This retrospective study was approved by the Institutional Ethic Committee of Fuda Cancer Hospital and conducted in accordance with the Declaration of Helsinki and the Declaration of Good Clinical Practice. Written informed consent was obtained from each participant. From July 2015 to June 2021, all patients with LAPC underwent either a combination of IRE with chemotherapy and PD-1/L1 blockade (group A) or a combination of IRE with chemotherapy (group B). The inclusion criteria were as follows: (1) histologically confirmed pancreatic adenocarcinoma and radiologically verified LAPC [LAPC was defined in the 8th edition of the American Joint Committee on Cancer (AJCC) staging system for pancreatic cancer] (28); (2) patients over 18 years; (3) sufficient bone marrow, liver, kidney, and coagulation function; and (4) performance status (PS) scores under 2. The exclusion criteria were as follows: (1) patients with distant metastases; (2) patients who underwent other forms of treatment; and (3) patients with incomplete or missing data.

Before the IRE procedure, patients received 1,000 mg/m² of gemcitabine intravenously (over 30 min). During the IRE, all patients were placed in a supine position under general anesthesia. Non-depolarizing neuromuscular blocking agents (vecuronium bromide and rocuronium bromide) were administered to reduce muscle contraction. A CT scanner and ultrasound system were used to confirm the morphology and surrounding relationships of the tumor. The number of electrode needles (2–6) and the approach route were determined according to tumor size and location. The exposed length of the probe tip was approximately 10–20 mm, and the needle distance was 15–25 mm. The ablation parameters were as follows: voltage 1,500 V/cm, number of pulses 90–100, and pulse width 70–90 μs. After releasing a group of pulses, the current rise was adjusted at 12–15 A, and the maximum was close to 50 A. For larger diameter tumors, ablation was repeated after retreating the needle by 1 cm until the ablation area covered the whole lesion. A CT scan or ultrasonography of the abdomen was performed again after the ablation to determine whether the ablation had been completed, whether bleeding had occurred, and whether essential structures had been damaged. When the disease was stable, oral S1 or albumin combined with paclitaxel was administered as maintenance therapy.

Patients in group A received PD-1/PD-L1 blockade therapy, including camrelizumab (200 mg/2 weeks), toripalimab (240 mg/2 weeks), nivolumab (100 mg/2 weeks), pembrolizumab (100 mg/3 weeks), and atezolizumab (1,200 mg/3 weeks). The first dose of PD-1/PD-L1 blockade was administered within 2 weeks after IRE. Patients received continuous PD-1/PD-L1 blockade treatment until intolerance toxicity or progressive disease. The treatment protocol is illustrated in Figure 1B.

Blood samples of each patient’s blood (2 ml) were collected just before the IRE procedure (Pre-IRE) and 30 days (Post-IRE) after treatment. Flow cytometry techniques (FACS caliber, four-color system, BD Bioscience, CA, US) were used to analyze the absolute number of CD3⁺ T cells (CD3⁺), CD4⁺ T cells (CD3⁺CD4⁺), CD8⁺ T cells (CD3⁺CD8⁺), NK cells (CD16⁺CD56⁺), T-lymphocytes PD-1 (CD3⁺PD-1⁺), T helper lymphocytes PD-1 (CD4⁺PD-1⁺), cytotoxic T lymphocytes PD-1 (CD8⁺PD-1⁺), and Treg cells (CD4⁺CD25⁺CD127dim). A similar method was used to measure serum cytokines, including interleukin-6 (IL-6), interleukin-10 (IL-10), tumor-necrosis factor-beta (TNF-β), and interferon-γ (IFN-γ). A record of the data was then made.

The clinical information and demographic were collected, including age, gender, lesion size, tumor location, performance status (PS), carbohydrate antigen 19-9 (CA19-9), preoperative therapy, PD-L1 expression, and PD-1/L1 blockade. The tumor response was evaluated based on the mRECIST (29). The objective response rate (ORR) was calculated as follows: ORR = (CR+PR cases)/total cases × 100%. OS was defined as the time from diagnosis to death or until the last follow-up. PFS was defined as the time from diagnosis to disease progression until the last follow-up.

All patients in this study were followed up at 1, 3, and 6 months post-IRE and then every 3 months. Two radiologists independently interpreted all follow-up scan results. Adverse events were evaluated and classified according to the National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.0. The last follow-up was completed on 25 February 2022.

All statistical analyses were performed and compiled using GraphPad Prism (version 8.01, GraphPad Software, San Diego, CA, USA) and SPSS software (version 20.0, SPSS Inc., Chicago, IL, USA). Categorical variables were compared using the chi-square or Fisher exact test. The continuous data, categorical data, and survival curves of the two groups were compared using the Mann–Whitney test, Fisher’s exact test, and log-rank (Mantel-Cox) tests individually. The Cox regression model was used in survival analysis to assess the influence factor of OS and PFS. All statistical tests were two-sided; a p-value < 0.05 was considered statistically significant.

Between July 2015 and June 2021, a total of 169 LAPC patients were included in this study (Figure 1). Based on the exclusion criteria, 66 patients were excluded owing to receiving other treatment (n = 28), tumor metastasis (n = 22), incomplete data, or missing (n = 16). Finally, a total of 103 LAPC patients were analyzed in this study: group A (n = 25) and group B (n = 78) (Figure 1A). The clinical characteristics of the patients were well-balanced between the two groups (Table 1). The median ages were 55 (26-80) and 62 (34-78) years in groups A and B, respectively. Both groups had a higher proportion of male patients than female patients. The median tumor sizes were 4.1 cm (range 2.4–6.7) and 3.8 cm (range 2.1–6.2) for patients in groups A and B, respectively. Group A had a higher tumor burden, but the difference was not statistically significant.

As shown in Table 2, there was no treatment-related mortality in either group in our study. The overall adverse events rate was similar between the two groups. The most common immune-related adverse events were pruritus (24%), hypothyroidism (16%), increased bilirubin (16%), and ALT increase (16%) in group A. The major (grade 3–4) immune-related adverse events were hypothyroidism (8%), ALT increased (4%), and colitis (4%) in group A, which was resolved after the treatment with a corticosteroid. There were no immune-related adverse events in group B. The most common IRE treatment-related adverse events were pain (72%), fatigue (64%), diarrhea cardiac (56%), arrhythmias (48%), and hypertension (48%) in group A. The major (grade 3–4) IRE-related adverse events were cardiac arrhythmias (48%), hypertension (48%), pancreatitis (28%), and hemorrhage (16%). All adverse events were relieved or improved after symptomatic treatment.

Overall, 4 (16.0%) patients achieved CR, 12 (48.0%) patients received PR, 6 (24.0%) patients developed SD, and 3 (12.0%) patients suffered PD in group A. The tumor response in group B was 8 (10.3%) patients with CR, 35 (44.8%) patients with PR, 22 (28.2%) patients with SD, and 13 (16.7%) with PD. The ORR was 64.0% vs. 55.1% (p = 0.787, Figure 2) in groups A and B, respectively.

We compared the absolute number of immune cells after combined IRE, chemotherapy, and PD-1/PD-L1 blockade to investigate post-treatment immune effects. The results demonstrated that the absolute number of CD3⁺ T cells, CD4⁺ T cells (p = 0.038), CD8⁺ T cells (p = 0.024), and NK cells steadily increased (Figures 3A–D), while the proportion of CD3⁺PD-1⁺, CD4⁺PD-1⁺, CD8⁺PD-1⁺, and Tregs cells (p = 0.023) decreased after treatment in group A compared to group B (Figures 3E–H).

To further compare the variations of the immune response, we compared the levels of cytokines in both groups. The data suggested that the IL-6, IL-10, TNF-β, and IFN-γ levels significantly increased after treatment in group A, while the change in IL-10 and IFN-γ was not significant after treatment in group B (Figure 4).

The median follow-up time was 18.2 months (3.0–38.6 months). The median OS from diagnosis was significantly longer in group A than in group B, respectively (23.6 months vs. 19.4 months, p = 0.001, Figure 5A). Similarly, the median PFS from diagnosis was significantly longer in group A than in group B, respectively (18.2 months vs. 14.7 months, p = 0.022, Figure 5B).

As shown in Table 3, tumor size (HR: 1.022, 95% CI: 1.001–1.043, p = 0.045), CA 19-9 (HR: 1.021, 95% CI: 0.095–1.024, p = 0.017), and IRE plus PD-1/PD-L1 blockade treatment (HR: 2.217, 95% CI: 1.161–5.034, p = 0.018) were significantly correlated with OS in the univariate Cox regression analysis. Multivariate analysis showed that the combination of IRE with chemotherapy and PD-1/PD-L1 blockade (HR: 3.605, 95%, 95% CI: 1.417–9.175, p = 0.007) were independent predictors of OS (Table 3). Furthermore, tumor number (HR: 1.756, 95% CI: 0.956–1.327, p = 0.024) and IRE plus PD-1/PD-L1 blockade treatment (HR: 0.459, 95% CI: 0.248–0.824, p = 0.006) were also significant prognostic factors for PFS (Table 4).