Prognostic performance of the nutritional risk screening 2002 and the global leadership initiative on malnutrition in patients with lung cancer treated with immune checkpoint inhibitors

Abstract

Background

The Nutritional Risk Screening 2002 (NRS 2002) is used to identify patients at risk who may benefit from nutritional intervention, while the Global Leadership Initiative on Malnutrition (GLIM) criteria serve as a diagnostic tool for malnutrition. However, their impacts on response to immunotherapy in patients with lung cancer are unknown. Therefore, this study investigated the role of NRS 2002 and GLIM criteria in predicting overall survival (OS) in patients with lung cancer receiving immune checkpoint inhibitors (ICIs).

Methods

This retrospective study included 146 patients with lung cancer treated with ICIs. Nutritional risk was evaluated using the NRS 2002, while nutritional status was assessed using the GLIM criteria. The OS rates were analyzed using Kaplan-Meier curves and Cox proportional hazard analyses. Two nomograms based on NRS 2002 and GLIM criteria were established to predict OS. Kendall’s Tau was used to determine the relationship between NRS 2002 scores and GLIM criteria.

Results

The Kaplan-Meier survival revealed poor survival times in patients with nutritional risk (NRS 2002 score ≥ 3) and malnutrition as defined by the GLIM criteria. The multivariable Cox analyses identified that nutritional risk (hazard ratio [HR], 2.00, 95% confidence interval [CI]: 1.06–3.81, P = 0.033) and malnutrition (HR, 2.46, 95%CI: 1.28–4.77, P = 0.007) are independent factors predicting OS. Age, stage, NRS 2002 score, and GLIM criteria were used to develop the nomograms. Higher nomogram scores predicted significantly reduced OS (P < 0.0001). NRS 2002 scores and GLIM criteria showed a moderately strong positive correlation, with Kendall’s tau = 0.709 (95% CI: 0.629–0.785; P < 0.001).

Conclusions

Nutritional risk and GLIM criteria-defined malnutrition are prognostic factors in patients with lung cancer receiving immunotherapy. NRS 2002 scores were significantly correlated with GLIM criteria in predicting diagnostic outcomes in lung cancer.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12885-025-15258-7.

Background

Lung cancer is the most common cancer worldwide [1], with 2.2 million diagnoses and 1.8 million deaths annually [2]. According to the stage and pathological type of lung cancer, the five-year survival rate of lung cancer with distant metastases is 8% [2]. The mean five-year survival rate of patients with non-small cell lung cancer (NSCLC) is only 15% [3]. In recent years, advances in immunotherapy have greatly improved the treatment outcomes in lung cancer, owing to its favorable safety profile and durable therapeutic response [4]. Compared with traditional treatments such as chemotherapy, immunotherapy significantly improves patient clinical outcomes [5]. The aggregated research data of CheckMate 017 and 057 confirmed the long-term overall survival (OS) and progression-free survival (PFS) benefits of nivolumab in a broad population of patients with NSCLC [6].

However, not all patients benefit from immunotherapy as the therapeutic effect of immune checkpoint inhibitors (ICIs) varies greatly among patients and many patients develop drug resistance and serious adverse reactions during treatment [7]. Although programmed cell death ligand 1 (PD-L1) expression and tumor mutational burden (TMB) are effective predictors of ICI response, they are unstable [8]. Moreover, the predictive effect of genetic mutations on the prognosis of individual patients is not entirely accurate. Therefore, additional factors associated with immunotherapy efficacy are needed.

Nutritional status plays an essential role in the prognosis of patients with cancer. Numerous studies have demonstrated poor prognosis in patients with malnutrition [9, 10]. In clinical practice, screening for nutritional risk using Nutritional Risk Screening 2002 (NRS 2002) is the essential initial step for preventing and treating malnutrition in hospitalized patients [11]. This tool is widely adopted, especially for medical inpatients, and can help to identify patients at risk for long-term mortality [12]. NRS 2002 scores are also associated with poor prognosis in patients with cancer [13]. Several studies have investigated the predictive value of NRS 2002 in immunotherapy. Tang et al. reported poorer prognosis among patients with solid tumors and nutritional risk according to NRS 2002 at treatment initiation compared with patients without nutritional risk [14]. However, conflicting results emerged from a larger study involving 1,340 patients with advanced esophageal cancer, which reported no significant correlation between NRS2002 scores and therapeutic response or prognosis following ICI therapy [15].

The Global Leadership Initiative on Malnutrition (GLIM) criteria provide a consensus-based framework for diagnosing malnutrition in adults, incorporating phenotypic criteria like muscle mass loss and etiologic factors such as inflammation [16]. While some studies have shown that malnutrition as defined by GLIM criteria is associated with a poor prognosis in patients with cancer, the results are inconsistent [17]. In one retrospective analysis of 117 patients with esophageal cancer, Okada et al. reported that five-year mortality was not significantly associated with moderate or severe malnutrition as defined by the GLIM criteria [18].

To our knowledge, only one retrospective study has examined the predictive role of NRS 2002 in patients with lung cancer receiving immunotherapy. However, that study did not specifically analyze the prognostic impact of different NRS 2002 scores. Moreover, whether NRS 2002 was assessed before or during immunotherapy was not clear [19]. Furthermore, no study has explored the predictive value of malnutrition as defined by GLIM criteria in patients with lung cancer receiving immunotherapy. Therefore, the present study investigated the predictive role of NRS2002 scores and GLIM criteria-defined malnutrition on immunotherapy efficacy in patients with lung cancer. This study also explored the relationship between NRS2002 scores and GLIM criteria scores in lung cancer.

Methods

Study design and patient selection

This single-center retrospective study enrolled patients with lung cancer treated at one institution between May 2018 and October 2020. The inclusion criteria were: (1) histopathologically confirmed lung cancer treated with immunotherapy including programmed cell death 1 (PD-1) programmed death-ligand 1 (PD-L1) inhibitors; (2) age 18–80 years; (3) availability of NRS2002 and GLIM criteria results within 2 weeks before treatment initiation; and (4) complete clinical datasets. The exclusion criteria were: (1) prior immunotherapy completion, (2) incomplete medical records, and (3) multiple primary malignancies.

Among 260 patients initially identified through electronic medical record searches, 146 who met all eligibility criteria constituted the study cohort. The Research Ethics Committee of Henan Provincial People’s Hospital approved this study (No. 2022 − 220) and waived the requirement for informed consent due to its retrospective nature. The study adhered to the principles of the Declaration of Helsinki.

Immunotherapy regimen, data collection, and follow-up

All patients with TNM stage II–Ⅳ lung cancer were treated with immunotherapy with PD-1/PD-L1 inhibitors as the main component of their treatment regimen. Both PD-1 and PD-L1 inhibitors were administered in accordance with instructions and guidelines for lung cancer treatment. Eligible patients received concurrent chemotherapy or targeted therapy drugs.

We retrospectively collected pretreatment clinical characteristics including sex, age, smoking history, height, weight, TNM stage, and pathological type. Treatment details (specific PD-1/PD-L1 inhibitors and regimens), pretreatment NRS 2002 scores, pretreatment GLIM criteria results, and last follow-up date were also recorded.

All patients were followed up every three months after immunotherapy. The follow-up methods included telephone consultation, outpatient examination, and hospitalization. The deadline for follow-up was September 30, 2023. OS was defined as the interval from immunotherapy initiation to death from any cause. Patients who survived were censored at the final follow-up.

NRS2002 and GLIM criteria

According to the clinical practice guidelines, nutritional risk screening and nutritional assessment, including GLIM diagnosis, must be completed for inpatients within 48 h of admission in the cancer ward of the hospital. NRS2002 was performed according to the European Society for Clinical Nutrition and Metabolism (ESPEN) guidelines. This tool assesses two main dimensions—impaired nutritional status and disease severity (stress metabolism)—with an adjustment for age. An initial pre-screen identifies patients needing further evaluation based on low body mass index (BMI), recent weight loss, reduced intake, or severe illness. The final screen assigns points for impaired nutritional status (0–3 points based on weight loss, BMI, or reduced intake), disease severity (0–3 points reflecting increased metabolic stress), and age (+ 1 point if ≥ 70 years). The total score classified patients as at nutritional risk (positive, ≥ 3 points), and with no risk (negative, < 3 points).

Malnutrition was diagnosed using the GLIM criteria, which require ≥ 1 phenotypic criterion (non-volitional weight loss of > 5% within 6 months, low BMI (< 20 kg/m² if age < 70 years; <22 kg/m² if age ≥ 70 years), or reduced muscle mass as measured by handgrip strength (< 28 kg for men and < 18 kg for women) and ≥ 1 etiologic criterion (reduced food intake/assimilation of ≤ 50% of energy requirements for > 1 week, or inflammation/disease burden including acute/infectious, chronic disease-related, or injury).

Statistical analysis

Continuous variables are expressed as means ± standard deviation (SD) or median (interquartile range). Categorical variables are expressed as the number of cases and percentages. Survival analysis was conducted using Kaplan–Meier curves with log-rank tests. Cox proportional hazard models were used to identify prognostic factors. We fitted multivariable Cox models defined a priori based on clinical relevance (age, sex, surgery, TNM stage, type of pathology, treatment regimen, and BMI), NRS 2002 scores and GLIM criteria results. Significant predictors from multivariable analysis were incorporated into prognostic nomograms for OS. The discriminative ability of the nomograms was assessed using the optimism-corrected C-index, computed through bootstrap validation (1,000 resamples) with the rms package in R. The 95% confidence interval was estimated using the percentile bootstrap method. Receiver operating characteristic curves were used to determine the best cut-off values for nomogram-derived points. The patients were stratified into high- and low-point groups based on the cut-off values. The correlation between NRS 2002 scores and GLIM criteria results was determined using Kendall’s Tau. A mosaic plot was used to visualize the relationship between NRS 2002 scores and GLIM criteria results. All analyses were performed using R (version 4.3.3). Statistical significance set to P < 0.05.

Results

Patient characteristics

This study enrolled 146 patients who met the inclusion criteria. The total cohort had a median age of 64 (32.0–88.0) years and a mean BMI of 23.7 ± 3.27 kg/m². Among all the patients, 98 were male and 48 were female. Based on the GLIM criteria, malnutrition was diagnosed in 37 patients (25.3%; 10 females, 27 males). Nutritional risk was identified in 46 patients (31.5%). Among all patients, 112 (76.7%), 31 (21.2%), and three (2.1%) had stage Ⅳ, Ⅲ, and Ⅱ disease, respectively. The pathological type was NSCLC in 119 cases and small cell lung cancer (SCLC) in the rest. Sixteen patients had undergone surgical treatment. Among ICIs, 136 patients received PD-1 drugs, including camrelizumab, pembrolizumab, sintilimab, sugemalimab, and tislelizumab, while the remaining patients received PD-L1 drugs including durvalumab and atezolizumab. The basic characteristics of the study population are shown in Table 1.

Table 1.

Baseline characteristics of the study population

		Total
	level	(N = 146)
Sex (%)	Female	48 (32.9%)
	Male	98 (67.1%)
Age (median [IQR])		64.0 [53.0, 69.0]
BMI (mean (SD))		23.7 (3.27)
Smoking	Yes	75 (51.4%)
	No	71 (48.6%)
TNM stage (%)	Ⅱ~Ⅲ	34 (23.3%)
	Ⅳ	112 (76.7%)
T (%)	1	17 (11.6%)
	2	34 (23.3%)
	3	29 (19.9%)
	4	60 (41.1%)
	X	6 (4.1%)
N (%)	1	13 (8.9%)
	2	83 (56.8%)
	3	45 (30.8%)
	X	4 (2.7%)
M (%)	0	36 (24.7%)
	1	110 (75.3%)
Pathology (%)	NSCLC	119 (81.5%)
	SCLC	27 (18.5%)
Surgery (%)	No	130 (89.0%)
	Yes	16 (11.0%)
Type of ICIs (%)	PD-1	136 (93.2%)
	PD-L1	10 (6.8%)
NRS 2002 (%)	Mean	1.83
	≥ 3 points	46 (31.5%)
	< 3 points	100 (68.5%)
GLIM criteria (%)	Non-malnutrition	109 (74.7%)
	Moderate malnutrition (Stage 1)	37 (25.3%)
	Severe malnutrition (Stage 2)	16 (11.0%)

BMI Body mass index, NSCLC Non-Small cell lung Cancer, SCLC Small cell lung Cancer, ICIs Immune checkpoint inhibitors, PD-1 Programmed cell death protein 1, PD-L1 Programmed cell death ligand 1, NRS 2002 Nutritional risk screening 2002, GLIM criteria The global leadership initiative on malnutrition criteria

Predictive value of NRS 2002 on OS

Figure 1A presents the Kaplan–Meier survival curves of OS according to NRS 2002. In the total cohort, patients with negative NRS 2002 results had significantly better estimated three-year OS rates compared with patients with nutritional risk (≥ 3 points) (62.9% vs. 40.4%; P = 0.0208). Delving deeper into NRS 2002 scores, patients with low NRS 2002 scores manifested a notably superior OS compared with patients with high scores (P < 0.001). (Fig. 1B) Survival analysis of 119 patients with NSCLC demonstrated significantly poorer OS among NSCLC patients with nutritional risk compared with those without nutritional risk (estimated 3-year OS rates, 38.0% vs. 64.4%; P = 0.018) (Fig. 1C). Additionally, a significant association was observed between nutritional risk and OS in patients receiving PD-1 inhibitors (estimated 3-year OS rates, 35.6% vs. 61.5%; P = 0.009) (Fig. 1D).

Fig. 1.

Kaplan-Meier curves of survival time according to Nutritional Risk Screening 2002 (NRS 2002) for overall survival (OS) according to nutritional risk (≥ 3) (A) and NRS 2002 scores (B). Kaplan-Meier curves of survival time according to NRS 2002 for OS in patients with non-small cell lung cancer (NSCLC) (C) and receiving programmed cell death protein 1(PD-1) inhibitors (D).

Predictive value of GLIM criteria on OS

The Kaplan–Meier survival curves of OS according to the GLIM criteria are presented in Fig. 2. The estimated three-year OS rate was significantly lower in the malnourished group than in the non-malnourished group (28.1% vs. 63.7%, P = 0.005) (Fig. 2A).

Fig. 2.

Kaplan-Meier curves of survival time according to the Global Leadership Initiative on Malnutrition (GLIM) criteria for overall survival (OS) (A) in patients with non-small cell lung cancer (NSCLC) (B) and receiving programmed cell death protein 1(PD-1) inhibitors (C)

In patients with NSCLC, a pronounced enhancement in the estimated three-year survival rate was observed in the non-malnourished group compared with the malnourished group (64.1% vs. 31.6%, P = 0.020) (Fig. 2B). Similar results were obtained in patients receiving PD-1 inhibitors (estimated 3-year OS rates, non-malnourished group, 62.2% vs. malnourished group, 18.6%; P = 0.002) (Fig. 2C).

Univariate and multivariable analyses of OS

The results of Cox regression analyses are shown in Table 2. In univariate Cox proportional hazard regression analysis, NRS 2002 scores were significantly associated with OS (Table 2). The HR values suggested that NRS 2002 scores ≥ 2 had better performance on predicting OS than NRS 2002 scores ≥ 3 (HR, 2.46 vs. 1.97). Univariate analysis also revealed that malnutrition as defined by GLIM criteria was associated with worse OS (HR, 2.31, 95% CI: 1.28–4.17, P = 0.006).

Table 2.

Univariate and multivariable analyses of overall survival in total cohort

	Univariate analysis		Multivariable Cox regression analysis
	HR (95% CI)	P	HR (95% CI)	P
Age	1.03 (1.00, 1.06)	0.052	1.03 (1.01, 1.07)	0.021*
Sex
Female vs. Male	1.14 (0.61, 2.12)	0.690	1.29 (0.52, 3.22)	0.588
BMI	0.99 (0.90, 1.08)	0.730	1.05 (0.95, 1.17)	0.312
Smoking
Yes vs. No	1.15 (0.65, 2.05)	0.635	0.86 (1.17, 0.37)	0.725
Types of Pathology
SCLC vs. NSCLC	1.18 (0.59, 2.38)	0.642	1.20 (0.57, 2.54)	0.637
Surgery
No vs. Yes	0.34 (0.08, 1.38)	0.13	0.36 (0.09, 1.53)	0.167
TNM stage
Ⅳ vs. Ⅱ~Ⅲ	2.15 (0.96, 4.80)	0.063	2.52 (1.10, 5.77)	0.029 *
Type of ICIs
PD-L1 vs. PD-1	0.52 (0.13, 2.14)	0.362	0.43 (0.10, 1.96)	0.278
NRS 2002
≥ 3 vs.< 3	1.97 (1.10, 3.54)	0.023	2.00 (1.06, 3.81)	0.033 *
≥ 2 vs.< 2	2.46 (1.35, 4.47)	0.003	2.34 (1.23, 4.45)	0.010 *
Continuous	1.52 (1.16, 1.99)	0.003	1.57 (1.14, 2.18)	0.006 *
GLIM criteria
Malnutrition vs. Non-malnutrition	2.31 (1.28, 4.17)	0.006	2.46 (1.28, 4.77)	0.007 #

BMI Body mass index, NSCLC, Non-Small Cell Lung Cancer, SCLC Small Cell Lung Cancer, ICIs Immune checkpoint inhibitors, PD-1 Programmed cell death protein 1, PD-L1 Programmed cell death ligand 1, NRS 2002 Nutritional Risk Screening 2002, GLIM criteria The Global Leadership Initiative on Malnutrition criteria

* adjusted by age, TNM stage and NRS 2002

# adjusted by age, TNM stage and GLIM criteria

In the multivariable Cox regression analysis, OS was shorter in patients with malnutrition as diagnosed by GLIM criteria compared with patients without malnutrition (HR, 2.46, 95% CI: 1.28–4.77, P = 0.007). In the multivariable Cox regression adjusted for the same covariates, NRS 2002 scores were also significantly associated with worse OS (≥ 3 vs.<3 points: HR, 2.00, 95% CI: 1.06–3.81, P = 0.033; ≥2 vs. <2 points: HR, 2.34, 95% CI: 1.23–4.45, P = 0.010; continuous NRS 2002 scores, HR, 1.57, 95% CI: 1.14–2.18, P = 0.006) (Table 2).

Development of nomogram models for OS based on NRS 2002 scores and GLIM criteria

The prognostic nomograms were established using prognostic variables of OS identified in the multivariable Cox regression analyses, including age, TNM stages, and NRS 2002 scores/GLIM criteria results. Based on HR value reflecting each factor’s contribution to the model, we assigned points to factor levels. Summing these points yielded a total score, which was then converted to a predicted mortality rate using a predefined relationship in the nomograms (Supplemental files). Given the sample size (n = 146 with 47 events), internal validation was performed using bootstrap resampling with 1,000 replicates to correct for overoptimism. The optimism-corrected C-index of the nomogram incorporating the NRS 2002 score was 0.665 (95% CI: 0.581–0.750), whereas that of the model based on the GLIM criteria was 0.662 (95% CI: 0.579–0.745). Both values indicate modest discriminative performance for 3-year overall survival. Calibration plots derived from 1 000 bootstrap resamples demonstrated good agreement between the predicted and observed survival probabilities (Supplemental files). Receiver operating characteristic curves were used to determine the optimal nomogram score cutoffs. The cutoff values for NRS 2002 and GLIM criteria were 131.8 and 111.7 for nomogram points, respectively (Supplemental files). Patients were then stratified by these cutoffs into high-point and low-point groups. Kaplan–Meier analysis demonstrated that the nomogram points were significantly correlated with OS in the total cohort (log-rank P < 0.0001 for both nomograms) (Supplemental files).

Relationship between NRS 2002 scores and GLIM criteria

Kendall’s Tau was used to determine the relationship between NRS 2002 scores and GLIM criteria. A moderately strong positive correlation was observed between NRS 2002 scores and GLIM criteria (Kendall’s Tau = 0.709; 95% CI: 0.629–0.785; P < 0.001). The correlation analysis revealed a strong positive correlation between nutritional risk (NRS 2002 scores ≥ 3) and malnutrition as defined by GLIM criteria (Kendall’s Tau = 0.791; 95% CI: 0.651–0.879; P < 0.001). A mosaic plot was also created to show the relationship between NRS 2002 scores and GLIM criteria results (Fig. 3).

Fig. 3.

Mosaic plot for relationship between Nutritional Risk Screening 2002 scores and the Global Leadership Initiative on Malnutrition criteria results

Discussion

This retrospective study investigated the predictive value of pretreatment nutritional risk and malnutrition on survival time in patients with lung cancer treated with ICIs. Nutritional risk was identified using the NRS2002 while malnutrition was diagnosed using the GLIM criteria. The analysis included 46 (31.5%) and 100 (68.5%) patients with and without nutritional risk, respectively. The findings demonstrated poorer OS in patients with nutritional risk compared with those without nutritional risk. Among patients with lung cancer who received immunotherapy, every one-point increase in NRS 2002 score was associated with a 1.57-fold increased risk of mortality. Malnutrition as diagnosed by the GLIM criteria also predicted poor survival in patients with lung cancer treated with ICIs. Moreover, NRS 2002 scores were highly correlated with GLIM criteria results in this patient population.

NRS 2002 is widely used to screen patients at risk of poor prognosis in clinical practice. ESPEN guidelines recommend the use of this tool within 24 h of admission [11]. A retrospective analysis of 128 randomized controlled trials of nutritional support demonstrated the predictive validity of this tool, demonstrating that patients who met the risk criteria in the randomized controlled trials were more likely to achieve prior clinical outcomes from nutritional support compared with those without nutritional risk [12]. Moreover, a secondary analysis of a prospective randomized controlled trial in Switzerland reported that nutritional risk was a strong and independent factor for malnutrition-associated mortality (HR = 1.37) and adverse outcomes [20]. Thus, the application scope of NRS 2002 covers all adult patients, regardless of their diseases and ages, including all other patients in clinical settings [20]. NRS 2002 score is also closely related to the prognosis of patients with cancer. One meta-analysis of 22 studies showed that NRS 2002 score was independently related to increased risks of postoperative complications and poor OS [21]. However, the validation of NRS 2002 in specific disease populations and specific treatment regimens is limited, and the selection of cutoff values has not been thoroughly studied [22]. Peng et al. study reported significantly poorer survival outcomes among patients with nasopharyngeal carcinoma with NRS2002 scores ≥ 4 compared with those scores ≥ 3 [23]. Furthermore, few studies have investigated the predictive role of NRS 2002 in patients receiving immunotherapy. A real-world evidence analysis of 341 patients with solid tumors including esophageal, lung, gastric, and colorectal cancer demonstrated shorter OS in patients with higher pretreatment NRS 2002 compared with patients without nutritional risk [14].

The results of the present study demonstrated that high NRS 2002 scores predicted poor survival in patients with lung cancer treated with ICIs. These findings further confirmed that the pretreatment NRS 2002 scores can effectively predict the prognosis of these patients. Moreover, as the score increased, the risk of death significantly increased. Additionally, the risk of death in patients with NRS 2002 scores ≥ 2 was 2.34-fold higher than that in patients with a score < 2. These findings suggest that both a nutritional risk score ≥ 3 and an NRS-2002 score = 2 should be considered in clinical evaluation. A comprehensive assessment of nutritional status, along with appropriate nutrition therapy, should be provided for patients with NRS scores ≥ 2 who are receiving immunotherapy.

The GLIM criteria are a global consensus diagnostic tool for malnutrition [24]. Several studies have explored the relationship between GLIM criteria score and prognosis in patients with cancer. A meta-analysis of 10 studies reported that GLIM-defined malnutrition worsens OS (HR = 1.56) and increases postoperative complications in patients with cancer [17]. However, most study participants were patients who underwent surgery. Another study concluded that the GLIM criteria effectively predict survival in adult hematology patients receiving chimeric antigen receptors T-cell therapy [25]. However, few studies have explored the predictive value of GLIM criteria in patients receiving immunotherapy. Thus, the present study evaluated malnutrition as defined by GLIM criteria in predicting survival among patients with lung cancer receiving immunotherapy. Our results confirmed that, after adjusting for other factors, malnutrition was associated with poor OS during immunotherapy.

The proposed mechanism by which the presence of malnutrition before treatment can predict the prognosis of patients with lung cancer is as follows. First, malnutrition can increase the incidence of adverse reactions during anti-tumor treatment, including immunotherapy [26]. For patients undergoing combined chemotherapy, the toxicity of chemotherapy also significantly increases, thereby shortening the OS [27, 28]. Second, patients with malnutrition often have frailty, which can lead to dose reductions and treatment plan modifications [29]. Frailty also predicts poor clinical outcomes in patients with cancer [30]. Third, malnourished patients often experience muscle loss, and sarcopenia is also an important factor affecting the prognosis of patients with cancer [31, 32]. Fourth, malnourished patients are vulnerable to cancer cachexia, defined as disease-related malnutrition associated with chronic inflammation [33]. Patients with cancer cachexia show a significantly increased risk of mortality [34].

In this study, we also developed and preliminarily validated two nomograms for predicting OS based on the NRS 2002 scores and GLIM criteria. Our results also suggested the need for consideration of nutritional treatment in patients with high nomogram scores. Notably, one of the main research gaps in present research is that the results could not verify the predictive value of NRS 2002 and the GLIM criteria on the response to nutritional treatment. In future, more research to explore the predictive role of NRS 2002 and the GLIM criteria on response to nutritional support are needed.

The findings of this study also demonstrated the significant correlation between NRS 2002 scores and GLIM results in patients with lung cancer receiving immunotherapy. While this relationship has been explored in various studies, the results are inconsistent. A study in Norway showed that only 77 (54%) of the 143 patients who were at nutritional risk by NRS 2002 were also diagnosed as malnourished according to the GLIM criteria [35]. Another study demonstrated the high sensitivity and specificity between NRS 2002 and the GLIM criteria in patients with gastrointestinal stromal tumors [36]. The GLIM criteria are the widely recognized diagnostic standard for malnutrition. The results of the present study also suggested that NRS 2002 can effectively predict malnutrition. However, further research is needed to verify the application value of the NRS 2002 scores in diagnosing malnutrition and the first-step risk screening of the GLIM scheme.

This study has some limitations. First, this was a retrospective, single-center study, which may have introduced selection bias. Second, nutritional interventions administered throughout the treatment course could have influenced the clinical outcomes; however, relevant data were insufficiently documented in this study. Third, the reliance on telephone-based follow-up examinations for some patients may have resulted in inaccuracies in death ascertainment. Fourth, the limited sample size may have compromised the reliability of the findings. Fifth, the absence of detection of PD-L1 expression and TMB also impacted the reliability of the prognosis results. Finally, this study did not evaluate other clinically relevant endpoints, including PFS and treatment-related toxicity rates.

Conclusion

Nutritional risk and GLIM-defined malnutrition are prognostic factors in patients with lung cancer receiving immunotherapy. Additionally, NRS 2002 scores were significantly correlated with GLIM diagnostic outcomes in lung cancer. In clinical practice, the NRS2002 and the GLIM criteria should be used to evaluate the nutritional status of patients undergoing immunotherapy, and nutritional intervention should be provided as early as possible. Additional prospective studies with large sample sizes are needed to verify the predictive role of these tools in predicting the prognosis and response to nutritional therapy of lung cancer.

Abbreviations

NSCLC

Non-small cell lung cancer

OS

Overall survival

PFS

Progression-free survival

ICIs

Immune checkpoint inhibitors

PD-1

Programmed cell death protein 1

PD-L1

Programmed cell death ligand 1

NRS 2002

Nutritional Risk Screening 2002

GLIM

the Global Leadership Initiative on Malnutrition

Author’ contributions

Wen Wang, Conceptualization, Data curation, Investigation, Formal analysis; Xintian Xu, Data curation, Formal analysis, Investigation, Methodology, Software and Writing-review and editing; Mengxing Tian, Methodology, and Writing-review and editing; Qian Han, Investigation and Writing-review and editing; Tingting Yang, Writing-review and editing; Xin Jin, Conceptualization, Formal analysis, Supervision, Validation and Writing-original draft; Lei Lei, Supervision. All authors have read and approved the final manuscript.

Funding

This work was supported by the National Science Foundation of Hubei Province, China (grant No. 2024AFB1030), CIFST - Abbott Foundation of Food Nutrition and Safety (202305), Talent Project of Hubei Cancer Hospital (2025HBCHQHRC009) and Henan Provincial Department of Health, 2025 Science and Technology Research Project (LHGJ20250034).

Data availability

The data that support the findings of this study are available on request from the corresponding author.

Declarations

Competing interests

The authors declare no competing interests.