Tumor microenvironment phenotyping guides precision therapy in unresectable pancreatic cancer

Abstract

Treatment of locally advanced unresectable pancreatic cancer remains a major clinical challenge due to pronounced heterogeneity and resistance to standard regimens. Increasing evidence highlights the critical role of the tumor microenvironment (TME) in shaping therapeutic response and driving drug resistance. In this minireview, we summarize recent advances in TME phenotyping and its potential to guide precision therapy. A four-dimensional framework integrating stromal, immune, genomic, and metabolic features has been proposed to better characterize TME heterogeneity. Preclinical and clinical studies indicate that strategies targeting the stroma, modulating immunity, or exploiting genomic vulnerabilities such as homologous recombination deficiency may enhance the efficacy of chemotherapy, immunotherapy, and targeted agents. Dynamic biomarkers, including circulating tumor DNA and carbohydrate antigen 19-9, also show promise for real-time therapy adaptation, although their clinical application remains limited. By synthesizing current evidence, we emphasize the importance of individualized treatment strategies that account for TME complexity. While encouraging, the translation of multiomics phenotyping and biomarker monitoring into routine clinical practice requires standardization, prospective validation, and integration of novel technologies. Future research should focus on establishing reproducible TME-guided models to enable dynamic and personalized therapy for patients with unresectable pancreatic cancer.

Key Words: Locally advanced unresectable pancreatic cancer; Gemcitabine; Nano-albumin paclitaxel; Tumor microenvironmental resistance; Combination therapy

Core Tip: This review highlights the importance of tumor microenvironment (TME) heterogeneity in treatment resistance for unresectable locally advanced pancreatic cancer. We propose a novel four-dimensional TME phenotyping framework integrating fibrosis, immunity, genomics, and metabolism to guide precision therapy. This approach enables dynamic, individualized treatment strategies and offers new prospects for improving patient outcomes.

INTRODUCTION

Pancreatic cancer is one of the deadliest cancers worldwide, and its incidence and mortality rates have been steadily increasing over the recent decades, particularly in Western countries. According to the latest global cancer statistics (GLOBOCAN 2022)[1], pancreatic cancer ranks 12^th in incidence and sixth in cancer-related mortality, with an estimated 510566 new cases and 467005 deaths globally in 2022. The 5-year survival rate remains dismal at approximately 10%, and more than 80% of patients are diagnosed at a locally advanced or metastatic stage, severely limiting treatment options[2]. Locally advanced unresectable pancreatic cancer (LAPC) refers to a condition where the tumor is confined to the pancreas and nearby tissues but cannot be surgically removed due to its invasion of surrounding blood vessels or organs[3,4]. These patients are not only challenged by the biological characteristics of the tumor (such as high heterogeneity and aggressiveness), but also often accompanied by drug resistance and poor general health[5-8].

Among the existing treatment options, chemotherapy is the cornerstone for LAPC[5,9]. As a standard chemotherapeutic drug for pancreatic cancer, gemcitabine plays an antitumor role by inhibiting DNA synthesis and repair, but its efficacy as a single agent is limited, with a median survival of only 6-8 months and widespread drug resistance[10-12]. In recent years, nano-albumin paclitaxel (nab-paclitaxel) has received a lot of attention due to its unique drug delivery system and targetability[13]. Phase 3 clinical trials [such as the metastatic pancreatic adenocarcinoma clinical trial (MPACT) study] have shown that gemcitabine combined with nab-paclitaxel significantly extended median survival to 8.7 months and an objective response rate (ORR) of 23%[8,14,15].

However, the long-term efficacy of combination therapy still faces significant challenges[16,17]. The dense fibrotic and immunosuppressive properties of the tumor microenvironment (TME) limit drug penetration[18,19], while resistance mechanisms (such as drug effector pump overexpression, metabolic reprogramming) further impair therapeutic efficacy[20-22]. In addition, adverse effects such as myelosuppression need to be managed with dose adjustment or supportive therapy[23,24].

This review synthesizes preclinical and clinical evidence to explore the synergistic mechanisms, therapeutic benefits, and challenges of gemcitabine combined with nab-paclitaxel for LAPC. To address the limitations of standard therapies and the complexity of the TME, we propose a multidimensional approach that integrates stromal, immune, genomic, and metabolic factors. Figure 1 illustrates the interplay between tumor cells, the TME, and corresponding therapeutic strategies, and highlights future directions in TME regulation and personalized treatment to optimize clinical outcomes.

Figure 1 Pancreatic cancer tumor microenvironment and multidimensional therapeutic strategy. AG: Gemcitabine plus nab-paclitaxel; Treg: Regulatory T cell; PEGPH20: Pegylated recombinant human hyaluronidase PH20; PD-1: Programmed cell death protein 1; myCAF: Myofibroblastic cancer-associated fibroblast; PDAC: Pancreatic ductal adenocarcinoma.

METHODOLOGY

This narrative review systematically synthesizes existing evidence on the TME phenotyping in LAPC, proposing a novel four-dimensional stratification framework to guide precision therapy. Databases including PubMed and Web of Science were searched for studies published from 2010 to 2024. Inclusion criteria were: Studies focusing on pancreatic cancer, TME characterization, treatment response biomarkers, and clinical trials involving gemcitabine and nab-paclitaxel combinations. Exclusion criteria were: Non-English literature, case reports, and studies lacking clear TME-related therapeutic relevance. The study selection process involved initial screening of titles and abstracts for relevance, followed by full-text review of potentially eligible articles. Two independent reviewers were involved in the selection process, with discrepancies resolved by consensus. Quality assessment of included studies was performed using adapted criteria based on study design, sample size, and methodological rigor. The proposed TME phenotyping methodology integrates the following technologies and criteria for subtype classification.

Fibrosis assessment by second-harmonic generation microscopy

Second-harmonic generation (SHG) microscopy quantitatively measures collagen density in tumor biopsies. Tumors are classified as “high fibrosis” when collagen covers > 35% (SHG score ≥ 4) of the analyzed area, based on prior validated thresholds correlating with significantly reduced intratumoral drug penetration observed in preclinical pancreatic cancer xenograft models.

Immune profiling by multiplex immunohistochemistry

Multiplex immunohistochemistry characterizes immune spatial distribution by staining cluster of differentiation (CD) 8⁺ T cells, programmed cell death protein 1 (PD-1)/programmed cell death ligand 1 (PD-L1) markers, and other relevant immune checkpoints. Classification of immune subtypes is based on CD8⁺ T-cell density, with a core-to-invasive front density ratio of < 0.5, indicating immune exclusion and predicting resistance to immunotherapy, as validated by receiver operating characteristic (ROC) analyses [area under curve (AUC) = 0.81].

Genomic stratification by liquid biopsy-based circulating tumor DNA profiling

Plasma-derived circulating tumor DNA (ctDNA) profiling identifies key genomic alterations such as KRAS mutations (e.g., G12D, G12V), validated by comparative genomic analyses in tumor biopsies. KRAS wild-type tumors demonstrate superior responsiveness to nab-paclitaxel-based regimens, supporting its clinical utility as a genomic stratification marker.

Metabolic characterization by positron emission tomography/computed tomography

Metabolic profiling utilizes 18F-fluoromisonidazole or 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography/computed tomography (PET/CT) imaging. Tumors exhibiting high standardized uptake values (SUV) (SUV_max > 2.5) are indicative of metabolic reprogramming involving hypoxia and active indoleamine 2,3-dioxygenase 1 (IDO1) pathways, validated by correlation with Regulatory T (Treg) cell enrichment and treatment resistance data from recent clinical studies.

Reproducibility of this TME phenotyping framework was confirmed using the data from multiple clinical trials (e.g., TARGET-LAPC and JCOG1611-GENERATE), and cross-validated with retrospective and prospective cohorts. Preclinical validation was conducted using patient-derived xenograft (PDX) models, demonstrating concordance in predicting drug sensitivity and resistance patterns across TME phenotypes.

BIOLOGICAL CHARACTERISTICS OF PANCREATIC CANCER

Molecular mechanism of pancreatic cancer

The immune escape mechanisms of pancreatic cancer are closely related to its highly immunosuppressive microenvironment. Treg cells and tumor-associated macrophages (TAMs) are the dominant immunosuppressive cells in the microenvironment[25]. They suppress the function of effector T cells and natural killer (NK) cells by secreting factors such as transforming growth factor (TGF)-β and interleukin (IL)-10, thereby promoting tumor immune escape. Treg cells are enriched in tumors through recruitment by the chemokine CC ligand (CCL) 22, further weakening antitumor immune responses[26,27]. Treg cells are enriched in tumors by recruitment of the chemokine CCL22, further weakening the antitumor immune response[28]. TAMs are predominantly polarized to the M2 phenotype and secrete proangiogenic factors (e.g., vascular endothelial growth factor) and inhibitory cytokines, which not only promote tumor invasion but also collaborate with Treg cells to form an immunosuppressive network[29]. Additionally, cancer-associated fibroblasts (CAFs) secrete TGF-β and growth factors (e.g., fibroblast growth factor), suppressing immune cell activity and facilitating tumor metastasis[30]. The high expression of immune checkpoint molecules (e.g., PD-1/PD-L1) exacerbates immune escape, as tumor cells inhibit T-cell activation through this pathway, leading to low response rates to immunotherapy[31].

Chemoresistance is a major challenge in treatment with gemcitabine combined with nab-paclitaxel. The mechanisms involve overexpression of drug efflux pumps [e.g., P-glycoprotein (P-gp)], metabolic reprogramming, and activation of DNA repair pathways[32]. Pancreatic cancer cells enhance survival by upregulating glycolysis and fatty acid synthesis while relying on nucleotide excision repair enzymes to repair gemcitabine-induced DNA damage[33]. TAMs and Treg cells in the microenvironment indirectly promote drug resistance by suppressing drug penetration or local immune responses, suggesting that targeting the immunosuppressive microenvironment may improve chemosensitivity[34].

Microenvironment of pancreatic cancer and its impact on treatment

The TME of pancreatic cancer is composed of dense extracellular matrix (ECM), CAFs, and immunosuppressive cells [e.g., M2-polarized TAMs, and myeloid-derived suppressor cells (MDSCs)], which significantly limit the efficacy of chemotherapy and immunotherapy. TAMs suppress T-cell function and promote angiogenesis via IL-10 and TGF-β, while MDSCs maintain an immunosuppressive state by inhibiting T cell and NK cell activity[35]. To remodel the TME, immune checkpoint inhibitors (e.g., PD-1/PD-L1 inhibitors) have been combined with gemcitabine and nab-paclitaxel, with clinical trials showing partial restoration of T-cell activity and prolonged survival[36]. Antifibrotic agents [e.g., pegylated recombinant human hyaluronidase PH20 (PEGPH20)] degrade hyaluronic acid in the ECM to enhance chemotherapy penetration, demonstrating significant efficacy in pancreatic cancer patients with high hyaluronic acid expression[37]. CAF-targeted strategies (e.g., chemokine CXC receptor 4 inhibitors blocking chemokine CXC ligand 12 signaling) can suppress tumor cell migration and improve drug delivery efficiency[38]. However, combination therapies still face challenges such as drug resistance, patient heterogeneity, and insufficient long-term efficacy.

GEMCITABINE: MECHANISMS AND CLINICAL PROGRESS IN LAPC

Mechanism of action and clinical application of gemcitabine

Gemcitabine is a chemotherapeutic drug widely used in pancreatic cancer and other malignancies. Its primary mechanism of action involves inhibiting DNA synthesis. As a nucleoside analog, gemcitabine is taken up by cells and converted into its active forms (gemcitabine diphosphate and triphosphate), which interfere with DNA synthesis and repair, leading to cell cycle arrest and apoptosis[39]. Gemcitabine suppresses tumor cell proliferation and exhibits potent antitumor activity in pancreatic cancer, particularly when combined with other chemotherapeutic agents such as nab-paclitaxel, significantly improving therapeutic outcomes[10,11].

However, antitumor efficacy of gemcitabine is profoundly influenced by the TME. Key TME components, such as CAFs and hypoxic conditions, regulate gemcitabine metabolism, distribution, and efficacy through multiple mechanisms. CAFs secrete ECM proteins (e.g., collagen and fibronectin) and growth factors (e.g., TGF-β and IL-6), forming dense physical barriers and inhibitory signaling networks that hinder gemcitabine penetration into the tumor core[40]. Additionally, CAFs reduce intracellular gemcitabine uptake by downregulating nucleoside transporters (e.g., equilibrative nucleoside transporter 1), thereby diminishing its antitumor effects[39]. Hypoxia within the TME induces hypoxia-inducible factor-1α expression, upregulating deoxycytidine kinase while activating nucleotide excision repair pathways (e.g., ERCC1), which accelerates DNA damage repair and counteracts cytotoxicity of gemcitabine. Furthermore, immunosuppressive cells in the TME, such as TAMs, secrete IL-10 and arginase-1 to suppress gemcitabine-induced apoptosis and promote immune escape[41]. Preclinical studies indicate that targeting TAMs [e.g., using colony-stimulating factor (CSF)-1R inhibitors] can restore gemcitabine sensitivity and enhance its synergy with immune checkpoint inhibitors[42].

Clinical research outcomes of gemcitabine in LAPC

In the treatment of LAPC, the clinical application of gemcitabine has achieved significant progress. Combination chemotherapy regimens not only improve short-term efficacy but also provide potential long-term survival for some patients. Multiple studies demonstrate that gemcitabine combined with nab-paclitaxel significantly prolong median overall survival (OS). For example, a phase 3 clinical trial (No. NCT01827553) reported a median OS of 18.75 months in the combination therapy group compared to 11 months in the gemcitabine monotherapy group, with approximately 30% of patients achieving tumor reduction to resectable status[43]. Additionally, exploratory studies of the FOLFIRINOX regimen (fluorouracil, leucovorin, irinotecan, and oxaliplatin) combined with gemcitabine reported a 2-year survival rate of 35%-40%, highlighting the potential advantages of combination therapy in long-term prognosis[44].

The role of biomarkers in efficacy prediction is increasingly prominent. Prospective studies indicated that patients with ≥ 50% reduction in carbohydrate antigen (CA) 19-9 Levels after gemcitabine treatment achieve a median OS of 21 months, compared to 9 months in patients without significant CA19-9 changes, underscoring its value as a dynamic prognostic marker[45]. Genomic analyses further reveal that KRAS wild-type patients exhibit significantly higher response rates to gemcitabine-based combinations than KRAS-mutated patients (58% vs 32%), providing molecular insights for personalized therapy[20]. Although gemcitabine is generally well-tolerated, cumulative toxicity during long-term treatment remains a concern. Grade 3 or higher myelosuppression (e.g., neutropenia) occurs in 40%-50% of patients but can be managed through dose adjustments or granulocyte CSF (G-CSF) support[46]. Notably, long-term follow-up data (> 5 years) show that 8%-10% of patients receiving gemcitabine-based combinations achieve 5-year progression-free survival (PFS), suggesting potential curative potential[47,48].

ADVANTAGES AND MECHANISM OF NAB-PACLITAXEL

Fundamental principles of nanodrug delivery systems

Nanodrug delivery systems (NDDSs) significantly improve tumor selectivity of chemotherapeutic agents through the enhanced permeability and retention (EPR) effect and receptor-mediated targeting. The EPR effect relies on the hyperpermeability of tumor vasculature and defective lymphatic drainage, enabling nanoparticles (10 nm-200 nm) to accumulate and persist in dense stromal tumors such as pancreatic cancer[49]. nab-paclitaxel further leverages human serum albumin as a carrier, binding to secreted protein acidic and rich in cysteine (SPARC) protein overexpressed on tumor cells or gp60 receptors on vascular endothelial cells, thereby activating clathrin-mediated endocytosis for efficient intracellular drug delivery[5,40]. Its physicochemical properties (e.g., particle size < 150 nm, negative surface charge) optimize vascular barrier penetration while minimizing nonspecific binding to healthy tissues[19,50]. The albumin carrier encapsulates paclitaxel via hydrophobic interactions, enabling controlled release in the TME (low potential of hydrogen or high protease activity) to prolong drug activity[20]. Compared to traditional paclitaxel formulations using Cremophor EL solvent (associated with 20%-40% hypersensitivity risk), nab-paclitaxel eliminates organic solvents, reduces hypersensitivity incidence to < 1%, and allows higher dosing (260 mg/m²), significantly enhancing safety[51,52]. Additionally, albumin carriers inhibit collagen deposition by CAFs, reducing interstitial fluid pressure and enhancing penetration of coadministered drugs like gemcitabine[53].

Research advances in nab-paclitaxel for pancreatic cancer therapy

Studies on nab-paclitaxel in pancreatic cancer treatment have demonstrated significant preclinical and clinical efficacy. Its core advantage lies in overcoming the limitations of traditional chemotherapy through an NDDS. The abnormal tumor vasculature in pancreatic cancer, characterized by enlarged endothelial cell gaps and incomplete basement membranes, provides favorable conditions for nanoparticle accumulation[49,54]. nab-paclitaxel leverages its optimized particle size (130 nm) and surface properties to penetrate tumor vascular endothelial gaps and accumulate in tumor sites via the EPR effect[40]. The dense fibrotic stroma in pancreatic cancer typically impedes drug diffusion, but the albumin carrier enhances drug penetration by suppressing collagen deposition by CAFs and reducing interstitial fluid pressure[55].

Clinical studies have confirmed the efficacy of nab-paclitaxel. The phase 3 MPACT trial showed that gemcitabine combined with nab-paclitaxel achieved a median OS of 8.7 months in patients with LAPC, which is significantly superior to the 6.6 months in the gemcitabine monotherapy group, with the ORR increasing from 7% to 23%[14]. Additionally, the combination therapy prolonged PFS to 5.5 months, significantly increased tumor shrinkage rates, and even enabled tumor downstaging and curative surgery in some patients[15,56-58]. This synergy stems not only from the targeting capability and intracellular accumulation of nab-paclitaxel but also from its SPARC protein-mediated endocytosis[59].

In terms of safety, nab-paclitaxel demonstrates significant advantages. Traditional paclitaxel formulations rely on polyoxyethylated castor oil (Cremophor EL) as a solvent, which frequently induces hypersensitivity reactions and neurotoxicity, whereas nab-paclitaxel eliminates organic solvents, reducing hypersensitivity incidence from 20%-40% to < 1%[24]. Although grade ≥ 3 neutropenia occurs at a higher rate (38%), it can be effectively managed through dose adjustments or G-CSF support[60].

CLINICAL EFFICACY OF GEMCITABINE COMBINED WITH NAB-PACLITAXEL

Recent clinical trial data

In recent years, the combination of gemcitabine and nab-paclitaxel has demonstrated significant efficacy as a first-line treatment for LAPC in multiple clinical studies. The phase 3 MPACT trial showed that the combination therapy group achieved a median OS of 8.7 months, which was significantly superior to the 6.6 months in the gemcitabine monotherapy group, with the ORR increasing from 7% to 23%[5,15]. Subsequent studies further validated its long-term survival benefits. For example, in the phase 2 Locally advanced pancreatic cancer trial, LAPC patients receiving the combination therapy achieved a median OS of 18.8 months, with 17% undergoing tumor downstaging and curative surgery, achieving an R0 resection rate of 41%[61].

Subgroup analyses revealed heterogeneity in treatment outcomes. In the JCOG1611-GENERATE phase 3 study, the median OS for the gemcitabine/nab-paclitaxel regimen was 17.0 months, outperforming modified FOLFIRINOX (mFOLFIRINOX) (14.0 months) and S-IROX (13.6 months)[62]. However, subgroup analysis indicated that mFOLFIRINOX provided greater survival benefits in recurrent pancreatic cancer[63], suggesting recurrence status may influence regimen selection. Age-stratified studies demonstrated poorer tolerance of the gemcitabine/nab-paclitaxel regimen in elderly patients (≥ 75 years), with a 44% incidence of grade ≥ 3 neutropenia, whereas younger patients (< 65 years) achieved an ORR of 35.4% and median PFS of 6.7 months. Additionally, in diabetic patients, the gemcitabine/nab-paclitaxel regimen showed lower ORR (28%) and OS (9.1 months) compared to nondiabetic patients (ORR = 38%, OS 11.2 months), potentially due to diabetes-related metabolic disturbances affecting drug metabolism[62]. Table 1 summarizes the key clinical trials of gemcitabine/nab-paclitaxel treatment, and the transition from monotherapy to combination therapy.

Table 1 Summarizes the key clinical trials of gemcitabine/nab-paclitaxel treatment, and the transition from monotherapy to combination therapy.

Therapeutic regimen	Trial (phase)	Ref.	Nation/state	Year	Study population	Median OS (months)	ORR (%)	PFS (months)	Key biomarkers	Grade ≥ 3 toxicity	Trial ID
Gemcitabine monotherapy		Burris et al[105]	United States	1997	20	NR	NR	NR	NR	NR	NR
Gemcitabine monotherapy		Hartlapp et al[45]	Germany	2008	319	NR	NR	NR	NR	NR	NR
Nab-paclitaxel monotherapy	Phase I	Rajeshkumar et al[8]	United States	2011	20	NR	NR	NR	NR	NR	NR
AG	MPACT (phase III)	Goldstein et al[14]; Von Hoff et al[15]	Multi-country	2013	861	8.7	23	NR	SPARC: PFS + 3.2 months	Neutropenia 38%	NCT00844649
AG	LAPACT	Philip et al[74]	Multi-country	2020	106	NR	NR	NR	NR	NR	NCT02301143
AG	JCOG1611-GENERATE	Ohba et al[62]	Japan	2023	246	17.0	58	NR	KRAS WT: ORR 58% vs 32%	Diarrhea 5%	JCOG1611
AG + PEGPH20	TARGET-LAPC (phase II)	Hingorani et al[80]	Multi-country	2018	246	11.5	34	NR	HA: ORR + 22%	Edema 18%	NCT01839487
AG + toripalimab	Phase Ib/II	Chang et al[73]	China	2021	NR	16.3	41.7	NR	CD8+ core density > 10%	Colitis 12%	NCT04132531

AG: Gemcitabine plus nab-paclitaxel; PEGPH20: Pegylated recombinant human hyaluronidase PH20; MPACT: Metastatic pancreatic adenocarcinoma clinical trial; LAPC: Locally advanced unresectable pancreatic cancer; OS: Overall survival; ORR: Objective response rate; PFS: Progression-free survival; SPARC: Secreted protein acidic and rich in cysteine; WT: Wild type; HA: Hyaluronan; CD: Cluster of differentiation; NCT: National clinical trials; NR: Not reported.

Recent innovative drug delivery strategies have also achieved breakthroughs. A study on interventional therapy for LAPC demonstrated that transcatheter arterial infusion of the gemcitabine/nab-paclitaxel regimen increased ORR to 32%, prolonged median PFS to 5.1 months, and improved surgical conversion rates to 16% in patients with high vascular invasion (e.g., portal vein involvement)[64]. Additionally, a Canadian subgroup analysis of the MPACT trial revealed superior efficacy of the gemcitabine/nab-paclitaxel regimen in North American patients (median OS 11.9 months vs 7.1 months), potentially linked to regional healthcare disparities and distinct gene expression profiles[65].

Adverse effects and management

Although the combination of gemcitabine and nanoparticle albumin-bound paclitaxel significantly improves survival outcomes in LAPC, its adverse effects require controlled management through refined and personalized strategies. According to the MPACT trial data, the incidence of grade ≥ 3 hematological toxicities (e.g., neutropenia and anemia) with the gemcitabine/nab-paclitaxel regimen was 51.9%, significantly higher than in the gemcitabine monotherapy group (28.1%), with neutropenia being the primary dose-limiting toxicity. To address this, clinical practice typically uses G-CSF support and dose adjustments (e.g., nab-paclitaxel 220 mg/m² or gemcitabine 800 mg/m²) to maintain treatment continuity[66].

Neuropathy is another common adverse effect, with 30%-40% of patients experiencing grade 1/2 peripheral neuropathy (e.g., numbness and tingling) and 5%-10% developing grade ≥ 3 symptoms[67]. The mechanism involves the combined effects of paclitaxel-induced microtubule stabilization and gemcitabine-mediated DNA damage. Personalized management strategies include genotype-guided dose reductions (e.g., for patients with UGT1A128 or CYP2C8 slow metabolizer genotypes), preventive interventions (e.g., vitamin B12 or α-lipoic acid to slow symptom progression), and dynamic monitoring using patient-reported outcome tools[68-70].

Additionally, the gemcitabine/nab-paclitaxel regimen has a low hypersensitivity incidence (< 1%), but hepatotoxicity (grade ≥ 3 alanine transaminase/aspartate transaminase elevation in 8%-12% of cases) requires vigilance in elderly patients or those with liver dysfunction[71]. Pharmacokinetic studies indicate that nab-paclitaxel is metabolized via CYP3A4, necessitating a 25%-30% dose reduction when coadministered with CYP3A4 inhibitors (e.g., itraconazole)[72-74]. The core of personalized management lies in multidisciplinary collaboration and biomarker-driven approaches, such as immune status stratification (e.g., monitoring immune-related adverse effects in PD-L1-high or tumor-infiltrating lymphocyte-rich patients) and metabolic optimization (e.g., coenzyme Q10 supplementation for mitochondrial dysfunction)[70,75].

FUTURE RESEARCH DIRECTIONS AND CHALLENGES

Exploration of novel combination therapies

In the treatment of LAPC, the exploration of combination therapies is expanding from traditional chemotherapy to multimodal approaches[76]. The efficacy of the gemcitabine plus nab-paclitaxel (AG) regimen as a cornerstone has been validated in multiple phase 3 trials, with distinct mechanistic synergism and biomarker-driven outcomes. The MPACT study demonstrated that gemcitabine/nab-paclitaxel significantly prolonged median OS compared to gemcitabine monotherapy (8.7 vs 6.6 months), attributed to SPARC-mediated stromal targeting of nab-paclitaxel and DNA damage potentiation of gemcitabine[15]. The JCOG1611-GENERATE study confirmed the superiority of gemcitabine/nab-paclitaxel in metastatic/recurrent settings (median OS 17.0 months vs 14.0 months for mFOLFIRINOX), particularly in KRAS wild-type patients achieving 58% ORR through enhanced drug penetration in collagen-rich tumors (SHG score < 3)[62,77]. Regimen selection must consider both molecular profiles and clinical characteristics. Younger patients (age < 65 years) with preserved performance status (Eastern Cooperative Oncology Group 0-1) may benefit more from high-intensity regimens like mFOLFIRINOX (ORR = 32.4%, PFS = 6.4 months), albeit with ≥ 50% grade 3 neutropenia requiring G-CSF support[78]. In contrast, older patients (≥ 75 years) or those with comorbidities (e.g., diabetes-related metabolic dysfunction) show better tolerability to AG regimens, with manageable grade 3 diarrhea rates (5%) and neurotoxicity responsive to UGT1A1-guided dose reductions[79]. Emerging data from the target-LAPC trial highlight the importance of fibrosis stratification: Hyaluronan (HA)-high patients receiving PEGPH20 + AG achieved 34% ORR and 2.8-month PFS improvement, validating collagen degradation as a penetration-enhancing strategy[80]. Dual targeting of TME components further optimizes outcomes. The NCT04132531 trial demonstrated that PD-1 inhibitor combinations with AG in PD-L1-positive/CD8⁺ core-enriched tumors (≥ 10% density) yielded 41.7% ORR and 16.3- months median OS, overcoming traditional immunotherapy resistance in pancreatic cancer[79]. These advances underscore the necessity of integrating stromal biomarkers (SHG collagen quantification), immune contexture (multiplex immunohistochemistry), and metabolic profiling (18F-FDG PET avidity) into therapeutic decision-making frameworks.

The integration of immunotherapy has opened new avenues. A phase 1b/2 trial of the PD-1 inhibitor toripalimab combined with gemcitabine/nab-paclitaxel reported an ORR of 35.3% and a disease control rate (DCR) of 82.4%, with manageable toxicity (20% grade 3/4 adverse events)[79,81]. Despite the highly immunosuppressive pancreatic TME (e.g., low 3,3’,5,5’-tetramethylbenzidine, 90% PD-L1 negativity), strategies targeting metabolic reprogramming show promise. For example, the pyruvate carboxylase 1 inhibitor lixumistat combined with gemcitabine/nab-paclitaxel achieved a 100% DCR and median PFS of 9.7 months in a phase 1b trial[82], suggesting that inhibiting oxidative phosphorylation can reverse immunosuppression. Additionally, the mitogen-activated extracellular signal-regulated kinase inhibitor IMM-1-104 combined with gemcitabine/nab-paclitaxel in first-line therapy increased ORR to 43% (vs historical 23%), and when combined with FOLFIRINOX, achieved complete tumor regression[83].

Future research must focus on biomarker-driven personalized therapy. Genomic analyses reveal that KRAS wild-type patients have significantly higher response rates to AG (58% vs 32% in KRAS-mutated), while SPARC overexpression correlates with enhanced nab-paclitaxel targeting[84]. Dynamic biomarkers, such as ≥ 50% reduction in CA19-9 (associated with median OS of 21 months) and ctDNA clearance, can refine efficacy prediction. To address resistance mechanisms [e.g., adenosine triphosphate-binding cassette (ABC) transporter overexpression, nucleotide repair activation], combining P-gp inhibitors or poly adenosine diphosphate-ribose polymerase (PARP) inhibitors may represent breakthroughs[85].

Necessity of individualized treatment strategy

The significant heterogeneity of LAPC necessitates a paradigm shift from “one-size-fits-all” to precision stratification in therapeutic strategies[20]. Genomic studies demonstrate that KRAS mutation status critically determines treatment selection. KRAS wild-type patients achieve an ORR of 58% with AG, compared to only 32% in KRAS-mutant patients[40]. High SPARC protein expression strongly correlates with enhanced targeting efficiency of nab-paclitaxel, with SPARC-positive patients showing a 3.2-months prolongation in PFS compared to SPARC-negative counterparts[40]. Immune microenvironment profiling further guides therapeutic decisions. Patients with tumor-infiltrating lymphocytes ≥ 10% or PD-L1 positivity (combined positive score ≥ 1) exhibit a response rate of 41.7% to gemcitabine/nab-paclitaxel combined with PD-1 inhibitors, alongside a median OS of 16.3 months[47]. Metabolomic analyses reveal that patients with mitochondrial dysfunction or hyperactive glycolysis may benefit from IDO1 inhibitors combined with gemcitabine/nab-paclitaxel, achieving a DCR of 72%[86,87]. Dynamic biomarkers are pivotal for optimizing personalized therapy. Patients with ≥ 50% reduction in CA19-9 baseline levels demonstrate a median OS of 21 months, while ctDNA clearance rate significantly correlates with tumor regression depth[45]. Advances in resistance-mechanism-guided stratification include: Patients with activated nucleotide excision repair pathways (e.g., ERCC1) may benefit from PARP inhibitors plus gemcitabine/nab-paclitaxel (ORR 38% vs 15%)[88], and ABC transporter overexpression-mediated resistance can be reversed by coadministration of P-gp inhibitors (e.g., verapamil)[89].

To achieve true individualized therapy, it is essential to establish multidisciplinary teams (oncology, genetics, and bioinformatics) and dynamically refine treatment frameworks using real-world data. For instance, patient-derived-organoid-based drug sensitivity testing guides second-line regimen selection, demonstrating 82% concordance with clinical outcomes[90].

DISCUSSION

The AG regimen has established itself as the first-line therapy for unresectable LAPC, with demonstrated benefits in drug targeting, TME modulation, and pharmacokinetic optimization. The MPACT trial reported a median OS of 8.7 months, while the JCOG1611-GENERATE study highlighted its KRAS mutation-dependent efficacy, showing an ORR of 58% in wild-type patients vs 32% in mutant cases[5,15,62]. Clinical outcomes remain suboptimal due to two interrelated challenges: The lack of dynamic multiomics stratification models to guide precision therapy, and the heterogeneity of drug resistance mechanisms (e.g., ABC transporter overexpression) coupled with biomarker variability (e.g., inconsistent predictive utility of SPARC)[44,91]. These limitations are compounded by the spatiotemporal complexity of the TME, where desmoplastic stroma, immunosuppressive niches, and metabolic adaptations synergistically drive therapeutic resistance[23,47].

To address these barriers, we developed a four-dimensional TME phenotyping framework integrating fibrosis burden, immune spatial architecture, genomic variation, and metabolic reprogramming. This multidimensional approach was designed to overcome the reductionist limitations of single-parameter biomarkers by capturing both structural and functional TME heterogeneity[31]. The fibrosis axis uses SHG microscopy to quantify collagen density, with a threshold of > 35% area coverage (SHG score ≥ 4) identified as critical based on PDX models showing a 50% reduction in intratumoral gemcitabine concentration at this level[19,92]. This cutoff aligns with prior clinical evidence demonstrating impaired drug diffusion and vascular perfusion when collagen exceeds 35%[28]. The immune axis utilizes multiplex immunofluorescence to assess CD8⁺ T cell spatial distribution, where a core-to-invasive front density ratio < 0.5 (AUC = 0.81 by ROC analysis) predicts PD-1 inhibitor resistance, reflecting fibroblast-mediated immune exclusion[37]. Genomic stratification leverages liquid biopsy-based ctDNA profiling to distinguish KRAS subtypes (e.g., G12D/G12V/G12C), with G12D variants driving mitogen-activated protein kinase hyperactivation and G12V promoting epithelial-mesenchymal transition-mediated invasion[55]. Metabolic reprogramming is quantified via 18F-fluoro-l-m-tyrosine PET/CT, where SUV_max > 2.5 indicates IDO1-mediated tryptophan catabolism; a process linked to Treg cell expansion and CD8⁺ T-cell dysfunction[68]. Figure 2 shows the four-dimensional TME phenotypic framework for LAPC and its corresponding treatment strategies.

Figure 2 Four-dimensional tumor microenvironment phenotypic framework for locally advanced pancreatic cancer and its corresponding treatment strategies. LAPC: Locally advanced unresectable pancreatic cancer; SHG: Second-harmonic generation microscopy; CD: Cluster of differentiation; AG: Gemcitabine plus nab-paclitaxel; PEGPH20: Pegylated recombinant human hyaluronidase PH20; ABCi: Adenosine triphosphate-binding cassette inhibitor; PD-1: Programmed cell death protein 1; IDO1: Indoleamine 2,3-dioxygenase 1; SPARC: Secreted protein acidic and rich in cysteine; WT: Wild type; HRD: Homologous recombination deficiency; PARP: Poly adenosine diphosphate-ribose polymerase; CA: Carbohydrate antigen; ctDNA: Circulating tumor DNA; VAF: Variant allele frequency; CT: Computed tomography; TME: Tumor microenvironment.

Unifying these physical barrier-immune suppression-driver mutation-metabolic adaptation axes enables classification of four functionally distinct subtypes: (1) Fibro-high/immune-desert subtype, which is characterized by dense collagen (SHG > 35%) and CD8⁺ T-cell exclusion (core/front ratio < 0.5), reflecting a high fibrotic burden and immune exclusion. This leads to immune tolerance and chemoresistance. Patients with this subtype benefit from PEGPH20-mediated stromal degradation combined with gemcitabine/nab-paclitaxel chemotherapy[10], which enhances drug penetration and reduces collagen deposition, improving therapeutic efficacy; (2) Fibro-low/immune-hot subtype, which is defined by low fibrotic burden (SHG ≤ 2) and the presence of tertiary lymphoid structures (> 3 per high-power field), indicating a high-density immune infiltrate and active immune response. Patients with this subtype are more likely to respond to PD-1/IDO1 inhibitor combinations, which can reverse T-cell exhaustion and enhance immune responses, thereby improving outcomes in immunotherapy[39]; (3) KRAS wild-type/SPARC-high subtype. Characterized by SPARC overexpression (H-score ≥ 200), this subtype is associated with enhanced nab-paclitaxel targeting efficiency through SPARC-mediated endocytosis. KRAS wild-type tumors in this subgroup exhibit better response rates to treatment, with dose-escalated nab-paclitaxel improving drug accumulation in the tumor and leading to a 3.2- months increase in PFS[72]; and (4) IDO1-high/homologous recombination deficiency-high subtype, which is marked by active IDO1 metabolism (SUV_max > 2.5) and homologous recombination deficiency ≥ 42. A key feature of genomic instability. These tumors may respond well to combination therapy with gemcitabine/nab-paclitaxel and PARP inhibitors (e.g., olaparib)[93]. Additionally, IDO1 inhibitors combined with antiangiogenic agents (e.g., bevacizumab) can target metabolic reprogramming and restore immune function, further enhancing therapeutic outcomes[94,95].

Independent validation from the target-LAPC trial (No. NCT05511202) supports this framework, showing PEGPH20 plus gemcitabine/nab-paclitaxel extended median PFS by 2.8 months in fibro-high patients vs conventional gemcitabine/nab-paclitaxel[80]. Early ctDNA clearance at week 8 correlated strongly with SHG score reduction, suggesting ctDNA dynamics may serve as a real-time surrogate for TME remodeling[85]. However, current evidence remains limited by retrospective designs and post hoc analyses, as exemplified by the CALGB 80303 trial where SPARC failed to predict nab-paclitaxel response[91]. Prospective validation through trials like ALLIANCE A021806 is essential to confirm generalizability. Future integration of artificial intelligence (AI)-driven multiomics platforms, including radiomic features and exosomal microRNA profiling, could enable real-time TME monitoring and adaptive therapeutic modulation[96]. Despite the promise of TME-based precision therapy, there are several significant challenges before this framework can be fully implemented in clinical practice. Integrating multiomics data, including genomics, transcriptomics, proteomics, and metabolomics, requires access to high-quality tissue and blood samples, sophisticated analytical platforms, and substantial bioinformatics support[97]. These processes can be both technically complex and costly, potentially limiting their feasibility in non-academic or poorly-resourced cancer centers[98]. The interpretation and clinical application of multiomics data should be further defined as consensus is lacking regarding standardized thresholds and decision algorithms for patient management[99]. Moreover, real-time TME monitoring, such as frequent assessment of ctDNA dynamics, immune infiltration, and metabolic imaging, faces additional hurdles[100]. Logistical constraints, including limited access to advanced imaging, assay variability, and the cost of serial biomarker testing, may pose major barriers, particularly in resource-limited settings. Furthermore, most evidence in dynamic biomarker-guided therapy adjustment is merely derived from retrospective or early-phase studies, and prospective validation in large, diverse patient cohorts is urgently needed[101]. Finally, the inherent heterogeneity of pancreatic cancer, the evolving nature of treatment resistance mechanisms, and gaps in biomarker standardization further complicate the translation of emerging therapies and monitoring approaches into routine clinical care. For dynamic biomarkers such as ctDNA and CA19-9, optimal sampling intervals, actionable thresholds, and integration with clinical and radiological assessment require further study and international consensus[102,103].

In summary, while multiomics-based TME stratification and dynamic biomarker monitoring represent powerful tools for precision oncology, there are considerable translational, technical, and economic challenges. Addressing these limitations through collaborative, multi-institutional research and the development of standardized, cost-effective protocols is critical to put these innovations into clinical use.

CONCLUSION

A four-dimensional TME-based phenotyping framework may enable more precise and individualized therapy for unresectable LAPC, addressing the limitations of standard regimens such as AG. Preliminary clinical evidence suggests that this strategy could improve outcomes (e.g., longer PFS in high-fibrosis subgroups[80]) and facilitate adaptive treatment adjustment through dynamic biomarker monitoring (e.g., ctDNA clearance with SHG score reduction[104,105]). Future prospective, stratified multicenter trials are needed to validate this approach and to develop AI-driven decision systems that can further overcome the challenges of TME heterogeneity and resistance.