Safety of bronchial artery infusion immunotherapy: from comparative analysis in beagle canines to clinical validation

Abstract

Background

Despite advancements in systemic chemotherapy and immune checkpoint inhibitors (ICIs), advanced non-small cell lung cancer (NSCLC) continues to exhibit poor prognosis, underscoring an urgent need for safer and more effective therapeutic strategies. This study investigates the safety profile and biological effects of bronchial arterial infusion (BAI)-administered anti-PD-1 monoclonal antibody (aPD-1 mAb) using a preclinical beagle model and a clinical cohort of advanced NSCLC patients.

Methods

In preclinical evaluations, male beagles (n = 3/group) were randomized to receive 5 mg/kg aPD-1 mAb via BAI or intravenous routes (Venous group). Safety assessments included longitudinal imaging, biochemical analyses, and histopathological evaluation. Clinically, patients with advanced NSCLC meeting stringent inclusion criteria underwent BAI immunotherapy, with systematic monitoring of adverse events (AEs).

Results

Both administration routes demonstrated comparable safety in canines, with no evidence of immune-related pneumonitis or structural lung alterations on CT or histology. Transient AEs (e.g., hematoma, lameness) resolved spontaneously. Pharmacokinetic analysis revealed similar systemic drug concentrations and tissue distribution between BAI and Venous groups (all p > 0.05). Biochemical profiling identified isolated mild LDH elevation in one BAI-treated canine. Notably, the BAI group exhibited significantly enhanced systemic IL-2 levels (80.15 ± 5.24 pg/mL vs. 66.47 ± 5.24 pg/mL in Venous groups, p = 0.001) at day 28, paralleled by elevated pulmonary IL-2 expression (626.90 ± 18.49 vs. 559.18 ± 45.61 pg/mg, p = 0.03). In the clinical cohort (n = 17; 94.1% male, mean age 61.6 ± 7.1 years), BAI immunotherapy was well-tolerated with mild AEs including nausea (n = 1), dyspnea (n = 1), atrial fibrillation (n = 1), and puncture-site hematoma (n = 1). No severe immune-related toxicities (e.g., pneumonitis) emerged during follow-up.

Conclusion

Our study suggest the preliminary safety and feasibility of delivering aPD-1 mAb via BAI in both canine models and NSCLC patients.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12672-025-02398-2.

Introduction

Primary lung cancer remains the leading cause of cancer-related mortality worldwide, with non-small cell lung cancer (NSCLC) accounting for 85% of these fatalities [1, 2]. The insidious onset and nonspecific symptomatology result in most cases being diagnosed at advanced stages [3]. For inoperable advanced lung cancer, chemotherapy or chemoradiotherapy continues to serve as primary therapeutic modalities [4]. Tyrosine kinase inhibitors (TKIs) have been extensively employed in managing advanced NSCLC, demonstrating substantial prognostic improvements [5]. In patients lacking actionable oncogenic drivers, immune checkpoint inhibitors (ICIs) targeting programmed death 1 (PD-1) or its ligand PD-L1 have emerged as promising therapeutic options [6].

Recent clinical investigations revealed that pembrolizumab monotherapy achieved a 5-year survival rate of 31.9% in PD-L1-high patients (tumor proportion score [TPS] ≥ 50%), significantly surpassing the 16.3% observed in chemotherapy cohorts [7]. Furthermore, overall survival (OS) benefits were observed across PD-L1 expression subgroups, with pembrolizumab-chemotherapy combination demonstrating a 5-year OS rate of 19.4% versus 11.3% for placebo-chemotherapy controls [8]. Meta-analyses corroborated that PD-1/PD-L1 inhibitors substantially prolong survival in advanced NSCLC patients [9, 10].

ICI-based regimens, whether as monotherapy or combined with chemotherapy, have shown potential in enhancing advanced NSCLC outcomes [8, 11]. Nevertheless, a significant proportion of patients exhibit limited therapeutic response or develop acquired resistance to ICIs [12, 13]. Clinical data indicate objective response rates (ORR) of 33–46% for ICI monotherapy in PD-L1-high populations [14, 15], while combination immunochemotherapy achieves ORRs of 45–75% across broader populations [16, 17]. Systemic administration may lead to suboptimal intratumoral drug concentrations while increasing risks of systemic toxicities including diarrhea, hepatic dysfunction, and pneumonitis [18, 19]. These limitations underscore the urgent need for novel therapeutic strategies that optimize efficacy while mitigating adverse effects.

The bronchial artery, serving as a primary vascular supply for pulmonary malignancies, provides a rational anatomical target for bronchial artery infusion (BAI) chemotherapy [20]. Pharmacokinetic studies demonstrate that BAI administration achieves superior intratumoral cytotoxic drug concentrations compared to systemic delivery, yielding favorable clinical outcomes in advanced NSCLC patients [20, 21]. This regional approach concurrently reduces systemic exposure and associated toxicities, thereby improving quality of life [22]. These findings highlight the critical importance of exploring complementary therapeutic approaches like BAI to enhance treatment efficacy while minimizing adverse events.

The mechanistic complexity of ICIs extends beyond simple PD-1/PD-L1 blockade. Preclinical models reveal that sustained antitumor immunity requires coordinated interactions between dendritic cell-mediated antigen presentation, CD8⁺ T cell clonal expansion, and dynamic remodeling of the immunosuppressive tumor microenvironment (TME) [23]. Systemic ICI administration risks off-target T cell activation in peripheral tissues, potentially exacerbating immune-related adverse events (irAEs) [24, 25]. Intriguingly, murine models of breast cancer and melanoma demonstrate that localized aPD-1 mAb delivery significantly modulates immune cell infiltration patterns within both tumor and peritumoral microenvironments, suggesting enhanced therapeutic specificity [23]. Notably, there are currently no animal studies investigating local treatment via bronchial artery infusion of ICIs. The anatomical dominance of bronchial arteries in tumor neovascularization positions BAI as a spatially optimized delivery strategy, potentially amplifying intratumoral CD8⁺ T cell infiltration while attenuating systemic toxicity [26]. However, this hypothesis remains speculative and currently lacks robust clinical validation.

Therefore, our study aimed to explore the safety and possible biological impacts of Tislelizumab (an anti-PD-1 monoclonal antibody, aPD-1 mAb) administered via BAI by utilizing animal models. Furthermore, we examined the safety of administering aPD-1 mAb through BAI in patients with advanced NSCLC.

Methods

Animals and experimental models

All animal experiments were conducted in strict accordance with the National Institutes of Health (NIH) guidelines for the care and use of laboratory animals, and the study was approved by the Institutional Animal Care and Use Committee (IACUC) of The First Affiliated Hospital of Chongqing Medical University (Approval No. K2023-505). The protocols followed the ARRIVE guidelines to ensure the welfare of the animals. Healthy adult male beagle canines, provided by Chongqing Watson Biotechnology Company Limited, with body weights ranging from 10 to 12 kg, were selected for this study. All animals were fasted for 12 h prior to the experiment, and water was withheld for 4 h. The Beagles were randomly assigned to two experimental groups: the BAI group and the Venous group, with n = 3 for each group.

Prior to the experiment, anesthesia was induced via intramuscular injection of Zoletil at a dosage of 10 mg per kilogram of body weight. Once anesthesia was achieved, the Beagles were placed on the surgical table in the Digital Subtraction Angiography (DSA) suite in a supine position, with their heads and limbs secured. Femoral artery puncture was performed, and a 2.1 F bronchial artery microcatheter was used to locate the bronchial artery that shares a common trunk with the intercostal artery. After identifying the target artery, BAI was conducted: aPD-1 mAb, dissolved at a dosage of 5 mg per kilogram of body weight, was slowly infused over 30 min. Control animals were administered aPD-1 mAb via intravenous drip through the saphenous vein of the hind limb (5 mg per kilogram of body weight), with a cycle every 21 days for a total of two cycles.

Comprehensive safety assessments were conducted at predefined timepoints throughout the study. Thoracic computed tomography (CT) scans were performed at baseline (Day 0) and following two immunotherapy cycles (Day 42) to evaluate pulmonary integrity. Systemic safety parameters including complete blood count, hepatic function (alanine aminotransferase [ALT], aspartate aminotransferase [AST]), renal profile (blood urea nitrogen [BUN], creatinine), myocardial enzymes (creatine kinase [CK], lactate dehydrogenase [LDH]), and glucose metabolism were analyzed pre-treatment and 7 days post-treatment for each immunotherapy cycle. The flowchart is shown in Fig. 1.

Fig. 1.

A Flowchart of animals experiment and B BAI immunotherapy. aPD-1 mAb anti-PD-1 monoclonal antibody, BAI bronchial artery infusion, BR blood routine, LFTs liver function tests, RF renal function, HE hematoxylin and eosin

Bronchial artery infusion immunotherapy in patients

This study involving human participants was conducted in accordance with the World Medical Association Declaration of Helsinki (2013) and approved by the Ethics Committee of The First Affiliated Hospital of Chongqing Medical University (Approval No. 2020-132). Written informed consent was obtained from all participants prior to enrollment after screening hospitalized patients against predefined criteria between July 2023 and March 2024.

Inclusion criteria included: (1) age between 18 and 75 years; (2) histopathologically or cytologically confirmed NSCLC, with stage III unresectable or stage IV disease confirmed by chest CT, head CT or MRI, whole-body bone scan, and abdominal ultrasound; if stage III NSCLC, the lesion was assessed as unresectable by an experienced surgeon; (3) at least one measurable lesion as a target lesion (according to the Response Evaluation Criteria in Solid Tumors version 1.1, RECIST 1.1); (4) lack of EGFR sensitizing mutations (i.e., exon 19 deletion, or exon 21 L858R, exon 21 L861Q, exon 18 G719X, or exon 20 S768I mutations), ALK rearrangement, and ROS1 positivity; (5) confirmed independent blood supply to the primary lesion via target bronchial arteries by bronchial artery CTA or enhanced CT; (6) patients who did not achieve partial response(PR) after 2 or 4 cycles of standard first-line immunotherapy combined with chemotherapy or had progressive disease(PD); (7) estimated survival time of at least 12 weeks; (8) Eastern Cooperative Oncology Group (ECOG) performance status score of 0–2; (9) no severe heart, brain, lung, or kidney diseases that may affect the patient's survival time; (10) agreement to participate in this clinical study and signed informed consent.

Exclusion criteria included: (1) inability to tolerate surgery due to cardiac or pulmonary dysfunction; (2) abnormal liver or kidney function, making the patient unable to tolerate surgery; (3) presence of severe heart, brain, lung, or kidney diseases that may affect the patient's survival time; (4) presence of other severe diseases that may prevent the implementation of bronchial artery infusion treatment, such as unstable angina, acute myocardial infarction within the past six months, or congestive heart failure; (5) history of other malignant tumors within the past five years; (6) contraindications to bronchial artery infusion treatment, such as severe coagulation dysfunction, uncontrolled local infection at the puncture site, or contrast agent allergy; (7) contraindications to PD-1 inhibitor treatment; (8) patients with symptomatic brain metastases, or a history of spinal cord compression and meningeal metastases; (9) disagreement to participate in this clinical study.

During the surgical procedure, patients underwent femoral artery puncture using the modified Seldinger technique. A 5 F-Cobra catheter (Terumo Corporation, Tokyo, Japan) was used for bronchial artery angiography to identify the blood supply artery of the tumor. Vascular angiography was performed on other possible blood supply arteries, such as the internal thoracic artery, inferior phrenic artery, and esophageal intrinsic artery, to confirm the tumor blood supply artery. The target artery was coaxially inserted through a 2.7-F microcatheter (APT Medical, Shenzhen, China). aPD-1 mAb was dissolved in 50 ml of 0.9% saline and infused at a rate of 2.5 ml per minute using an infusion pump, with a treatment cycle every 3–4 weeks. All surgical procedures were performed by physicians with more than ten years of experience in interventional treatment. Study data included demographic information, details of BAI surgery, imaging data, interventional surgery records, and follow-up results.

Hematoxylin–eosin (H&E)

Following a 12-h fasting period, experimental beagles underwent general anesthesia via intravenous administration of appropriate anesthetic agents. Humane euthanasia was performed through carotid artery exsanguination under confirmed pain-free conditions. Thoracoabdominal organs (heart, liver, spleen, lungs, kidneys) were rapidly harvested postmortem and immersion-fixed in 4% neutral-buffered formalin for ≥ 24 h. Fixed tissues underwent standardized processing involving graded ethanol dehydration (70–100%), xylene clearing, paraffin impregnation, and embedding. Tissue blocks were sectioned at 5 μm thickness using a rotary microtome (Leica RM2235) and mounted on charged glass slides. After sequential dewaxing in xylene and rehydration through descending ethanol series, sections were stained with hematoxylin (5 min) and eosin (1 min) for morphological evaluation. Brightfield microscopy (Olympus BX53) with × 100–400 magnification was employed for systematic assessment of tissue architecture, cellular integrity, and pathological alterations.

Immunohistochemistry

Paraffin-embedded sections were deparaffinized, rehydrated, underwent antigen retrieval, and were then incubated with primary and secondary antibodies, followed by observation with 3′-diaminobenzidine. Subsequently, the sections were counterstained with hematoxylin, differentiated, dehydrated, and imaged under a light microscope (Nikon). The antibodies are summarized in Supplementary file: Table 1.

Quantitative real-time PCR (qRT-PCR)

Total RNA was extracted from individual lung tissues and was reverse-transcribed to complementary DNA using a Reverse Transcription Kit (Takara, China). The synthesized cDNA was then subjected to qRT-PCR using a SYBR Green PCR kit (Takara, China) with the following program: 95 °C for 30 s, 40 cycles of 95 °C for 5 s, and 60 °C for 30 s. GAPDH was used as the internal reference. The relative RNA level was evaluated by the 2^{− ΔΔCt} method. The primer sequences are shown in Supplementary file: Table 2.

Enzyme-linked immunosorbent assay (ELISA)

Serum and lung tissue levels of interferon-gamma (IFN-γ) and interleukin-2 (IL-2) in Beagles were quantified using ELISA kits procured from Jiubang Biotech (China), according to the manufacturer’s recommended protocols. Additionally, the systemic distribution of tislelizumab was assessed by measuring its concentrations in the serum, lung, heart, liver, spleen, and kidney of Beagles, using a specific ELISA kit provided by Jiubang Biotech.

Statistical analysis

All data were analyzed using Graphpad Prism 10.0 software and are presented as the mean ± SD. Data obeying normal distribution and homogeneity of variance between two groups were compared using unpaired t-tests. p < 0.05 was considered significant. *p < 0.05, **p < 0.01, and ***p < 0.001, ns p > 0.5.

Results

Bronchial artery angiography and physiological monitoring

Angiography revealed that most canine bronchial arteries originated from a common trunk with the sixth intercostal artery, displaying characteristic tortuous trajectories along pulmonary hila and primary bronchi (Fig. 2). Systemic monitoring revealed stable thermoregulation in all subjects post-immunotherapy, with body temperatures consistently maintained within the physiological range (37.9–39.3 °C; Fig. 3A). Baseline body weights showed no intergroup differences (Venous: 11.07 ± 0.67 kg vs BAI: 10.87 ± 0.23 kg; p = 0.669), a pattern that persisted after two treatment cycles (Day 42: Venous 11.57 ± 0.06 kg vs BAI 11.03 ± 0.45 kg; p = 0.112; Fig. 3B).

Fig. 2.

Bronchial artery angiography analysis. A–C Depict the first bronchial artery angiography, revealing slender and tortuous bronchial arteries primarily distributed along the hilum of the lung and around the main bronchi. D–F Show the second bronchial artery angiography, highlighting similar characteristics

Fig. 3.

Dynamic monitoring of body temperature and body weight in beagles. A Changes in body temperature and B body weight following aPD-1 mAb administration via intravenous injection and bronchial artery infusion. aPD-1 mAb anti-PD-1 monoclonal antibody, BAI bronchial artery infusion

Treatment-related adverse events in canine models

Following initial aPD-1 mAb administration, transient adverse events were observed including self-limiting diarrhea in the Venous group (n = 1), while the BAI cohort exhibited lameness (n = 1) and unilateral conjunctival erythema (n = 1) that resolved spontaneously with topical erythromycin ointment. Post-second cycle evaluation revealed procedure-related complications consisting of a puncture-site hematoma with concurrent appetite suppression in the Venous group (n = 1), contrasted by an isolated injection-site hematoma in the BAI group (n = 1). Comprehensive documentation of these events is presented in Table 1.

Table 1.

Adverse events in BAI immunotherapy in Beagle canines

Characteristics	Cycle 1		Cycle 2
	Venous	BAI	Venous	BAI
Nausea/vomiting	0/3	0/3	0/3	0/3
Diarrhea	1/3	0/3	0/3	0/3
Decreased appetite	0/3	0/3	1/3	0/3
Cough	0/3	0/3	0/3	0/3
Fever	0/3	0/3	0/3	0/3
lameness	0/3	1/3	0/3	0/3
Puncture site hematoma	0/3	0/3	1/3	1/3
Increased AST	1/3	0/3	2/3	0/3
Increased ALT	1/3	0/3	1/3	0/3
Increased CK	0/3	0/3	0/3	0/3
Increased LDH	1/3	1/3	2/3	0/3

BAI bronchial artery infusion, AST aspartate aminotransferase, ALT alanine aminotransferase, CK creatine kinase, LDH lactate dehydrogenase

Blood biochemical profiles, imaging, and histological assessment

All Beagles in both cohorts demonstrated baseline values within normal reference ranges for hematological parameters (complete blood count), hepatic function (ALT, AST), renal function (BUN, creatinine), myocardial enzymes (CK, LDH), and glucose metabolism. Following the first treatment cycle, transient hepatic transaminase elevations were observed in the Venous group (concurrent AST/LDH elevation: n = 1; isolated ALT elevation: n = 1), while the BAI group exhibited a single case of LDH elevation (n = 1). After the second cycle, the Venous group developed multicomponent enzyme elevation (AST/ALT/LDH: n = 1), though all values remained below 2 × upper limit of normal (ULN). Notably, all biochemical parameters in the BAI group persisted within normal thresholds throughout the study (Fig. 4 and Table 1).

Fig. 4.

Hematological and biochemical parameters following aPD-1 mAb administration via intravenous injection and bronchial artery infusion in beagles. Statistical analysis was performed by student’s t-test. Statistical significance: ***p < 0.001. BAI bronchial artery infusion, RBC red blood cell count, HGB hemoglobin, WBC white blood cell count, PLT platelet count, T-protein total protein, AST aspartate aminotransferase, ALT alanine aminotransferase, ALP alkaline phosphatase, LDH lactate dehydrogenase, CK creatine kinase, SCr serum creatinine, BUN blood urea nitrogen, FBS fasting blood sugar

In this study, aPD-1 mAb was administered to Beagles via both venous and BAI routes. Longitudinal thoracic CT imaging was systematically performed pre- and post-intervention to evaluate potential treatment-related pulmonary structural alterations, with a focus on detecting immune-mediated pneumonitis. Baseline CT scans demonstrated normal pulmonary architecture in all subjects. Post-treatment imaging following two cycles of aPD-1 mAb administration revealed preserved lung morphology across all animals, with no radiographic evidence of alveolar infiltrates, nodular lesions, or interstitial abnormalities (Fig. 5A).

Fig. 5.

Chest imaging and histology of major organs post-aPD-1 mAb administration via intravenous injection and BAI in beagles. A Thoracic CT imaging of the lungs before treatment (Day 0) and after treatment (Day 42). Representative coronal/axial reconstructions demonstrate preserved pulmonary architecture in both Venous and BAI groups, with no radiographic evidence of fibrosis or inflammatory infiltration. B Histopathological evaluation by H&E staining. Representative micrographs from lung, heart, liver, spleen, and kidney specimens (Day 42) show intact tissue morphology. Scale bar represents 20 μm. Data representative of n = 3 biological replicates per group. BAI bronchial artery infusion, aPD-1 mAb anti-PD-1 monoclonal antibody, CT computed tomography, HE hematoxylin and eosin

Complementary histopathological analysis of multiple organs (heart, liver, spleen, lungs, kidneys) via H&E staining confirmed physiological tissue integrity. Pulmonary sections exhibited intact alveolar septa with physiological thickness, absence of inflammatory cell infiltration, and normal interstitial architecture without fibrotic remodeling. Systemic evaluation further revealed unremarkable histological profiles in cardiac, hepatic, splenic, and renal tissues. Critically, characteristic histopathological markers of immune-mediated pulmonary injury were universally absent across all specimens (Fig. 5B).

Pharmacokinetic and pharmacodynamic profiling

During initial treatment cycle, serum aPD-1 mAb concentrations showed comparable pharmacokinetics between administration routes, with marginally higher levels in the BAI group at 1-h post-dose (20.05 ± 1.07 vs 18.22 ± 0.71 μg/ml; p = 0.07). By 24-h post-administration, systemic exposure increased comparably in both groups (Venous: 22.60 ± 1.02 μg/ml vs BAI: 23.54 ± 0.67 μg/ml; p = 0.253), a pattern maintained during subsequent treatment cycles (Fig. 6A). Biodistribution analysis at study termination (Day 42) revealed equivalent aPD-1 mAb accumulation across examined organs (all p > 0.05; Fig. 6B).

Fig. 6.

Pharmacokinetic and cytokine profiling after aPD-1 mAb delivery through intravenous injection versus BAI. A Serum aPD-1 mAb concentrations measured by ELISA at 1 h and 24 h post-administration. B Tissue distribution of aPD-1 mAb in major organs (heart, liver, spleen, lung, kidney) on day 42, comparing the i.v. and BAI groups. C Time-dependent changes in serum IFN-γ levels quantified by ELISA. D IFNG mRNA expression in lung tissue (qRT-PCR) and E corresponding protein levels (ELISA) on day 42. F Temporal dynamics of serum IL-2 levels (ELISA). G IL-2 mRNA expression in lung tissue (qRT-PCR) and H protein quantification (ELISA) on day 42. Data are presented as mean ± SD. Statistical analysis was performed by student’s t-test. Statistical significance: *p < 0.05, **p < 0.01, ns, not significant (p > 0.05). BAI bronchial artery infusion, aPD-1 mAb anti-PD-1 monoclonal antibody, IFN-γ interferon-gamma, ELISA enzyme-linked immunosorbent assay, IL-2 interleukin-2, mRNA messenger RNA, qRT-PCR quantitative reverse transcription polymerase chain reaction, SD standard deviation

Cytokine analyses demonstrated route-dependent IL-2 modulation despite conserved IFN-γ responses. Systemic IFN-γ levels showed non-significant elevation with BAI administration (all p > 0.05; Fig. 6C), while pulmonary IFN-γ expression remained comparable at both transcriptional and translational levels (Fig. 6D, E). Strikingly, BAI-treated animals exhibited enhanced IL-2 bioavailability, demonstrating 20.5% higher serum concentrations versus venous controls at Day 28 (80.15 ± 5.24 vs 66.47 ± 5.24 pg/ml; p = 0.001; Fig. 6F). This differential persisted in lung tissue at Day 42 (626.90 ± 18.49 vs 559.18 ± 45.61 pg/mg; p = 0.03), despite equivalent IL-2 mRNA expression (Fig. 6G, H).

Molecular profiling of apoptosis-related effectors revealed conserved expression patterns across administration routes. Immunohistochemical quantification showed comparable protein levels of caspase-3, BAX, BCL2, HIF-1α, and α-SMA (all p > 0.05; Fig. 7A–E), corroborated by concordant mRNA expression profiles for corresponding genes (Fig. 7A–E). Functional assessment through TUNEL assays confirmed equivalent apoptotic activity in pulmonary parenchyma (Venous: 0.51 ± 0.13 vs BAI: 0.91 ± 0.46 fluorescence units; p = 0.216; Fig. 7F).

Fig. 7.

aPD-1 mAb treatment preserves pulmonary apoptotic homeostasis across administration routes. A–E Comparative analysis of apoptosis-related biomarkers through immunohistochemistry (IHC, left panels) and transcriptional profiling (qRT-PCR, right panels): A Caspase-3 protein expression and CASP3 mRNA levels. B BAX protein expression and BAX mRNA levels. C BCL2 protein expression and BCL2 mRNA levels. D HIF-1α protein expression and HIF1 A mRNA levels. E α-SMA protein expression and ACTA2 mRNA levels. F Apoptotic cell quantification by TUNEL assay. Green fluorescence: TUNEL-positive nuclei. Blue: DAPI counterstain (scale bar = 20μm). Data represent mean ± SD (n = 3/group). Statistical analysis was performed by student’s t-test. Statistical significance: ns, not significant (p > 0.05). BAI bronchial artery infusion, aPD-1 mAb anti-PD-1 monoclonal antibody, CASP3 caspase-3, BAX Bcl-2-associated X protein, BCL2 B-cell lymphoma 2, ACTA2 actin alpha 2 smooth muscle, HIF-1α hypoxia-inducible factor 1-alpha, α-SMA α-smooth muscle actin, mRNA messenger RNA, IHC immunohistochemistry, qRT-PCR quantitative reverse transcription polymerase chain reaction, TUNEL terminal deoxynucleotidyl transferase dUTP nick end labeling, DAPI 4',6-diamidino-2-phenylindole, SD standard deviation

Patient characteristics and adverse events in bronchial artery infusion immunotherapy

The study cohort comprised 17 advanced NSCLC patients (94.1% male; mean age 61.6 ± 7.1 years), including nine patients (52.9%) aged ≥ 60 years, with comorbidities documented in 52.9% of cases (hypertension, coronary artery disease, and chronic obstructive pulmonary disease). Histopathological evaluation revealed squamous cell carcinoma in 15 patients (88.2%) and adenocarcinoma in 2 patients (11.8%), staged as III (52.9%) and IV (47.1%). PD-L1 expression (tumor proportion score ≥ 1%) was observed in 76.5% of cases (n = 13). Patients underwent 23 BAI procedures, with representative treatment workflows illustrated in Fig. 8. Over a median follow-up of 6 months (range: 3–11 months), transient adverse events included nausea/vomiting (n = 1), dyspnea (n = 1), atrial fibrillation (n = 1), and puncture-site hematoma (n = 1). No delayed immune-related adverse reactions, including pneumonitis, were observed during longitudinal monitoring (Table 2).

Fig. 8.

Clinical outcomes of BAI combined with immunochemotherapy in advanced squamous cell carcinoma. Two male patients (66-year-old with stage IIIB; 65-year-old with stage IVB) with progressive squamous cell carcinoma after first-line immunochemotherapy received BAI therapy. Patient 1 presented with a right hilar mass causing lower lobe bronchial stenosis, atelectasis, and obstructive pneumonia (A). Interventional imaging (B) documented selective catheterization of bronchial arteries for targeted immunochemotherapy delivery (red arrow). Following 2 cycles of BAI combined with immunochemotherapy, 2-month follow-up CT (C) showed > 50% tumor volume reduction and complete resolution of atelectasis, achieving partial response (RECIST v1.1). Patient 2 exhibited left lower lobe bronchial obstruction with ipsilateral lung collapse and contralateral metastases (D). Procedure imaging (E) confirmed the presence of tumor-feeding vessels (red arrow). After 3 cycles of BAI combined with immunochemotherapy, 3-month follow-up CT (F) demonstrated significant regression of both primary and metastatic lesions with lung re-expansion, also classified as partial response (RECIST v1.1). BAI bronchial artery infusion, CT computed tomography, RECIST v1.1 response evaluation criteria in solid tumors version 1.1

Table 2.

Patient characteristics and adverse events in BAI immunotherapy

Variables	No. of patients
Patient characteristics
Gender, male (%)	16 (94.1%)
Age > 60 years	9 (52.9%)
Comorbidity	9 (52.9%)
Squamous cell carcinoma	15 (88.2%)
Adenocarcinoma	2 (11.8%)
Stage III	9 (52.9%)
Stage IV	8 (47.1%)
PD-L1 positive	13 (76.5%)
Adverse events
Nausea/vomiting	1 (5.9%)
Chest Pain	0 (0%)
Cough	0 (0%)
Fever	0 (0%)
Dyspnea	1 (5.9%)
Atrial fibrillation	1 (5.9%)
Puncture site hematoma	1 (5.9%)
checkpoint inhibitor pneumonitis	0 (0%)

BAI bronchial artery infusion, No number, PD-L1 programmed cell death ligand 1

Discussion

Our study shows that in a beagle canine model, BAI of aPD-1 mAb exhibits similar stability and safety profiles in terms of biochemical indicators and pulmonary imaging, when compared to traditional systemic administration. Moreover, there were no significant differences observed between the two groups in drug concentration, cytokine levels, and histological assessments, further affirming the safety of BAI-administered aPD-1 mAb. These findings align with previous studies demonstrating good tolerability and low toxicity of aPD-1 mAb therapy [27–29]. Therefore, our results endorse BAI as a safe therapeutic approach, offering valuable insights into targeted immunotherapy for advanced lung cancer.

BAI, as a method of directly infusing drugs into the tumor-feeding arteries, effectively overcomes the physiological barriers faced by systemic chemotherapy, achieving the"first-pass effect"of the drug. This strategy has been proven to significantly increase the local drug concentration in tumors, thereby enhancing therapeutic effects [30, 31]. In the treatment of NSCLC, BAI chemotherapy has become an important adjunct, especially for advanced lung cancer patients who have lost surgical options or are refractory to systemic chemotherapy [20, 32]. Additionally, studies have shown that in animal models of melanoma and breast cancer, local administration of aPD-1 mAb can more effectively activate CD8⁺ T cells, improving antitumor efficacy while offering better safety [23]. The BAI route, as a form of localized drug delivery, holds promise for stimulating a more robust immune response through the delivery of drugs via the nutritional arteries of lung cancer.

However, concerns regarding immune checkpoint inhibitor-related pneumonia (CIP) have hindered the widespread use of the BAI approach in NSCLC treatment. Despite CIP's low incidence rate, its poor prognosis remains a significant concern [33, 34]. Through CT imaging and histological H&E staining observations in our study, we found no notable changes in pulmonary structure and morphology, with histological features appearing normal. This provides compelling evidence that BAI does not cause the common pulmonary complications linked to other treatment methods. This discovery holds substantial significance in immunotherapy, as the risk of immune-related adverse events (irAEs), particularly pneumonia, represents a pivotal aspect to consider during treatment.

No fever was observed after the administration of aPD-1 mAb, and only mild, transient adverse events were documented. This suggests good overall tolerability of BAI immunotherapy. In terms of biochemical indicators after treatment, there were slight increases in LDH, ALT, and AST. While these minor elevations may indicate some level of tissue response or cellular damage, they remained within twice the upper limit of normal values and did not result in significant clinical symptoms. This implies that BAI did not cause systemic organ toxicity, which is a frequent concern with various chemotherapeutic drugs. For patients who have compromised organ function or have had adverse reactions to systemic treatments in the past, BAI could potentially provide a safer therapeutic alternative.

The pharmacokinetic analysis from our study yielded a notable finding: when comparing intravenous injection to BAI, the two distinct administration methods exhibited no statistically significant variation in the peripheral serum concentration of the aPD-1 mAb. This outcome strongly substantiates the efficacy of BAI as a targeted approach for drug delivery, affirming that the overall drug exposure remains unaffected by circumventing systemic administration. Furthermore, according to previous studies, in a mouse model of subcutaneous tumors, the local administration of the aPD-1 mAb at merely 1/10 th of the standard dosage can attain therapeutic outcomes on par with the full dose. This opens up the potential for reducing immunotherapy drug dosages, subsequently lowering the risks of adverse events linked to high-dose administrations [35]. However, it should be noted that our study did not include a quantitative assessment of drug concentration within local tissues, leaving it undetermined whether significant differences exist in this aspect. This gap in knowledge highlights a critical avenue for future research to investigate.

Cytokine analysis and apoptosis-related indicators offered insights into the immunological shifts following aPD-1 administration. While the variations in IFN-γ and IL-2 levels were not statistically significant, the observed patterns hinted at a potential modulation of the immune response. The equilibrium between these cytokines holds significant importance, as it has the capability to influence T-cell polarization and the overall efficacy of the immune system in targeting the tumor. Immunohistochemical findings revealed no notable discrepancies in the expression of proteins pertaining to apoptosis and tissue repair among the intravenous and BAI groups after aPD-1 administration, thus reinforcing the idea that BAI immunotherapy does not cause considerable tissue harm. Excessive apoptosis has the potential to result in tissue necrosis and undesirable outcomes [36, 37].

In our clinical study, we observed the relative safety of BAI immunotherapy. Out of the 17 patients enrolled, the majority were elderly males with at least one comorbidity, classifying them as a high-risk group for traditional systemic therapy. Notably, despite this high-risk profile, only minor adverse events were reported post-BAI immunotherapy, including nausea, vomiting, dyspnea, and small hematomas at the puncture site. Importantly, no serious adverse events, such as immune checkpoint inhibitor-related pneumonia, occurred. These findings align with safety observations from animal experiments and indicate that BAI administration of aPD-1 mAb could offer a viable and safe therapeutic alternative for patients unsuitable for traditional intravenous administration.

Our study provides preliminary evidence supporting the safety profile of BAI immunotherapy, yet several critical limitations warrant cautious interpretation of the findings. First, the restricted sample sizes in both preclinical (n = 3 per group) and clinical cohorts (n = 17 patients) may introduce selection bias and reduce statistical power, potentially inflating type II error rates. Second, the short follow-up durations (42 days in animals; 6 months clinically) preclude assessment of delayed toxicities, such as immune-related pneumonitis or endocrine disorders, which often manifest beyond 12 months in systemic immunotherapy [38]. Third, while healthy beagle models are pragmatic for initial safety screening, they fail to recapitulate the immunosuppressive tumor microenvironment (TME) of NSCLC—a factor known to alter drug pharmacokinetics and immune cell trafficking [39]. Fourth, the absence of local drug concentration measurements (e.g., tumor vs. serum levels) leaves unresolved whether BAI achieves its hypothesized advantage: enhancing intratumoral exposure while sparing systemic toxicity.

Compared to traditional intravenous administration, BAI offers the potential for improved antitumor efficacy by more precisely targeting the tumor microenvironment. Animal model studies have demonstrated that local administration can achieve superior antitumor effects compared to systemic administration, while significantly reducing systemic toxicity and side effects [23, 35]. Our study also preliminarily confirms these findings. However, as a localized delivery method, BAI has certain limitations. It is only applicable to lung cancer types with well-defined blood supply patterns [20, 40]. For example, BAI can only be used when the feeding arteries supplying the tumor are clear and well-defined. This restricts its application in other types of lung cancer or in cases where the tumor vasculature is less predictable.

Conclusions

In summary, our study demonstrates the preliminary safety and feasibility of BAI for delivering aPD-1 mAb in both beagle models and patients with NSCLC, offering a novel strategy for locoregional immunotherapy. Further large-scale clinical studies, head-to-head comparisons with systemic therapies, and quantitative pharmacokinetic analyses of target-site drug exposure are required to comprehensively evaluate the clinical applicability and translational value of BAI in human populations.

Author contributions

B.L. and J.Z. wrote the main manuscript text. W.H. prepared Figs. 1–4, while B.X. contributed to statistical analysis. R.Z. and X.C. were responsible for data collection. Y.Z. and L.X. performed the experiments. S.G. supervised the project and provided funding acquisition. All authors reviewed and approved the final manuscript.

Funding

There is no funding for this study .

Data availability

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

Animal experiments adhered to NIH guidelines and were approved by the IACUC of The First Affiliated Hospital of Chongqing Medical University. The human-subject study followed the Declaration of Helsinki, was approved by the Ethics Committee of The First Affiliated Hospital of Chongqing Medical University, and obtained informed consent from all participants.

Competing interests

The authors declare no competing interests.