Efficacy and safety of nivolumab plus chemotherapy in patients with advanced gastric cancer with massive ascites

Abstract

BACKGROUND

Chemotherapy with an immune checkpoint inhibitor is one of the standard regimens for treating advanced gastric cancer (AGC). Ascites and peritoneal dissemination are common complications and poor prognostic factors of AGC; however, reports regarding its efficacy and safety in patients with AGC and massive ascites are limited.

AIM

To evaluate the safety and efficacy of nivolumab combined with chemotherapy in patients with AGC and ascites.

METHODS

We retrospectively collected clinical data from 124 patients with AGC who received chemotherapy plus nivolumab as first-line treatment from July 2017 to December 2024. Based on computed tomography scans, massive or moderate ascites were classified as high ascites burden (HAB), whereas mild or no ascites were classified as low ascites burden.

RESULTS

Ascites was detected in 47 patients (38%); 26 (21%) were classified into the HAB group. Patients in the HAB group exhibited a significantly poorer performance status, a higher prevalence of diffuse-type histology, and lower programmed cell death ligand 1 (PD-L1) expression. Combination therapy with FOLFOX and neutropenia was significantly more common in the HAB group. Progression-free survival (PFS) (4.4 months vs 9.3 months, P = 0.0012) and overall survival (OS) (7.3 months vs 21.2 months, P < 0.0001) were significantly poorer in the HAB group. However, an improvement in ascites was observed in 61.5% of patients in the HAB group. PD-L1 expression did not correlate with either PFS or OS in the HAB group.

CONCLUSION

Nivolumab plus chemotherapy demonstrated modest efficacy and acceptable toxicity in patients with AGC and HAB.

Key Words: Gastric cancer; Ascites; Nivolumab; Chemotherapy plus nivolumab; Immune checkpoint inhibitor

Core Tip: The combination of chemotherapy and immune checkpoint inhibitors has become standard therapy for advanced gastric cancer (AGC). However, data on efficacy in patients with AGC and high ascites burden (HAB) remain limited. Chemotherapy combined with nivolumab has demonstrated modest efficacy in AGC patients with HAB (progression-free survival: 4.4 months; overall survival: 7.3 months). Chemotherapy with nivolumab improved ascites in 56% of patients. Furthermore, these results showed no correlation with programmed cell death ligand 1 (PD-L1) expression. These results suggest that nivolumab chemotherapy could be a therapeutic option for patients with AGC and HAB, regardless of PD-L1 expression.

INTRODUCTION

Palliative chemotherapy is the standard of care for unresectable, recurrent, advanced gastric cancer (AGC). In recent years, the addition of immune checkpoint inhibitors to combination therapy with 5-fluorouracil (5-FU) and platinum-based agents has been reported to improve prognosis in several trials. In the CHECKMATE-649 trial of human epidermal growth factor receptor 2 (HER2)-negative AGC, chemotherapy combined with nivolumab significantly prolonged overall survival (OS) compared with chemotherapy alone (13.8 months vs 11.6 months; HR: 0.80; P < 0.001)[1]. In addition, pembrolizumab and tislelizumab show additive effects when combined with chemotherapy, making the combination of chemotherapy and immune checkpoint inhibitors the standard treatment for HER2-negative AGC[2,3].

Peritoneal dissemination is one of the most common forms of metastasis in gastric cancer and can lead to complications such as bowel obstruction and ascites. Massive ascites is observed in approximately 3%-10% of AGC cases, resulting in a poor prognosis[4,5]. Patients with AGC and massive ascites are often excluded from phase III trials. The JCOG1108/WJOG7312G phase II/III trial is the only phase III trial evaluating the efficacy of 5-FU/L-leucovorin (FL) plus paclitaxel (FLTAX) in patients with AGC and massive ascites or impaired oral intake due to peritoneal dissemination. However, the FLTAX regimen did not demonstrate superiority to FL regarding OS (7.3 months vs 6.1 months; P = 0.14)[4]. Similarly, WJOG10517G, a single-arm phase II study evaluating the efficacy of mFOLFOX6 (oxaliplatin plus leucovorin plus 5-FU) in the same AGC patient population, showed that OS was almost identical at 7.4 months[6,7]. However, there are currently no data on the efficacy and safety of combination therapy with chemotherapy and immune checkpoint inhibitors in patients with AGC and massive ascites.

Massive ascites can lead to a poor quality of life due to symptoms such as abdominal pain, abdominal fullness, and vomiting. From a clinical perspective, managing these complications with palliative care is crucial, and new approaches are needed. Therefore, this study aimed to retrospectively evaluate the safety and efficacy of nivolumab combined with chemotherapy in patients with AGC and ascites.

MATERIALS AND METHODS

Patients

This retrospective study included consecutive patients with AGC who received chemotherapy plus nivolumab as first-line treatment at Ichinomiyanishi Hospital and Kobe City Medical Center General Hospital between July 2017 and December 2024. All data were retrospectively collected from the patients’ medical records. The inclusion criteria were as follows: (1) Unresectable or metastatic gastric or gastroesophageal junction cancer; (2) Histologically or cytologically confirmed adenocarcinoma; (3) Confirmed HER2-negative cancer; (4) Received nivolumab plus platinum doublet chemotherapy as the first-line treatment; and (5) Evaluable lesions (whether lesions are measurable is irrelevant). This study was approved by the Institutional Review Board of Kobe City Medical Center General Hospital (No. zn250106) and conformed to the guidelines of the Declaration of Helsinki. The requirement for informed consent was waived because of the retrospective nature of the study and the opportunity to opt out. The data cutoff date was April 8, 2025.

Biomarker analysis

The levels of programmed cell death ligand 1 (PD-L1) were measured using an immunohistochemistry PD-L1 antibody assay (Dako 28-8). The combined positive score (CPS) was calculated as the number of PD-L1-staining tumour and immune cells divided by the total number of viable tumour cells, multiplied by 100, with a maximum score of 100. Microsatellite instability was measured by multiplex PCR fragment analysis (MSI Test Kit, FALCO). Claudin (CLDN) 18 expression was assessed using the VENTANA CLDN18 (43-14A) RxDx assay, with positivity defined as 75% of tumour cells showing moderate to strong membranous staining.

Evaluation of ascites

Based on studies by Nakajima et al[4] and Masuishi et al[5], we classified ascites using computed tomography (CT) scans as follows: Massive, extending throughout the abdominal cavity; moderate, neither mild nor massive; mild, located only in the upper or lower abdominal cavity; and no ascites, ascites not detected. Cases with massive or moderate ascites were defined as having a high ascites burden (HAB), and cases with mild or no ascites were defined as having a low ascites burden (LAB). An improvement in ascites was defined as a decrease in one or more levels of ascites on the CT scan at the time of evaluation.

Evaluation and statistical analysis

The Eastern Cooperative Oncology Group performance status (PS) was defined by clinical oncologists. Tumour response was assessed according to the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1. Toxicity was assessed using the CTCAE version 5.1. Categorical variables were compared using Fisher’s exact test, and continuous variables were compared using two-sample t-tests. In patients with at least one measurable lesion, the tumour response was retrospectively reassessed by the investigator according to the RECIST version 1.1. In cases without evaluable lesions, the treatment efficacy was defined as not being evaluated. OS was assessed from the start of chemotherapy plus nivolumab until death. Patients who were alive or had missing data at the data cutoff point were censored. Progression-free survival (PFS) was assessed from the date of initiation of chemotherapy plus nivolumab until disease progression was confirmed. Patients with no information regarding disease progression were treated as censored cases. OS and PFS were estimated using the Kaplan-Meier method. Statistical analyses were performed using JMP, version 17 (SAS Institute Inc., Cary, NC, United States).

RESULTS

Patient characteristics

Among the 124 patients with AGC who received chemotherapy plus nivolumab as a first-line regimen, 98 (79%) had LAB and 26 (21%) had HAB (Figure 1). The patient characteristics are presented in Table 1. Poor PS, diffuse histology, and peritoneal dissemination were significantly more common in the HAB group than in the LAB group. PD-L1 expression was predominantly < 1% in the HAB group, whereas expression levels of ≥ 5% were more common in the LAB group. Regarding combination chemotherapy regimens, FOLFOX was administered more frequently in the HAB group, whereas SOX was administered more frequently in the LAB group.

Figure 1 Study flow. AGC: Advanced gastric cancer.

Table 1 Patients’ characteristics, n (%).

		All (n = 124)	HAB (n = 26)	LAB (n = 98)	P value
Age (years)	Median (range)	70 (23-87)	71.5 (38-86)	70 (23-87)	0.9211
Sex	Male	82 (66)	14 (54)	68 (69)	0.1641
PS	0	44 (35)	1 (4)	43 (44)	< 0.0001
	1	60 (48)	16 (62)	44 (45)
	2 or 3	20 (16)	9 (35)	11 (11)
Location	Esophago-gastric junction	20 (16)	1 (4)	19 (19)	0.0716
	Gastric	104 (84)	25 (96)	79 (81)
Gastrectomy	Yes	34 (27)	5 (19)	29 (30)	0.3345
Pathology	Diffuse type	75 (60)	24 (92)	51 (52)	0.0009
Number of metastatic organs	2 or more	62 (50)	15 (58)	47 (48)	0.5086
Liver metastasis	Yes	38 (31)	4 (15)	34 (35)	0.0921
Peritoneum dissemination	Yes	75 (60)	25 (96)	50 (51)	< 0.0001
Ascites	Yes	47 (38)	26 (100)	21 (21)	< 0.0001
High ascites burden	Yes	26 (21)	26 (100)	0	< 0.0001
Regimen	SOX	67 (54)	8 (31)	59 (60)	0.0037
	FOLFOX	51 (41)	18 (69)	33 (34)
	CAPOX	6 (5)	0	6 (6)
PD-L1 CPS	Less than 1	24 (19)	10 (38)	14 (14)	0.0082
	1 ≤ CPS < 5	36 (29)	6 (23)	29 (30)
	5 or more	45 (36)	5 (19)	41 (42)
	Unknown	19 (15)	5 (19)	14 (14)
MSI	MSI-high	5 (4)	0	5 (5)	0.3489
	MSS	74 (60)	18 (69)	56 (57)
	Unknown	45 (36)	8 (31)	37 (38)
CLDN 18.2	Positive	14 (11)	4 (15)	10 (10)	0.0767
	Negative	29 (23)	2 (8)	27 (28)
	Unknown	81 (65)	20 (77)	61 (62)

PS: Performance status; CPS: Combined positive score; MSI: Micro satellite instability; MSS: Microsatellite stable; HAB: High ascites burden; LAB: Low ascites burden.

Efficacy

The median follow-up duration was 20.9 months (range: 3.5-94.3 months). The overall response rate (ORR) was 42.3% and the disease control rate (DCR) was 83.9%. The median PFS was 8.7 months (95%CI: 6.8-11.0), and the median OS was 19.6 months (95%CI: 16.0-22.5 months) (Table 2 and Figure 2).

Figure 2 Kaplan-Meier curves of progression-free survival and overall survival in all patients. A: Progression-free survival; B: Overall survival. PFS: Progression-free survival; OS: Overall survival.

Table 2 Response rate and disease control rate of patients with evaluable lesions.

	Total (n = 124)	HAB (n = 26)	LAB (n = 98)	P value
Best overall response
Complete response	2	0	2
Partial response	48	8	40
Stable disease	49	10	39
Progressive disease	19	6	13
Not evaluated	6	2	4
Objective response rate (%)	42.3	33.3	44.9	0.3615
Disease control rate (%)	83.9	75.0	86.2	0.2152
Improvement of ascites (%)	59.6	61.5	57.1	0.7745

HAB: High ascites burden; LAB: Low ascites burden.

When comparing the HAB and LAB groups, the median PFS (4.4 months vs 9.3 months, P = 0.0012) and OS (7.3 months vs 21.2 months, P < 0.0001) were significantly lower in the HAB group (Figure 3). There were no significant differences in the ORR (33.3% vs 44.9 %, P = 0.3615) and DCR (75% vs 86.2%, P = 0.2152) between the groups. Ascites showed an overall improvement rate of 59.6%, with rates of 61.5% and 57.1% in the HAB and LAB groups, respectively (P = 0.7745).

Figure 3 Kaplan-Meier estimates of progression-free survival and overall survival according to low ascites burden and high ascites burden. A: Progression-free survival; B: Overall survival. LAB: Low ascites burden; HAB: High ascites burden.

Association between PD-L1 expression and efficacy in the HAB group

We performed an exploratory analysis to investigate the relationship between PD-L1 expression and the treatment response. The CPS for PD-L1 was measured in 105 patients. When comparing CPS ≥ 5 and CPS < 5, no statistically significant differences were observed in PFS (9.2 months vs 7.5 months, P = 0.3557) and OS (17.4 months vs 19.6 months, P = 0.9501) in the overall population (Supplementary Figure 1). Similarly, in the HAB group, no statistically significant differences in PFS (5.4 months vs 4.1 months, P = 0.4251) and OS (7.1 months vs 8.3 months, P = 0.6742) were observed between CPS ≥ 5 and CPS < 5 (Figure 4). The ascites improvement rate was 56% in patients with a PD-L1 CPS of < 5, and 60% in patients with a CPS of ≥ 5 (P = 1.000).

Figure 4 Kaplan-Meier estimates of progression-free survival and overall survival according to programmed cell death ligand 1 expression in patients with high ascites burden. A: Progression-free survival; B: Overall survival. CPS: Combined positive score; PD-L1: Programmed cell death ligand 1.

Univariate and multivariate analysis of PFS and OS

We performed univariate and multivariate analyses of PFS and OS, using the following variables: Age, PS, prior gastrectomy, chemotherapy regimen, presence of HAB, and presence of liver metastases (Supplementary Table 1). In the multivariate analysis for PFS, PD-L1 ≥ 1 and HAB were identified as adverse prognostic factors. In the multivariate analysis for OS, PS ≥ 1, PD-L1 ≥ 1, and the presence of HAB were significant poor prognostic factors. Overall, PD-L1 ≥ 1 and HAB were consistently identified as poor prognostic factors for both PFS and OS.

Safety

The adverse events are listed in Tables 3 and 4. When comparing adverse events of all grades, peripheral neuropathy was significantly more common in the LAB group (23% vs 44%, P = 0.014). When comparing grade 3 or higher adverse events, neutropenia was significantly more common in the HAB group (46% vs 18%, P = 0.01). When comparing the immune-related adverse events between the two groups, skin reactions were significantly more common in the LAB group (0% vs 16%, P = 0.023), whereas no significant difference was observed in the incidence of grade 3 or higher immune-related adverse events. The overall transition rate to second-line treatment was 69%, with rates of 54% in the HAB group and 74% in the LAB group (P = 0.1255).

Table 3 Adverse events, n (%).

	All (n = 124)		HAB (n = 26)		LAB (n = 98)		P value
	All	Grade 3 or more	All	Grade 3 or more	All	Grade 3 or more	All	Grade 3 or more
	111 (90)	51 (41)	23 (88)	14 (54)	88 (90)	37 (38)	1.000	0.179
Neutropenia	44 (35)	30 (24)	12 (46)	12 (46)	32 (33)	18 (18)	0.258	0.010
Anaemia	18 (15)	7 (6)	4 (15)	3 (12)	14 (14)	4 (4)	0.546	0.159
Low platelet	17 (14)	6 (5)	3 (12)	2 (8)	14 (14)	4 (4)	1.000	0.281
Nausea	30 (24)	2 (2)	5 (19)	0	25 (26)	2 (2)	0.794	1.000
Fatigue	42 (34)	1 (1)	6 (23)	0	36 (37)	1 (1)	0.246	1.000
Anorexia	34 (27)	5 (4)	6 (23)	1 (4)	28 (29)	4 (4)	0.631	1.000
Peripheral sensory neuropathy	49 (40)	1 (1)	6 (23)	0	43 (44)	1 (1)	0.014	1.000
Stomatitis	6 (5)	0	0	0	6 (6)	0	0.342	-
Diarrhoea	20 (16)	3 (2)	2 (8)	1 (4)	18 (18)	2 (2)	0.240	1.000
Constipation	11 (9)	0	0	0	11 (11)	0	0.118	-
Oedema	6 (5)	0	0	0	6 (6)	0	0.583	-
Allergy	9 (7)	7 (6)	1 (4)	1 (4)	8 (8)	6 (6)	0.201	0.342
Dysgeusia	10 (8)	0	2 (8)	0	8 (8)	0	0.714	-
Infection	10 (8)	4 (3)	3 (12)	2 (8)	7 (7)	2 (2)	0.395	0.281
Hand-foot syndrome	5 (4)	0	0	0	5 (5)	0	0.583	-

HAB: High ascites burden; LAB: Low ascites burden.

Table 4 Immune related adverse events, n (%).

	All (n = 124)		HAB (n = 26)		LAB (n = 98)		P value
	All	Grade 3 or more	All	Grade 3 or more	All	Grade 3 or more	All	Grade 3 or more
Any adverse event	45 (36)	12 (10)	6 (23)	2 (8)	39 (40)	10 (10)	0.168	1.000
Skin reaction	16 (13)	0	0	0	16 (16)	0	0.023	-
Hypothyroidism	7 (6)	0	1 (4)	0	6 (6)	0	1.000	-
Adrenal insufficiency	5 (4)	2 (2)	0	0	5 (5)	2 (2)	0.583	1.000
Hypopituitary function	2 (2)	0	0	0	2 (2)	0	1.000	-
Pneumonitis	9 (7)	3 (2)	4 (15)	1 (4)	5 (5)	2 (2)	0.091	0.510
Colitis	6 (5)	2 (2)	0	0	6 (6)	2 (2)	0.342	0.579
Hepatitis	2 (2)	2 (2)	0	0	2 (2)	2 (2)	1.000	1.000
Nephritis	2 (2)	1 (1)	1 (4)	0	1 (1)	1 (1)	0.377	1.000
Myositis	1 (1)	1 (1)	0	0	1 (1)	1 (1)	1.000	1.000
Polymyalgia	1 (1)	0	0	0	1 (1)	0	1.000	-
Arthritis	1 (1)	0	0	0	1 (1)	0	1.000	-
Cholangitis	1 (1)	1 (1)	1 (4)	1 (4)	0	0	1.000	1.000

HAB: High ascites burden; LAB: Low ascites burden.

DISCUSSION

To our knowledge, this is the first retrospective observational study to evaluate the efficacy and safety of chemotherapy combined with nivolumab in patients with AGC and massive ascites. A higher proportion of patients in the HAB group received FOLFOX than in the LAB group (69% vs 34%). This is probably because the efficacy and safety of FOLFOX have been reported more extensively in patients with AGC accompanied by ascites[4,5]. Although chemotherapy plus nivolumab showed an OS comparable to that reported in previous studies, the ascites improvement rate (61.5%) was better than that reported for FLTAX (36.2%) and FOLFOX (30%)[4,5]. Regarding adverse events, grade ≥ 3 neutropenia was more common in the HAB group, while peripheral neuropathy and skin reactions were more common in the LAB group. No other toxicities were significantly different between groups. When adverse events were analysed by regimen, neutropenia was observed more frequently in the FOLFOX group. The higher prevalence of FOLFOX use in the HAB subgroup is considered to be associated with the increased incidence of neutropenia observed in this group. These results suggest that although OS remains poor and caution is warranted regarding neutropenia, chemotherapy plus nivolumab may be tolerable and potentially effective in reducing ascites in patients with AGC and HAB.

PD-L1 expression is known to be a predictive biomarker for the efficacy of immune checkpoint inhibitors[1,2], and previous reports have suggested that chemotherapy in combination with nivolumab further prolonged OS in patients with PD-L1 CPS ≥ 5[1]. In our study, most patients in the HAB group had a PD-L1 CPS of < 5; however, no correlation between PD-L1 expression and OS was observed. In contrast, the addition of nivolumab to chemotherapy has been reported to improve response rates, even in patients with low PD-L1 expression[1], and the response rate may be associated with a reduction in ascites. These findings suggest that the addition of nivolumab may improve ascites in patients with HAB regardless of PD-L1 expression status.

In the HAB group, diffuse-type histology was significantly more common than in the LAB group. Furthermore, an analysis of other biomarkers showed a trend towards a higher prevalence of increased CLDN18.2 expression in the HAB group. Several studies have reported that CLDN18.2 expression is higher in diffuse AGC than in intestinal AGC[8-14]. An integrated analysis of the SPOTLIGHT and GLOW trials in CLDN18.2-positive AGC suggested that zolubetuximab may show promising efficacy, even in the diffuse type[15]. These results suggest that treatment selection based on biomarkers other than PD-L1, such as CLDN18.2, may become increasingly important for improving OS in AGC with HAB.

Recently, several trials have investigated the efficacy of intraperitoneal chemotherapy for AGC with peritoneal dissemination. In the phase III PHOENIX-GC trial that compared intraperitoneal and intravenous paclitaxel plus S-1 with S-1 plus cisplatin, the benefit of intraperitoneal administration was not observed at the primary endpoint of OS. However, when adjusted for ascites, the HR for OS was 0.59 (95%CI: 0.39-0.87, P = 0.008), indicating significantly better outcomes in the intraperitoneal group[16]. In the phase III DRAGON-01 trial, which compared intraperitoneal and intravenous paclitaxel plus S-1 with S-1 plus paclitaxel alone, patients with AGC without distant metastases other than peritoneal dissemination were enrolled, with over 60% having ascites of 300 mL or more. At the primary OS endpoint, intraperitoneal administration significantly improved the OS (19.4 months vs 13.9 months, HR: 0.66, P = 0.0056)[17]. These results suggest that the combination of systemic chemotherapy and intraperitoneal administration may be beneficial in AGC with HAB, warranting further investigation.

This study has some limitations. First, it was a retrospective study with a relatively small sample size. The efficacy of chemotherapy combined with nivolumab for managing ascites is controversial, and prospective interventional studies are warranted. A further limitation is that, given the retrospective study design, we were unable to assess changes in quality of life. A multicentre phase II study of mFOLFOX6 plus nivolumab in gastric cancer with severe peritoneal metastases, WJOG16322G (Japan Registry of Clinical Trials 041220164.), is ongoing[18], and the results are eagerly awaited. Second, biomarker exploration in this trial was insufficient. Currently, treatment development targeting not only PD-L1 and microsatellite instability, but also CLDN18.2 and FGFR2, is progressing for gastric cancer. Future studies on AGC with HAB should include further biomarker analyses.

CONCLUSION

In conclusion, chemotherapy plus nivolumab showed a modest improvement in OS, tolerable safety, and improvement in ascites in patients with AGC with HAB. As this was a retrospective, small-scale study, future prospective studies are warranted to establish more effective treatments for AGC with HAB.

ACKNOWLEDGEMENTS

We gratefully thank the participants of this study and their families, as well as the study investigators, teams, and data managers (Mizuho Takahashi, Kie Takeda, and Ryo Ogura) for their contributions to the publication of this study.