Prognostic Nutrition Index as Predictor for Nivolumab Monotherapy in Advanced Gastric Cancer Patients with Peritoneal Metastases


 Although nivolumab shows survival benefits for patients with advanced gastric cancer (AGC), predictive biomarkers for nivolumab treatment in AGC remain unclear, especially in the case of peritoneal metastases. Therefore, this study investigated the clinical significance of the prognostic nutrition index (PNI), reflecting the host nutritional status and immunity, in AGC patients undergoing nivolumab monotherapy. This study retrospectively analyzed 53 AGC patients who received nivolumab between October 2017 and February 2021. Among them, 35 patients with peritoneal metastases were reviewed to investigate the relationship between the PNI and oncological outcomes. The PNI was calculated as 10 × serum albumin level (g/dl) + 0.005× total lymphocyte count (per mm3) at the first administration of nivolumab. With a median follow-up duration of 2.0 (0.3 − 13.5) months, the median overall survival (OS) was 2.0 months. The overall response and disease-control rates were 0.0% and 20.0%, respectively. Among the 35 patients, 13 patients were identified as a high-PNI group. In the univariate analysis, the high-PNI group showed a significantly longer PFS and OS than the low-PNI group. In the multivariate analysis, the high-PNI was independently associated with a longer PFS (p = 0.021) and OS (p = 0.022). The PNI can be useful for predicting PFS and OS in AGC patients with peritoneal metastases. However, further studies are required to validate these results in AGC and new strategies are needed to improve the outcome for AGC patients with peritoneal metastases.


Introduction
Despite improved outcomes for advanced gastric cancer (AGC) via the introduction of several effective combination chemotherapies and identi cation of immune checkpoint inhibitors (ICIs), distant metastases remain frequent and are associated with a dismal prognosis, where peritoneal implantation is the most common metastatic site, with an incidence of 53.5% [1,2]. Since peritoneal metastases exhibit aggressive behavior and biological resistance to chemotherapy, the treatment of patients with peritoneal metastases is rarely successful with only a 2% ve-year survival rate [3]. Thus, novel approaches are needed to overcome the limitation of conventional cytotoxic chemotherapy for AGC patients with peritoneal metastases.
ICIs are already recognized standard treatment for patients with recurrent or metastatic AGC [4]. For example, a phase III (ATTRACTION-2) trial that compared nivolumab targeting the programmed cell death protein-1 (PD-1) with a placebo in 493 Asian patients showed a survival bene t in third-or later line treatment [5]. Moreover, a recent global phase III (CheckMate 649) trial found that nivolumab in combination with chemotherapy was the rst PD-1 inhibitor to demonstrate superior overall survival (OS) and progression-free survival (PFS) as a rst-line treatment [6]. Plus, a phase II/III (ATTRACTION-4) trial conducted in Asia reported a signi cantly improved PFS [7]. Notwithstanding, subgroup analyses of these data show disappointing results for peritoneal metastases, although there have been a few case reports of successful treatment when using nivolumab for AGC with peritoneal metastases [8]. Yet, the effects of ICIs seem to vary depending on the tumor biology, with various clinical factors also in uencing the response to ICIs [4]. Thus, evaluating the clinical features and treatment outcomes for peritoneal metastases treated with nivolumab may help to provide more effective therapeutic strategies for AGC patients.
The prognostic nutrition index (PNI) is calculated based on the serum albumin level and peripheral blood lymphocyte count and was originally developed to predict the risk of postoperative complications mainly in surgical patients by assessing the preoperative nutritional status [9]. Notably, the total lymphocyte count can have a favorable impact on the tumor inhibiting effects of ICIs and be used as an index for evaluating the host immunity and response to ICIs [10]. Meanwhile, the serum albumin level can re ect the host immunologic status in AGC patients with peritoneal metastases, where cancer progression in the diminished the oral intake, leading to downregulation of the nutritional status of the patient [11]. Thus, there is increasing evidence that the PNI can be an effective prognostic marker, as well as a predictive indicator related to ICIs for various solid tumors [12][13][14]. Accordingly, this study investigated the clinical signi cance of the PNI for predicting the therapeutic effects of nivolumab in AGC with peritoneal metastases.

Study design and patients
This study retrospectively examined the medical records of all patients with unresectable advanced or recurrent gastric cancer who received nivolumab treatment at Kyungpook National University Chilgok Hospital (KNUCH) between October 2017 and February 2021. The clinical parameters, such as age, sex, performance status, histology, number of organs with metastases, and laboratory ndings at the time of the rst nivolumab administration were reviewed from the hospital database. Nivolumab was administered by intravenous infusion at a dose of 3mg/kg every 2 weeks until disease progression or unacceptable toxicity. The study was approved by the Institutional Review Board of KNUCH.

De nition of PNI
The PNI was calculated as 10 × serum albumin level (g/dl) + 0.005× total lymphocyte count (per mm 3 ) at the rst administration of nivolumab. The patients were classi ed as either low (< 40) or high (≥ 40) as the reference [15].
Statistical analysis PFS was measured from the time of commencing treatment to disease progression or death. OS was estimated from the date of diagnosis to death from any cause. The tumor response was evaluated according to the response evaluation criteria in solid tumors (RECIST) version 1.1. The survival analysis used the Kaplan-Meier method with a log-rank test. A multivariate analysis was performed using variables with a value of p < 0.1 in a univariate analysis using Cox's proportional hazards model to derive a potentially suitable set of predictors. Two-sided p-values of < 0.05 were considered signi cant. The statistical analyses were performed using SPSS software version18.0 (SPSS, Inc., Chicago, IL, USA).

Response and survival outcomes for nivolumab
No patient exhibited a complete response or partial response. 7 patients showed stable disease, giving a disease control rate of 20.0%. At the last follow-up, the median follow-up duration was 2.0 (0.-13.5) months. During the analyses, 31 (88.6%) patients experienced recurrence and 33 (94.3%) patients died.
The median PFS was 1.1 months and the median OS was 2.0 months (Fig. 1A and 1B).

Relationship between PNI and survival outcome
In the univariate analysis, the high-PNI group showed a signi cantly longer PFS and OS than the low-PNI group (Fig. 2). In the multivariate analysis using a Cox proportional hazard model adjusted for age, histologic differentiation, and ECOG, the high-PNI group was independently associated with a longer PFS (Hazard ratio = 0.366, 95% con dence interval (CI) = 0.155-0.861, p = 0.021) and OS (Hazard ratio = 0.349, 95% CI = 0.142-0.860, p = 0.022) ( Table 2).

Discussion
The clinical signi cance of the PNI was investigated in 35 patients with metastatic AGC who underwent nivolumab mostly as second-or third-line therapy. As a result, the PNI was identi ed as an independent predictive factor of PFS and OS, suggesting that the PNI may be a useful biomarker to predict the response to ICIs in AGC patients with peritoneal metastases.
The molecular mechanisms by which AGC undergoes peritoneal metastases are not completely clear and considered as a multistep process, including the detachment of cancer cells from the primary tumor, survival in the free abdominal cavity, attachment to the distant peritoneum, invasion into the subperitoneal space and proliferation with angiogenesis [16]. In particular, various molecules, such as Ecadherin, chemokines, growth factor receptors/ligands, immune cells, and extracellular matrix, broadly contribute during the invasion of the gastric wall and migration of the cancer cells [17]. These factors all play an essential role in the progression and chemoresistance of peritoneal metastases [18]. Although recent studies of AGC patients with peritoneal metastases have attempted to demonstrate improved survival with systemic chemotherapy and hyperthermic intraperitoneal chemotherapy (HIPEC)/peritonectomy, the long-term outcomes remain dismal [19]. In the present study, peritoneal metastases showed poor outcomes even after treatment with nivolumab, as consistent with previous study results. Subgroup analyses of the ATTRACTION-2 trial found no signi cant bene t from nivolumab in patients with peritoneal metastases. Similarly, Aarnink et al. reported that ICIs used in non-small cell lung cancer (NSCLC) patients with peritoneal metastasis were associated with poor PFS and OS [20].
Recent studies also showed that diffuse and signet ring cell histologies had poor outcomes with nivolumab treatment, indicating that these types seemingly promote AGC cell migration, invasion, and enhanced peritoneal metastases [6, [21][22][23]. Therefore, since these ndings and the current results suggest that peritoneal metastases have a relatively limited response to ICIs, the role of ICIs in AGC with peritoneal metastases requires further clari cation.
Recent research has been focused on identifying robust predictive biomarkers for AGC treated with ICIs. The PNI, rst reported by Onodera et al., is a well-known in ammatory prognostic marker for several solid tumors [24]. The PNI includes the serum lymphocyte and albumin levels. There is increasing evidence that the lymphocyte ratio can be an effective predictive indicator related to ICIs for various solid tumors, having a favorable effect on their tumor inhibiting properties [4]. Moreover, albumin is an acute-phase protein and decreases in response to in ammation [25]. Thus, low levels of albumin may re ect cancerinduced malnutrition and have a negative impact on prognosis. Therefore, indicating a poor diet in the case of AGC with peritoneal metastases, these factors may help to determine the predictive value of ICIs including nivolumab in these patients. Several studies covering a variety of cancers: gastric cancer, colorectal cancer, NSCLC, and genitourinary cancer treated with ICIs found that a low PNI resulted in worse OS and PFS across various types of malignancies, which is consistent with the current study results [9,12,26,27]. Plus, another recent study showed a statistically signi cant outcome with a large number of gastric cancer patients. Mohri et al. analyzed 365 CRC patients who underwent curative resection, and reported that a PNI < 45 independently affected OS [9]. This particular parameter also has several advantages for daily clinical practice: ready to use, easily measurable, repeatable, and relatively economical to evaluate [4]. Thus, considering its recognized in uence on host nutritional status, immunity, and cancer, the PNI can be effectively used to predict the therapeutic effects of nivolumab in AGC patients with peritoneal metastases.
In summary, the PNI can be useful for predicting PFS and OS in AGC patients with peritoneal metastases. However, further studies are required to validate these results in AGC and new strategies are needed to improve the outcome for AGC patients with peritoneal metastases.

Declarations
Funding The authors received no nancial support for the research, authorship, and/or publication of this article.
Con ict of interest The authors declare no con ict of interest.