Advancing precision medicine in human epidermal growth factor receptor 2 negative gastric cancer: Insights from a novel nomogram for immunochemotherapy prognosis

Abstract

Gastric cancer poses a significant global health burden, particularly in advanced human epidermal growth factor receptor 2-negative cases where prognosis remains poor despite advances in immunochemotherapy. The recent study by Yao et al introduces a nomogram model integrating programmed death ligand 1 expression, microsatellite status, tumor-node-metastasis stage, tumor differentiation, neutrophil-to-lymphocyte ratio, and C-reactive protein-albumin-lymphocyte index to predict progression-free and overall survival. This letter discusses the model’s strengths, limitations, and its alignment with recent developments in biomarkers and therapies, emphasizing its potential for personalized medicine.

Key Words: Gastric cancer; Human epidermal growth factor receptor 2; Immunochemotherapy; Nomogram; Prognostic model; Biomarkers; Precision oncology

Core Tip: The nomogram by Yao et al represents a comprehensive tool for prognosticating outcomes in advanced human epidermal growth factor receptor 2-negative gastric cancer patients on immunochemotherapy. By incorporating immune, clinicopathological, and inflammatory markers, it achieves high discriminative power. Recent advancements in circulatory tumor DNA, artificial intelligence-radiomics, and combination therapies like claudin 18.2-targeting complement this model, highlighting the need for prospective validation to enhance clinical utility.

TO THE EDITOR

The investigation by Yao et al[1], regarding survival risk assessment in human epidermal growth factor receptor 2 (HER2) negative gastric cancer patients receiving immunotherapy combined with chemotherapy is thought provoking. The recent study by Yao et al[1], published in the World Journal of Gastrointestinal Oncology, introduces a timely and innovative nomogram model for predicting progression-free survival (PFS) and overall survival (OS) in advanced HER2-negative gastric cancer patients treated with sintilimab [a programmed death protein 1 (PD-1) inhibitor] plus chemotherapy. From a retrospective cohort of 200 patients from one Chinese center (2021-2024), the authors determined six independent prognostic factors through multivariate Cox regression: Programmed death ligand 1 (PD-L1) expression (combined positive score ≥ 5 vs < 5), microsatellite instability (MSI, high vs stable), tumor-node-metastasis stage (advanced vs earlier), tumor differentiation (poor vs moderate/high), neutrophil-to-lymphocyte ratio (NLR; continuous), and C-reactive protein-albumin-lymphocyte (CALLY) index (continuous). These were combined into nomograms estimating 3- and 6-month PFS, and 12-, 15-, and 18-month OS.

Gastric cancer is the fifth most prevalent neoplasm and the fourth most frequent cause of cancer death globally, with more than 1 million new diagnoses and around 769000 deaths in 2020[2]. In advanced stages, particularly HER2-negative disease, prognosis is dismal, with median OS was historically < 12 months under chemotherapy[3]. The development of immune checkpoint inhibitors (ICIs), including PD-1 inhibitors, has transformed treatment models, providing long-lasting response in some patients when used together with chemotherapy.

The methodological strength of the study is impressive. Patients were split into training (n = 140) and validation (n = 60) sets, with no imbalances at baseline. Model performance was outstanding: C-indices were 0.78-0.82 for both PFS and OS, well above many previous instruments (usually 0.70-0.75)[4]. Area under the receiver operating characteristic curve values were strong (e.g., 0.89 for training 6-month PFS; 0.91 for 15-month OS), and calibration plots established high concordance of predicted vs observed outcomes. Logistic regression also associated these factors (in addition to systemic immune-inflammation index) with objective response rate (16.5% total), illustrating their diverse utility.

The strength of this nomogram is in its comprehensive incorporation of biomarkers. PD-L1 and MSI, proven immunotherapy predictors, conform to National Comprehensive Cancer Network and European Society for Medical Oncology guidelines[5,6]. In tumor biology, tumor-node-metastasis stage and differentiation are mainly discussed. But, NLR and CALLY which are measures of systemic inflammation and nutrition are frequently neglected in molecular-centered model for tumor. High NLR (> 3-5) is associated with immunosuppressive microenvironments through neutrophil cytokines such as interleukin-6/8, suppressing T-cell activity[7]. Likewise, low CALLY (< 1-2) indicates malnutrition and C-reactive protein-induced signal transducer and activator of transcription 3 activation, promoting regulatory T-cell dominance[8]. By measuring these, the model facilitates risk stratification: High-risk patients (e.g., low PD-L1, microsatellite stability, high NLR) may deserve alternative treatment strategies such as intensified monitoring or trials.

Limitations, as noted, are its single-center, retrospective design, potential biases, and sintilimab-specific nature, precluding generalizability to other ICIs. The cohort mixed unresectable locally advanced and metastatic cases, possibly obfuscating subgroup variations. There is no external validation, and advanced biomarkers (e.g., tumor-infiltrating lymphocytes, T-cell exhaustion markers such as T cell immunoglobulin and mucin domain-containing protein 3/Lymphocyte activation gene-3) were excluded, representing real-world limitations but opportunities for optimization[9].

Recent developments in gastric cancer immunotherapy and prognostic modeling

Since the study’s data cutoff (December 2024), the field has evolved rapidly. The 2025 American Society of Clinical Oncology conference featured the phase III RATIONALE-305 trial’s longer-term data, with tislelizumab (another PD-1 inhibitor) with chemotherapy demonstrating median OS of 15.0 months in PD-L1-high advanced gastric cancer, NLR and CALLY serving as exploratory predictors[10]. Likewise, the KEYNOTE-859 3-year follow-up (2025) reaffirmed pembrolizumab’s advantage but cited hyperprogression in 10%-15% of low-PD-L1/microsatellite stability, highlighting prognostic instruments[11].

Biomarker innovations encompass spatial transcriptomics revealing immune “cold” vs “hot” tumors, with MSI-high/PD-L1-high profiles predicting > 80% response rates[12]. Liquid biopsies for circulating tumor DNA (ctDNA) kinetics now complement NLR/CALLY; a 2025 meta-analysis linked ctDNA clearance post-ICI to doubled PFS[13]. Nomogram developments include artificial intelligence: A 2025 study employed machine learning to integrate radiomics, genomics, and inflammation indices, achieving C-indices > 0.85 in multicenter cohorts[14].

Combination therapies are on the rise. Claudin 18.2-targeted agents such as zolbetuximab (in combination with chemotherapy/ICI) improved OS by 3-4 months in CLDN18.2-high subsets[15]. Bispecific antibodies (e.g., PD-1/cytotoxic T-lymphocyte antigen 4) are promising in refractory disease, with trials running stratifying by NLR/PD-L1[16]. In the realm of sintilimab’s prevalence in China, real-world evidence from 2025 registries (n > 5000) confirms its effectiveness, which is consistent with Yao et al’s median OS (16.1 months)[1]. To contextualize, Table 1 compares recent nomograms for advanced gastric cancer prognosis under immunotherapy. These models underscore a shift toward multimodal integration[17-20], with Yao et al[1] emphasis on accessible indices like CALLY filling a gap in resource-limited settings.

Table 1 Comparison of recent nomograms for prognosis in advanced gastric cancer treated with immunotherapy.

Ref.	Key factors included	Cohort size/type	Endpoints	Performance (C-index/AUC)	Strengths/Limitations
Yao et al[1], 2025	PD-L1, microsatellite instability, tumor-node-metastasis stage, differentiation, neutrophil-to-lymphocyte ratio, C-reactive protein-albumin-lymphocyte	200/retrospective, single-center	PFS (3/6 months), OS (12/15/18 months)	C-index: 0.78-0.82; AUC: 0.79-0.91	Integrates inflammation-nutrition; internal validation
Wang et al[18], 2024	IrAE, clinical characters	158/retrospective	OS	C-index: 0.72-0.76; AUC: 0.78-0.85	Includes immune-related adverse events; single-center
Zhao et al[19], 2021	Radiomics features for EBV prediction	120/retrospective, multi-database	EBV status (surrogate for prognosis)	AUC: 0.82-0.89	Noninvasive CT-based; limited to EBV
He et al[20], 2025	Circulating tumor DNA dynamics	85/prospective cohort	PFS/OS	AUC: 0.80-0.87	Dynamic monitoring; small sample
Fu et al[14], 2025	Delta radiomics, AI features	200/multicenter retrospective	PFS in stage IV	C-index: 0.75; AUC: 0.81-0.88	AI-powered; focuses on immune checkpoint inhibitors

AUC: Area under the receiver operating characteristic curve; PD-L1: Programmed death ligand 1; PFS: Progression-free survival; OS: Overall survival; IrAE: Immune-related adverse event; EBV: Epstein-Barr virus; CT: Computed tomography; AI: Artificial intelligence.

Conclusion

Yao et al’s work[1] exemplifies precision oncology’s progress, offering a practical tool to optimize immunochemotherapy in HER2-negative gastric cancer. By bridging biomarkers and clinical reality, it paves the way for tailored interventions, potentially transforming outcomes in this lethal disease.

Future directions

Provision for prospective, multicenter validation of Yao et al’s nomogram[1] is essential, optimally including ctDNA, spatial omics, and new therapies. Dynamic modeling, repeatedly updating scores during treatment, might increase utility. Equity in worldwide availability of ICIs remains important, since inequality continues to exist in low-resource areas.

ACKNOWLEDGEMENTS

We acknowledge the moral support given by Centre for Integrated OMICS and Computational Biology, Utkal University, Vani Vihar, Bhubaneswar.