Immune therapy-related hyperprogressive disease: Molecular mechanisms, biomarkers, and clinical strategies

Abstract

Programmed death receptor-1 inhibitors have significantly improved the prognosis of various malignancies. Nevertheless, hyperprogressive disease (HPD), recognized as a severe adverse reaction to immunotherapy, causes a substantial surge in tumor burden and notably shortens the survival of 4%-29% of patients. This article comprehensively reviews the controversies regarding the clinical definition of HPD, its cross-cancer epidemiological features (encompassing gastric cancer, non-small cell lung cancer, head and neck squamous cell carcinoma, etc.), and potential molecular mechanisms (such as MDM2/MDM4 gene amplification, EGFR mutations, and reprogramming of the immune microenvironment). It further delves into biomarker-based predictive models, targeted combination therapy strategies, and salvage treatment alternatives. Ultimately, it puts forward future directions, including the establishment of a multicenter HPD registry database and organoid predictive models, aiming to offer evidence-based guidance for clinical practice.

Key Words: Hyperprogressive disease; Immune checkpoint inhibitors; Tumor microenvironment; Regulatory T cells; Predictive biomarkers; Combination therapy

Core Tip: Immune checkpoint inhibitor-induced hyperprogressive disease represents a paradoxical clinical phenomenon characterized by accelerated tumor progression mediated through complex immunosuppressive mechanisms involving regulatory T cells expansion, M2-type tumor-associated macrophages polarization, and myeloid-derived suppressor cells accumulation within the tumor microenvironment, coupled with oncogenic alterations such as MDM2 amplification and EGFR activation. Current predictive approaches integrate liquid biopsy profiling (e.g., circulating tumor DNA and blood tumor mutational burden dynamics) with systemic inflammatory markers (e.g., neutrophil-to-lymphocyte ratio and lactate dehydrogenase ratios), while emerging therapeutic strategies focus on combinatorial approaches targeting immunosuppressive networks, metabolic reprogramming, and checkpoint inhibition. This multifaceted pathogenesis underscores the critical need for mechanistic elucidation and translational development of precision management algorithms to mitigate this life-threatening complication of cancer immunotherapy.

INTRODUCTION

Programmed death receptor-1 (PD-1) inhibitors, as a significant breakthrough in the field of cancer immunotherapy[1], have been approved globally for the treatment of various malignancies, including non-small cell lung cancer (NSCLC) and melanoma[2,3]. Drugs such as nivolumab and pembrolizumab have significantly extended patient survival[4]. However, the emergence of hyperprogressive disease (HPD), a unique pattern of tumor progression during immunotherapy, has severely impacted clinical outcomes[5]. HPD is characterized by a dramatic increase in tumor volume by 20-30 times during the initial phase of treatment, a doubling of the tumor growth rate (TGR) within two months, and the rapid appearance of new metastatic lesions, leading to a sharp reduction in median survival to 3.4 months[6-8]. Although the diagnostic criteria for HPD are not yet fully standardized and its incidence varies widely across different tumor types, its association with poor patient prognosis has been widely confirmed[9]. A deeper understanding of the immunological mechanisms of HPD is not only crucial for the development of predictive biomarkers but also provides a theoretical basis for exploring combined targeted therapy strategies, thereby advancing cancer immunotherapy towards precision and personalization.

DIAGNOSTIC FEATURES AND EPIDEMIOLOGICAL OVERVIEW OF HPD

The diagnostic criteria for HPD demonstrate considerable diversity, with several widely accepted core standards currently based on three main aspects: (1) TGR increasing by > 50% compared to pretreatment levels[10]; (2) Time to treatment failure (TTF) of less than 2 months; and (3) Disease progression meeting Response Evaluation Criteria in Solid Tumours criteria accompanied by rapid clinical deterioration. Compared to the conventional method of calculating TGR based on tumor volume, the criterion of ≥ 2-fold increase in the sum of maximum diameters of target lesions (TGK) compared to baseline provides a more sensitive measure of tumor growth acceleration. Notably, this TGK-based definition can be effectively integrated with the evaluation of new lesions for comprehensive assessment[11,12], an approach that has demonstrated promising clinical utility in multiple studies. Yamada et al[13] proposed TTF < 2 months, combined with a > 50% increase in tumor burden and a ≥ 2-fold increase in progression speed. Lo Russo et al[14] included five criteria (e.g., new metastatic lesions, worsening Response Eastern Cooperative Oncology Group score), emphasizing the combination of early clinical deterioration and imaging progression.

The significant fluctuation in HPD incidence under different criteria is attributed to factors such as the mathematical difference between TGK and TGR (e.g., TGK = 2 corresponds to TGR = 8) and whether new lesions or clinical deterioration indicators are included. Currently, the definitions of HPD have certain limitations. The lack of a unified definition is a major bottleneck in HPD research, leading to heterogeneity in incidence rates, risk factors, and prognostic conclusions[15]. In the future, it is essential to establish a cross-cancer-type universal framework that integrates imaging, clinical, and molecular biomarkers to dynamically assess accelerated tumor progression. This will guide immunotherapy decision-making and mitigate the risk of overtreatment. Table 1 summarizes the incidence rates, key studies, and high-risk factors of HPD across different cancer types[10,13,14,16-22].

Table 1 Incidence, key studies, and risk factors of hyperprogressive disease.

Ref.	Cancer type	HPD incidence	Proposed risk factors
Ferrara et al[16], 2018	NSCLC	13.8%-26%	EGFR mutation, LDH > 250 U/L, liver metastasis, ≥ 2 metastatic sites
Lo Russo et al[14], 2020
Economopoulou et al[17], 2021	HNSCC	15.4%	11q13 chromosomal amplification (CCND1/FGF3), local recurrence
Kim et al[18], 2022	GC	9.2%-29.4%	EGFR/FGF4, MDM2 amplification, liver metastasis, high tumor burden, advanced age
Aoki et al[19], 2024
Hwang et al[22], 2020	Urothelial carcinoma	6.4%-8%	Elevated baseline NLR, high LDH levels
Abbas et al[10], 2019
Yamada et al[13], 2018	Melanoma	6%-42%	Advanced age, elevated baseline inflammatory markers (CRP/NLR)
Zhou et al[20], 2025
Şen et al[21], 2024	Bladder cancer	12.9%	Elevated baseline NLR, advanced age
	Renal cell carcinoma	4.8%

HPD: Hyperprogressive disease; NSCLC: Non-small cell lung cancer; HNSCC: Head and neck squamous cell carcinoma; GC: Gastric cancer; EGFR: Epidermal growth factor receptor; LDH: Lactate dehydrogenase; CCND1: Cyclin D1; FGF3: Fibroblast growth factor 3; FGF4: Fibroblast growth factor 4; MDM2: Mouse double minute 2 homolog; NLR: Neutrophil-to-lymphocyte ratio; CRP: C-reactive protein.

MECHANISMS OF HPD OCCURRENCE FOLLOWING IMMUNOTHERAPY

In-depth elucidation of the mechanisms underlying HPD not only explains why immunotherapy produces counterproductive effects in some patients but, more importantly, provides a theoretical basis for clinical prevention and treatment. It offers new insights for optimizing existing immunotherapy regimens, guiding clinicians in avoiding treatment risks, and enhancing the safety and efficacy of immunotherapy. The key molecular and cellular mechanisms of HPD are systematically summarized in Table 2.

Table 2 Key regulatory mechanisms and therapeutic strategies in hyperprogressive disease.

Regulatory factor	Core mechanism	Clinical significance	Potential intervention strategies
Treg cells	Abnormal activation post PD-1 blockade; inhibits APC function via CTLA-4/CD80/CD86; secretes IL-10/TGF-β to suppress effector T cells	Significant Treg infiltration in HPD patients, negatively correlated with treatment efficacy	Target CTLA-4 or deplete Tregs (e.g., anti-CCR4 antibodies)
M2-type TAM	Fc receptor-mediated pro-tumor phenotype conversion by anti-PD-1; recruited via CCR2/CCL2 and maintained by CSF-1R signaling	Forms synergistic network with Tregs/MDSCs to drive HPD	Combine CSF-1R inhibitors with PD-1 blockade (e.g., pexidartinib)
MDSC	Suppresses T cell function via IDO/VEGF/MMP9; promotes angiogenesis	High MDSC levels predict poor ICI response and HPD risk	Target CXCR2 or arginine metabolism (e.g., CB-1158)
T cell exhaustion	Compensatory upregulation of TIM-3/CTLA-4 post PD-1 blockade, leading to “secondary exhaustion”	Enriched in HPD patients with lost anti-tumor function	Multi-checkpoint targeting (e.g., anti-TIM-3 + anti-PD-1)
CAF	Upregulates PD-L1 via IL-6/STAT3; recruits Tregs via CXCL12; promotes M2-TAM differentiation	Promotes immunosuppressive microenvironment, positively correlated with HPD progression	Target TGF-β signaling (e.g., galunisertib) or CAF reprogramming
Inflammatory dysregulation	IL-10 inhibits CD28 signaling; IFN-γ induces PD-L1 upregulation; IL-6/TNF-α activate STAT3/NF-κB to promote proliferation	Cytokine profile predicts HPD risk	JAK inhibitors (e.g., ruxolitinib) or IL-6R blockade (tocilizumab)
Metabolic reprogramming	Tregs utilize lactate/OXPHOS for survival; CD36 mediates fatty acid metabolism adaptation	Metabolic competition exacerbates T cell dysfunction	Lactate dehydrogenase inhibitors or CD36 blockade
Genetic alterations	MDM2/MDM4 amplification (p53 degradation); EGFR activation (PD-L1 modification); DNMT3A mutations (epigenetic dysregulation)	11-gene mutation signature significantly associated with HPD	MDM2 inhibitors (e.g., idasanutlin) or EGFR-TKI combination therapy

TAM: Tumor-associated macrophages; Tregs: Regulatory T cells; APC: Antigen-presenting cell; MDSC: Myeloid-derived suppressor cell; CAF: Cancer-associated fibroblasts; CTLA-4: Cytotoxic T-lymphocyte-associated protein 4; IL-10/TGF-β: Interleukin-10/transforming growth factor-beta; HPD: Hyperprogressive disease; CCR4: C-C chemokine receptor type 4; CCR2: C-C chemokine receptor type 2; CCL2: C-C chemokine ligand 2; CSF-1R: Colony-stimulating factor 1 receptor; MDSCs: Myeloid-derived suppressor cells; ICI: Immune checkpoint inhibitors; PD-1: Programmed death receptor-1; CXCR2: C-X-C chemokine receptor type 2; CB-1158: Arginase inhibitor; TIM-3: T cell immunoreceptor with Ig and ITIM domains; STAT3: Signal transducer and activator of transcription 3; IDO: Indoleamine 2,3-dioxygenase; VEGF: Vascular endothelial growth factor; MMP9: Matrix metalloproteinase 9; CXCL12: C-X-C chemokine ligand 12; M2-TAM: M2-type tumor-associated macrophages; IFN-γ: Interferon-gamma; PD-L1: Programmed death-ligand 1; TNF-α: Tumor necrosis factor-Alpha; NF-κB: Nuclear factor-kappa B; OXPHOS: Oxidative phosphorylation; MDM2: Mouse double minute 2 homolog; EGFR-TKI: Epidermal growth factor receptor-tyrosine kinase inhibitor.

CHANGES IN IMMUNE CELLS WITHIN THE TUMOR MICROENVIRONMENT

Regulatory T cells

Studies have shown that the PD-1 signaling pathway plays a significant role in regulating the activity of regulatory T (Treg) cells[23]. Researchers such as Kamada et al[24] discovered through human and mouse models that PD-1 blockade or deficiency not only enhances the anti-tumor activity of conventional T cells (Tconv) but also promotes the activation of Treg) cells. Specifically, PD-1 blockade strengthens the transmission of TCR and CD28 signaling pathways in Treg cells, promoting their proliferation and suppressing the function of effector T cells (including CD4 T cells and CD8 T cells) through various mechanisms[25]. Therefore, the ultimate effect of PD-1 blockade therapy actually depends on the dynamic balance between Treg activation and Tconv activation which is influenced by factors such as tumor type and characteristics of the immune microenvironment. Multiple clinical studies on HPD have observed a significant increase in Treg cell infiltration in the tumor tissues of patients experiencing HPD[26]. Particularly in studies on gastric cancer immunotherapy, it was found that proliferative (Ki67+) effector Treg cells (eTreg) were markedly increased in the tumor tissues of HPD patients, while the opposite trend was observed in patients without HPD[24]. Further analysis revealed that PD-1-positive eTreg cells in both the circulatory system and tumor tissues of these patients were in a highly activated state, with their cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) expression levels significantly higher than those of PD-1-negative eTreg cells. These findings collectively suggest that abnormal activation of Treg cells and the enhancement of their immunosuppressive function may be one of the key mechanisms leading to HPD following immunotherapy. In clinical treatment, special attention should be paid to the dynamic changes of Treg cells in the tumor microenvironment (TME) and their impact on treatment outcomes[26].

Antigen-presenting cell

In the model of HPD following: PD-1 blockade therapy, antigen-presenting cell (APC) cells play a crucial role[27]. The increased Treg cells after PD-1 blockade therapy, particularly PD-1+ Treg cells with high CTLA-4 expression, can establish a robust immunosuppressive network by modulating the function of APCs. Specifically, Treg cells with high CTLA-4 expression bind to the co-stimulatory molecules CD80/CD86 on the surface of APCs, leading to the downregulation of these key co-stimulatory molecules[24]. This not only directly inhibits the activation of APCs but, more importantly, blocks the pathway through which CD8+ T cells receive co-stimulatory signals via the CD28 receptor, significantly impairing the anti-tumor function of CD8+ T cells[28,29]. Furthermore, within the tumor tissue, this interaction may result in increased accumulation of Treg cells on APCs. Combined with the downregulated CD80/CD86 co-stimulatory molecules, this limits the access and activation of Tconv cells by APCs[30].

Treg-macrophage immunosuppressive axis

Macrophages are a crucial immune cell population mediating the occurrence of HPD[31]. Analysis of pre-treatment tumor specimens from NSCLC patients revealed a characteristic infiltration pattern of tumor-associated macrophages (TAMs) in the tumor lesions of HPD patients. These macrophages exhibit a unique immune phenotype [CD163, CD33, programmed death-ligand 1 (PD-L1)][32], display an epithelial-like morphology (similar to alveolar macrophages), and tend to form dense clusters within the tumor foci. Animal experiments using immunodeficient mouse models also confirmed this finding. In HPD tumor tissues, there was a significant increase in leukocyte infiltration, primarily manifested as a marked rise in the number of intratumoral macrophages (F4/80+), while the infiltration levels of B lymphocytes (CD45R/B220+), granulocytes (Gr-1+), and natural killer (NK) cells (NKp46+) were comparable to those in the control group. Multiple studies support this observation: In lung adenocarcinoma models, reducing M2-type TAM recruitment through C-C chemokine receptor type 2/C-C chemokine ligand (CCL) 2 signaling blockade inhibited tumor growth[33]; in hepatocellular carcinoma research, macrophages were found to directly suppress T cell responses, and blocking the colony-stimulating factor 1 (CSF-1)/CSF-1 receptor (CSF-1R) pathway induced tumor rejection; pancreatic cancer models demonstrated that combining PD-1/CTLA-4 antibodies with CSF-1R blockade enhanced therapeutic efficacy[34].

Furthermore, researchers have discovered that the Fc domain of anti-PD-1 antibodies plays a critical role in the functional reprogramming of macrophages, and the signaling it mediates is a significant factor contributing to HPD[32]. Experiments revealed that using full-length antibodies to block the PD-1 receptor expressed on TAMs themselves promoted the acquisition of a pro-tumor phenotype, whereas using anti-PD-1 F(ab’)2 fragments lacking the Fc segment completely eliminated this pro-tumor effect. Researchers proposed the hypothesis that the interaction between Fc receptors and the immune phenotype specific to immune checkpoint inhibitors (ICIs) might trigger the aggregation of M2 TAMs[32], reprogramming them into a more aggressive pro-tumor phenotype. These findings suggest that HPD may occur in patients with specific genetic and immune characteristics, and the underlying mechanisms still require further elucidation.

Myeloid-derived suppressor cell

Myeloid-derived suppressor cells (MDSCs), as a heterogeneous population of immunosuppressive cells, play a pivotal role in the immune regulation of the TME[35]. These cells not only directly participate in promoting tumor angiogenesis, metastasis, and other malignant processes but are also closely associated with clinical outcomes in patients. Multiple clinical studies have confirmed a significant correlation between MDSC levels and treatment efficacy: In prostate cancer, breast cancer, and colorectal cancer patients, lower MDSC infiltration often predicts better overall survival[36].

In the field of ICI therapy, the regulatory role of MDSCs is particularly prominent. A significant finding by Meyer et al[37] demonstrated that melanoma patients with lower MDSC frequencies exhibited better responses to ipilimumab treatment, suggesting that MDSC levels may serve as a predictive biomarker for ICI efficacy. This finding has been validated in numerous other studies[38,39], which showed that MDSCs negatively impact survival rates and are inversely correlated with the presence of functional antigen-specific T cells in advanced melanoma patients. This further underscores the critical role of MDSCs in immunotherapy responses. Notably, in mouse models of renal cell carcinoma and Lewis lung cancer, strategies combining MDSC blockade with ICI therapy demonstrated more significant survival benefits compared to ICI monotherapy[40].

T CELL EXHAUSTION AND COMPENSATORY UPREGULATION OF IMMUNE CHECKPOINTS

T cell exhaustion is a critical mechanism of tumor immune evasion, characterized by the persistent overexpression of multiple inhibitory receptors, including PD-1, lymphocyte activation gene-3, T cell immunoreceptor with ig and ITIM domains, and mucin domain-containing molecule 3[41,42]. In patients with HPD, T cell exhaustion is particularly pronounced. Studies have found that due to continuous exposure to high concentrations of tumor neoantigens, these exhausted T cells cannot restore their function like normal memory T cells but instead fall into a deeper state of functional suppression, becoming more significantly enriched in tumor tissues[43,44]. Specifically, PD-1 blockade therapy leads to a significant expansion of PD-1/CTLA-4 double-positive Treg cells, which secrete inhibitory cytokines such as interleukin (IL)-10 and transforming growth factor-beta (TGF-β). These cytokines not only directly suppress effector T cell function but may also cause T cell dysfunction. Although PD-1 blockade initially activates some Tconv, under the persistent suppression of Treg cells, these T cells often exhibit rapid exhaustion after short-term activation, ultimately losing their anti-tumor capabilities entirely. Notably, abnormal activation of the PD-1/PD-L2 pathway also plays a role: PD-L2 (primarily expressed on endothelial cells, macrophages, and dendritic cells) inhibits T cell function through an atypical pathway (independent of CD80)[45]. This multi-layered T cell dysfunction forms a critical basis for the occurrence of HPD and explains why ICI therapy in some patients not only fails to restore T cell function but may further worsen the immune state[46].

Additionally, a noteworthy phenomenon during ICI therapy is the compensatory upregulation of immune checkpoints. In NSCLC patients and lung adenocarcinoma mouse models, tumor-infiltrating CD8+ T cells exhibit excessive T cell immunoreceptor with ig and ITIM domains expression during tumor progression. Meanwhile, ovarian cancer models show that PD-1 blockade significantly upregulates CTLA-4 and lymphocyte activation gene-3 expression on CD8+ T cells[47]. This compensatory upregulation leads to severe consequences: Further reduction in cytokine production, severe impairment of proliferation and migration functions, and the formation of a “secondary exhaustion” state. Understanding these mechanisms is crucial for developing intervention strategies targeting HPD.

TAM

M2-type TAMs play a pivotal role in the development and progression of HPD[48]. Multiple studies have shown that PD-1 antibody treatment leads to the characteristic expansion of M2-type TAMs, which exhibit a unique CD163+CD33+PD-L1+ immunophenotype and cluster-distributed epithelial-like morphological features. Mechanistically, M2-type TAMs are extensively recruited to the TME through the C-C chemokine receptor type 2/CCL2 signaling pathway[49,50] and exert potent immunosuppressive functions via CSF-1/CSF-1R interactions. Notably, the Fc segment of anti-PD-1 antibodies binding to Fc receptors triggers abnormal aggregation and functional reprogramming of M2-type TAMs[32,51], endowing them with a more pro-tumorigenic phenotype. These reprogrammed M2-type TAMs not only directly suppress T cell function[52] but also form a synergistic network with immunosuppressive cells such as Tregs and MDSCs[53], collectively promoting tumor immune evasion and the development of HPD.

Of particular significance, preclinical studies have confirmed that the combined use of PD-1/CTLA-4 antibodies and CSF-1R inhibitors significantly enhances anti-tumor efficacy, providing a potential therapeutic strategy to overcome HPD. Furthermore, the development of novel treatments that specifically eliminate or reprogram M2-type TAMs, as well as the exploration of their interaction networks with other immune cells, will open new avenues for the prevention and treatment of HPD (Figure 1).

Figure 1 Molecular mechanisms of hyperprogressive disease. CAF: Cancer-associated fibroblasts; TGF-β: Transforming growth factor-beta; CXCL: C-X-C chemokine ligand; IL: Interleukin; TAM: Tumor-associated macrophages; PD-L1: Programmed death-ligand 1; CTLA-4: Cytotoxic T-lymphocyte-associated protein 4; APC: Antigen-presenting cell; MDSC: Myeloid-derived suppressor cells; MDM2: Mouse double minute 2 homolog.

CANCER-ASSOCIATED FIBROBLASTS

Cancer-associated fibroblasts (CAFs), as a critical component of the TME, contribute to the development and progression of HPD through multiple mechanisms[54]. These activated stromal cells expressing α-smooth muscle actinnot only directly promote tumor growth and metastasis but also establish an immunosuppressive microenvironment through a complex immune regulatory network. Studies have shown that CAFs can recruit monocytes and promote their differentiation into M2-type TAMs[55], significantly enhancing immunosuppressive effects. In liver cancer, CAFs induce neutrophil activation and IL-6 release, triggering the IL-6-signal transducer and activator of transcription 3 (STAT3)-PD-L1 signaling pathway, ultimately leading to T cell suppression. Additionally, CAFs promote Treg cell differentiation through the TGF-β signaling pathway[56,57] while inhibiting the activation of cytotoxic T cells and NK cells[58]. Notably, CAFs can directly upregulate PD-L1 expression on tumor cells[59] and induce PD-1 expression on immune cells. These mechanisms collectively explain the pivotal role of CAFs in promoting HPD during ICI therapy.

CHANGES IN INFLAMMATORY FACTORS WITHIN THE IMMUNE MICROENVIRONMENT

In patients experiencing HPD, the TME exhibits characteristic dysregulation of inflammatory factors, forming a unique immunosuppressive network. IL-10, as a key inhibitory factor, promotes HPD through a dual mechanism: On one hand, it inhibits CD28 tyrosine phosphorylation[60], blocking T cell co-stimulatory signals; on the other hand, it induces T cell unresponsiveness in an IL-2-deficient environment. This effect is particularly pronounced in microenvironments with increased Treg cells. Simultaneously, abnormal activation of the CCL17/CCL22-CCR4/CCR8 axis promotes Treg cell expansion and IL-10 secretion through the STAT3-FoxP3 pathway, creating a positive feedback loop. The interferon-gamma (IFN-γ) pathway exhibits a paradoxical role[61]: While it enhances antigen presentation and T cell recruitment, persistent stimulation leads to tumor cells upregulating PD-L1 expression via the Janus kinase (JAK)-signal transducer and activator of transcription and phosphoinositide 3-kinase-protein kinase b pathways, ultimately resulting in treatment resistance[62]. Notably, in the HPD microenvironment, pro-inflammatory factors such as IL-6[63,64] and tumor necrosis factor-alpha[65] are elevated alongside anti-inflammatory factors like IL-10 and TGF-β. These factors activate the STAT3/nuclear factor-kappa B signaling pathway, promoting tumor proliferation and recruiting immunosuppressive cells such as TAMs and MDSCs. Additionally, proliferating PD-1+ Treg cells rapidly consume IL-2, depriving tumor-reactive effector T cells of this critical cytokine. These changes collectively construct an immune environment conducive to tumor hyperprogression. Understanding the characteristics of these inflammatory factors in the TME is crucial for future efforts to predict and treat HPD-related inflammation.

TME

Research has found that the TME of HPD patients exhibits significant metabolic and immunosuppressive abnormalities. In a hypoxic, glucose-deprived, and acidic TME, tumor cells rapidly consume oxygen and glucose through aerobic glycolysis, producing large amounts of lactate and creating a harsh environment unfavorable for the survival of conventional immune cells. However, Tregs undergo FoxP3-mediated metabolic reprogramming, inhibiting glycolysis and enhancing oxidative phosphorylation[66], enabling them to utilize lactate as an energy source to sustain their survival and immunosuppressive functions[67]. Additionally, Tregs upregulate the fatty acid transporter CD36, adapting to the hostile microenvironment by utilizing fatty acids released by tumor and stromal cells, and the efficacy of ICIs is further diminished in Treg-dominated tumors. This abnormal metabolic adaptation is closely linked to the synergistic effects of various immunosuppressive cells. Furthermore, CAFs recruit Tregs by secreting C-X-C chemokine ligand 12, IL-6, and release TGF-β to promote the differentiation of M2-type TAMs[68]. M2 TAMs further secrete CCL22 and IL-6, forming a positive feedback loop that enhances immunosuppression. Meanwhile, MDSCs inhibit the function of effector T cells by producing molecules such as IDO, VEGF, and MMP9[69], promoting tumor metastasis. These abnormal interactions among cells collectively construct the unique immunosuppressive network characteristic of HPD. The above studies indicate that the occurrence of HPD is closely related to the metabolic adaptation and abnormal interactions of immune cells in the TME. Suppressive cells like Tregs survive and function in the harsh environment through unique metabolic reprogramming, while the synergistic effects of CAFs, M2 TAMs, and MDSCs further strengthen the immunosuppressive microenvironment[54]. These findings provide a theoretical basis for developing combination therapeutic strategies targeting TME metabolism and cellular interactions, which may help prevent or reverse the occurrence of HPD. Future research should focus on how to disrupt this abnormal metabolic-immune network and restore anti-tumor immune responses.

SPECIFIC GENETIC ALTERATIONS ASSOCIATED WITH HPD

Specific genomic alterations have been confirmed to be closely associated with the occurrence of HPD. Multiple studies have found that mutations or amplifications of certain key genes significantly increase the risk of HPD. The amplification of the mouse double minute 2 homolog (MDM2)/MDM4 gene family and mutations in the DNMT3A gene have been identified by several research teams as important molecular markers of HPD[8,70]. Mechanistic studies reveal that the MDM2 protein can directly bind to the p53 protein, promoting its degradation and thereby weakening p53’s tumor-suppressive function[71]. Additionally, the JAK-signal transducer and activator of transcription signaling pathway activated by ICIs upregulates the expression of interferon regulatory factor IRF-8, which in turn positively feedbacks to activate MDM2 transcription. MDM2 also participates in tumor angiogenesis and metastasis by regulating VEGF expression. The activation of another key gene, EGFR, not only promotes the expression of immune checkpoints such as CTLA-4, PD-1, and PD-L1, but its signaling pathway can also induce post-translational modifications of the PD-L1 protein[72,73]. Furthermore, Xiong et al[74] identified through whole-exome sequencing analysis that deleterious mutations in 11 genes, including TRPC4, POTEE, and FBN2, are significantly associated with HPD.

These changes collectively form the molecular basis of tumor immune evasion. These findings provide crucial insights into understanding the genetic mechanisms of HPD and also point the way for developing targeted predictive and intervention strategies.

PREDICTIVE AND DIAGNOSTIC VALUE OF HPD-RELATED CELLS AND MOLECULAR MARKERS

Currently, the prediction research of HPD still faces significant challenges. Conventional imaging assessments exhibit notable delays, while existing molecular biomarkers such as PD-L1 demonstrate substantial limitations in predictive efficacy: Clinical data reveal insufficient accuracy (with only 40%-45% response rates in high-expression patients), compounded by multiple constraints including tumor spatial and temporal heterogeneity (30%-40% discrepancies between primary and metastatic lesions), standardization issues in detection (60%-70% concordance rates across different platforms), and complex microenvironmental regulation (e.g., the bypass effects of immunosuppressive cells like Tregs and MDSCs)[75,76]. These factors collectively hinder the ability to meet clinical demands for precise early warning. With advancements in genomics and liquid biopsy technologies, researchers are actively working to establish more sensitive and specific molecular prediction systems.

Liquid biopsy, as an emerging non-invasive detection technology, has demonstrated significant value in monitoring the efficacy of tumor immunotherapy and providing early warning of HPD. This technology comprehensively evaluates blood-based tumor mutation burden (bTMB) and tumor genomic instability characteristics by detecting circulating tumor DNA (ctDNA) and cell-free DNA in peripheral blood. Multiple clinical studies have confirmed that bTMB levels are significantly positively correlated with immunotherapy response[77,78], offering a new molecular marker for pre-treatment efficacy prediction. Notably, the dynamic changes in ctDNA exhibit higher sensitivity compared to traditional imaging examinations. For example, a study involving 125 melanoma patients showed that ctDNA levels in patients with pseudoprogression decreased significantly by 10-fold during treatment, while those in true progression patients continued to rise[79]. The Weiss team further discovered that cell-free DNA chromosomal instability scores could predict the occurrence of HPD 6-9 weeks before conventional imaging detection[80], a breakthrough finding that provides a valuable time window for clinical intervention.

In addition to liquid biopsy, routine inflammatory indicators and metabolic markers in peripheral blood also show promising predictive potential. In the analysis of immune cell subsets, the dynamic changes in the neutrophil-to-lymphocyte ratio (NLR) are closely related to disease progression. Kiriu et al[81] found that patients with an NLR increase of more than 30% had significantly shorter TTF. However, it should be noted that research teams such as Ferrara et al[16] did not observe significant correlations between NLR, LDH, and HPD[67], a discrepancy that may stem from differences in study populations, detection methods, and evaluation criteria. Based on the current evidence, establishing a multi-parameter joint prediction model that integrates genomic features (ctDNA/bTMB), immune-inflammatory indicators (NLR), and metabolic markers (LDH) may be an important direction for improving the accuracy of HPD prediction. With advancements in detection technology and the development of large-scale studies in the future, the clinical application value of these biomarkers will be more clearly defined.

THERAPEUTIC ADVANCES AND POTENTIAL STRATEGIES TARGETING HPD MECHANISMS

Based on the immunometabolic mechanisms of HPD, a variety of targeted therapeutic strategies have emerged in recent years. Firstly, targeting immunosuppressive cells is a crucial direction: Anti-CCR4/CCR8 antibodies can block Treg recruitment; CSF-1R inhibitors (such as pexidartinib) or CD47-targeting drugs can eliminate M2-type macrophages[82]; indoleamine 2,3-dioxygenaseinhibitors (such as epacadostat) or arginase inhibitors can reverse the immunosuppressive functions of MDSCs[83]. Secondly, drugs that intervene in metabolic abnormalities can disrupt the adaptive survival of Tregs. Additionally, combination immunomodulatory therapies may restore effector T cell function. Preclinical studies have shown that the combination of PD-1 inhibitors with the aforementioned targeted drugs can significantly delay HPD progression. Currently, multiple clinical trials are underway. In the future, it will be necessary to further optimize treatment timing and combination strategies, and to screen beneficiary populations through biomarkers to achieve precise intervention in HPD.

Combination therapy strategies represent a critical direction for enhancing the efficacy of ICIs and reducing the risk of HPD. In hepatocellular carcinoma treatment, the combination of atezolizumab and bevacizumab has demonstrated significant advantages, with a lower incidence of HPD compared to sorafenib monotherapy, and the risk of poor prognosis being independent of the treatment approach. Studies indicate that tumor plasticity is a key mechanism of ICI-mediated HPD, where IFN-γ and PD-L1 promote malignant transformation of tumor cells by regulating CD44 expression and the MAPK/STAT3 signaling pathway. This provides new insights for predicting HPD (e.g., monitoring plasticity markers such as CD44) and treating it (targeting the IFN-γ/PD-L1 pathway). Furthermore, immune combination therapy models show potential: The combination of NK cells with ICIs may serve as an alternative option for HPD patients[84]; data from the CheckMate 227 trial confirmed that nivolumab combined with chemotherapy significantly reduces the risk of disease progression compared to its combination with ipilimumab[85]. These findings offer important evidence for optimizing combination therapy regimens to avoid HPD.

CONCLUSION

HPD, as a significant challenge in PD-1 inhibitor immunotherapy, involves multiple mechanisms, including abnormal activation of Treg cells, recruitment of M2-type macrophages and MDSCs, T cell exhaustion, and dysregulation of the inflammatory cytokine network within the TME. Clinically, HPD is characterized by accelerated tumor growth during the early stages of treatment, explosive emergence of new metastatic lesions, and extremely poor prognosis, with significant variations in incidence and risk factors across different cancer types. Currently, diagnostic criteria remain inconsistent, and future research should focus on establishing a universal diagnostic framework across cancer types to enable early warning and personalized intervention for HPD.