Cryoablation remodels the immune microenvironment in hepatocellular carcinoma: From mechanistic insights to clinical translation in combination immunotherapy

Table 1 Dual immunomodulatory effects of cryoablation on the hepatocellular carcinoma microenvironment

Immunomodulatory direction	Core mechanisms and events	Key effects/findings
Positive immune activation	Induction of immunogenic cell death: Releases intact tumor antigens and DAMPs via unique physical injury (ice crystal formation, cell rupture)[30-33,35]	Functions as an in situ vaccine, initiating systemic anti-tumor immunity. Antigen accumulation within DCs may be higher than with thermal ablation
	DAMP release and antigen-presenting cell activation: Releases key DAMPs (high mobility group box 1, adenosine triphosphate, calreticulin, heat shock proteins), promoting DC maturation and antigen presentation via receptors like Toll-like receptor 4 and P2X7[19,32,34,36]	Provides “danger signals”, bridging innate and adaptive immunity
	Initiation of adaptive immunity: DCs capture antigens, migrate to lymph nodes, and activate tumor-specific CD4+ and CD8+ T cells[13,14,20,35]	Preclinical models confirm increased numbers and activity of T cells and NK cells locally and systemically, with tumor-specific antibody production
	Effector T cell infiltration and abscopal effect: Activated CTLs home to tumors. Enrichment of CD8+ T cells and NK cells is also observed in untreated distant lesions[27,37-39]	Combination with immune adjuvants or anti-PD-L1 significantly enhances T cell infiltration and function, yielding synergistic activity
	Formation of immunological memory: Induces expansion of effector memory T cell and central memory T cell, particularly potent neoantigen-specific CD4+ T-cell responses[40-46]	Clinical studies show increased memory T cell subsets in peripheral blood post-procedure. Combination with immune checkpoint inhibitors or adjuvants significantly enhances the strength and durability of memory responses
Potential immunosuppression	Release of immunosuppressive cytokines: Necrotic areas release TGF-β, interleukin-10, vascular endothelial growth factor, etc. May induce more intense pro-inflammatory cytokine release compared to MWA/radiofrequency ablation[30,47-50]	TGF-β strongly inhibits CTL/NK function, promotes Treg differentiation, and M2 macrophage polarization
	Recruitment of immunosuppressive cells: Recruits myeloid-derived suppressor cells, M2-type tumor-associated macrophages, and regulatory T cells to the tumor site. May polarize tumor-associated neutrophils to an immunosuppressive N2 phenotype, forming physical barriers via NETs[20,37,50-53]	These cells inhibit effector T cells via mechanisms like depletion of essential amino acids and secretion of inhibitory factors. NET formation is a novel mechanism limiting efficacy
	Upregulation of immune checkpoints: As an adaptive resistance mechanism, induces upregulation of checkpoint molecules (e.g., programmed death-1/PD-L1) in distant tumor tissues[54]	May lead to rapid T cell exhaustion and limit the abscopal effect. Studies show it may induce higher PD-L1 upregulation than MWA
	Antigen overload and immune tolerance[55,56]	Massive short-term antigen release in a suppressive milieu may lead to T cell deletion/anergy, inducing immune tolerance

CTL: Cytotoxic T lymphocyte; DAMP: Damage-associated molecular pattern; DC: Dendritic cell; MWA: Microwave ablation; NET: Neutrophil extracellular trap; NK: Natural killer; PD-L1: Programmed death-ligand 1; TGF-β: Transforming growth factor-beta; Treg: Regulatory T cell.

Table 2 Comparison of immunomodulatory and clinical-technical properties among radiofrequency ablation, microwave ablation, and cryoablation

Comparison parameter		RFA	MWA	Cryoablation	Putative mechanisms and clinical implications
Immunological parameter	Primary damage-associated molecular pattern release	HMGB1, heat shock proteins	HMGB1, ATP	HMGB1, ATP, CALR[23,33] (possibly a broader spectrum)	The freezing process better preserves antigens and facilitates the in situ exposure of “eat-me” signals like CALR[30]
	Antigen preservation integrity	Moderate/high (high temperature may denature some epitopes)	Moderate (rapid high temperature may alter antigen conformation)	High (low temperature better maintains native antigen conformation[25,70]	More intact antigens may help elicit high-affinity T-cell responses
	T-cell infiltration trend	Can increase, but may coincide with substantial Treg recruitment	Can increase, with heterogeneous effects	Significant increase in CD8+ T cells, some studies show Treg reduction[25,37]	May be related to a more favorable cytokine/chemokine profile
	Immune checkpoint induction	Induces PD-L1 upregulation	Induces PD-L1 upregulation	Potently induces PD-L1 upregulation[54,58]	Provides a clear target for combination with immune checkpoint inhibitors, but also suggests limited efficacy as monotherapy
	Risk to adjacent vasculature/bile ducts	High (thermal injury)	High (thermal injury)	Low (less affected by “heat sink” effect, collagen structure preserved)[24,76]	Makes cryoablation more suitable for tumors in high-risk locations, allowing more aggressive ablation for antigen release[80,81]
	Reported clinical abscopal effect	Case reports[65]	Rare	Relatively more preclinical evidence[28,38] and case reports[82,83]	Suggests its potential for inducing systemic immunity might be more pronounced
Technical and clinical parameters	Local efficacy for large tumors (> 3-4 cm)	Limited by “heat sink” effect	Generally more effective, less susceptible to “heat sink” effect[76,77]	Technically challenging; ice ball growth constrained by perfusion, leading to less predictable margins[76,77]	Durable local control is the prerequisite for any immune benefit. MWA may offer advantages in ablating large volumes
	Comparative LTP rate	Baseline standard, variable	Often comparable or superior to RFA in studies	Some studies suggest a potentially higher LTP rate, though data are inconsistent and influenced by learning curve[78,79]	Highlights the critical importance of operator expertise and optimal technique in cryoablation for foundational oncologic outcomes
	Technical complexity and cost	Relatively simple, lower cost	Moderately complex, intermediate cost	More complex and costly; requires specialized gas systems and often multi-probe setups[21,22]	Influences the learning curve and may limit widespread availability in resource-constrained settings

It summarizes key comparative findings from preclinical studies; clinical validation of these observations is ongoing. Representative references are indicated by superscript numbers. Risk assessments (low/high) denote the relative probability of major complications. The “putative mechanisms” column integrates hypotheses from the cited literature to explain the differences observed. When used in this table and the text, the terms “M1/M2” or “M2-like” serve as simplified descriptive shorthand for pro-inflammatory/anti-tumor and immunosuppressive/pro-tumor functional states, respectively. It is acknowledged that macrophage phenotypes in vivo are heterogeneous, plastic, and context-dependent. ATP: Adenosine triphosphate; CALR: Calreticulin; HMGB1: High mobility group box 1; LTP: Local tumor progression; MWA: Microwave ablation; PD-L1: Programmed death-ligand 1; RFA: Radiofrequency ablation; Treg: Regulatory T cell.

Table 3 Summary of selected clinical studies on cryoablation combined with immunotherapy for liver cancers

Design	Patient population (line of therapy)	Intervention details	Key efficacy and survival outcomes (assessment criteria)	Key baseline confounders (tumor burden, PVTT, liver function)	Key immunological findings	Ref.
Case report	Advanced HCC, post-multiline therapy (post-resection, post-lenvatinib failure)	Bevacizumab + immune checkpoint inhibitors + intratumoral cryoablation (of a single metastasis)	mRECIST, sustained CR lasting > 24 months	Tumor burden: High (multifocal intrahepatic metastases + lymph node involvement); PVTT: Yes (noted in primary tumor); liver function: NR (HBV-positive, well-controlled on therapy)	Tumor mutational burden increased from 3 Muts/Mb to 18.67 Muts/Mb post-cryoablation (biopsy of a separate, non-ablated lesion)	Li et al[91], 2021
Case report	Advanced, multifocal HCC, post-sorafenib and nivolumab (human immunodeficiency virus/HBV co-infection)	Cryoablation (partial, of 2 sites) + nivolumab - liver transplant	The mRECIST, durable CR; bridged to successful liver transplantation > 4 years post-treatment	Tumor burden: High (multifocal, bilobar, infiltrative progression); PVTT: NR; liver function: Cirrhosis (Child-Pugh score NR, HBV-DNA undetectable)	Potent “abscopal effect”: Cryoablation of partial lesions triggered a systemic immune response, leading to regression of untreated, multifocal intrahepatic tumors and reversal of immunotherapy resistance	Lucas et al[92], 2024
Phase II study (preliminary results) NCT04724226	Advanced HCC with PVTT; first-line therapy	Cryoablation (to inactivate lesions as much as possible) –within 48 hours, start combination therapy with camrelizumab (anti-PD-1) + apatinib	ORR (mRECIST): 71.4%; ORR (RECIST v1.1): 14.3%; mPFS (mRECIST): 4.63 months; mOS: 19.0 months	Tumor burden: Limited (“up-to-7” criteria); PVTT: Yes (100% of patients, 57.1% with Vp3/4); liver function: Child-Pugh A (all patients, scores of 5 or 6)	Increase in CD8+ T cells; decrease in regulatory T cells and myeloid-derived suppressor cells	Gao et al[82], 2025
CASTLE-01 single-arm, phase II NCT05010668	Locally advanced or metastatic ICC; post first-line gemcitabine + cisplatin (second-line)	Partial cryoablation of one intrahepatic lesion, followed by combination therapy with sintilimab (anti-PD-1) + lenvatinib	ORR: 75.0% (21/28, including 2 CR); disease control rate: 100%; mPFS: 16.8 months; mOS: 25.4 months (RECIST v1.1)	Tumor burden: High (79% TNM stage IV, advanced/metastatic); PVTT: Not specifically reported; liver function: Predominantly well-preserved (all enrolled had Child-Pugh class A; 11% had cirrhosis)	Cryoablation triggered recruitment and clonal expansion of CD8+ PD-1hi effector T cells into the TME and enhanced tumor immunogenicity (increased antigen presentation and IFN signaling). Lenvatinib promoted tumor vasculature normalization (increased postcapillary venule ECs, decreased angiogenic tip/stalk ECs), facilitating the influx of novel T cell clones. Combination therapy reshaped a “cold” TME into an inflamed “hot” one, increased CD4+ CXCL13+ T follicular helper cells, and was associated with tertiary lymphoid structures formation	Gu et al[83], 2026
Single-arm study NCT03183219	Advanced HCC/ICC	Locoregional therapy (incl. cryoablation) + allogeneic γδ T-cell adoptive transfer	HCC: The mPFS 8.0 months vs 4.0 months (combo vs mono), mOS 13.0 months vs 8.0 months; ICC: MPFS 8.0 months vs 4.0 months	Mixed population (HCC and ICC); detailed burden/function: NR	Expansion/persistence of donor γδ T cells; elevated serum IFN-γ, tumor necrosis factor-alpha	Zhang et al[93], 2022
Retrospective study	Metastatic HCC; advanced patients unsuitable for or refusing surgery/chemotherapy (not first-line, mostly after prior treatments)	Four groups compared: (1) Comprehensive cryoablation + immunotherapy (cryo-immunotherapy): Complete cryoablation of intrahepatic primary and accessible extrahepatic metastatic lesions, followed by adoptive dendritic cell and cytokine-induced killer cell immunotherapy cell immunotherapy (4 infusions); (2) Cryotherapy only group; (3) Immunotherapy only group; and (4) Untreated/supportive care group	The mOS: Cryo-immunotherapy group: 32.0 months; cryotherapy only group: 17.5 months; immunotherapy only group: 4.0 months; untreated group: 3.0 months; statistical significance: Overall survival in the cryo-immunotherapy group was significantly longer than in the cryotherapy only group (P = 0.024) and the untreated group (P < 0.01; assessment criteria: Revised RECIST 1.1)	Tumor burden: High. All patients had metastatic disease (bone, lung, or multiple organs). High intrahepatic primary tumor burden (24 cases with single lesion, avg. diameter 6.5 cm; 21 cases with multiple lesions, total 71 lesions). PVTT: Not explicitly mentioned. Liver function: All patients were Child-Pugh class A (25 cases) or B (18 cases); 43/45 (95.6%) had cirrhosis	In the combination therapy group (cryo-immunotherapy), patients showed an increased proportion of CD3+ CD4+ T cells in peripheral blood and elevated serum levels of interleukin-2 and IFN-γ	Niu et al[94], 2013

Efficacy was assessed according to modified Response Evaluation Criteria in Solid Tumors for hepatocellular carcinoma unless stated otherwise. Sample size (n) indicates the number of patients or evaluable lesions. Survival outcomes (median progression-free survival, median overall survival) are expressed in months. CR: Complete response; HBV: Hepatitis B virus; HCC: Hepatocellular carcinoma; ICC: Intrahepatic cholangiocarcinoma; IFN: Interferon; mOS: Median overall survival; mPFS: Median progression-free survival; mRECIST: Modified Response Evaluation Criteria in Solid Tumors; NR: Not reported; ORR: Objective response rate; PD-1: Programmed death-1; PVTT: Portal vein tumor thrombus; RECIST: Response Evaluation Criteria in Solid Tumors; TME: Tumor microenvironment.

Table 4 Summary of key preclinical studies on cryoablation combined with immunomodulatory strategies

Ref.	Disease model	Ablation completeness	Combination regimen and sequencing	Sample size	Key immunological and therapeutic outcomes
Mandt et al[27], 2023	Murine HCC	Incomplete	CpG (Toll-like receptor 9 agonist) + αPD-1, administered after cryoablation	n = 63 total	Synergistically enhanced systemic antitumor immunity, characterized by increased intratumoral CD8+ T cell infiltration and a robust Th1-type cytokine response, leading to inhibited distal tumor growth and prolonged survival
Gu et al[28], 2024	Murine HCC	Complete	Anti-PD-1 + anti-CTLA-4 (dual immune checkpoint inhibitor), concurrent with/after cryoablation	n = 40 total, n = 10/group	Most effectively remodeled the TME by promoting CD8+ and CD4+ T cell infiltration while reducing regulatory T cells and myeloid-derived suppressor cells, resulting in potent inhibition of distant tumors and a synergistic effect
Qian et al[29], 2024	Bilateral subcutaneous HCC	Complete	αPD-1 + αCTLA-4, after cryo-thermal ablation	Not specified	Converted immunologically “cold” tumors to “hot” by inducing immunogenic cell death and enhancing DC activation, with the triple combination eliciting the strongest abscopal effect
Shewarega et al[37], 2024	Murine HCC	Incomplete	Matrix metalloproteinase inhibitor, administered after cryoablation	n = 40 total, n = 6/group	Specifically increased intratumoral CD8+ T cell infiltration, providing a model to study the immune effects of subtotal ablation
Ghani et al[38], 2023	Murine HCC	Incomplete	CPMV priming, followed by cryoablation, and then a CPMV boost	n = 44-56 total, n = 11-14/group	Functioned as an in situ vaccine, where CPMV acted as a potent innate immune stimulant, synergizing with cryoablation-released antigens to enhance antigen-presenting cell recruitment and cross-priming
Yang et al[89], 2025	Murine cervical cancer	Complete	αPD-1, administered after cryoablation	Bilateral model; typical n = 3-4/group for assays	Induced an effective abscopal effect by remodeling the TME, increasing cytotoxic lymphocyte activity, and upregulating programmed death-ligand 1 expression, thereby sensitizing tumors to checkpoint blockade
Wang and Guo[90], 2025	Murine colorectal cancer liver metastasis	Complete (described as tumor fully ice-covered)	Cryoablation + granulocyte-macrophage colony-stimulating factor (administered immediately post-ablation)	n = 80 total, n = 20/group	Enhanced antitumor immune response: Significantly and persistently increased intratumoral infiltration of CD11c+ DCs; promoted a Th1-biased immune response with sustained high serum interferon-gamma levels and suppressed interleukin-4 levels; associated with prolonged survival in the combination group

CPMV: Cowpea mosaic virus; CTL: Cytotoxic T lymphocyte; DC: Dendritic cell; HCC: Hepatocellular carcinoma; PD-1: Programmed death-1; TME: Tumor microenvironment.

Table 5 Summary of the multifactorial determinants influencing outcomes and the current translational hurdles for cryoablation combined with immunotherapy in hepatocellular carcinoma

Category	Core variable/challenge	Specific elements and impact	Clinical implication/optimization direction
Key efficacy variables	Technology-related variables[37,78,94,103-111]	Ablation mode and completeness: Contradiction between preclinical “incomplete ablation” for immune stimulation and clinical need for local control. Current consensus favors radiologically complete ablation when combined with effective systemic immunotherapy. Technical parameters: Probe layout, freezing rate, minimum temperature, and number of freeze-thaw cycles influence cell death mode, damage-associated molecular pattern release, and the intensity/quality of the immune response	Paradigm shift from “radical ablation” to “strategic immuno-ablation”. Standardize and optimize parameters to maximize the in situ vaccine effect and immune benefits
	Baseline host and tumor variables[103,112-117]	Host status: Systemic immune capacity, immunodeficiency, severity of background liver disease (Child-Pugh grade, albumin-bilirubin score), cirrhosis, and hepatitis B virus/hepatitis C virus infection status. Tumor features: Intrinsic immunogenicity (tumor mutational burden, neoantigen load), baseline immune cell infiltration level (immune score), programmed death-ligand 1 expression. Conventional metrics like tumor burden and alpha-fetoprotein levels also correlate with efficacy	Foundation for personalized therapy. Pre-identification of “advantaged populations” via molecular imaging or liquid biopsy is key. The combination may potentiate the conversion of “immune-cold” tumors
	Treatment strategy and dynamic monitoring variables[27,94,118]	Timing and sequence of combination: The sequence (neoadjuvant, concurrent, adjuvant) and interval are core points of controversy for strategy optimization. Management of immunosuppressive complications: Complications (e.g., massive pleural effusion, “cryoshock”) may induce systemic immunosuppression. Dynamic peripheral blood immune monitoring: Dynamic changes in lymphocyte subsets, neutrophil-to-lymphocyte ratio, T-cell receptor clonal expansion, and cytokine profiles serve as a real-time, non-invasive “liquid biopsy” window	Dynamic strategy adjustment based on baseline conditions. Active complication management to preserve efficacy. Utilize blood-based monitoring for early response assessment and efficacy prediction
Current major challenges	Lack of high-level prospective evidence[78,118,119]	Current evidence is primarily from retrospective analyses or small single-arm studies, limited by selection bias, confounding factors, and protocol heterogeneity. Conclusions are hypothesis-generating	Future phase III randomized controlled trials must employ prospective randomized design, stratified randomization, standardized protocols, and appropriate control arms (e.g., atezolizumab + bevacizumab) to confirm survival benefit
	Immunosuppressive risk and combination safety[118]	Ablation itself may stimulate immunosuppression. Overlapping toxicities (e.g., immune-related hepatitis vs post-ablation injury) require careful management. Significant heterogeneity exists in patient selection	Need for strategies to counteract ablation-induced immunosuppression and to differentiate/manage overlapping toxicities. Establishing universal patient selection criteria is difficult
	Evolution of efficacy evaluation standards[120]	Traditional size-based criteria (e.g., Response Evaluation Criteria in Solid Tumors) may fail to capture delayed responses or pseudoprogression induced by immunotherapy	Comprehensive evaluation incorporating functional imaging (e.g., contrast-enhanced magnetic resonance imaging, positron emission tomography) and immune-related response criteria is necessary

Table 6 Proposed minimal marker set and endpoints for assessing macrophage reprogramming in preclinical and translational studies

Category	Key markers / endpoints	Rationale and interpretation	Suggested supporting reference
Phenotypic and characteristic markers	Pro-inflammatory/anti-tumor (M1-like) phenotype: Surface: CD86, MHC-II (human leukocyte antigen-DR isotype); Cytokine: IL-12high, TNF-αhigh	Indicates antigen-presenting capacity and Th1-type immune activation. Increased ratio of these markers to M2 markers suggests successful reprogramming	CD86/MHC-II: Noy et al[63] and De Palma et al[64]; IL-12/TNF-α: Mills et al[59] and Wynn et al[60]
	Immunosuppressive/pro-tumor (M2-like) phenotype: Surface: CD206 (MRC1), CD163; enzyme: Arg1; immune checkpoint: PD-L1	Associated with T-cell suppression, tissue repair, and angiogenesis. Reduction post-treatment indicates attenuation of immunosuppressive tumor microenvironment	CD206/Arg1: Mills et al[59] and Wynn et al[60]; PD-L1: Tan et al[54]
	Phagocytic and “eat-me” signals: Phagocytosis receptor: FcγR; pro-phagocytic signal: CALR exposure on tumor cells	Essential for antibody-dependent cellular phagocytosis and immunogenic cell death. Upregulation enhances tumor cell clearance	CALR/phagocytosis: Fucikova et al[32] and Zhou et al[34]
Functional and prognostic endpoints	In vitro co-culture suppression assay: Inhibition of T-cell proliferation or interferon-gamma production	Direct functional readout of macrophage-mediated immunosuppression. Decreased suppression indicates functional reprogramming	Mandt et al[27], Shewarega et al[37], and Wang et al[50]
	Spatial context (multiplex immunohistochemistry/immunofluorescence): Co-localization with CD8+ T cells (permissive vs excluded)	Defines the physical interaction between macrophages and effector cells, critical for predicting immunotherapy response	Shewarega et al[37] and Santana et al[121]
	Correlation with treatment outcome: Inverse correlation with CD8+ T-cell infiltration: Direct correlation with tumor growth or survival in vivo	Validates the clinical relevance of the macrophage phenotype. Successful reprogramming should correlate with improved T-cell infiltration and survival	Mauda-Havakuk et al[20], Tan et al[54], and Gu et al[83]

It proposes a framework for standardizing the assessment of macrophage polarization states beyond the simplistic M1/M2 dichotomy, focusing on functional relevance in the context of cryoablation and immunotherapy combinations. These markers and endpoints are recommended for use in preclinical studies (murine models) and translational research (analysis of human biopsy samples pre- and post-treatment) to evaluate the efficacy of combination strategies in reprogramming the immunosuppressive macrophage compartment. A shift in the balance or ratio of pro-inflammatory to immunosuppressive markers is often more informative than absolute changes in a single marker. Spatial analysis (e.g., via multiplex imaging) is crucial for understanding the functional impact within the tissue architecture. Arg1: Arginase-1; CALR: Calreticulin; IL: Interleukin; MHC: Major histocompatibility complex; PD-L1: Programmed death-ligand 1; TNF-α: Tumor necrosis factor-alpha.