Peripheral blood marker changes predict response and prognosis in locally advanced gastric cancer with neoadjuvant sintilimab plus XELOX

Abstract

BACKGROUND

Neoadjuvant immunotherapy with chemotherapy (NICT) is a relatively advanced treatment modality for locally advanced gastric cancer (LAGC). However, reliable peripheral blood biomarkers for predicting the efficacy and prognosis of this therapy remain lacking. We aimed to investigate and evaluate whether baseline and post-treatment changes in peripheral blood inflammatory markers and T lymphocyte subsets could serve as potential predictive and prognostic indicators for patients with LAGC receiving NICT.

AIM

To investigate the association between changes in peripheral blood markers, treatment response, and prognosis in patients with LAGC.

METHODS

We conducted a retrospective analysis of clinical data from 39 patients with LAGC who received neoadjuvant therapy with sintilimab and XELOX chemotherapy. Peripheral blood samples were collected at baseline and after treatment completion for complete blood count and T-cell subset analyses. The patients were divided into the major pathological response group (tumor regression grade = 0-1) and non-major pathological response group (tumor regression grade = 2-3). The independent predictors of therapeutic efficacy and factors related to progression-free survival were explored.

RESULTS

Multivariate analysis revealed that low-adhesion gastric cancer (non-poorly vs poorly cohesive gastric adenocarcinoma) (odds ratio = 8.522, P = 0.036) and a higher post-treatment monocyte-to-lymphocyte ratio (MLR) (odds ratio = 2.479, P = 0.035) were associated with a favorable response to treatment. A higher relative change rate in CD8+ T lymphocytes (hazard ratio = 0.970, P = 0.024) and diffuse gastric cancer (hazard ratio = 5.545, P = 0.011) may be potential prognostic factors for progression-free survival. The lymphocyte count (significant decrease from 1.44 × 10⁹/L to 1.18 × 10⁹/L, P < 0.001) and the platelet-to-lymphocyte ratio showed significant decreases. By contrast, the monocyte count, CD3+ and CD8+ T-cell counts, and MLR all showed significant increases (all P < 0.005).

CONCLUSION

Post-treatment MLR and relative change in CD8+ T-cell proportion are potential prognostic biomarkers for patients with LAGC receiving NICT, whereas diffuse-type gastric cancer may be related to poor prognosis.

Key Words: Gastric cancer; Peripheral blood; Neoadjuvant therapy; Immunotherapy; Sintilimab; Gastric signet ring cell carcinoma; Gastric carcinoma subtypes

Core Tip: This study investigated the changes in peripheral blood inflammatory markers and T-cell subsets in patients with gastric cancer at baseline and after neoadjuvant chemotherapy with sintilimab combined with XELOX. Results indicate that high post-treatment monocyte-to-lymphocyte ratio levels are associated with a favorable treatment response, a greater relative change in the proportion of CD8+ T cells may be associated with better prognosis, and diffuse-type gastric cancer may be associated with poorer clinical outcomes. These easily detectable biomarkers offer a new perspective for predicting treatment response and assessing prognosis; however, further validation is required.

INTRODUCTION

Gastric cancer is the most commonly occurring cancer of the gastrointestinal tract. Its incidence rate is relatively high worldwide, posing a serious threat to human health and safety[1]. According to 2022 GLOBOCAN data, there are approximately 1.09 million new confirmed cases and about 770000 deaths globally each year, making this disease a serious threat to human health and safety[2]. Each year, gastric cancer deaths in China represent nearly half of the global total; its incidence ranks second only to lung cancer, and it stands as the third leading cause of cancer mortality nationwide[3]. Because early-stage gastric cancer often shows no obvious symptoms, and symptoms such as vague upper abdominal pain and bloating are frequently confused with gastritis or gastric ulcers, patients are often not diagnosed in the early stages. Consequently, most patients are not diagnosed until the disease has reached a locally advanced stage. In locally advanced gastric cancer (LAGC), the tumor has typically invaded the depth of the gastric wall or is accompanied by lymph node metastasis, making complete resection difficult to achieve through surgery alone. Patients continue to carry a high postoperative risk of recurrence and metastasis; thus, comprehensive perioperative multidisciplinary treatment is crucial for the management of LAGC[4].

For a long time, platinum-based double-agent chemotherapy was the standard first-line treatment for LAGC; unfortunately, its efficacy in patients was less than ideal. With the advancement of cancer immunotherapy, immune checkpoint inhibitors (ICIs) have emerged, completely transforming the treatment landscape[5,6]. The MATTERHORN study found that the perioperative combination of durvalumab and the FLOT chemotherapy regimen reduced the risk of disease progression or recurrence by 29% and the risk of death by 22%, with a 3-year overall survival (OS) rate of 68.6% among patients[7]. Growing evidence indicates that for patients with LAGC, neoadjuvant chemotherapy combined with immunotherapy can activate immune responses and generate synergistic antitumor effects, thereby significantly improving disease response rates and prolonging progression-free survival (PFS)[8-10]. Although neoadjuvant chemoimmunotherapy offers promising efficacy and prospects, not all patients benefit from this approach alone. Therefore, the main challenge in current clinical practice lies in accurately identifying the patients with the greatest likelihood of benefiting from this treatment.

Inflammation plays a significant role in tumor progression. In neoadjuvant chemotherapy and immunotherapy regimens for gastric cancer, easily obtainable peripheral blood inflammatory markers including the monocyte-to-lymphocyte ratio (MLR) and systemic immune-inflammatory indices, among others, have been repeatedly shown to correlate with treatment efficacy. Furthermore, peripheral blood immune cell subsets, particularly CD8+ T cells, have a significant predictive value in immunotherapy[11-13]. However, few studies have investigated peripheral blood inflammatory markers and immune cell subsets in the context of neoadjuvant chemoimmunotherapy for gastric cancer. Moreover, the dynamic changes in these markers and their associations with treatment response and prognosis have not yet been elucidated.

In light of the above considerations, this retrospective study enrolled LAGC patients who underwent neoadjuvant therapy with the XELOX regimen plus sintilimab. Serial peripheral blood samples were assayed for inflammatory markers and lymphocyte subsets. The primary objective was to evaluate changes before and after treatment and to determine the relationship between these changes and pathological tumor regression and PFS.

MATERIALS AND METHODS

Study population

The retrospectively conducted study included 48 patients with gastric cancer who received neoadjuvant immunotherapy combined with chemotherapy at the First Affiliated Hospital of Zhejiang University School of Medicine between August 2019 and October 2020. After applying eligibility criteria, 39 patients were finally analyzed (Figure 1). This retrospective, single-center study received approval from the Clinical Research Ethics Committee of the First Affiliated Hospital, Zhejiang University School of Medicine. Given that this study is a retrospective, non-interventional study, the ethics committee waived the need for written informed consent.

Figure 1 Flow chart of patients selection. PD-1: Programmed death-1.

The inclusion criteria were as follows: (1) Gastric adenocarcinoma pathologically confirmed by preoperative gastroscopy; (2) cT3-4 stage regardless of N stage or M0 status; (3) Complete clinicopathological and postoperative pathological records; (4) Received at least three cycles of neoadjuvant therapy with sintilimab combined with the XELOX regimen, with the last treatment cycle completed within 4 weeks before surgery; and (5) Underwent radical gastrectomy after neoadjuvant therapy.

The exclusion criteria were as follows: (1) Unknown or different neoadjuvant chemoimmunotherapy regimen; (2) Presence of other concurrent malignancies; (3) Absence of complete blood count data during the period from completion of neoadjuvant therapy to within 1 week before surgery, specifically within 1 month after the last treatment cycle; (4) Disease progression confirmed during neoadjuvant chemoimmunotherapy requiring a switch to a first-line chemotherapy regimen; (5) Incomplete pathological data precluding tumor regression grade (TRG) assessment; and (6) Missing follow-up data precluding survival analysis.

Data collection

The general clinical data collected included the patients’ sex, age, primary tumor location, surgical procedure, clinical stage, and pathological stage. The histological classification, histological differentiation, and Lauren classification were also recorded. Routine blood parameters, tumor markers, and T lymphocyte subsets were measured at two time points: Within 1 week before the initiation of neoadjuvant chemoimmunotherapy (baseline) and within 1 week before surgery (post-treatment). Peripheral blood inflammatory markers were calculated as follows: Neutrophil-to-lymphocyte ratio, neutrophil count divided by lymphocyte count; platelet-to-lymphocyte ratio (PLR), platelet count divided by lymphocyte count; MLR, monocyte count divided by lymphocyte count. Two types of change indicators between baseline and post-treatment were predefined for all continuous laboratory parameters, with unified abbreviations and calculation formulas: (1) Absolute change (denoted as Δ; unit: Consistent with the original indicator): This was calculated as post-treatment value minus baseline value. For T lymphocyte subsets (original unit: %), the unit of Δ is also the %, which reflects the net absolute change of the cell proportion before and after treatment; and (2) Relative percentage change rate (denoted as Δ%; unit: %): This was calculated as [(post-treatment value-baseline value)/baseline value] × 100%. This rate reflects the relative change in the amplitude of the indicator relative to its baseline level and is expressed as a percentage.

Treatment regimen and follow-up

Before surgery, all patients underwent three cycles of neoadjuvant immunotherapy with chemotherapy. On the first day of each treatment cycle, intravenous sintilimab (3 mg/kg for body weight < 60 kg, or 200 mg for body weight ≥ 60 kg) and oxaliplatin (130 mg/m²) were given. From days 1 through 14, oral capecitabine (1000 mg/m²) was given twice a day. The patients underwent D2 lymphadenectomy and radical gastrectomy two weeks following the third cycle.

All patients received surgical resection and were followed up for at least 18 months. Regular follow-up assessments were conducted until March 1, 2022, including gastroscopy and contrast-enhanced abdominal computed tomography or magnetic resonance imaging at intervals of 3-6 months to evaluate disease progression. PFS refers to the interval from the start of neoadjuvant therapy to the first occurrence of disease progression, death from any cause, or the date of the last follow-up, whichever comes first.

Statistical analysis

IBM SPSS Statistics Version 26.0 was used for statistical analysis. Two-sided tests were used, and statistical significance was defined as P < 0.05. The Shapiro-Wilk test was initially employed to evaluate the normality of continuous variables. As the data did not conform to a normal distribution, medians and interquartile ranges were used to describe the continuous variables. Post-treatment comparisons were analyzed using the Wilcoxon signed-rank test for paired samples. The Mann-Whitney U test was used to compare the major pathological response (MPR) group (TRG = 0-1) with the non-MPR group (TRG = 2-3). Classified variables are presented as n (%), and group comparisons are conducted using either the χ² test or Fisher’s exact test, depending on the specific circumstances. This study employed binary logistic regression analysis to identify independent predictive factors for a favorable response to neoadjuvant therapy. Candidate independent variables were identified based on the primary endpoint, statistically significant variables (P < 0.10) identified in univariate analyses, and clinically relevant confounding factors widely recognized in gastric cancer research. All predefined variables were simultaneously entered into the multivariate logistic regression model using the Enter method. PFS was plotted using Kaplan-Meier survival curves, and univariate survival analysis was performed using the log-rank test and Cox regression model. Incorporate variables with P values < 0.10 from the univariate analysis using the Enter method into a Cox proportional hazards regression model for multivariate analysis. We used the bootstrapping to verify the model’s stability. The results are expressed as hazard ratios (HRs) and 95% confidence intervals (CIs). These analyses are exploratory, and no adjustment for multiple testing was applied.

RESULTS

Baseline clinical characteristics

On the basis of the eligibility criteria, 39 patients were finally enrolled, comprising 25 males (64.1%) and 14 females (35.9%), with a median age of 64 years (range, 35-77 years). Regarding histological differentiation, 25 patients (64.1%) had poorly differentiated adenocarcinoma, and 14 (35.9%) had moderately differentiated adenocarcinoma. Regarding the histological subtype, 24 patients (61.5%) had non–poorly cohesive carcinoma, and 15 (38.5%) had poorly cohesive gastric cancer. According to the Lauren classification, 21 patients (53.8%) had diffuse-type gastric cancer, and 18 (46.2%) had intestinal-type gastric cancer. The primary tumor locations included the cardia in 9 patients (23.1%), gastric body in 14 (35.9%), and gastric antrum in 16 (41.0%). Regarding the pre-treatment clinical T stage, there were 16 patients (41.0%) with cT3 stage and 23 patients (59.0%) with cT4 stage. The clinical N stage distribution was as follows: CN₀ in 5 patients (12.8%), cN₁ in 17 (43.6%), cN₂ in 9 (23.1%), and cN₃ in 8 (20.5%). The postoperative pathological T stage was pT₀ in 6 patients (15.4%), pT₁ in 4 (10.3%), pT₂ in 8 (20.5%), pT₃ in 11 (28.2%), and pT₄ in 10 (25.6%). The pathological N stage included pN₀ in 21 patients (53.8%), pN₁ in 7 (17.9%), pN₂ in 1 (2.6%), and pN₃ in 10 (25.6%). For TRG assessed on postoperative pathological specimens, 6 patients (15.4%) achieved TRG = 0, 8 (20.5%) achieved TRG = 1, 22 (56.4%) achieved TRG = 2, and 3 (7.7%) achieved TRG = 3 (Table 1).

Table 1 Baseline clinical and pathological characteristics of patients (n = 39).

Characteristic	Median (range) or n (%)
Age (years)	64 (35-77)
Gender
Male	25 (64.1)
Female	14 (35.9)
Tumor differentiation
Poor	25 (64.1)
Moderate	14 (35.9)
Histological type
PC	15 (38.5)
Non-PC	24 (61.5)
Tumor location
Cardia	9 (23.1)
Gastric antrum	16 (41.0)
Gastric body	14 (35.9)
Lauren classification
Intestinal	18 (46.2)
Diffuse	21 (53.8)
cT
cT3	16 (41.0)
cT4	23 (59.0)
cN
cN0	5 (12.8)
cN1	17 (43.6)
cN2	9 (23.1)
cN3	8 (20.5)
pT
pT0	6 (15.4)
pT1	4 (10.3)
pT2	8 (20.5)
pT3	11 (28.2)
pT4	10 (25.6)
pN
pN0	21 (53.8)
pN1	7 (17.9)
pN2	1 (2.6)
pN3	10 (25.6)
TRG
Grade 0	6 (15.4)
Grade 1	8 (20.5)
Grade 2	22 (56.4)
Grade 3	3 (7.7)

PC: Poorly cohesive gastric adenocarcinoma; TRG: Tumor regression grade.

Changes in peripheral blood parameters before and after neoadjuvant chemoimmunotherapy

Following neoadjuvant chemoimmunotherapy, significant decreases were observed in lymphocyte counts and platelet counts (both P < 0.001), as well as in the PLR (P = 0.005). Conversely, monocyte counts, CD3+ T-cell counts, CD8+ T-cell counts, and MLR increased significantly (all P < 0.005). We observed no statistically significant differences in neutrophil counts, CD4+ T-cell proportions, or neutrophil-to-lymphocyte ratio between baseline and after treatment (all P > 0.05; Table 2).

Table 2 Changes in peripheral blood parameters at baseline and post-treatment, median (range).

	Baseline	Post-treatment	P value
Laboratory examination
Neutrophil count (× 109/L)	3.25 (2.70, 4.22)	3.21 (2.60, 4.22)	0.653
Lymphocyte count (× 109/L)	1.44 (1.14, 1.91)	1.18 (0.88, 1.68)	0.000
Monocyte count (× 109/L)	0.39 (0.29, 0.46)	0.51 (0.42, 0.65)	0.000
Blood platelet count (× 109/L)	245.00 (204.00, 315.00)	168.00 (111.00, 202.00)	0.000
T lymphocyte subsets
CD3+ T lymphocytes (%)	72.00 (64.00, 80.40)	80.50 (72.60, 84.10)	0.000
CD4+ T lymphocytes (%)	41.60 (33.50, 46.10)	41.30 (34.90, 50.50)	0.171
CD8+ T lymphocytes (%)	28.70 (19.10, 35.80)	32.70 (21.60, 41.50)	0.002
Inflammatory indicators
NLR	2.35 (1.79, 3.17)	2.68 (2.11, 4.02)	0.070
MLR	0.28 (0.21, 0.35)	0.43 (0.32, 0.64)	0.000
PLR	182.05 (131.82, 207.24)	121.05 (92.19, 211.97)	0.005

NLR: Neutrophil-to-lymphocyte ratio; PLR: Platelet-to-lymphocyte ratio; MLR: Monocyte-to-lymphocyte ratio.

Predictors of treatment response

According to the TRG assessed on postoperative pathological specimens, the 39 patients were divided into the MPR group (n = 14, 35.9%) and non-MPR group (n = 25, 64.1%). Univariate analysis revealed that histological subtype, clinical N stage, pathological T stage, pathological N stage, pre-treatment CD3+ T-cell count, pre-treatment CD4+ T-cell count, post-treatment MLR, and rate of change in CD3+ T-cell count were associated with treatment response (Table 3). Variables (including histological classification, post-treatment MLR, and clinically significant clinical T stage) were included in a multivariate binary logistic regression model. The results demonstrated that patients with poorly cohesive gastric adenocarcinoma (PC) had a lower likelihood of achieving favorable efficacy than patients without PC (odds ratio = 8.522, 95%CI: 1.146-63.374, P = 0.036). By contrast, higher post-treatment MLR was potentially correlated with better efficacy (odds ratio = 2.479, 95%CI: 1.064-5.788, P = 0.035; Table 4).

Table 3 Univariate analysis of factors associated with treatment response.

	MPR (n = 14)	Non-MPR (n = 25)	χ2/Z	P value
Sex			1.987	0.187
Male	11 (78.6)	14 (56.0)
Female	3 (21.4)	11 (44.0)
Age			0.223	0.733
≤ 60 years	4 (28.6)	9 (36.0)
> 60 years	10 (71.4)	16 (64.0)
Tumor differentiation			1.888	0.297
Poor	7 (50.0)	18 (64.1)
Moderate	7 (50.0)	7 (35.9)
Histological type			5.393	0.038
PC	2 (14.3)	13 (52.0)
Non-PC	12 (85.7)	12 (48.0)
Tumor location			1.988	0.370
Cardia	5 (35.7)	4 (16.0)
Gastric body	4 (28.6)	10 (40.0)
Gastric antrum	5 (35.7)	11 (44.0)
Lauren classification			2.889	0.108
Intestinal	9 (64.3)	9 (36.0)
Diffuse	5 (35.7)	16 (64.0)
cT			2.345	0.179
cT3	8 (57.1)	8 (32.0)
cT4	6 (42.9)	17 (68.0)
cN			10.242	0.003
cN0	5 (35.7)	0 (0)
cN1-3	9 (64.3)	25 (100.0)
pT			19.168	0.000
pT0-2	13 (92.9)	5 (20.0)
pT3-4	1 (7.1)	20 (80.0)
pT			13.374	0.000
pN0	13 (92.9)	8 (32.0)
pN1-3	1 (7.1)	17 (68.0)
Baseline CD3+ T lymphocytes (%)	66.05 (51.98, 79.23)	73.80 (70.30, 80.45)	-2.255	0.024
Baseline CD4+ T lymphocytes (%)	34.00 (32.60, 42.28)	42.70 (35.55, 48.15)	-2.108	0.035
Baseline CD8+ T lymphocytes (%)	29.05 (16.55, 36.53)	28.40 (21.15, 35.70)	-0.293	0.770
Baseline NLR	2.53 (1.84, 3.13)	2.30 (1.78, 3.41)	-0.088	0.930
Baseline PLR	189.27 (109.94, 209.08)	181.48 (132.58, 202.99)	-0.263	0.792
Baseline MLR	0.29 (0.22, 0.33)	0.27 (0.20, 0.35)	-0.234	0.815
Post-treatment CD3+ T lymphocytes (%)	78.15 (67.43, 83.53)	82.00 (73.00, 85.25)	-1.098	0.272
Post-treatment CD4+ T lymphocytes (%)	40.90 (31.60, 52.88)	43.40 (35.55, 50.10)	-0.381	0.703
Post-treatment CD8+ T lymphocytes (%)	31.00 (15.90, 42.50)	32.70 (23.40, 41.50)	-0.454	0.650
Post-treatment NLR	2.62 (2.31, 4.18)	2.68 (1.89, 3.96)	-0.351	0.725
Post-treatment PLR	185.07 (90.05, 194.27)	124.14 (89.37, 220.30)	-0.263	0.792
Post-treatment MLR	0.57 (0.42, 0.74)	0.39 (0.29, 0.54)	-2.196	0.028
ΔCD3+ T cell (%)	11.41 (5.37, 23.05)	5.06 (-4.09, 12.32)	-2.108	0.035
ΔCD4+ T cell (%)	5.96 (-8.10, 42.41)	2.16 (-8.62, 16.26)	-0.908	0.364
ΔCD8+ T cell (%)	15.79 (-6.19, 38.67)	9.44 (-1.57, 23.01)	-0.351	0.725
ΔNLR (%)	38.16 (-17.95, 64.18)	29.48 (-28.99, 55.18)	-0.716	0.447
ΔPLR (%)	-18.50 (-38.86, 10.81)	-30.63 (-47.68, -6.89)	-0.732	0.464
ΔMLR (%)	114.43 (50.51, 170.73)	57.53 (17.98, 89.64)	-2.576	0.010
ΔCD3+ T cell (%)	6.65 (3.05, 13.50)	3.70 (-3.10, 8.95)	-1.874	0.061
ΔCD4+ T cell (%)	1.70 (-3.10, 17.40)	1.00 (-4.15, 6.50)	-0.952	0.341
ΔCD8+ T cell (%)	3.30 (-1.85, 7.00)	2.90 (-0.45, 6.65)	-0.132	0.895
ΔNLR	0.68 (-0.54, 1.44)	0.65 (-0.61, 1.13)	-0.322	0.747
ΔPLR	-27.31 (-90.98, 10.66)	-51.53 (-117.17, -7.75)	0.878	0.380
ΔMLR	0.27 (0.14, 0.48)	0.11 (0.04, 0.22)	-2.752	0.006

NLR: Neutrophil-to-lymphocyte ratio; PLR: Platelet-to-lymphocyte ratio; MLR: Monocyte-to-lymphocyte ratio; PC: Poorly cohesive gastric adenocarcinoma; MPR: Major pathological response.

Table 4 Multivariate logistic regression analysis of predictors of treatment response.

	OR	P value	95%CI
cT stage (cT4 vs cT3)	0.425	0.283	0.089-2.028
Histological type (non-PC vs PC)	8.522	0.036	1.146-63.374
Post-treatment MLR	2.479	0.035	1.064-5.788

OR: Odds ratio; CI: Confidence interval; PC: Poorly cohesive gastric adenocarcinoma; MLR: Monocyte-to-lymphocyte ratio.

Predictors of PFS

With a minimum follow-up period of 18 months, the overall median follow-up duration for all 39 patients, calculated using the reverse Kaplan-Meier method, was 645 days (95%CI: 594-696 days). Among the 13 patients (33.3%) who experienced disease progression or death, the median time to progression was 305 days. Among the 26 censored patients (66.7%) who remained progression-free at the time of analysis, the median follow-up duration was 748 days (95%CI: 665-831 days). Univariate analysis using the Kaplan-Meier method for categorical variables revealed that clinical T stage, clinical N stage, pathological T stage, pathological N stage, histological classification, and Lauren classification were associated with PFS. For continuous variables, univariate Cox regression analysis demonstrated that ΔCD8+ T cells and the relative change in CD8+ T-cell proportion were associated with PFS. No significant associations were observed between PFS and other peripheral blood inflammatory markers or T lymphocyte subsets. Variables with P value < 0.10 were then entered into a multivariable Cox regression model. The results identified Lauren classification and the relative change in CD8+ T-cell proportion (ΔCD8+ T cell%) were associated with PFS. Specifically, diffuse-type gastric cancer was associated with an increased risk of disease progression or death (HR = 5.545, 95%CI: 0.568-12.955, P = 0.011), whereas a higher ΔCD8+ T cell% proportion correlated with favorable PFS (HR = 0.970, 95%CI: 0.934-0.998, P = 0.024). Table 5 presents the detailed results.

Table 5 Univariate and multivariate Cox regression analyses of prognostic factors for progression-free survival.

	HR (95%CI)	P value	HR (95%CI)	P value
Sex
Male
Female	0.871 (0.268-2.830)	0.818
Age
≤ 60 years
> 60 years	0.696 (0.228-2.131)	0.526
Tumor differentiation
Poor
Moderate	0.267 (0.059-1.206)	0.086
Histological type
PC
Non-PC	0.653 (0.219-1.945)	0.444
Tumor location
Cardia
Gastric body	2.425 (0.489-12.034)	0.279
Gastric antrum	1.484 (0.288-7.650)	0.637
Lauren classification
Intestinal
Diffuse	6.284 (1.389-28.431)	0.017	5.545 (0.568-12.955)	0.011
cT
cT3
cT4	4.854 (1.074-21.936)	0.040
cN
cN0
cN1-3	1.951 (0.254-15.009)	0.521
pT
pT0-2
pT3-4	3.756 (1.030-13.688)	0.045
pT
pN0
pN1-3	3.333 (1.025-10.839)	0.045
Baseline CD3+ T lymphocytes (%)	1.008 (0.959-1.059)	0.753
Baseline CD4+ T lymphocytes (%)	0.990 (0.930-1.055)	0.764
Baseline CD8+ T lymphocytes (%)	1.017 (0.961-1.075)	0.561
Baseline NLR	1.035 (0.722-1.485)	0.852
Baseline PLR	1.002 (0.995-1.008)	0.622
Baseline MLR	0.679 (0.152-3.201)	0.642
Post-treatment CD3+ T lymphocytes (%)	0.977 (0.928-1.028)	0.373
Post-treatment CD4+ T lymphocytes (%)	1.003 (0.959-1.049)	0.894
Post-treatment CD8+ T lymphocytes (%)	0.982 (0.939-1.026)	0.408
Post-treatment NLR	0.975 (0.754-1.261)	0.847
Post-treatment PLR	1.001 (0.994-1.009)	0.760
Post-treatment MLR	1.000 (0.076-13.126)	1.000
ΔCD3+ T cell (%)	0.971 (0.928-1.017)	0.216
ΔCD4+ T cell (%)	1.817 (0.270-12.233)	0.540
ΔCD8+ T cell (%)	0.971 (0.946-0.998)	0.032	0.970 (0.934-0.998)	0.024
ΔNLR (%)	0.976 (0.584-1.629)	0.925
ΔPLR (%)	0.961 (0.320-2.888)	0.944
ΔMLR (%)	1.000 (0.991-1.009)	0.948
ΔCD3+ T cell (%)	0.960 (0.901-1.023)	0.205
ΔCD4+ T cell (%)	1.012 (0.959-1.069)	0.659
ΔCD8+ T cell (%)	0.906 (0.830-0.981)	0.016
ΔNLR	0.965 (0.760-1.225)	0.770
ΔPLR	0.999 (0.993-1.005)	0.842
ΔMLR	1.535 (0.302-7.801)	0.606

HR: Hazard ratio; CI: Confidence interval; NLR: Neutrophil-to-lymphocyte ratio; PLR: Platelet-to-lymphocyte ratio; MLR: Monocyte-to-lymphocyte ratio; ΔCD8+ T cells: Relative change in CD8+ T-cell proportion; PC: Poorly cohesive gastric adenocarcinoma.

DISCUSSION

Neoadjuvant chemotherapy combined with immunotherapy has emerged as an important component of the multimodal treatment for LAGC. All 39 patients in this study with LAGC received neoadjuvant therapy consisting of the XELOX regimen combined with sintilimab. We investigated the correlation between these parameters and treatment response and prognosis by dynamically monitoring the changes in peripheral blood inflammatory markers and T lymphocyte subset proportions. The study design ensured a high degree of homogeneity in the treatment regimen, thereby excluding heterogeneity in immune marker alterations that may arise from different chemotherapeutic agents or ICIs. These observations provide preliminary support for candidate biomarkers that could aid in identifying patients with the highest likelihood of deriving treatment benefit. Nonetheless, the results should be interpreted cautiously given the exploratory nature of this study and the small sample size.

In this study, LAGC patients treated with XELOX in combination with sintilimab as neoadjuvant therapy experienced significant decreases in both peripheral blood lymphocyte and platelet counts (both P < 0.001), and the PLR also decreased accordingly (P = 0.005). The platelet count decreased by 31.43%, and the lymphocyte count decreased by 18.06%; these results are consistent with the hematologic toxicity profile associated with ICI therapy. Previous studies have indicated that patients treated with programmed death 1 inhibitors for advanced gastric cancer are more prone to treatment-related lymphocytopenia and thrombocytopenia, a finding consistent with the conclusions of this study[14]. These findings may be partly attributable to T-cell activation and subsequent immune responses induced by ICIs; however, the primary contributing factor is likely the myelosuppressive effects of chemotherapy. The decline in platelet counts was significantly greater than that in lymphocyte counts. Consequently, the decrease in PLR values was primarily driven by the drop in platelet counts rather than by relative changes in lymphocyte counts. This finding suggests that PLR values may reflect the severity of hematologic toxicity associated with the XELOX regimen, since both oxaliplatin and capecitabine can induce cytopenia by directly suppressing bone marrow function[15]. In this study, after neoadjuvant chemotherapy combined with immunotherapy, we found that the monocyte count was significantly elevated (P < 0.005), with a growth rate of 30.77%; lymphocyte counts showed a downward trend; therefore, we speculate that the elevation in MLR is primarily due to the increase in monocyte count rather than lymphocytopenia. A key component in the innate immune system, monocytes can differentiate into tumor-associated macrophages upon entering the tumor microenvironment, where they play a dual role in promoting tumor growth and mediating immune suppression[16]. However, accumulating evidence suggests that the increase in the number of monocytes after ICI therapy may reflect a state of systemic immune activation. Previous studies have found that, among patients with resectable esophageal adenocarcinoma receiving neoadjuvant programmed death ligand-1 (PD-L1) checkpoint inhibitor therapy, higher post-treatment monocyte counts are associated with better treatment outcomes[17]. Our observations echo this finding: Elevated post-treatment MLR levels were associated with good therapeutic responses. This association is consistent with the hypothesis that an early rise in monocytes may reflect systemic immune activation. However, given the limited sample size and the lack of functional immune assessments, definitive conclusions regarding the underlying mechanisms cannot yet be drawn.

Regarding T lymphocyte subsets, this study demonstrated that the proportions of CD3+ and CD8+ T cells increased significantly following neoadjuvant therapy (both P < 0.005). The proportion of CD4+ T cells did not show any significant changes. These findings are congruent with previous studies. Laza-Briviesca et al[18] observed in non-small cell lung cancer patients that an increased proportion of CD8+ T cells in peripheral blood following neoadjuvant chemoimmunotherapy was correlated with pathological complete response. As cytotoxic T lymphocytes (CTLs), CD8+ T cells are the key effector cells within the anti-tumor immune response. The observed increase in the proportion of peripheral blood may reflect the clonal expansion and activation of tumor-specific T cells induced by sintilimab[19]. We observed an intriguing paradox in this study: While the total lymphocyte count decreased after treatment, the proportions of CD3+ and CD8+ T cells increased significantly. This seemingly contradictory phenomenon actually reflects the unique immune-activating effects of sintilimab. The decrease in total lymphocyte count coupled with an increase in the proportion of CD8+ T cells is likely the result of an interaction between chemotherapy-induced bone marrow suppression and immune activation triggered by immunotherapy. Specifically, the relative increase in CD8+ T cells may stem from the clonal expansion and activation of tumor-specific CTLs induced by sintilimab; however, this hypothesis requires validation through T-cell receptor sequencing or functional assays in future studies.

In this study, patients were divided into MPR (TRG = 0-1) and non-MPR (TRG = 2-3) on the basis of the TRG. Both univariate and multivariate analyses were followed to identify predictors of treatment response. Univariate analysis revealed that pathological T stage, pre-treatment CD3+ T-cell proportion, pre-treatment CD4+ T-cell proportion, and rate of change in CD3+ T-cell proportion were associated with treatment response. This suggests that higher pre-treatment proportions of CD3+ and CD4+ T cells may reflect a favorable baseline immune status in patients, thus conferring greater anti-tumor immune potential. Furthermore, the rate of change in CD3+ T-cell proportion reflects the degree of immune activation induced by XELOX combined with sintilimab, and its association with treatment response aligns with immunological expectations. Multivariate analysis revealed that patients with PC had a lower likelihood of achieving favorable efficacy than patients without PC. PC is a type of gastric adenocarcinoma characterized by tumor cells lacking intercellular adhesion and exhibiting monolithic or small cluster-like infiltrative growth; it primarily comprises signet ring cell carcinoma and its related subtypes[20]. However, signet ring cell carcinoma exhibits lower sensitivity to chemotherapy than other types, and a higher proportion of signet ring cells is correlated with poorer response to chemotherapy[21]. Furthermore, the unique immunosuppressive and stromal-enriched microenvironmental characteristics of signet ring cell carcinoma induce treatment resistance and immune evasion, thereby reducing the efficacy of chemotherapy and ICIs[22].

Higher post-treatment MLR was potentially correlated with better efficacy. Although this result appears to contradict the conventional understanding of MLR as a pro-tumor inflammatory marker, it can be interpreted in the context of our observation that the elevation in MLR was primarily driven by an increase in monocyte count. Specifically, the post-treatment increase in monocytes may reflect a systemic immune activation response to sintilimab. Some studies have suggested that gastric cancer patients with low MLR levels at baseline and after two cycles of treatment may achieve longer OS, and that patients with low baseline MLR levels may benefit more from first-line immunotherapy combined with chemotherapy than from chemotherapy alone[23]. It should be specifically noted, however, that the study included advanced gastric cancer patients who had completed multiple treatment cycles; the dynamic decline in their MLR values is more likely to reflect the resolution of the inflammatory response following long-term treatment. By contrast, we focused on patients with LAGC undergoing neoadjuvant treatment, and blood samples were collected at an early post-treatment time point (2 weeks after the final cycle). In this context, an increase in the number of monocytes may reflect the acute-phase immune activation response induced by sintilimab. Therefore, the opposite trends in MLR observed between these two studies likely represent distinct patterns of dynamic immune evolution across different treatment phases and disease settings rather than contradictory findings. However, this explanation remains speculative and warrants further investigation through immune profiling and long-term follow-up.

Multivariate analysis ultimately revealed that the Lauren classification and the relative change in CD8+ T-cell proportion may be potential independent predictors of PFS. Notably, although the absolute change in CD8+ T-cell proportion was associated with PFS in univariate analysis, the final multivariate model included the relative change in CD8+ T-cell proportion rather than their absolute values. This is because absolute changes often convert dynamic changes into static numerical values, thereby failing to capture the dynamic evolution of the immune system. Baseline biomarker levels vary significantly among patients, and the sole reliance on absolute change values would systematically be influenced by baseline levels. Dividing the change by the baseline level as a relative change effectively eliminates this confounding effect. Studies have demonstrated that relative changes more accurately reflect the magnitude of biomarker alterations in the context of immunotherapy, thereby exhibiting stronger associations with prognosis[24,25].

Diffuse gastric cancer may result in shorter PFS (P = 0.006, log rank test; Figure 2A). This is compatible with the biological characteristics associated with this histological subtype. As illustrated by the survival curves, patients with diffuse-type tumors experienced significantly shorter PFS than those with intestinal-type tumors. Diffuse-type gastric cancer typically presents as a poorly differentiated or signet ring cell carcinoma, exhibits higher immune cell infiltration but reduced functional activity, and is characterized by high invasiveness and a propensity for early dissemination, all of which contribute to poor prognosis[26,27]. Furthermore, diffuse-type gastric cancer is characterized by a “cold” tumor immune microenvironment, featuring the enrichment of cancer-associated fibroblasts and reduced infiltration of CTLs, which contributes to its poor response to ICIs[28]. For this patient subset, the current neoadjuvant chemoimmunotherapy regimen may not achieve optimal treatment outcomes. Future efforts should focus on exploring individualized therapeutic strategies tailored to this specific subtype.

Figure 2 Kaplan-Meier curves. A: Kaplan-Meier curves for progression-free survival stratified by Lauren classification; B: Kaplan-Meier curves for progression-free survival plotted by stratification on the basis of the relative change in the proportion of CD8+ T cells. HR: Hazard ratio; CI: Confidence interval.

An increase in the relative change in CD8+ T-cell proportion may correlate with improved PFS (P = 0.028, log rank test; Figure 2B), further substantiating the central role of CD8+ T cells in anti-tumor immune responses. We hypothesized that this elevation reflected treatment-induced clonal expansion and the functional activation of CD8+ T cells. These activated effector T cells subsequently migrate to the tumor tissue, where they exert sustained anti-tumor effects, thereby delaying disease progression. This hypothesis is supported by Li et al[29], who demonstrated that the high baseline expression of interferon gamma in CD8+ T cells predicted favorable treatment response, and by Kim et al[30], who found that a significant increase in CD8+ T cells within the first week was associated with improvement in PFS.

The exploratory study demonstrates the potential of routine peripheral blood parameters to serve as readily available biomarkers for neoadjuvant therapy. If validated in larger cohorts in the future, post-treatment MLR values derived from preoperative complete blood counts could aid in patient stratification or serve as an option for guiding the intensity of adjuvant therapy. Meanwhile, the relative change in CD8+ T-cell proportion could help identify patients who, despite a favorable pathological response, remain at high risk of recurrence, thereby enabling the development of follow-up regimens tailored to their individual risk profiles. For patients with diffuse-type gastric cancer, our data indicate that current treatment regimens are ineffective in this subgroup, suggesting a need to explore more personalized treatment approaches. Of course, these exploratory findings should be interpreted with caution and must be validated through prospective studies before they can be considered for clinical application.

This study has a few limitations. First, given that this study was a single-center retrospective study with a relatively small sample size (n = 39) and a limited number of PFS events, the statistical power of the multivariate analysis remained insufficient. Second, this study did not analyze the tumor simultaneously; therefore, the direct correlation between peripheral blood markers and intratumoral immune responses requires further validation. Third, the PD-L1 combined positive score (CPS), which is a widely used predictive biomarker for ICI therapy, was not included as a covariate in the analysis. PD-L1 CPS levels themselves may be highly correlated with post-treatment immune response status; patients with CPS ≥ 5 are more likely to exhibit an immune-activated phenotype, and the benefit of combination immunotherapy is significantly greater in this subgroup than in those with CPS < 5[31,32]. Fourth, postoperative adjuvant treatment regimens were individually tailored based on each patient’s performance status and pathological high-risk factors. The heterogeneity of subsequent treatment regimens may have introduced confounding bias into the PFS analysis. Therefore, the association between post-treatment MLR and favorable treatment response may mask subgroup differences in PD-L1 CPS, whereas variations in CD8+ T-cell counts may also correlate with baseline CPS levels. Finally, peripheral blood samples were collected at only two time points, which precluded a comprehensive characterization of dynamic immune changes throughout the treatment course. The findings of this study warrant further validation in large-scale, multicenter, prospective cohorts.

CONCLUSION

This exploratory study suggests that in patients with LAGC receiving neoadjuvant therapy (sintilimab combined with the XELOX regimen), histological type and post-treatment MLR may be associated with treatment response. By contrast, Lauren classification and the relative change in CD8+ T-cell proportion may be associated with PFS. These findings highlight the potential value of dynamic immune monitoring in the context of neoadjuvant therapy and provide a rationale for the further investigation of peripheral blood parameters as readily available biomarkers. However, given the limitations of this small-sample retrospective study, our results should be regarded as preliminary findings and require validation in future large-scale, prospective, multicenter cohort studies.