Clinical value of geriatric nutritional risk index and pan-immune-inflammation value in locally advanced gastric cancer receiving neoadjuvant chemotherapy

Abstract

BACKGROUND

Metabolism and nutrition status play an important role in the development of cancer. However, whether inflammation and malnutrition related indicators can predict the efficacy of neoadjuvant chemotherapy (NACT) and the prognosis of gastric cancer has not been addressed.

AIM

To evaluate the predictive value of malnutrition as determined by the geriatric nutritional risk index (GNRI) and inflammation represented by the pan-immune-inflammation value (PIV) for the response to NACT patients’ prognosis in locally advanced gastric cancer (LAGC).

METHODS

We retrospectively analyzed 147 LAGC patients who underwent radical resection after NACT. The PIV, and GNRI were compared according to whether receiving nutritional intervention during NACT. The prognostic values of GNRI and PIV were assessed using time-dependent receiver operating characteristic curve analysis, log-rank tests, and Cox regression analysis.

RESULTS

Nutritional intervention could improve nutrition status and reduce inflammation during NACT in LAGC patients. Multivariate analysis showed that GNRI (hazard ratio = 0.956, 95% confidence interval: 0.923-0.991, P = 0.013), PIV (hazard ratio = 1.002, 95% confidence interval: 1-1.005, P = 0.041) were independent predictors for OS. Significant differences of overall survival and disease-free survival according to GNRI (P < 0.001) and PIV (P < 0.001) were observed between the low and high groups. The GNRI-PIV score constructed with GNRI and PIV had a higher area under the curve and was significantly associated with pathological tumor regression response.

CONCLUSION

GNRI and PIV are useful predictive biomarkers in patients with LAGC receiving NACT, and nutritional supplement can improve patients’ status. The GNRI-PIV score may contribute to a more personalized and holistic approach for LAGC patients underwent NACT.

Key Words: Prognosis; Neoadjuvant chemotherapy; Locally advanced gastric cancer; Pan-immune-inflammation value; Geriatric nutritional risk index

Core Tip: This is a retrospective single-center observational study to investigate the value of geriatric nutritional risk index and pan-immune-inflammation value in locally advanced gastric cancer patients. We found that the geriatric nutritional risk index and pan-immune-inflammation value serve as prognostic indictors that reflects nutrition and inflammation status and predict the response to neoadjuvant therapy in locally advanced gastric cancer. Additionally, the results also emphasized the necessity of nutritional support during neoadjuvant therapy for patients.

INTRODUCTION

The incidence of gastric cancer ranks fifth globally. Early gastric cancer is typically asymptomatic, and diagnostic methods are limited, resulting in the majority of patients being diagnosed at a locally advanced stage[1]. Clinically, neoadjuvant chemotherapy (NACT) offers several potential benefits, including tumor down-staging, control of micrometastasis, and an increased rate of radical resection. Moreover, previous clinical trials have demonstrated that NACT provides a long-term survival advantage over surgery alone[2], and it remains one of the primary treatments for locally advanced gastric cancer (LAGC) according to several guidelines[3,4]. As more patients undergo NACT, the demand for more accurate long-term prognostic predictions has become a growing challenge.

Patients with gastrointestinal tumors often experience impaired nutritional status and weight loss due to cancer-related metabolic changes and reduced food intake[5]. Additionally, commonly used NACT regimens, such as S-1 plus oxaliplatin, are frequently associated with gastrointestinal adverse events, including anorexia, nausea, vomiting, stomatitis, and diarrhea. These side effects can further exacerbate nutritional depletion, increasing the risk of malnutrition[6]. Previous studies have shown that greater weight loss and a lower body mass index (BMI) prior to NACT are associated with poorer treatment responses and worse outcomes[7]. Most gastric cancer patients suffer from cancer-related inflammation and malnutrition, which are linked to tumor onset, progression, and prognosis[8]. The geriatric nutritional risk index (GNRI) can be accessed easily in routine clinical practice and was first introduced to assess nutritional status in patients[9]. Recent studies have identified GNRI as a prognostic biomarker for various cancers[10,11]. The pan-immune inflammation value (PIV) has emerged as a promising indicator of systemic inflammation and offers a more comprehensive reflection of overall inflammatory status than other markers, such as the neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), lymphocyte-to-monocyte ratio, and systemic immune-inflammation index[12,13]. PIV has also been found to be associated with prognosis in cancer patients[14-16]. Given that both nutritional and inflammatory status independently influence survival, a combined indicator may provide more accurate predictions than either marker alone.

Considering the changes in nutritional and inflammatory status during NACT, this study aimed to investigate the predictive utility of GNRI and PIV, and to determine the prognostic significance of their combined index for long-term outcomes and response to NACT in LAGC.

MATERIALS AND METHODS

Study design and patients

This retrospective study includes 147 LAGC patients who received NACT for radical gastrectomy as the Peking Union Medical College Hospital from July 2016 to February 2020. The inclusion criteria were as follows: (1) Gastric adenocarcinoma was identified by pathological evidence; (2) Age ≥ 18 years; (3) NACT performed for at least 2 cycles; (4) No other treatment had been given before and after NACT; and (5) Gastrectomy performed with D2 lymphadenectomy. The exclusion criteria were as follows: (1) Incomplete clinical and laboratory data; (2) Neoadjuvant treatments other than chemotherapy; (3) Emergency surgery performed in case of digestive bleeding or perforation; (4) Combined with other malignancy or distant metastasis of the tumor; and (5) Combined with other autoimmune or hematologic diseases. This retrospective study was approved by the Ethics Committee of the Peking Union Medical College Hospital (Approval No. I-25PJ0504). This study was performed in accordance with the guidelines of the Helsinki Declaration.

Data collection

Clinical and pathological informations, including gender, age, BMI, histories of diabetes, hypertension, smoking, and drinking, tumor location, TNM stage, vascular invasion, nerve invasion, and other relevant data, were extracted from patients’ medical records. Blood lab results were collected before NACT and again within two weeks prior to surgery.

Calculation of indicators

The BMI was calculated as follows: Weight (kg)/[height (m)]². Patients were stratified into four BMI groups according to the standards set by the World Health Organization. Ideal body weight and GNRI was calculated via the Lorentz formula for men: Ideal body weight (men) = height (cm) - 100 – {[height (cm) – 150]/4}, for women: Ideal body weight (women) = height (cm) - 100 – {[height (cm) – 150]/2}, and GNRI = (1.489 × serum albumin (g/L) + 41.7 × (present weight/ideal body weight)[9]. The blood-based PIV was calculated based on peripheral blood data using the formula: PIV = (neutrophil count × platelet count × monocyte count)/Lymphocyte count.

Assessment of treatment response

Based on postoperative pathological report, tumor regression was evaluated according to the College of American Pathologists (CAP) guidelines[17]: CAP 0, complete tumor regression with no microscopic tumor cells; CAP 1, scattered single cells or small clusters of cancer cells; CAP 2, marked tumor regression with residual tumor cells; and CAP 3, extensive residual viable cancer. For further analyses, patients were classified as responders (CAP 0-1) or non-responders (CAP 2-3).

Treatments and follow-ups

The treatment regimens and cycles were determined by the oncologists, with NACT administered for at least two cycles. Assessments were conducted every 2-3 cycles based on the patient’s physical condition, tumor markers, and computed tomography (CT) results. During the treatment, for patients with Nutritional Risk Screening 2002 score ≥ 3 or those with deficient oral intake for more than 3 days, we gave oral or eternal nutritional supplementation in addition to their regular daily diet. All patients received treatment for no more than 4 cycles. Surgery, which involved the resection of at least two-thirds of the stomach and D2 lymphadenectomy, was performed between the fourth and sixth weeks following the completion of NACT.

All patients were routinely followed every 3 months during the first 2 years and every 6 months during the subsequent 3 years. Overall survival (OS) was defined as the interval from the start of NACT to death from any cause or to the last follow-up (in months). Disease-free survival (DFS) was defined as the period following treatment during which no signs of cancer were observed, or until the last follow-up.

Statistical analysis

Statistical analyses were performed by R (version 4.2.1) for survival analysis, GraphPad Prism 9 (version 9.0.0) and SPSS (version 26.0) for descriptive statistics and Cox regression analysis. Descriptive statistics were presented as mean ± SD for continuous variables, and as n (%) for categorical variables. Categorical data were compared using Fisher’s exact test or the χ² test. For normally distributed variables, the independent sample t-test was applied, while the Mann-Whitney U test was used for non-normally distributed data. The optimal cut-off values for the GNRI and PIV were identified by analyzing receiver operating characteristic curves. Univariate and multivariate Cox regression analyses were performed to identify independent risk factors associated with prognosis, with results expressed as hazard ratios (HR) and 95% confidence intervals (CI). The overall significance of the regression model was assessed using the omnibus test.

RESULTS

Demographic and pathological characteristics

This study enrolled 147 patients with LAGC who underwent NACT followed by surgery. Their clinical, pathological, and laboratory characteristics are summarized in Table 1. The cohort comprised 107 males (72.79%) and 40 females (27.21%). At diagnosis, the patients’ ages ranged from 26 years to 82 years, with a mean of 59.53 ± 9.91 years. The mean BMI for the entire cohort was 23.94 ± 3.09 kg/m². 54 (36.73%) patients were at clinical stage 2, 93 (63.27%) at stage 3. As for NACT regimens, 137 (93.2%) patients used S-1 plus oxaliplatin, and 10 patients received others regimen, including capecitabine and oxaliplatin, folinic acid, fluorouracil and oxaliplatin and S-1 capsules. The tumor located in 1/3 lower, 1/3 middle, 1/3 upper and total part was 60.54%, 21.77%, 7.48%, and 10.2%, respectively. Postoperative infectious complications occurred in 49 (33.33%) patients.

Table 1 Patient characteristics and demographics in this study, n (%)/mean ± SD.

Characteristics	Total (n = 147)	Mean (SD)	Range
Gender
Male	107 (72.79)
Female	40 (27.21)
Age, years		59.53 ± 9.91	26-82
< 65	101 (68.71)
≥ 65	46 (31.29)
BMI (kg/m2)		23.94 ± 3.09	15.04-32.39
Underweight	4 (2.72)
Normal	72 (48.98)
Overweight	56 (38.10)
Obese	15 (10.20)
Smoking history
No	67 (45.58)
Yes	80 (54.42)
Drinking history
No	85 (57.82)
Yes	62 (42.18)
NACT regimen
SOX	137 (93.20)
Others	10 (6.80)
Diabetes
No	122 (82.99)
Yes	25 (17.01)
Hypertension
No	103 (70.07)
Yes	44 (29.93)
Location
1/3 lower	89 (60.54)
1/3 middle	32 (21.77)
1/3 upper	11 (7.48)
Total	15 (10.20)
Clinical T stage
T2	15 (10.20)
T3	65 (44.22)
T4	67 (45.58)
Clinical N stage
N-	37 (25.17)
N+	110 (74.83)
Clinical stage
2	54 (36.73)
3	93 (63.27)
CEA, ng/mL		3.3 ± 3.52	0.56-33.31
CA19-9, U/mL		16.6 ± 20.63	0.6-159.5
hsCRP, mg/L		2.56 ± 4.75	0.14-28.43
GNRI		107.17 ± 10.56	56.23-162.83
PIV		170.35 ± 127.5	16.66-823.23
NRS2002 score
< 3	103 (70.07)
≥ 3	44 (29.93)
Vascular invasion
No	99 (67.35)
Yes	48 (32.65)
Perineural invasion
No	108 (73.47)
Yes	39 (26.53)
Pathological T stage
T0	22 (14.97)
T1	25 (17.01)
T2	35 (23.81)
T3	36 (24.49)
T4	29 (19.73)
Pathological N stage
N-	92 (62.59)
N+	55 (37.41)
Pathological stage
PCR	21 (14.29)
1	53 (36.05)
2	35 (23.81)
3	38 (25.85)
CAP
0	22 (14.97)
1	28 (19.05)
2	47 (31.97)
3	50 (34.01)
Complication
No	98 (66.67)
Yes	49 (33.33)

BMI: Body mass index; NACT: Neoadjuvant therapy; SOX: S-1 plus oxaliplatin; CEA: Carcinoembryonic antigen; CA: Carbohydrate antigen; hsCRP: High-sensitivity C-reactive protein; GNRI: Geriatric nutritional risk index; PIV: Pan-immune-inflammation value; NRS2002: Nutritional Risk Screening 2002; PCR: Pathological complete response; CAP: College of American Pathologists.

As for pathological features, pathological complete response was acquired in 21 (14.29) patients, the CAP of 1, 2, and 3 grade was acquired in 28 (19.05%), 47 (31.97%), and 50 (34.01%) patients, respectively. The pathological stage of 1, 2, and 3 was diagnosed for 53 (36.05), 35 (23.81), and 38 (25.85) patients.

Changes of the PIV and GNRI before and after NACT

The level of PIV and GNRI were significantly reduced (all P < 0.01) after NACT (Figure 1A and B). Since some patients received nutritional intervention during NACT, including oral or enteral nutrition, particularly those with nutrition risk, we conducted a further analysis stratified by the presence or absence of nutritional supplementation during NACT. We found that in the nutritional intervention group, both GNRI and PIV were significantly reduced after NACT (Figure 1C and D), while no similar outcome was observed in the other group (Figure 1E and F).

Figure 1 Evaluation the variation of geriatric nutritional risk index and pan-immune-inflammation value during neoadjuvant chemotherapy. A and B: The changes in geriatric nutritional risk index (GNRI) and pan-immune-inflammation value (PIV) during neoadjuvant chemotherapy; C and D: The influence of nutritional intervention on GNRI and PIV; E and F: The influence of no nutritional supplementation on GNRI and PIV.

Prognostic value of the PIV and GNRI

Univariate analysis revealed that drinking history, clinical stage, carbohydrate antigen 19-9, GNRI, PIV, Nutritional Risk Screening 2002 score, vascular invasion, perineural invasion, pathological T stage, pathological N stage, pathological stage and complication were significantly influenced the OS time (Table 2). Then the variables with P < 0.05 were included in the multivariate Cox regression analysis. The omnibus test for the Cox regression model showed a statistically significant overall model fit (χ² = 64.047, degrees of freedom = 15, P < 0.001), indicating that the set of independent variables collectively had a significant effect on patient survival time. Subsequent analysis revealed that drinking history (HR = 0.504, 95%CI: 0.27-0.939, P = 0.031), GNRI (HR = 0.956, 95%CI: 0.923-0.991, P = 0.013), PIV (HR = 1.002, 95%CI: 1-1.005, P = 0.041), and pathological stage (HR = 3.72, 95%CI: 1.061-13.044, P = 0.04) were independent predictors of prognosis in patients receiving NACT (Table 2). The univariate and multivariate analyses for DFS established GNRI (HR = 0.962, 95%CI: 0.928-0.997, P = 0.035) and pathological stage (HR = 4.257, 95%CI: 1.242-14.596, P = 0.021) as independent prognostic factor (Table 3).

Table 2 Univariate and multivariate Cox regression analysis for overall survival of clinicopathologic characteristics in the cohorts, median (interquartile range).

Variables	Univariate analysis		Multivariate analysis
	HR (95%CI)	P value	HR (95%CI)	P value
Gender
Male vs female	0.752 (0.436-1.295)	0.304
Age (years)	1.014 (0.989-1.041)	0.280
BMI (kg/m2)
Underweight	1 (reference)	NA
Normal	0.387 (0.118-1.266)	0.116
Overweight	0.308 (0.092-1.040)	0.058
Obese	0.316 (0.079-1.267)	0.104
Smoking history
No vs yes	0.693 (0.417-1.150)	0.156
Drinking history
No vs yes	0.564 (0.327-0.973)	0.039	0.504 (0.270-0.939)	0.031
Diabetes
Yes vs no	1.238 (0.643-2.385)	0.522
Hypertension
Yes vs no	1.293 (0.760-2.198)	0.343
Location
1/3 lower	1 (reference)	NA
1/3 middle	1.373 (0.746-2.528)	0.309
1/3 upper	1.011 (0.358-2.854)	0.984
Total	1.655 (0.764-3.584)	0.201
Nutrition supplement
Yes vs no	1.153 (0.664-2.003)	0.613
Clinical T stage
T2	1 (reference)	NA
T3	0.648 (0.257-1.633)	0.358
T4	1.533 (0.646-3.639)	0.333
Clinical N stage
Positive vs negative	1.166 (0.641-2.121)	0.615
Clinical stage
Stage 3 vs stage 2	1.836 (1.036-3.255)	0.038	0.868 (0.432-1.741)	0.690
CEA, ng/mL	1.048 (0.984-1.115)	0.145
CA19-9, U/mL	1.011 (1.002-1.019)	0.016	1.004 (0.992-1.015)	0.536
hsCRP, mg/L	1.022 (0.970-1.078)	0.407
GNRI	0.953 (0.927-0.980)	< 0.001	0.956 (0.923-0.991)	0.013
PIV	1.002 (1.000-1.004)	0.019	1.002 (1-1.005)	0.041
NRS2002 score
≥ 3 vs < 3	1.768 (1.055-2.964)	0.031	0.745 (0.358-1.55)	0.431
Vascular invasion
Yes vs no	2.096 (1.259-3.489)	0.004	1.32 (0.732-2.38)	0.356
Perineural invasion
Yes vs no	2.122 (1.253-3.593)	0.005	1.058 (0.568-1.968)	0.860
Pathological T stage
T0	1 (reference)	NA	1 (reference)	NA
T1	2.750 (0.555-13.626)	0.215	2.957 (0.575-15.202)	0.194
T2	3.871 (0.866-17.296)	0.076	2.061 (0.421-10.086)	0.372
T3	7.216 (1.673-31.133)	0.008	2.082 (0.353-12.265)	0.418
T4	15.956 (3.738-68.105)	< 0.001	2.888 (0.475-17.565)	0.250
Pathological N stage
Positive vs negative	3.849 (2.279-6.500)	< 0.001	1.154 (0.519-2.565)
pathological stage
Stage 2-3 vs stage 1-0	5.459 (2.941-10.132)	< 0.001	3.720 (1.061-13.044)	0.040
Complication
Yes vs no	1.823 (1.096-3.033)	0.021	1.450 (0.823-2.551)	0.198

BMI: Body mass index; CEA: Carcinoembryonic antigen; CA: Carbohydrate antigen; hsCRP: High-sensitivity C-reactive protein; GNRI: Geriatric nutritional risk index; PIV: Pan-immune-inflammation value; NRS2002: Nutritional Risk Screening 2002; HR: Hazard ratio; CI: Confidence interval; NA: Not available.

Table 3 Univariate and multivariate Cox regression analysis for disease-free survival of clinicopathologic characteristics in the cohorts, median (interquartile range).

Variables	Univariate analysis		Multivariate analysis
	HR (95%CI)	P value	HR (95%CI)	P value
Gender
Male vs female	0.717 (0.414-1.24)	0.234
Age (years)	1.008 (0.983-1.035)	0.522
BMI (kg/m2)
Underweight	1 (reference)	NA
Normal	0.367 (0.112-1.207)	0.099
Overweight	0.312 (0.093-1.052)	0.060
Obese	0.322 (0.08-1.29)	0.110
Smoking history
No vs yes	0.758 (0.453-1.268)	0.291
Drinking history
No vs yes	0.642 (0.373-1.104)	0.109
Diabetes
Yes vs no	1.09 (0.551-2.156)	0.804
Hypertension
Yes vs no	1.174 (0.678-2.032)	0.567
Location
1/3 lower	1 (reference)	NA
1/3 middle	1.455 (0.785-2.697)	0.233
1/3 upper	1.056 (0.373-2.993)	0.918
Total	1.823 (0.838-3.967)	0.130
Nutrition supplement
No vs yes	1.097 (0.629-1.914)	0.743
Clinical T stage
T2	1 (reference)	NA
T3	0.586 (0.229-1.497)	0.264
T4	1.553 (0.654-3.686)	0.319
Clinical N stage
Positive vs negative	1.259 (0.679-2.335)	0.464
Clinical stage
Stage 3 vs stage 2	2.086 (1.143-3.809)	0.017	0.821 (0.394-1.71)	0.599
CEA, ng/mL	1.052 (0.985-1.123)	0.132
CA19-9, U/mL	1.010 (1.002-1.019)	0.018	1.005 (0.994-1.016)	0.374
hsCRP, mg/L	1.022 (0.968-1.08)	0.424
GNRI	0.955 (0.929-0.981)	0.001	0.962 (0.928-0.997)	0.035
PIV	1.002 (1.000-1.004)	0.040	1.001 (0.999-1.004)	0.209
NRS2002 score
≥ 3 vs < 3	1.904 (1.129-3.212)	0.016	0.872 (0.417-1.825)	0.717
Vascular invasion
Yes vs no	2.426 (1.448-4.064)	0.001	1.258 (0.699-2.264)	0.444
Perineural invasion
Yes vs no	2.148 (1.262-3.655)	0.005	0.977 (0.534-1.79)	0.941
Pathological T stage
T0	1 (reference)	NA	1 (reference)	NA
T1	4.619 (0.54-39.542)	0.162	5.224 (0.597-45.681)	0.135
T2	7.987 (1.038-61.429)	0.046	4.727 (0.583-38.327)	0.146
T3	13.694 (1.821-102.995)	0.011	3.158 (0.344-28.949)	0.309
T4	33.792 (4.55-250.976)	0.001	6.105 (0.655-56.896)	0.112
Pathological N stage
Positive vs negative	4.632 (2.685-7.992)	< 0.001	1.21 (0.544-2.691)	0.640
pathological stage
Stage 2-3 vs stage 1-0	6.459 (3.336-12.506)	< 0.001	4.257 (1.242-14.596)	0.021
Complication
Yes vs no	1.952 (1.165-3.272)	0.011	1.642 (0.92-2.93)	0.094

BMI: Body mass index; CEA: Carcinoembryonic antigen; CA: Carbohydrate antigen; hsCRP: High-sensitivity C-reactive protein; GNRI: Geriatric nutritional risk index; PIV: Pan-immune-inflammation value; NRS2002: Nutritional Risk Screening 2002; HR: Hazard ratio; CI: Confidence interval; NA: Not available.

The optimal cutoff values for GNRI and PIV were determined to be 104.56 and 282.74, respectively, based on the maximum Youden index from the receiver operating characteristic curve. Patients were then classified into low and high groups for GNRI and PIV based on these cutoff values. The 5-year OS and DFS rate of the overall study cohort was 59.18% and 60.54%. Kaplan-Meier survival curves for OS, stratified by GNRI and PIV, are presented in Figure 2A and B. The group with a high GNRI and the group with a low PIV demonstrated 5-year OS rates of 70.65% and 64.8%, respectively. Statistically significant differences in OS were observed between the low and high groups for both GNRI (P < 0.001) and PIV (P < 0.001). Similarly, there was a significant difference in DFS between the GNRI and PIV groups (Figure 2C and D).

Figure 2 Survival analysis in different group. A and B: Kaplan-Meier plot lines of overall survival for different geriatric nutritional risk index (GNRI) and pan-immune-inflammation value (PIV) groups; C and D: Disease-free survival for different GNRI and PIV groups; E and F: Overall survival and disease-free survival for different GNRI-PIV score groups. OS: Overall survival; GNRI: Geriatric nutritional risk index; PIV: Pan-immune-inflammation value; DFS: Disease-free survival.

Construct GNRI-PIV score and its relationship with clinicopathological features

As both GNRI and PIV were identified as independent prognostic factors in the multivariate analysis, a composite GNRI-PIV score was established. Patients were assigned a score of 2 for low GNRI and high PIV, a score of 1 for either high GNRI or low PIV (but not both), and a score of 0 for high GNRI and low PIV. This scoring system distributed 82 (55.78%), 53 (36.05%), and 12 (8.16%) patients into scores 0, 1, and 2, respectively. The composite score demonstrated predictive ability for 5-year OS with an area under the curve of 0.693. Kaplan-Meier curves in Figure 2E and F revealed that higher GNRI-PIV scores were inversely associated with 5-year OS and DFS.

The association between the GNRI-PIV score and pathological response is shown in Figure 3. Among the 147 patients, the proportion of patients achieving a response (CAP 0-1) was 42.68% in the score 0 group, which decreased to 22.64% in the score 1 group and 25.00% in the score 2 group. Conversely, the proportion of non-responders (CAP 2-3) increased accordingly with higher scores. Then univariate and multivariate logistic regression showed that GNRI-PIV score associated with tumor regression though it was not an independent factor, which might due to the sample is not sufficient. The changes in pathological response differed significantly across groups with different GNRI-PIV scores (P = 0.044; Table 4). Additionally, the proportion of patients over the BMI of 25 with a GNRI-PIV score of 0 was significantly higher compared to those with scores of 1 and 2 (P < 0.001) and there were less patients with nutrition risk in GNRI-PIV score 0 group (P = 0.002). Furthermore, GNRI-PIV score 0 was significantly associated with early clinical stage and less postoperative complications (P = 0.028 and P = 0.035, respectively; Table 5).

Figure 3 Postoperative pathologic response between different geriatric nutritional risk index-pan-immune-inflammation value score groups. A: Distribution of College of American Pathologists (CAP) grading in patients with a geriatric nutritional risk index (GNRI)-pan-immune-inflammation value (PIV) score of 0; B: CAP grading in patients with a GNRI-PIV score of 1; C: CAP grading in patients with a GNRI-PIV score of 2. GNRI: Geriatric nutritional risk index; PIV: Pan-immune-inflammation value; CAP: College of American Pathologists.

Table 4 Univariate and multivariate logistic regression analysis for pathological response of clinicopathologic characteristics in the cohorts, median (interquartile range).

Variables	Univariate analysis		Multivariate analysis
	OR (95%CI)	P value	OR (95%CI)	P value
Gender
Male vs female	1.098 (0.507-2.376)	0.813
Age (years)	0.999 (0.965-1.034)	0.964
Smoking history
Yes vs no	0.764 (0.385-1.514)	0.440
Drinking history
Yes vs no	1.267 (0.636-2.522)	0.501
Diabetes
Yes vs no	1.672 (0.696-4.018)	0.250
Hypertension
Yes vs no	1.005 (0.477-2.117)	0.990
Location
1/3 lower	1 (reference)	NA
1/3 middle	1.28 (0.559-2.934)	0.559
1/3 upper	0.702 (0.174-2.836)	0.619
Total	0.468 (0.123-1.783)	0.266
Nutrition supplement
Yes vs no	1.788 (0.828-3.861)	0.139
Clinical T stage
T2	1 (reference)	NA	1 (reference)	NA
T3	0.662 (0.215-2.044)	0.473	1.071 (0.313-3.662)	0.913
T4	0.231 (0.072-0.747)	0.014	0.520 (0.132-2.051)	0.350
Clinical N stage
Positive vs negative	1.1 (0.498-2.429)	0.814
Clinical stage
Stage 3 vs stage 2	0.288 (0.141-0.589)	0.001	0.419 (0.178-0.986)	0.046
CEA, ng/mL	1 (0.907-1.102)	0.997
CA19-9, U/mL	0.997 (0.979-1.015)	0.714
hsCRP, mg/L	0.905 (0.795-1.03)	0.132
NRS2002 score
≥ 3 vs < 3	0.641 (0.296-1.392)	0.261
GNRI-PIV score
0	1 (reference)	NA	1 (reference)	NA
1	0.393 (0.181-0.856)	0.019	0.501 (0.221-1.137)	0.098
2	0.448 (0.113-1.776)	0.253	0.446 (0.104-1.909)	0.277

CEA: Carcinoembryonic antigen; CA: Carbohydrate antigen; hsCRP: High-sensitivity C-reactive protein; NRS2002: Nutritional Risk Screening 2002; GNRI: Geriatric nutritional risk index; PIV: Pan-immune-inflammation value; OR: Odds ratio; CI: Confidence interval; NA: Not available.

Table 5 Relationship between clinical characteristics and the different score group, n (%)/mean ± SD.

Clinical characteristics	GNRI-PIV score				P value
	Total (n = 147)	0 (n = 82)	1 (n = 53)	2 (n = 12)
Gender					0.274
Male	107 (72.79)	64 (78.05)	35 (66.04)	8 (66.67)
Female	40 (27.21)	18 (21.95)	18 (33.96)	4 (33.33)
Age					0.555
< 65	101 (68.71)	59 (71.95)	35 (66.04)	7 (58.33)
≥ 65	46 (31.29)	23 (28.05)	18 (33.96)	5 (41.67)
BMI					< 0.001
< 25	97 (65.99)	40 (48.78)	46 (86.79)	11 (91.67)
≥ 25	50 (34.01)	42 (51.22)	7 (13.21)	1 (8.33)
Smoking history					0.935
No	67 (45.58)	37 (45.12)	25 (47.17)	5 (41.67)
Yes	80 (54.42)	45 (54.88)	28 (52.83)	7 (58.33)
Drinking history					0.295
No	85 (57.82)	43 (52.44)	35 (66.04)	7 (58.33)
Yes	62 (42.18)	39 (47.56)	18 (33.96)	5 (41.67)
Diabetes					0.635
No	122 (82.99)	66 (80.49)	46 (86.79)	10 (83.33)
Yes	25 (17.01)	16 (19.51)	7 (13.21)	2 (16.67)
Hypertension					0.285
No	103 (70.07)	59 (71.95)	38 (71.7)	6 (50)
Yes	44 (29.93)	23 (28.05)	15 (28.3)	6 (50)
Location					0.64
1/3 lower	89 (60.54)	46 (56.1)	35 (66.04)	8 (66.67)
1/3 middle	32 (21.77)	19 (23.17)	10 (18.87)	3 (25)
1/3 upper	11 (7.48)	9 (10.98)	2 (3.77)	0 (0)
Total	15 (10.2)	8 (9.76)	6 (11.32)	1 (8.33)
Clinical T stage					0.489
T2	15 (10.2)	10 (12.2)	4 (7.55)	1 (8.33)
T3	65 (44.22)	40 (48.78)	20 (37.74)	5 (41.67)
T4	67 (45.58)	32 (39.02)	29 (54.72)	6 (50)
Clinical N stage					0.744
N-	37 (25.17)	22 (26.83)	13 (24.53)	2 (16.67)
N+	110 (74.83)	60 (73.17)	40 (75.47)	10 (83.33)
Clinical stage					0.028
Stage 2	54 (36.73)	37 (45.12)	12 (22.64)	5 (41.67)
Stage 3	93 (63.27)	45 (54.88)	41 (77.36)	7 (58.33)
CEA, ng/mL	3.30 (3.52)	3.05 (2.62)	3.70 (4.79)	3.20 (2.10)	0.883
CA19-9, U/mL	16.59 (20.63)	15.11 (18.56)	19.56 (24.91)	13.62 (10.52)	0.807
hsCRP, mg/L	2.56 (4.75)	2.48 (4.92)	2.94 (4.97)	1.36 (1.12)	0.416
NRS2002 score					0.002
< 3	103 (70.07)	67 (81.71)	30 (56.6)	6 (50)
≥ 3	44 (29.93)	15 (18.29)	23 (43.4)	6 (50)
Vascular invasion					0.223
No	99 (67.35)	59 (71.95)	31 (58.49)	9 (75)
Yes	48 (32.65)	23 (28.05)	22 (41.51)	3 (25)
Perineural invasion					0.401
No	108 (73.47)	60 (73.17)	41 (77.36)	7 (58.33)
Yes	39 (26.53)	22 (26.83)	12 (22.64)	5 (41.67)
Pathological stage					0.347
PCR	21 (14.29)	12 (14.63)	8 (15.09)	1 (8.33)
1	53 (36.05)	35 (42.68)	16 (30.19)	2 (16.67)
2	35 (23.81)	18 (21.95)	12 (22.64)	5 (41.67)
3	38 (25.85)	17 (20.73)	17 (32.08)	4 (33.33)
CAP					0.044
CAP 0-1	50 (34.01)	35 (42.68)	12 (22.64)	3 (25)
CAP 2-3	97 (65.99)	47 (57.32)	41 (77.36)	9 (75)
Complication					0.035
No	98 (66.67)	62 (75.61)	29 (54.72)	7 (58.33)
Yes	49 (33.33)	20 (24.39)	24 (45.28)	5 (41.67)

BMI: Body mass index; CEA: Carcinoembryonic antigen; CA: Carbohydrate antigen; hsCRP: High-sensitivity C-reactive protein; NRS2002: Nutritional Risk Screening 2002; PCR: Pathological complete response; CAP: College of American Pathologists; GNRI: Geriatric nutritional risk index; PIV: Pan-immune-inflammation value.

DISCUSSION

Most inflammatory and nutritional markers are derived from easily accessible data and commonly conducted tests; however, their clinical utility and relevance remain unclear for patients undergoing NACT. This study evaluated the GNRI and PIV - easily accessible clinical metrics of nutrition and inflammation - as prognostic indicators in 147 LAGC patients receiving NACT. Pre-treatment GNRI and PIV were independent predictors for OS and DFS. Furthermore, nutritional intervention was associated with improved nutritional status and reduced inflammation. Moreover, a combined prognostic index, the GNRI-PIV score, demonstrated superior predictive ability for survival and treatment response compared to either GNRI or PIV alone. These findings suggest that preoperative nutritional and inflammatory status are strongly associated with prognosis and treatment response to NACT in LAGC.

The impact of systemic inflammation and nutritional status on pathological response and prognosis in LAGC patients after NACT remains incompletely characterized. Cancer development leads to systemic inflammation and impaired nutritional status in patients. Malnutrition is closely linked to metabolic dysfunction, immune impairment, and gut microbiota disturbances. It is estimated that 15%-40% of cancer patients are malnourished or at nutritional risk at the time of diagnosis, and 40%-80% become malnourished during treatment[18]. Previous studies have shown that malnutrition contributes to up to 20% of cancer-related deaths, surpassing tumor-related mortality[19]; however, there is currently no standardized method for identifying malnutrition.

Sarcopenia and cachexia are common manifestations of malnutrition in cancer patients, particularly those with gastric cancer. Zou et al[20] measured the cross-sectional area of the third lumbar skeletal muscle, and multivariate analysis revealed that, compared to non-sarcopenic patients, those in the sarcopenic group experienced higher rates of postoperative complications and longer hospital stays. Sarcopenia also emerged as an independent risk factor for overall health status and survival at 6 months post-surgery. However, due to challenges in measurement and potential inaccuracies, sarcopenia is not a reliable marker for malnutrition. While patients with a BMI < 18.5 are typically considered malnourished[21], our study found that BMI was not a significant risk factor for OS or DFS, which is consistent with previous research[22,23]. Serum albumin serves as both an important marker of nutritional status and a significant predictor of tumor prognosis[24,25]. Additionally, hypoalbuminemia has been shown to be associated with rapid clearance of therapeutic drugs and an increased risk of adverse events during adjuvant therapy in advanced cancer[26,27]. However, both BMI and albumin have limitations, as they focus solely on one aspect of nutrition while neglecting the broader metabolic changes and the complex pathophysiological effects of malnutrition. The GNRI offers a more comprehensive assessment of nutritional status, calculated using serum albumin levels, current weight, ideal weight, and height. Previous studies have demonstrated that GNRI is associated with sarcopenia, muscle strength, and muscle area, and is useful in identifying patients who would benefit from nutritional support[28,29]. GNRI has also been identified as a prognostic biomarker and a predictor of treatment response in various cancers[30-32]. Matsunaga et al[33] included a panel of inflammation and nutritional markers for prognostic analysis and found that GNRI exhibited the best performance for elderly patients with gastric cancer following curative gastrectomy. However, the value of GNRI in LAGC patients receiving NACT remains unclear. In our study, we found that low GNRI was associated with poor prognosis and more advanced disease stage, which aligns with previous studies[34,35]. This association may reflect greater tumor aggressiveness and a decline in physical condition among patients with low GNRI.

Although inflammation plays a crucial role in cancer prevention, uncontrolled inflammation is considered a contributor to tumor development through several mechanisms, including DNA damage caused by pro-inflammatory cytokines and chemokines, as well as an increased risk of genomic alterations and instability[36,37]. The PIV score was developed to incorporate all pro-inflammatory cells in the immune system, offering a more precise reflection of inflammatory pressure compared to markers such as the NLR or PLR[12]. In a breast cancer study, the PIV demonstrated superior predictive performance compared to NLR, PLR, monocyte-to-lymphocyte ratio, and the systemic immune-inflammation index, and was significantly associated with patient prognosis[38]. Additionally, in patients receiving immune checkpoint inhibitors, an increase in PIV was linked to poor response to immune checkpoint inhibitors and adverse survival outcomes in colorectal cancer patients[39]. Moreover, Pérez-Martelo et al’s research[40] suggests that PIV monitoring may be useful in predicting disease progression. However, it remains unclear whether inflammatory factors directly promote cancer progression and poor prognosis, and further research is needed to establish PIV as a prognostic indicator in clinical practice. In this research, we found that PIV was one of the risk factors for prognosis in LAGC patients, but its CI was narrow. Although the prognostic utility of PIV alone appears limited due to its minimal effect size, it may still contribute to a larger prognostic model. Future studies with larger sample sizes are needed to determine whether the association between PIV and the outcome is robust and clinically meaningful. For patients undergoing neoadjuvant therapy, existing research supports that nutritional intervention can reduce the incidence of adverse events during treatment while significantly improving preoperative nutritional status[41]. In this study, we also examined the impact of nutritional intervention on the GNRI and PIV. We found that providing oral nutritional supplements or enteral nutrition for LAGC patients during NACT increased GNRI and decreased PIV, indicating that nutritional supplementation can improve nutritional status and reduce inflammation, though the changes of PIV and GNRI did not have an impact on prognosis in this study, which might need more investigation.

The composite GNRI-PIV score demonstrated a higher prognostic value than either marker alone, underscoring the potential of integrated nutritional-inflammatory indices in enhancing prognostic prediction for LAGC patients undergoing NACT. A higher proportion of patients with underweight and nutritional risk were observed in the group with lower nutritional status and higher inflammatory levels, possibly due to increased energy expenditure in patients with elevated inflammation, which exacerbates malnutrition. Additionally, the GNRI-PIV score of 0 was associated with a lower clinical tumor-node-metastasis stage, improved treatment response, and a reduced complication rate. These findings indicate that neoadjuvant therapy is more strongly recommended for LAGC patients with better nutritional status and lower inflammation levels. Nutritional support is also essential, and further investigation into its effects on survival in LAGC patients undergoing NACT is warranted.

Our study has several limitations. First, the patients included in the analysis may have received different numbers of cycles of NACT, which could affect their immune and nutritional status in varying ways. Second, the retrospective nature of this study, combined with its single-institution design, may introduce potential bias. Lastly, the exact implementation of nutritional interventions during treatment presented heterogeneity, which may introduce uncertainty regarding changes in GNRI and PIV. Future prospective research is required to further validate our findings.

CONCLUSION

In summary, we estimated the clinical value of GNRI and PIV in LAGC patients receiving NACT, which are easy and inexpensive markers to assess. We found that GNRI and PIV are reliable biomarkers for prognosis in LAGC. The GNRI-PIV score can be an integral part of treatment decision-making, providing a more reliable therapeutic suggestions for patients with LAGC.

ACKNOWLEDGEMENTS

We extend our gratitude to the patients and their families for their dedication.