survival is an important prognostic measure for more advanced-stage cancers [33]. However, although both the OS and PFS are considered meaningful efficacy endpoints of anticancer therapy, including traditional immunotherapy, the correlation between these two has not been sufficiently demonstrated. According to Blumenthal et al., there was no association between the OS and PFS (coefficient determination, R 2 = 0.08) in the case of targeted and standard therapies in advanced non-small-cell lung cancer treatment [34]. Similarly, Mushti et al. also demonstrated a weak association between the PFS and OS in both trial and individual-level analyses in immunotherapy, particularly with regard to PD-1 and PD-L1 inhibitors (R 2 = 0.1303 and R 2 = 0.1277, respectively). Thus, it was concluded that PFS could not prove a sufficient surrogate endpoint to replace the OS in assessing the clinical benefit in immuno-oncology trials [35]. Furthermore, pathophysiological differences in the amplification of chromosome 11q13 and microsatellite instability in advanced esophageal cancer might contribute to beneficial effects of the OS rather than the PFS parameters, when treated with PD-1/PD-L1 inhibitors [9]. As the most direct measure of clinical benefit, outcomes of the OS post-treatment with PD-1/PD-L1 inhibitors in trials with esophageal or G/GEJ cancer patients were incongruous. Although previous metaanalyses consistently exhibited improvements in the OS rather than the PFS in esophageal or G/GEJ cancer treatment with PD-1/PD-L1 inhibitor, limitations such as the lack of controlled data or the possibility of known/unknown biases increasing heterogeneity in the analysis might result in less statistically powerful conclusions [6,9]. Furthermore, a previous study included only two studies in the analysis demonstrating improvement of the OS and PFS in the control group, rather than the anti-PD-1 or anti-PD-L1 groups, to treat esophageal or G/GEJ cancer patients [36]. One meta-analysis study that included data but only provided by abstract limited the preciseness of the outcomes [13]. Thus, in terms of clinically meaningful surrogate endpoints to estimate the activity of anti-PD-1 and anti-PD-L1 in treating advanced cancers such as esophageal or G/GEJ cancer, as the current study showed, improving the OS post-treatment with PD-1/PD-L1 inhibitors could provide important information on optimal treatment modalities.
In addition, the current study demonstrated that PD-1 inhibitor usage had a stronger association with improved OS in esophageal or G/GEJ cancer treatment than did PD-L1 inhibitors. In addition to blocking PD-1 and PD-L1 as well as reinvigorating T-cell activity, PD-1 inhibitors competitively bind to PD-1 along with PD-L1 by sharing overlapping binding surfaces [37]. Compared to PD-1 inhibitors, PD-L1 inhibitors are usually not associated with significant conformational changes in PD-L1 [13]. Moreover, such discrepancies might be related to the substantial variations in the mechanisms of action of a single PD-1/PD-L1 blockade agent [15,38]. Based on discrepancies in the mechanisms of action between PD-1/PD-L1 inhibitors, more attention is being paid to the differences between anti-PD-1 and anti-PD-L1 treatments in clinical practice, and evidence-based analysis to understand their comparable efficacies is urgently required [15]. Taking the above into consideration, an evaluation of the differences in response efficacy of anti-PD-1 and anti-PD-L1 usage in various cancer types was conducted in the present study. To this end, better objective response rates and durations of response were observed when treatment with PD-1 inhibitor was carried out [35]. Although Koneru et al. indirectly compared the safety and efficacy of PD-1 and PD-L1 antibodies across solid tumors and reported no significant differences [39], a prior meta-analysis with head-to-head comparison nonetheless demonstrated favorable OS outcomes of PD-1 inhibitors in cancer treatment [40]. Although the current outcome specifically supports PD-1 inhibitors as optimal treatment for esophageal or G/GEJ cancer rather than PD-L1, considering the limited number of studies conducted with PD-L1, more caution should be applied in clinical practice [41].
As to the safety issues associated with the administration of PD-1/PD-L1 inhibitors in esophageal or G/GEJ cancer patient treatment, the current study elucidated no significant differences of developing SAEs in the anti-PD-1/PD-L1 therapy group as compared to control groups. Enhancing the immune response against cancer cells, particularly in the case of adverse events-called immune-related adverse events (irAEs)-resulting from immune checkpoint inhibitors, such as PD-1/PD-L1 inhibitors, should be a primary focus [42]. Such toxicities are different from those caused by traditional chemotherapy, which affects whole cells in the body, resulting in significant reactions such as blood cell reduction or abnormal gastrointestinal, liver, and renal functions [43]. Although PD-1/PD-L1 inhibitors as immune checkpoint blockers can cause a widespread immune response affecting several organs, targeting only the immune system without affecting other cells in the body could possibly reduce the severity of these adverse events [44]. A prior metaanalysis involving 3450 patients also demonstrated a higher prevalence of irAEs (such as all-grade rash, pruritus, or hyperthyroidism) with a generally lower risk of adverse events and treatment discontinuation compared to chemotherapy [45]. When evaluating the risks and benefits of SAEs, PD-1/PD-L1 inhibitor use could be more beneficial in treating patients with esophageal or G/GEJ cancer.
The present study had several limitations. Firstly, to compare efficacy, the OS, and the PFS, more studies need to be performed with each subgroup. A lack of studies may cause uncertain biases while interpreting results from the analyses [11]. However, recently, more studies in various settings have been performed, which could not meet the inclusion criteria of the current study. Thus, we expect that more updated information of OS or PFS relevant to using anti-PD-1 and PD-L1 will be provided by pivotal studies soon [4,28,46]. Secondly, the control groups of the included studies received different types of chemotherapies. Since different SOCs were indicated in clinical practice guidelines such as National Comprehensive Cancer Network (NCCN) guidelines depending on the patients' status, various chemotherapies were dosed to the control groups in the included studies.
Furthermore, this study did not assess the cost-effectiveness of PD-1/PD-L1 inhibitor usage in treating patients with esophageal or G/GEJ cancer. However, this topic is beyond the scope of this work and can be further investigated in subsequent studies. Finally, ethnic or regional variations were not evaluated in the current study. Although such differences play a role in the development of gastric cancer [1], in general, the efficacy and toxicities of PD-1/PD-L1 inhibitors do not vary drastically across geographic regions or races [41]. Hence, more studies evaluating these differences in the use of PD-1/PD-L1 inhibitors are necessary.

Conclusions
In conclusion, our analysis revealed that PD-1/PD-L1 inhibitors significantly prolonged the OS as compared to the control, while no significant effect on the PFS was observed in patients with esophageal or G/GEJ cancer. Compared to the PD-L1 inhibitors, PD-1 inhibitors were more strongly associated with improved OS in these patients. Coming to the development of SAEs, PD-1/PD-L1 inhibitor usage did not contribute to an increase in SAEs in esophageal or G/GEJ cancer. Although limited improvement of the PFS was observed, as a measure of efficacy, the OS increment associated with PD-1/PD-L1 inhibitor use might provide important information for determining the optimal treatment for esophageal or G/GEJ cancer. Furthermore, considering the significant interventionspecific differences shown in the subgroup analysis, more individualized therapies should be implemented in clinical practice.