Gastrointestinal tumor personalized immunotherapy: an integrated analysis from molecular genetics to imaging biomarkers

Abstract

The immunotherapy landscape for gastrointestinal (GI) tumors is rapidly evolving. There is an urgent need for reliable biomarkers capable of predicting treatment outcomes to optimize therapeutic strategies and enhance patient prognosis. This review presents a comprehensive overview of biomarkers associated with the immunotherapy response of GI tumors, covering advances in molecular genetics, histopathological markers, and imaging. Key molecular biomarkers, such as microsatellite instability, tumor mutational burden, and programmed death-ligand 1 expression, remain critical for identifying patients likely to benefit from immune checkpoint inhibitors. The significance of tumor-infiltrating lymphocytes, notably the CD8+ T cell to regulatory T cell ratio, as a predictor of immunotherapy response is explored. In addition, advanced imaging techniques, including computed tomography (CT), magnetic resonance imaging, and positron emission tomography-CT, facilitate the noninvasive evaluation of tumor biology and therapeutic response. By bridging molecular and imaging data, this integrated strategy enhances precision in patient selection, treatment monitoring, and adaptive therapy design. Future studies should aim to validate these biomarkers in larger, multicenter cohorts and focus on clinical translation to advance precision medicine in GI oncology.

Keywords

biomarkers gastrointestinal tumors imaging immunotherapy tumor immunology

Introduction

Gastrointestinal (GI) cancers, including esophageal, gastric, and colorectal cancers, are major causes of cancer-related deaths globally.¹ Despite progress in surgery, chemotherapy, and radiotherapy, advanced GI cancers still have poor outcomes due to genetic complexity and tumor microenvironment (TME) heterogeneity,^2,3 sparking increasing interest in immunotherapy as a novel treatment approach. Immunotherapy, particularly through immune checkpoint inhibitors (ICIs),^4,5 has revolutionized the treatment landscape for these malignancies.^6,7 Nevertheless, variability in patient responses to immunotherapy underscores the urgent need for predictive biomarkers capable of identifying patients most likely to benefit from such treatments.^8–10

Predictive biomarkers are now a key focus in GI oncology research.^11,12 Key biomarkers have under investigated, including molecular genetic factors such as microsatellite instability (MSI) and mismatch repair deficiency (dMMR),¹³ protein expression levels such as programmed death 1 (PD-1) and programmed death-ligand 1 (PD-L1).¹⁴ In recent years, imaging-based biomarkers have gained significant attention.^15–17 This review aims to integrate current research findings on predictive biomarkers for immunotherapy in GI tumors, assess their role in elucidating tumor–immune interactions, emphasize their predictive potential and clinical application, and explore their potential in guiding individualized treatment strategies (Figure 1).

Figure 1.

The main content of this review (By Figdraw).

TME and mechanisms of immunotherapy

The TME is a complex and dynamic ecosystem composed of cancer cells, immune cells, fibroblasts, stromal elements, and vascular components,^18,19 which play crucial roles in tumor progression and therapeutic response.^20,21 Immune cells, such as CD8+ cytotoxic T lymphocytes and natural killer (NK) cells, help eliminate malignant cells. However, tumors can evade immune surveillance by recruiting regulatory T cells (Tregs) and myeloid-derived suppressor cells (MDSCs), fostering an immunosuppressive environment,^22–24 and upregulating immune checkpoint molecules such as PD-1/PD-L1.^25–27

Interaction between immune cells and the TME

The interaction between immune cells and the TME plays a pivotal role in determining the success of immunotherapy.^28,29 For instance, high densities of effector T cells, particularly CD8+ T cells, which can recognize and kill tumor cells, are often associated with a favorable prognosis and robust responses to ICIs.^30,31 By contrast, the recruitment of Tregs and MDSCs, which suppress T-cell proliferation and activation, contributes to an immunosuppressive TME that promotes immune escape by tumor cells, often correlating with resistance to immunotherapy.^32–34 Understanding these interactions is critical for developing therapeutic strategies that can modulate the TME to enhance immunotherapy efficacy.

Mechanisms of immune escape

Tumors employ various mechanisms to evade immune detection and destruction, including (1) the upregulation of immune checkpoint molecules (e.g., PD-1, PD-L1), which inhibit T-cell activation and proliferation²⁷; (2) the secretion of immunosuppressive cytokines, which alter the phenotype of immune cells within the TME^35,36; and (3) the induction of immune tolerance via the recruitment of immunosuppressive cells such as Tregs and MDSCs.^37,38 Targeting these mechanisms has become a primary focus of immunotherapeutic strategies, including the development of ICIs that block the inhibitory signals of the PD-1/PD-L1 and CTLA-4 pathways, thereby restoring T-cell activity and promoting antitumor responses.

As our understanding of the TME and its role in immune evasion grows, there is a rising interest in exploring molecular and histopathological biomarkers for their potential to predict responses to immunotherapy in GI tumors.

Novel immunotherapies

Beyond immune checkpoint blockade, novel immunotherapies such as CAR T-cell therapy, PI3K inhibitors (TQ-B3525), EZH2 inhibitors, anti-CD20/79B antibodies, and dual-targeting agents (e.g., CD20/CD3 bispecific antibodies) have shown efficacy in follicular lymphoma by remodeling the tumor microenvironment. Importantly, these therapies hold emerging potential for gastrointestinal cancers.^39,40 Immunomodulators like lenalidomide further expand therapeutic options by enhancing antitumor immunity through cytokine regulation.

Future research should prioritize integrating novel agents (e.g., dual-targeting antibodies, CAR T cells) with conventional immunotherapies to overcome resistance mechanisms. In EBV-associated tumors (e.g., cholangiocarcinoma), EBV-driven PD-L1 overexpression alters the immune microenvironment, impacting checkpoint blockade. Strategies combining pembrolizumab/rituximab (effective in relapsed lymphoma) and CD79a/CD79b co-targeting in B-cell malignancies may guide gastrointestinal cancer therapy.^41–43

Molecular and histopathological biomarkers for predicting immunotherapy response in GI tumors

Multiple histopathological and molecular biomarkers have been explored to predict the efficacy of immunotherapy in GI tumors.^44–47 This section categorizes these biomarkers and emphasizes their clinical relevance and potential applications.

Microsatellite instability and mismatch repair deficiency

MSI and dMMR are critical biomarkers for immunotherapy sensitivity, particularly in colorectal cancer.^48–50 MSI-High (MSI-H) and dMMR tumors generally exhibit a high mutational burden, leading to increased neoantigen production, which enhances tumor visibility to the immune system. This makes MSI-H and dMMR tumors particularly suitable for immune checkpoint blockade (ICB) therapy.^51,52

In clinical settings, MSI-H and dMMR statuses are typically determined via polymerase chain reaction or immunohistochemistry (IHC) methods.⁵³ Numerous studies have demonstrated that MSI-H patients exhibit significantly higher response rates to PD-1 inhibitors compared to MSS patients.^54–56 Consequently, the MSI and dMMR statuses are crucial in selecting GI cancer patients for ICI therapy.

Although MSI and dMMR are strong predictive factors in certain tumor types, such as colorectal cancer, their predictive value in other gastrointestinal cancers may not be well-defined. In addition, testing heterogeneity and the need for standardized assessment also limit their broader clinical application.

TME and immune cell infiltration

The TME, which consists of cancer cells, immune cells, fibroblasts, and vascular elements, influences tumor response to therapy through its inherent heterogeneity and dynamic nature.^18,19 The infiltration pattern of various immune cells, particularly CD8+ T cells, Tregs, and MDSCs, is a key determinant of patient outcomes. High densities of CD8+ T cells are associated with favorable immunotherapy responses,⁵⁷ while increased levels of Tregs and MDSCs often indicate an immunosuppressive TME, leading to resistance to therapy.^37,38

Research has shown that TME subtypes are crucial in determining immunotherapy outcomes. For example, “hot” tumors, characterized by high infiltration of CD8+ T cells and other effector immune cells, are more likely to respond to ICIs, while “cold” tumors, lacking these effector cells, are typically resistant.^58,59 Thus, evaluating TME status, including the density of CD8+ T cells and the distribution of suppressive cells, can serve as a powerful indicator of treatment efficacy.

Although TME and immune cell infiltration are promising areas of research, their clinical application requires further standardization and validation. A more profound understanding of intratumoral and intertumoral heterogeneity is also needed, as this variability can lead to inconsistent predictions of treatment responses.

Tumor mutational burden

TMB, representing the total number of somatic mutations within a tumor genome, is a significant predictor of immunotherapy response.^60,61 Tumors with high TMB tend to produce more neoantigens, which can be recognized by T cells, leading to a stronger antitumor immune response. High TMB has been associated with improved responses to ICIs and favorable prognoses in multiple solid tumors, including GI cancers.^62,63

However, the clinical implementation of TMB as a biomarker remains challenging due to variations in TMB thresholds across different cancer types and inconsistencies in its predictive value.⁶⁴ Standardization of TMB assessment and the development of consensus guidelines are necessary to fully realize its potential in clinical practice.

Although TMB has the potential to predict immunotherapy response, its clinical application remains limited by a lack of standardization and technical challenges in assessment.

PD-1 and PD-L1 expression

PD-1 and PD-L1 expression, assessed via IHC, are studied as immunotherapy biomarkers.^65–67 High PD-L1 expression is frequently associated with favorable responses to ICIs.^68–71 However, the predictive role of PD-L1 in GI tumors remains controversial, not only due to the variability in its expression but also because of multiple factors, including TME, genetic heterogeneity, types of immunotherapy, and patient immune status. The comprehensive impact of these complex factors means that the predictive value of PD-L1 needs a more holistic assessment. Furthermore, PD-L1 expression levels are often quantified using different scoring systems (e.g., combined positive score and tumor proportion score), which increases the complexity of interpretation.^72–74

PD-1 and PD-L1 expression have become key biomarkers in the field of immunotherapy, but they still face challenges in clinical practice. For example, accurate PD-L1 assessment often requires specialized techniques and expertise. In addition, the use of different scoring systems complicates the interpretation of PD-L1 staining results.

Tumor spatial heterogeneity

Tumor spatial heterogeneity (ITH) describes the diversity within a tumor across various regions in terms of genetics, immune cell composition, microenvironmental characteristics, and drug delivery.⁷⁵ This heterogeneity influences tumor response to immunotherapy, as different areas may have varying mutational burdens, immune cell densities, and microenvironmental conditions, such as hypoxia and differences in nutrient supply (Figure 2). For example, areas with greater infiltration of CD8+ T cells may be more sensitive to ICIs, whereas hypoxic regions may be more immunosuppressive. In addition, vascular and stromal differences within the tumor can lead to uneven drug distribution, affecting treatment outcomes.^76–78 Therefore, understanding ITH is crucial for developing personalized treatment strategies that target the specific characteristics of different tumor regions, which can improve the efficacy and selectivity of immunotherapy. Radiomics technology, through the extraction of imaging features (e.g., texture and density), can quantify this heterogeneity and provide insights into underlying genetic and immune patterns.^79,80

Figure 2.

Schematic of tumor spatial heterogeneity (By Figdraw). Region A is enriched with CD8+ T cells and oxygenated, rendering it sensitive to ICIs. Region B is hypoxic and dominated by immunosuppressive cells (Tregs and MDSCs), leading to therapy resistance. Uneven vascular networks further contribute to inefficient drug delivery.

Assessing ITH requires the use of advanced technologies, such as next-generation sequencing and multiparametric imaging, and standardized methods for evaluation. Moreover, the complexity of ITH data can be challenging to interpret. Continued research and technological advancements are needed to overcome these challenges and fully realize the potential of ITH in clinical oncology.

Imaging technologies for predicting the immunotherapy response

Medical imaging is one of the key factors informing medical science. Its superiority lies in its ability to noninvasively assess the characteristics of human tissue and organs.^81–83 Thus, it is routinely used in clinical practice for oncologic diagnosis as well as predicting immunotherapy responses in GI tumors.^84–86 Computed tomography (CT) has the advantages of fast imaging speed and high spatial resolution of images (Figure 3). Magnetic resonance imaging (MRI) has a high resolution of soft tissue, to reflect tumor functional status and microenvironmental changes (Figure 4). Positron emission tomography-CT (PET-CT) employs metabolic tracers (e.g., 18F18F-FDG, 68Ga68Ga-FAPI-04) to map tumor activity and immune cell infiltration by targeting glucose metabolism or fibroblast activation protein (FAP).

Figure 3.

Male, 48-year-old, diagnosed with Gastric Mucosa-associated low-differentiated adenocarcinoma (immunohistochemistry results: pMMR, HER2 (0), PD-L1 22C3 (combined positive score < 1)). Compared to pre-treatment portal venous phase CT image (a), the range of the lesion has significantly decreased after three cycles of immunochemotherapy (b).

Figure 4.

Advanced rectal cancer with positive circumferential resection margin (CRM+), invasion of the prostate, seminal vesicles, and posterior bladder wall, bilateral lateral lymph node metastasis (a, b), and significantly restricted diffusion on DWI (c). NGS: Microsatellite instability MSI-High; Germline MSH6 mutation p.Y969C (Deficient Mismatch Repair, dMMR). Compared to the baseline high-resolution MRI of the rectum (a, b, c), there is significant regression of the rectal tumor with a small amount of fibrosis, and the lateral lymph nodes have decreased in size (d, e), with no significant restriction on DWI diffusion (f) after immunotherapy combined with radiotherapy.

Recent advances in medical image acquisition and analysis have allowed for high-quality images with isotropy to be obtained. These images contain not only general morphological information but also rich data. Radiomics has been recognized as an important imaging technology in oncology. It can convert CT images into high-throughput quantitative data, which can reflect intratumor heterogeneity and be associated with tumor molecular and immune profiles.⁸⁷ Radiogenomics, the linkage of imaging features to gene expression, further enhances precision. Such integration bridges macroscopic imaging findings with microscopic genetic alterations, enabling dynamic monitoring of tumor evolution during therapy.^88–90

Imaging characterization of the molecular and histological features

Advanced imaging techniques characterizing molecular genetics have opened new avenues for predicting the efficacy of immunotherapy in GI tumors.⁹¹ Table 1 encapsulates various studies utilizing imaging biomarkers to characterize the molecular genetics of GI tumors, underscoring their correlation with treatment response.^85,92–107 These studies highlight the potential of imaging to reflect underlying molecular-genetic traits such as the TME, MSI, TMB, and PD-L1 expression.

Table 1.

Correlation studies of imaging characterization for molecular genetics.

Author/year	Tumor type	Imaging technique	Biomarker type	Sample size	Method	Findings	Key metrics
Jiang et al.⁹² (2023)	GC	CT	TME	2686	Deep learning radiomic features	The deep learning model could accurately assess TME status.	AUC = 0.937
Sun et al.⁹³ (2024)	GC	CT	TME	2600	Radiomics	LRS and MRS are associated with IHC-derived immune context, predicting DFS and OS.	AUC(LRS) = 0.765–0.773; AUC(MRS) = 0.736–0.750
Huang et al.⁹⁴ (2022)	GC	CT	TME	2272	Radiomics	Radiomics biomarker correlates with NLR, predicting DFS and OS.	AUC = 0.795–0.861
Saber et al.⁹⁵ (2023)	CRLM	CT	CD73	160	Radiomics	Preoperative CT radiomics predicts CD73 expression in CRLM, correlating with histological CD73 levels.	AUC = 0.95
Gao et al.⁹⁶ (2020)	GC	CT	TITreg	165	Radiomics	CT radiomics can predict gastric cancer prognosis via TITreg cell assessment.	AUC = 0.884
Zhao et al.⁹⁷ (2023)	GC	CT	MSI	396	Radiomic	The radiomics clinic combined model with “Radscore” optimally predicted MSI status across cohorts.	AUC = 0.836
Li et al.⁹⁸ (2021)	RC	MRI	MSI	90	Radiomics	Radiomics models are significantly associated with MSI status.	AUC = 0.908–0.926
Zhang et al.⁹⁹ (2021)	RC	MRI	MSI	491	Deep learning	Deep learning models can accurately predict MSI status in RC patients.	AUC = 0.868
Horvat et al.¹⁰⁰ (2019)	RC	MRI	Genetic mutations	65	Radiomics	Quantitative MRI features may be linked to genetic mutations, but qualitative features are not.
Lee et al.¹⁰¹ (2022)	CC	18F-FDG PET/CT	KRAS; MSI	195	Semi-quantitative analysis	SUVmax significantly correlated with KRAS mutation and MSI status.	AUC(KRAS) = 0.64;AUC(MSI) = 0.63
Hoshino et al.¹⁰² (2022)	CRC	CT	TMB	24	Radiogenomics	Radiomics analysis can accurately predict TMB differences in primary and metastatic lesions.	AUC = 0.732–0.812
Yang et al.⁸⁵ (2023)	GC	CT	TMB	377	Radiomics	High EMVI score linked to low MSI, TMB, ICI response, and high immune escape.	AUC: 0.668 (at 1 year); 0.744 (at 2 years); 0.714 (at 3 years)
Wang et al.¹⁰³ (2022)	GC	CT	PD-L1	153	Radiomics	AGC patients with CPS ⩾ 10 had more enhanced arterial phase and larger lymph nodes.	AUC = 0.671
Xie et al.¹⁰⁴ (2023)	GC	CECT	PD-L1	217	Radiomics	The radiomics model can accurately predict high PD-L1 expression levels.	AUC = 0.806
Qiao et al.¹⁰⁵ (2023)	CC	18F-FDG PET/CT	PD-L1	72	Semi-quantitative analysis	SUVmax on 18F-FDG PET/CT correlates positively with PD-L1 expression.	AUC = 0.745
Zhao et al.¹⁰⁶ (2023)	ESCC	18F-FAPI-04 PET/CT	PD-L1	24	Semi-quantitative analysis	18F-FAPI-04 PET/CT parameters are linked to PD-L1 expression, with SUVsd being the best predictor for PD-L1 positivity.	AUC: 0.882 (SUVmean), 0.874 (SUVsd),0.840 (SUVpeak),0.765 (SUVmax)).
Cytryn et al.¹⁰⁷ (2024)	GEC	PET	PD-L1	10	Semi-quantitative analysis	88% concordance between PD-L1 PET imaging and pathologic assessment.	r_s = 0.64

18F-FDG PET/CT, 2-deoxy-2-[18F] fluoro-D-glucose positron emission tomography/CT; ADC, apparent diffusion coefficient; AGC, advanced gastric cancer; AUC, area under the receiver operating characteristic curve; CC, colon cancer; CECT, contrast-enhanced CT; CPS, combined positive score; CRC, colorectal cancer; CRLM, colorectal cancer liver metastases; CT, computed tomography; DFS, disease-free survival; EMVI, extramural venous invasion; ESCC, esophageal squamous cell carcinoma; FAP, fibroblast activation protein; FAPI, fibroblast activation protein inhibitor; FDG, fluorodeoxyglucose; GC, gastric cancer; GEC, gastroesophageal cancer; ICI, immune checkpoint inhibitor; IHC, immunohistochemistry; KRAS, Kirsten ras sarcoma viral oncogene; LRS, lymphoid radiomic score; MRI, magnetic resonance imaging; MRS, myeloid radiomic score; MSI, microsatellite instability; MTV, metabolic tumor volume; NLR, neutrophil-to-lymphocyte ratio; OS, overall survival; PD-L1, programmed death-ligand 1; PET, positron emission tomography; RC, rectal cancer; r_s, Spearman rank correlation coefficient; SUV, standardized uptake value; SUVmax, maximum standardized uptake value; SUVmean, mean standardized uptake value; SUVpeak, peak standardized uptake value; SUVsd, standardized uptake value standard deviation; TITreg, tumor-infiltrating regulatory T cells; TLF, total lesion FAP; TMB, tumor mutational burden; TME, tumor microenvironment.

Jiang et al.⁹² developed a radiomics model using CT images to assess the TME in gastric cancer patients. The model achieved an AUC of 0.937 in the training cohort and 0.909 in the validation cohort, demonstrating its potential as a noninvasive predictor of TME composition. Similarly, Sun et al.⁹³ established radiomic scores for lymphoid and myeloid contexts in gastric cancer, which correlated with IHC-derived immune contexts and were predictive of disease-free survival (DFS) and overall survival (OS). Huang et al.⁹⁴ demonstrated a correlation between a CT-based radiomic score and DFS and OS, achieving AUCs of 0.795 and 0.861, respectively. Collectively, these studies suggest that radiomics may effectively capture a tumor’s immune profile, thereby enhancing treatment prediction for GI cancers.

Recently, researchers developed a “Radscore” model combining radiomics and clinical features to predict the MSI status in gastric cancer patients, achieving an AUC of 0.836 in the training cohort and 0.834 in the validation cohort. The model may serve as a noninvasive biomarker.⁹⁷ Li Z et al. revealed that magnetic resonance imaging (MRI)-based provides greater accuracy in assessing MSI in rectal cancer than T2-weighted imaging or apparent diffusion coefficient (ADC) imaging alone, with AUCs ranging from 0.908 to 0.926.⁹⁸ In addition, Lee et al.¹⁰¹ demonstrated that 18F-FDG positron emission tomography CT (PET/CT) could assess kirsten ratsarcoma viral oncogene homolog mutations and MSI status in colorectal cancer with moderate accuracy. Collectively, these findings underscore the expanding role of advanced imaging in predicting treatment response in patients with GI tumors.

Hoshino et al. demonstrated the potential of radiomic analysis to predict TMB variations between primary and metastatic lesions in colorectal cancer, providing valuable insights into the role of radiogenomics in assessing TMB status.¹⁰² Yang et al. identified a CT-detected extramural venous invasion-related gene signature, associating a high extramural venous invasion score with reduced MSI, lower TMB, and an increased tendency for immune escape, which correlates with a poorer response to ICI treatment in gastric cancer patients.⁸⁵ These studies highlight the potential of imaging as a predictive tool for immunotherapy response, complementing molecular genetic data.

Traditionally, PD-L1 expression has been evaluated via IHC, which involves the assessment of staining intensity and distribution in tumor biopsies. Complementary molecular techniques, such as Western blotting for direct protein quantification and RT-PCR for mRNA analysis, have also been routinely employed to validate PD-L1 levels. While IHC remains the clinical gold standard for PD-L1 assessment, its interpretation is complicated by variability in scoring systems and tissue heterogeneity. Noninvasive imaging approaches, such as radiomics and molecular PET/CT, are now being explored to overcome these limitations and provide spatially resolved insights into PD-L1 expression across entire tumors. Xie et al.¹⁰⁴ utilized contrast-enhanced CT in gastric cancer patients to build a radiomics model, achieving an AUC of 0.806 for predicting high PD-L1 levels and thus aiding immunotherapy decision-making. Qiao et al.¹⁰⁵ reported a positive correlation between 18F-FDG PET/CT maximum standardized uptake value (SUVmax) and PD-L1 expression in colon cancer liver metastases. Zhao et al.¹⁰⁶ identified standardized uptake value standard deviation (SUVsd) as the best predictor for PD-L1 positivity via 18F-FAPI-04 PET/CT, with an AUC of 0.882, underscoring the potential of molecular imaging in assessing PD-L1 expression for immunotherapy selection. Although current imaging techniques cannot directly quantify PD-L1, advanced modalities such as PET/CT and MRI provide valuable insights into the mutational landscape and immune cell infiltration of tumors.

By integrating various imaging modalities (e.g., CT, MRI, and PET/CT), researchers have preliminarily achieved noninvasive prediction of key molecular features in gastrointestinal tumors, such as microsatellite instability (MSI), tumor mutational burden (TMB), and PD-L1 expression. These imaging biomarkers not only overcome the sampling bias and invasiveness associated with traditional biopsies but also enable dynamic monitoring of tumor microenvironment evolution through multi-timepoint imaging analyses.

The integration of imaging biomarkers with molecular and histopathological data can refine predictive models. Future research should focus on validating these imaging biomarkers in larger-scale, multicenter cohorts to ensure their reproducibility and external validity.

Imaging prediction for immunotherapy outcomes

The efficacy of immunotherapy in GI tumors has been a significant focus of oncological research, with ICIs at the forefront. However, variability in patient response to these therapies necessitates the development of predictive biomarkers to enhance treatment personalization. Imaging techniques have emerged as valuable tools in this domain, providing insights into treatment outcomes and guiding patient selection for immunotherapy (Table 2).^108–125

Table 2.

Correlative studies of imaging prediction for immunotherapy outcomes.

Author/year	Tumor type	Imaging technique	Treatment	Sample size	Method	Findings	Key metrics
Fucà et al.¹⁰⁸ (2021)	CRC	CT	ICIs	169	RECIST V.1.1 criteria	ETS and DoR are significantly associated with OS and PFS.	HR(OS) = 0.35 and 0.14; HR(PFS) = 0.26 and 0.13
Huang et al.¹⁰⁹ (2023)	GC	CT	ICIs	294	Radiomics	Longer irPFS in patients with low RS.	AUC: 0.787–0.810 (12-month irPFS), 0.805 (24-month irPFS)
Wang et al.¹¹⁰ (2023)	GC	CT	ICIs	249	Semi-supervised Deep Learning	The deep learning model combined with PD-L1 expression improves prediction accuracy.	AUC = 0.952
Ruan et al.¹¹¹ (2024)	ESCC	CT	NICT	192	Semi-quantitative analysis	Patients in the high-ECPI group and those without vascular signs have a higher rate of pCR.	AUC = 0.918
Guo et al.¹¹² (2024)	ESCC	CT	NICT	158	Semi-quantitative analysis, Nomogram	cT stage, cN stage, and tumor length are associated with treatment response.	AUC = 0.813
Zhu et al.¹¹³ (2021)	ESCC	CT	NICT	64	Radiomics	The 2D radiomics model showed the best performance in predicting the response to ICI + CT.	AUC = 0.843
Zhan et al.¹¹⁴ (2024)	GC	CT	Immunotherapy	457	Radiomics	Radiomic features are associated with MSI status and immunotherapy outcomes.	AUC = 0.851
Kim et al. ¹¹⁵ (2020)	MSS GC	CT	PD-1 inhibitors	149	Semi-quantitative analysis	Sarcopenia is associated with shorter PFS.	HR = 1.79
Kim et al.¹¹⁶ (2021)	GC	CT	IORT	185	Semi-quantitative analysis	Median OS is shorter in a high-risk group (SAR/hNLR).
Lin et al.¹¹⁷ (2022)	GC	CT	NCI	81	Semi-quantitative analysis	Low SMI and ΔSMI ⩾ 1.8 are independent risk factors for poor tumor regression.	OR: 3.23 (Low SMI), 1.45 (ΔSMI ⩾ 1.8), 14 (High SAI)
Wang et al.¹¹⁸ (2024)	ESCC	CECT	NIT	82	Radiomics	Radiomics combined with clinical data can accurately predict NIT response.	AUC = 0.93
Fox et al.¹¹⁹ (2023)	dMMR CRC	CT/MRI	ICIs	38	Semi-quantitative analysis	Discordance between endoscopy and imaging responses to PD-1 therapy.
Xu et al.¹²⁰ (2024)	CC	IVIM-DWI and BOLD-MRI	Anti-PD-1 + VEGFR-2 inhibitors	48	Quantitative analysis	Combination therapy enhanced tumor inhibition and correlated MRI parameters and pathological markers.
Qi et al.¹²¹ (2024)	ESCC	18F-FDG PET and CECT	nCRT and anti-PD-1 inhibitors	126	Radiomics	Machine learning models combining multimodal radiomic features and clinical features aid in predicting pCR.	AUC = 0.852
Wang et al.¹²² (2022)	ESCC	18F-FDG PET/CT	NICT	58	Semi-quantitative analysis	The combination therapy group showed a higher tumor inhibition rate.	AUC = 0.848–0.860
Rong et al.¹²³ (2021)	GC	68Ga-FAPI-04 PET/CT	ICIs	21	Semi-quantitative analysis	High 68Ga-FAPI-04 uptake, correlating with TMEscore, indicated reduced ICB therapy efficacy.	AUC = 0.733
Hartimath et al.¹²⁴ (2022)	CC	PET	αPD1 + CpG-ODN	60	Quantitative analysis	Granzyme B PET imaging can differentiate treatment responders from non-responders.
Goggi et al.¹²⁵ (2021)	CC	PET	NICT	N/A	Quantitative analysis	[18F]AlF-mNOTA-GZP uptake correlates with treatment response.	r = 0.7139

[18F]AlF-mNOTA-GZP, [¹⁸F]Aluminum Fluoride-Conjugated Modified NOTA Chelator with Targeting Ligand GZP; 2D, two-dimensional; AUC, area under the receiver operating characteristic curve; BOLD-MRI, blood oxygenation level-dependent magnetic resonance imaging; CC, colon cancer; CECT, contrast-enhanced CT; CpG-ODN, CpG-oligodeoxynucleotides; CRC, colorectal cancer; CT, computed tomography; dMMR, mismatch repair deficiency; DoR, depth of response; ECPI-Score, ESCC preoperative imaging score; EMVI, extramural venous invasion; ETS, early tumor shrinkage; FAPI, fibroblast activation protein inhibitor; FDG, fluorodeoxyglucose; GC, gastric cancer; HR, hazard ratio; ICB, immune checkpoint blockade; ICIs, immune checkpoint inhibitors; IORT, immunotherapy combined with radiotherapy; irAEs, immune-related adverse events; irPFS, immunotherapy-related progression-free survival; IVIM-DWI, intravoxel-incoherent-motion diffusion-weighted imaging; MRI, magnetic resonance imaging; MSI, microsatellite instability; MSS, microsatellite stability; N/A, not available; NCI, neoadjuvant chemotherapy combined with immunotherapy; nCRT, neoadjuvant chemoradiotherapy; NICT, neoadjuvant immunochemotherapy; NIT, neoadjuvant immunotherapy; NLR, neutrophil-to-lymphocyte ratio; OR, odds ratio; OS, overall survival; pCR, pathological complete response; PD-1, programmed death 1; PET, positron emission tomography; PFS, progression-free survival; RECIST, response evaluation criteria in solid tumors; RFS, recurrence-free survival; RS, radiomic score; SAI, subcutaneous adipose index; SAR, sarcopenia; SMI, skeletal muscle index; TME, tumor microenvironment; VAI, visceral adipose index. VEGFR-2, vascular endothelial growth factor receptor 2.

Fucà et al.¹⁰⁸ found that early tumor shrinkage and the depth of response in 169 CRC patients treated with ICIs were associated with improved survival, thereby guiding treatment strategies. Huang et al.¹⁰⁹ showed that radiomic features could predict 12-month and 24-month irPFS in patients with gastric cancer with high accuracy. Rong et al.¹²³ demonstrated that 68Ga-FAPI-04 PET/CT could noninvasively predict the efficacy of ICIs in patients with gastric cancer, facilitating pretreatment patient stratification. Conversely, Fox et al.¹¹⁹ reported that CT, MRI, and endoscopy were insufficient for predicting outcomes in dMMR CRC patients, underscoring the need for better biomarkers. These findings suggest that imaging biomarkers may optimize treatment strategies and outcomes for GI tumors treated with ICI therapy, warranting further validation in future studies.

Combining immunotherapy with chemotherapy and radiotherapy shows promise in improving cancer treatment efficacy.¹²⁶ Imaging plays a crucial role in predicting responses to these combined therapies, providing valuable information for treatment planning and patient outcomes. Ruan et al. included 192 patients with esophageal squamous cell carcinoma undergoing neoadjuvant immunochemotherapy which were analyzed to assess dynamic radiologic features for predicting pathologic complete response (pCR). Notably, higher pCR rates were observed in patients without vascular signs.¹¹¹ Kim et al. identified an association between sarcopenia, as measured by the skeletal muscle index, and shorter PFS in gastric cancer patients on PD-1 inhibitors, underscoring the skeletal muscle index`s value as a prognostic marker.¹¹⁶ Xu et al.¹²⁰ utilized intravoxel incoherent motion diffusion-weighted imaging (IVIM-DWI) and blood oxygenation level-dependent MRI (BOLD-MRI) to demonstrate increased tumor inhibition with combined vascular endothelial growth factor receptor -2 and PD-1 blockade, emphasizing MRI`s role in assessing treatment efficacy. Qi et al. built a machine-learning model to predict pCR in patients with esophageal squamous cell carcinoma by integrating radiomics from 18F-FDG PET/CT and clinical data, achieving an AUC of 0.852.¹²¹

Dynamic imaging biomarkers, such as alterations in metabolic activity or tumor morphology reflected by standardized uptake values (SUVs) from PET scans, hold the potential for predicting responses to ICIs.^122,127 Particularly, 68Ga-FAPI-04 PET imaging, which targets fibroblast-activation protein, represents a novel approach that offers insights into the tumor immune microenvironment (TIME). Elevated 68Ga-FAPI-04 uptake is associated with poorer outcomes and diminished ICI therapy efficacy, suggesting its utility as an independent predictor of treatment outcomes.¹²³ However, this technology faces challenges regarding accessibility and cost and requires further validation before broader patient implementation. In clinical settings, 68Ga-FAPI-04 PET has the potential to complement existing biomarkers, enabling personalized treatment strategies for patients most likely to benefit from ICIs.

Recent studies indicate that body composition, particularly adipose tissue distribution, significantly influences the success of immunotherapy. Imaging techniques have revealed a correlation between an individual’s adipose composition and the effectiveness of ICIs.^117,128 These findings may help predict which patients are likely to respond well to specific immunotherapies, potentially personalizing treatment plans and enhancing the effectiveness of immunotherapy. In addition, the influence of adipose tissue on the TME might uncover new targets for treatment and the role of lifestyle changes in improving immunotherapy outcomes.

Innovative imaging technologies such as immuno-PET and hyperpolarized (HP) 13C-MRI show considerable potential for noninvasively monitoring the effects of immunotherapy on tumors. Immuno-PET uses monoclonal antibodies to track specific immune receptors or cells, predicting immunotherapy responses by quantifying tracer uptake.^129,130 By contrast, HP 13C-MRI enhances the signal of 13C-labeled molecules, enabling real-time tracking of metabolic processes. This technique facilitates the study of cancer metabolism and early detection of responses to immunotherapy.^131,132

Through fusion of these multidimensional data, imaging not only provides real-time guidance for immunotherapy initiation (e.g., metabolic decline preceding morphological changes) but also dynamically optimizes combination strategies (e.g., timing anti-angiogenic agents or radiotherapy), driving a paradigm shift in imaging from “post-treatment evaluation” to an “integrated tool for prediction, monitoring, and intervention.”

Future studies should prioritize validating these imaging biomarkers in prospective trials and integrating them into clinical workflows. Furthermore, longitudinal imaging studies are needed to assess dynamic changes in tumor characteristics (e.g., metabolic activity, immune cell infiltration) and correlate these changes with adaptive immune responses (e.g., T-cell receptor diversity, cytokine profiles), thereby uncovering resistance mechanisms and informing personalized therapeutic adjustments.

Challenges and future directions

Despite the significant potential of integrating molecular genetics and advanced imaging biomarkers into personalized immunotherapy for GI tumors, challenges remain in their clinical application: (1) many studies rely on single-center data, limiting generalizability due to homogeneous patient populations and institution-specific protocols. Multi-institutional collaborations are needed to enhance reproducibility and external validity. (2) Radiomic models are sensitive to differences in imaging equipment, reconstruction algorithms, and segmentation methods. Standardized guidelines are essential to harmonize radiomic analyses. (3) Limited cohorts reduce statistical power and increase overfitting risks, particularly in machine learning models. Larger, prospective datasets are required to validate findings. (4) Intratumoral and intertumoral heterogeneity challenge both molecular and imaging biomarkers. For instance, radiomic features averaged across entire tumors may obscure regional differences. Longitudinal studies with dynamic observation are needed to capture changes in the tumor microenvironment during therapy. (5) Discrepancies in endpoint definitions and follow-up durations hinder unified interpretation of biomarker efficacy. Long-term survival data are often lacking, limiting insights into durable immunotherapy benefits.

To overcome these challenges, future research should prioritize the following: (1) multi-center collaborations to ensure diverse and representative patient cohorts. (2) Standardization of imaging protocols and radiomic feature extraction. (3) Large-scale, prospective trials to validate integrated biomarker models. (4) Dynamic monitoring of tumor biology using advanced imaging techniques such as hyperpolarized 13C-MRI, immuno-PET.

Conclusion

This review highlights the transformative potential of integrating molecular genetics and advanced imaging biomarkers into personalized immunotherapy for GI tumors. Molecular biomarkers, such as MSI, TMB, and PD-L1 expression, have become indispensable tools for patient selection and treatment optimization. Meanwhile, advanced imaging technologies, including radiomics and PET/CT, provide noninvasive tools to assess tumor biology and predict therapeutic responses. Despite the challenges, through multi-center collaborations, standardized protocols, and large-scale prospective trials, these biomarkers are expected to play a significant role in future clinical practice, ultimately improving patient outcomes.

Footnotes

Appendix

Acknowledgements

None.

Declarations

ORCID iDs

Hongfeng Gou

Bin Song

References

Siegel

Miller

Fuchs

, et al. Cancer statistics, 2022. CA Cancer J Clin 2022; 72: 7–33.

Chen

Wang

RH.

[Updates on immunotherapy of gastrointestinal cancers and practical challenges]. Zhonghua Wei Chang Wai Ke Za Zhi 2024; 27: 24–34.

Wang

Pan

, et al. Immunotherapy in gastrointestinal cancers: advances, challenges, and countermeasures. Sci Bull (Beijing) 2023; 68: 763–766.

DeCarli

Strosberg

Almhanna

Immune checkpoint inhibitors for gastrointestinal malignancies: an update. Cancers (Basel) 2022; 14: 4201.

Noori

Jafari-Raddani

Davoodi-Moghaddam

, et al. Immune checkpoint inhibitors in gastrointestinal malignancies: an umbrella review. Cancer Cell Int 2024; 24: 10.

Chong

Madeti

Cai

, et al. Recent developments in immunotherapy for gastrointestinal tract cancers. J Hematol Oncol 2024; 17: 65.

Zhou

Zhang

Immunotherapy-based combination strategies for treatment of gastrointestinal cancers: current status and future prospects. Front Med 2019; 13: 12–23.

Mariam

Kamath

Schveder

, et al. Biomarkers for response to anti-PD-1/anti-PD-L1 immune checkpoint inhibitors: a large meta-analysis. Oncology (Williston Park) 2023; 37: 210–219.

Kovacs

Fekete

Gyorffy

Predictive biomarkers of immunotherapy response with pharmacological applications in solid tumors. Acta Pharmacol Sin 2023; 44: 1879–1889.

10.

Nowak

Chetty

Predictive and prognostic biomarkers in gastrointestinal tract tumours. Pathology 2024; 56: 205–213.

11.

Kong

Lee

, et al. Network-based machine learning approach to predict immunotherapy response in cancer patients. Nat Commun 2022; 13: 3703.

12.

Qin

Huo

Liu

, et al. Biomarkers and computational models for predicting efficacy to tumor ICI immunotherapy. Front Immunol 2024; 15: 1368749.

13.

Bilal

Raza

SEA

Azam

, et al. Development and validation of a weakly supervised deep learning framework to predict the status of molecular pathways and key mutations in colorectal cancer from routine histology images: a retrospective study. Lancet Digit Health 2021; 3: e763–e772.

14.

Han

Wang

Qiao

, et al. Association of PD-L1 expression with clinicopathologic characters in gastric cancer: a comprehensive meta-analysis. Curr Med Chem 2024; 31: 3198–3216.

15.

Zhang

AFFL

Shi

, et al. Systematic review of radiomic biomarkers for predicting immune checkpoint inhibitor treatment outcomes. Methods 2021; 188: 61–72.

16.

Hosny

Schabath

, et al. Artificial intelligence in cancer imaging: clinical challenges and applications. CA Cancer J Clin 2019; 69: 127–157.

17.

Sun

, et al. Epstein-Barr virus positive gastric cancer: the pathological basis of CT findings and radiomics models prediction. Abdom Radiol (NY) 2024; 49: 1779–1791.

18.

Galli

Aguilera

Palermo

, et al. Relevance of immune cell and tumor microenvironment imaging in the new era of immunotherapy. J Exp Clin Cancer Res 2020; 39: 89.

19.

Zavros

Merchant

JL.

The immune microenvironment in gastric adenocarcinoma. Nat Rev Gastroenterol Hepatol 2022; 19: 451–467.

20.

Kang

Qiu

Luo

, et al. Application of radiomics-based multiomics combinations in the tumor microenvironment and cancer prognosis. J Transl Med 2023; 21: 598.

21.

de Visser

Joyce

. The evolving tumor microenvironment: from cancer initiation to metastatic outgrowth. Cancer Cell 2023; 41: 374–403.

22.

Kim

Hashimoto

Markosyan

, et al. Ferroptosis of tumour neutrophils causes immune suppression in cancer. Nature 2022; 612: 338–346.

23.

Vilbois

PC.

Metabolic interplay: tumor macrophages and regulatory T cells. Trends Cancer 2024; 10: 242–255.

24.

Zhang

Yang

, et al. CTHRC1 promotes liver metastasis by reshaping infiltrated macrophages through physical interactions with TGF-beta receptors in colorectal cancer. Oncogene 2021; 40: 3959–3973.

25.

Immune checkpoint signaling and cancer immunotherapy. Cell Res 2020; 30: 660–669.

26.

Topalian

Forde

Emens

, et al. Neoadjuvant immune checkpoint blockade: a window of opportunity to advance cancer immunotherapy. Cancer Cell 2023; 41: 1551–1566.

27.

Lee

Kim

Woo

, et al. Prognostic significances of PD-L1- and CTLA-4-positive T cells and positive correlations of immunosuppressive marker expression between cancer tissue and peripheral blood in patients with gastric cancer. Front Immunol 2023; 14: 1138743.

28.

Lin

, et al. Single-cell landscape reveals active cell subtypes and their interaction in the tumor microenvironment of gastric cancer. Theranostics 2022; 12: 3818–3833.

29.

Kang

Zappasodi

Modulating Treg stability to improve cancer immunotherapy. Trends Cancer 2023; 9: 911–927.

30.

Zurlo

Schino

Strippoli

, et al. Predictive value of NLR, TILs (CD4+/CD8+) and PD-L1 expression for prognosis and response to preoperative chemotherapy in gastric cancer. Cancer Immunol Immunother 2022; 71: 45–55.

31.

Giatromanolaki

Kavazis

Gkegka

, et al. Tumor-infiltrating lymphocytes, PD-L1, and MMR-deficiency combined characterization may identify subgroups of rectal cancer patients who would benefit from immunotherapy. Immunobiology 2023; 228: 152756.

32.

Belle

Lonie

Brosda

, et al. Tumour microenvironment influences response to treatment in oesophageal adenocarcinoma. Front Immunol 2023; 14: 1330635.

33.

Tan

Sahin

IH.

Defining the current role of immune checkpoint inhibitors in the treatment of mismatch repair-deficient/microsatellite stability-high colorectal cancer and shedding light on future approaches. Expert Rev Gastroenterol Hepatol 2021; 15: 735–742.

34.

Kumagai

Itahashi

Nishikawa

Regulatory T cell-mediated immunosuppression orchestrated by cancer: towards an immuno-genomic paradigm for precision medicine. Nat Rev Clin Oncol 2024; 21: 337–353.

35.

Zhang

Cao

, et al. Poor clinical outcomes and immunoevasive contexture in intratumoral IL-10-producing macrophages enriched gastric cancer patients. Ann Surg 2022; 275: e626–e635.

36.

Han

Fang

Wang

, et al. TGFbeta2 is a prognostic biomarker for gastric cancer and is associated with methylation and immunotherapy responses. Front Genet 2022; 13: 808041.

37.

Wang

Zhang

Liu

, et al. Targeting regulatory T cells in gastric cancer: pathogenesis, immunotherapy, and prognosis. Biomed Pharmacother 2023; 158: 114180.

38.

Zhou

Fang

Liu

, et al. PMN-MDSCs accumulation induced by CXCL1 promotes CD8(+) T cells exhaustion in gastric cancer. Cancer Lett 2022; 532: 215598.

39.

Watanabe

Recent advances in treatment of follicular lymphoma: efficacy of PI3Kalpha/delta inhibitor (TQ-B3525). Signal Transduct Target Ther 2024; 9: 134.

40.

Watanabe

Gene targeted and immune therapies for nodal and gastrointestinal follicular lymphomas. World J Gastroenterol 2023; 29: 6179–6197.

41.

Chiang

Hou

Tan

, et al. The immune microenvironment features and response to immunotherapy in EBV-associated lymphoepithelioma-like cholangiocarcinoma. Hepatol Int 2022; 16: 1137–1149.

42.

Nastoupil

Chin

Westin

, et al. Safety and activity of pembrolizumab in combination with rituximab in relapsed or refractory follicular lymphoma. Blood Adv 2022; 6: 1143–1151.

43.

Seon

Okazaki

Duzen

, et al. Identification of unique molecular heterogeneity of human CD79, the signaling component of the human B cell antigen receptor (BCR), and synergistic potentiation of the CD79-targeted therapy of B cell tumors by co-targeting of CD79a and CD79b. Leuk Res 2024; 136: 107436.

44.

Rocken

Predictive biomarkers in gastric cancer. J Cancer Res Clin Oncol 2023; 149: 467–481.

45.

Petrillo

Smyth

EC.

Biomarkers for precision treatment in gastric cancer. Visc Med 2020; 36: 364–372.

46.

Sun

Chen

, et al. Dawning precision treatment for gastric cancer: the latest biomarkers. J Transl Int Med 2021; 9: 228–230.

47.

Liu

Predictive molecular markers for the treatment with immune checkpoint inhibitors in colorectal cancer. J Clin Lab Anal 2022; 36: e24141.

48.

Shimozaki

Nakayama

Hirota

, et al. Current strategy to treat immunogenic gastrointestinal cancers: perspectives for a new era. Cells 2023; 12: 1049.

49.

Parente

Grillo

Vanoli

, et al. The day-to-day practice of MMR and MSI assessment in colorectal adenocarcinoma: what we know and what we still need to explore. Dig Dis 2023; 41: 746–756.

50.

Lin

Quintanilha

JCF

Danziger

, et al. Pan-tumor validation of a NGS fraction-based MSI analysis as a predictor of response to Pembrolizumab. NPJ Precis Oncol 2024; 8: 204.

51.

Golshani

Zhang

Advances in immunotherapy for colorectal cancer: a review. Therap Adv Gastroenterol 2020; 13: 1756284820917527.

52.

Gallois

Landi

Taieb

, et al. Transcriptomic signatures of MSI-high metastatic colorectal cancer predict efficacy of immune checkpoint inhibitors. Clin Cancer Res 2023; 29: 3771–3778.

53.

Shimozaki

Hayashi

Tanishima

, et al. Concordance analysis of microsatellite instability status between polymerase chain reaction based testing and next generation sequencing for solid tumors. Sci Rep 2021; 11: 20003.

54.

Cui

The mechanisms leading to distinct responses to PD-1/PD-L1 blockades in colorectal cancers with different MSI statuses. Front Oncol 2021; 11: 573547.

55.

Xue

, et al. Efficacy and safety of anti-PD-1/PD-L1 therapy in the treatment of advanced colorectal cancer: a meta-analysis. BMC Gastroenterol 2022; 22: 431.

56.

Shek

Akhuba

Carlino

, et al. Immune-checkpoint inhibitors for metastatic colorectal cancer: a systematic review of clinical outcomes. Cancers (Basel) 2021; 13: 4345.

57.

Tong

Zhu

Chen

, et al. Intratumoral CD8(+) T cells as a potential positive predictor of chemoimmunotherapy response in PD-L1-negative advanced gastric cancer patients: a retrospective cohort study. J Gastrointest Oncol 2022; 13: 1668–1678.

58.

Wei

Wang

, et al. Tumor-associated macrophage clusters linked to immunotherapy in a pan-cancer census. NPJ Precis Oncol 2024; 8: 176.

59.

Ouyang

Wang

, et al. Overcoming cold tumors: a combination strategy of immune checkpoint inhibitors. Front Immunol 2024; 15: 1344272.

60.

Cao

Yang

Chen

, et al. The predictive efficacy of tumor mutation burden in immunotherapy across multiple cancer types: a meta-analysis and bioinformatics analysis. Transl Oncol 2022; 20: 101375.

61.

Huang

Chen

Zhang

, et al. Prognostic role of tumor mutational burden in cancer patients treated with immune checkpoint inhibitors: a systematic review and meta-analysis. Front Oncol 2021; 11: 706652.

62.

Marques

Cavaco

Torre

, et al. Tumor mutational burden in colorectal cancer: Implications for treatment. Crit Rev Oncol Hematol 2024; 197: 104342.

63.

, et al. Tumor mutational burden predicting the efficacy of immune checkpoint inhibitors in colorectal cancer: a systematic review and meta-analysis. Front Immunol 2021; 12: 751407.

64.

Marabelle

Fakih

Lopez

, et al. Association of tumour mutational burden with outcomes in patients with advanced solid tumours treated with pembrolizumab: prospective biomarker analysis of the multicohort, open-label, phase 2 KEYNOTE-158 study. Lancet Oncol 2020; 21: 1353–1365.

65.

Placke

Kimmig

Griewank

, et al. Correlation of tumor PD-L1 expression in different tissue types and outcome of PD-1-based immunotherapy in metastatic melanoma - analysis of the DeCOG prospective multicenter cohort study ADOREG/TRIM. EBioMedicine 2023; 96: 104774.

66.

Paderno

Petrelli

Lorini

, et al. The predictive role of PD-L1 in head and neck cancer: a systematic review and meta-analysis. Oral Oncol 2024; 153: 106799.

67.

Niu

, et al. Predictive biomarkers of anti-PD-1/PD-L1 therapy in NSCLC. Exp Hematol Oncol 2021; 10: 18.

68.

Noori

Yousefi

Zali

, et al. Predictive value of PD-L1 expression in response to immune checkpoint inhibitors for esophageal cancer treatment: a systematic review and meta-analysis. Front Oncol 2022; 12: 1021859.

69.

Sangani

Yazdani

Khalili-Tanha

, et al. The therapeutic impact of programmed death - 1 in the treatment of colorectal cancer. Pathol Res Pract 2024; 259: 155345.

70.

Noori

Fayyaz

Zali

, et al. Predictive value of PD-L1 expression in response to immune checkpoint inhibitors for gastric cancer treatment: a systematic review and meta-analysis. Expert Rev Anticancer Ther 2023; 23: 1029–1039.

71.

Szeles

Fazekas

Vancsa

, et al. Pre-treatment soluble PD-L1 as a predictor of overall survival for immune checkpoint inhibitor therapy: a systematic review and meta-analysis. Cancer Immunol Immunother 2023; 72: 1061–1073.

72.

Huo

Liu

Chen

Efficacy of PD-1/PD-L1 inhibitors in gastric or gastro-oesophageal junction cancer based on clinical characteristics: a meta-analysis. BMC Cancer 2023; 23: 143.

73.

Ruschoff

Kumar

Badve

, et al. Scoring PD-L1 expression in urothelial carcinoma: an international multi-institutional study on comparison of manual and artificial intelligence measurement model (AIM-PD-L1) pathology assessments. Virchows Arch 2024; 484: 597–608.

74.

Nomoto

Baba

Okadome

, et al. Prognostic impact of PD-1 on tumor-infiltrating lymphocytes in 433 resected esophageal cancers. Ann Thorac Surg 2022; 113: 286–294.

75.

Xiao

You

, et al. Radiogenomic-based multiomic analysis reveals imaging intratumor heterogeneity phenotypes and therapeutic targets. Sci Adv 2023; 9: eadf0837.

76.

Marusyk

Janiszewska

Polyak

Intratumor heterogeneity: the rosetta stone of therapy resistance. Cancer Cell 2020; 37: 471–484.

77.

Zhou

Peterson

Serritella

, et al. Spatial and temporal heterogeneity of PD-L1 expression and tumor mutational burden in gastroesophageal adenocarcinoma at baseline diagnosis and after chemotherapy. Clin Cancer Res 2020; 26: 6453–6463.

78.

Smyth

Gambardella

Cervantes

, et al. Checkpoint inhibitors for gastroesophageal cancers: dissecting heterogeneity to better understand their role in first-line and adjuvant therapy. Ann Oncol 2021; 32: 590–599.

79.

Prior

Macarro

Navarro

, et al. Identification of precise 3D CT radiomics for habitat computation by machine learning in cancer. Radiol Artif Intell 2024; 6: e230118.

80.

Sun

Sundahl

Hecht

, et al. Radiomics to predict outcomes and abscopal response of patients with cancer treated with immunotherapy combined with radiotherapy using a validated signature of CD8 cells. J Immunother Cancer 2020; 8: e001429.

81.

Nakata

Isohashi

Oka

, et al. Imaging Assessment of Tumor Response in the Era of Immunotherapy. Diagnostics (Basel) 2021; 11: 1041.

82.

Jin

, et al. Noninvasive imaging in cancer immunotherapy: the way to precision medicine. Cancer Lett 2019; 466: 13–22.

83.

Wang

Wahid

van Dijk

, et al. Radiomic biomarkers of tumor immune biology and immunotherapy response. Clin Transl Radiat Oncol 2021; 28: 97–115.

84.

Fernandes

Gollub

Brown

The importance of MRI for rectal cancer evaluation. Surg Oncol 2022; 43: 101739.

85.

Yang

Gou

Feng

, et al. Computed tomography-detected extramural venous invasion-related gene signature: a potential negative biomarker of immune checkpoint inhibitor treatment in patients with gastric cancer. J Transl Med 2023; 21: 4.

86.

Schwenck

Sonanini

Cotton

, et al. Advances in PET imaging of cancer. Nat Rev Cancer 2023; 23: 474–490.

87.

Mao

Zhou

Zhao

, et al. Role of radiomics in the diagnosis and treatment of gastrointestinal cancer. World J Gastroenterol 2022; 28: 6002–6016.

88.

Wesdorp

Hellingman

Jansma

, et al. Advanced analytics and artificial intelligence in gastrointestinal cancer: a systematic review of radiomics predicting response to treatment. Eur J Nucl Med Mol Imaging 2021; 48: 1785–1794.

89.

Cui

Peng

, et al. Use of radiomics to predict response to immunotherapy of malignant tumors of the digestive system. Med Sci Monit 2020; 26: e924671.

90.

Chen

Sun

, et al. Predicting gastric cancer response to anti-HER2 therapy or anti-HER2 combined immunotherapy based on multi-modal data. Signal Transduct Target Ther 2024; 9: 222.

91.

Wahl

RL.

The interaction of genomics, molecular imaging, and therapy in gastrointestinal tumors. Semin Nucl Med 2020; 50: 471–483.

92.

Jiang

Zhou

Sun

, et al. Non-invasive tumor microenvironment evaluation and treatment response prediction in gastric cancer using deep learning radiomics. Cell Rep Med 2023; 4: 101146.

93.

Sun

Zhang

Ahmad

, et al. Comprehensive assessment of immune context and immunotherapy response via noninvasive imaging in gastric cancer. J Clin Invest 2024; 134: e175834.

94.

Huang

Jiang

Xiong

, et al. Noninvasive imaging of the tumor immune microenvironment correlates with response to immunotherapy in gastric cancer. Nat Commun 2022; 13: 5095.

95.

Saber

Henault

Messaoudi

, et al. Radiomics using computed tomography to predict CD73 expression and prognosis of colorectal cancer liver metastases. J Transl Med 2023; 21: 507.

96.

Gao

Bai

, et al. A CT-based radiomics signature for evaluating tumor infiltrating Treg cells and outcome prediction of gastric cancer. Ann Transl Med 2020; 8: 469.

97.

Zhao

Gao

Bai

, et al. Development and external validation of a non-invasive imaging biomarker to estimate the microsatellite instability status of gastric cancer and its prognostic value: the combination of clinical and quantitative CT-imaging features. Eur J Radiol 2023; 162: 110719.

98.

Dai

Liu

, et al. Radiomics analysis of multi-sequence MR images for predicting microsatellite instability status preoperatively in rectal cancer. Front Oncol 2021; 11: 697497.

99.

Zhang

Yin

Huang

, et al. Development and validation of MRI-based deep learning models for prediction of microsatellite instability in rectal cancer. Cancer Med 2021; 10: 4164–4173.

100.

Horvat

Veeraraghavan

Pelossof

, et al. Radiogenomics of rectal adenocarcinoma in the era of precision medicine: a pilot study of associations between qualitative and quantitative MRI imaging features and genetic mutations. Eur J Radiol 2019; 113: 174–181.

101.

Lee

Choi

Park

JS.

Correlations among KRAS mutation, microsatellite instability, and 18F-FDG uptake in colon cancer. Asian Pac J Cancer Prev 2022; 23: 3501–3506.

102.

Hoshino

Yokota

Iwatate

, et al. Prediction of the differences in tumor mutation burden between primary and metastatic lesions by radiogenomics. Cancer Sci 2022; 113: 229–239.

103.

Wang

, et al. Correlation between imaging features on computed tomography and combined positive score of PD-L1 expression in patients with gastric cancer. Chin J Cancer Res 2022; 34: 510–518.

104.

Xie

Jiang

Zhou

, et al. Quantitative radiological features and deep learning for the non-invasive evaluation of programmed death ligand 1 expression levels in gastric cancer patients: a digital biopsy study. Acad Radiol 2023; 30: 1317–1328.

105.

Qiao

, et al. Relationship between SUVmax on 18F-FDG PET and PD-L1 expression in liver metastasis lesions after colon radical operation. BMC Cancer 2023; 23: 535.

106.

Zhao

Ren

(18)F-FAPI-04 PET/CT parameters predict PD-L1 expression in esophageal squamous cell carcinoma. Front Immunol 2023; 14: 1266843.

107.

Cytryn

Pandit-Taskar

Lumish

, et al. (18)F-BMS-986229 PET to assess programmed-death ligand 1 status in gastroesophageal cancer. J Nucl Med 2024; 65: 722–727.

108.

Fuca

Corti

Ambrosini

, et al. Prognostic impact of early tumor shrinkage and depth of response in patients with microsatellite instability-high metastatic colorectal cancer receiving immune checkpoint inhibitors. J Immunother Cancer 2021; 9: e002501.

109.

Huang

Xiong

Tang

, et al. Non-invasive CT imaging biomarker to predict immunotherapy response in gastric cancer: a multicenter study. J Immunother Cancer 2023; 11: e007807.

110.

Wang

Jiang

Chen

, et al. Cancer immunotherapy response prediction from multi-modal clinical and image data using semi-supervised deep learning. Radiother Oncol 2023; 186: 109793.

111.

Ruan

, et al. Dynamic radiological features predict pathological response after neoadjuvant immunochemotherapy in esophageal squamous cell carcinoma. J Transl Med 2024; 22: 471.

112.

Guo

Zhou

Gao

, et al. A computed tomography-based nomogram for neoadjuvant chemotherapy plus immunotherapy response prediction in patients with advanced esophageal squamous cell carcinoma. Front Oncol 2024; 14: 1358947.

113.

Zhu

Yao

, et al. Predicting response to immunotherapy plus chemotherapy in patients with esophageal squamous cell carcinoma using non-invasive radiomic biomarkers. BMC Cancer 2021; 21: 1167.

114.

Zhan

Yang

Liu

, et al. A radiomics signature derived from CT imaging to predict MSI status and immunotherapy outcomes in gastric cancer: a multi-cohort study. BMC Cancer 2024; 24: 404.

115.

Kim

Lee

Jeong

, et al. Prognostic significance of sarcopenia in microsatellite-stable gastric cancer patients treated with programmed death-1 inhibitors. Gastric Cancer 2021; 24: 457–466.

116.

Kim

Lim

, et al. Prognostic impact of sarcopenia and radiotherapy in patients with advanced gastric cancer treated with anti-PD-1 antibody. Front Immunol 2021; 12: 701668.

117.

Lin

Huang

Lin

, et al. Body composition parameters for predicting the efficacy of neoadjuvant chemotherapy with immunotherapy for gastric cancer. Front Immunol 2022; 13: 1061044.

118.

Wang

Tang

Zhong

, et al. A machine learning radiomics based on enhanced computed tomography to predict neoadjuvant immunotherapy for resectable esophageal squamous cell carcinoma. Front Immunol 2024; 15: 1405146.

119.

Fox

Bhamidipati

Konishi

, et al. Endoscopic and imaging outcomes of PD-1 therapy in localised dMMR colorectal cancer. Eur J Cancer 2023; 194: 113356.

120.

, et al. Magnetic resonance imaging-based approaches for detecting the efficacy of combining therapy following VEGFR-2 and PD-1 blockade in a colon cancer model. J Transl Med 2024; 22: 198.

121.

Xiao

, et al. A machine learning approach using (18)F-FDG PET and enhanced CT scan-based radiomics combined with clinical model to predict pathological complete response in ESCC patients after neoadjuvant chemoradiotherapy and anti-PD-1 inhibitors. Front Immunol 2024; 15: 1351750.

122.

Wang

Yang

Zhou

, et al. The role of (18)F-FDG PET/CT in predicting the pathological response to neoadjuvant PD-1 blockade in combination with chemotherapy for resectable esophageal squamous cell carcinoma. Eur J Nucl Med Mol Imaging 2022; 49: 4241–4251.

123.

Rong

Liu

, et al. PET/CT imaging of activated cancer-associated fibroblasts predict response to PD-1 blockade in gastric cancer patients. Front Oncol 2021; 11: 802257.

124.

Hartimath

Ramasamy

Xuan

, et al. Granzyme B PET imaging in response to in situ vaccine therapy combined with alphaPD1 in a murine colon cancer model. Pharmaceutics 2022; 14: 150.

125.

Goggi

Hartimath

Xuan

, et al. Granzyme B PET imaging of combined chemotherapy and immune checkpoint inhibitor therapy in colon cancer. Mol Imaging Biol 2021; 23: 714–723.

126.

Zhu

Liu

Zhu

, et al. Immune checkpoint inhibitors plus chemotherapy for HER2-negative advanced gastric/gastroesophageal junction cancer: a cost-effectiveness analysis. Therap Adv Gastroenterol 2023; 16: 17562848231207200.

127.

Wang

, et al. (18) F-FDG PET/CT predicts the role of neoadjuvant immunochemotherapy in the pathological response of esophageal squamous cell carcinoma. Thorac Cancer 2023; 14: 2338–2349.

128.

Pei

Xie

Liu

, et al. Imaging-based adipose biomarkers for predicting clinical outcomes of cancer patients treated with immune checkpoint inhibitors: a systematic review. Front Oncol 2023; 13: 1198723.

129.

Liberini

Laudicella

Capozza

, et al. The future of cancer diagnosis, treatment and surveillance: a systemic review on immunotherapy and immuno-PET radiotracers. Molecules 2021; 26: 2201.

130.

Glazer

Kummar

Mittra

Illuminating immunotherapy response via precision T cell-targeted PET imaging. Front Med (Lausanne) 2024; 11: 1233913.

131.

Hoffmann

Masthoff

Kunz

, et al. Multiparametric MRI for characterization of the tumour microenvironment. Nat Rev Clin Oncol 2024; 21: 428–448.

132.

Saida

Brender

Yamamoto

, et al. Multimodal molecular imaging detects early responses to immune checkpoint blockade. Cancer Res 2021; 81: 3693–3705.