Abbreviations

LAPC — Locally advanced pancreatic cancer

IRE — irreversible electroporation

TTP — time to progression

OS — overall survival

SMA — superior mesenteric artery

CT — celiac trunk

CHA — common hepatic artery

SMV — superior mesenteric vein

PV — portal vein

GEM — gemcitabine

PD — pancreatoduodenal resection

Introduction.

Locally advanced pancreatic cancer (LAPC) is diagnosed in 35% of patients [1]. Standard therapy includes systemic chemotherapy with reassessment of resectability. Meta-analysis by Suker M. et al. showed that induction chemotherapy with FOLFIRINOX made tumour resectable in 28% of cases. Although currently there is no evidence of improved long-term outcomes when radiation therapy is added to systemic one, two-thirds of patients receive radiation therapy for local control [2].

Other methods for local control are used less frequently. One of them is irreversible electroporation (IRE). It uses non-thermal energy of high-intensity, ultrashort, electric fields between probes to create nanopores in the cell wall which disrupts cell homeostasis and initiates apoptosis [3, 4]. Some authors support the method and report good long-term outcomes with median OS increase to nearly 25 months and low postoperative mortality of 1.5% [5]. The Asian multicenter study results with a 5-year survival rate of 31.2% are impressive [6]. However, there are many less optimistic papers that report a high complication rate (up to 53% of Clavien–Dindo grade III or higher), 90-day mortality of 4–13% with modest long-term outcomes (median OS, 16–17 months) [7, 8]. Maurice D et al. draw cautious conclusions based on the systematic review data and believe that IRE may be performed in carefully selected patients. But they highlight that not all analyzed studies show a positive effect on long-term outcomes prior to standard treatment of LAPC: one third of patients have complications which are fatal in 2% of cases [9].

Patients and methods

From 2015 to 2022, 23 patients underwent IRE for LAPC. The study included 8 men and 15 women. The median age was 60 years. In all cases, the tumour affected more than one region of the pancreas, with the body being the most common. Obstructive jaundice was diagnosed in five patients and required percutaneous transhepatic cholangiostomy in one patient; biliodigestive anastomoses were formed in four patients.

All patients were diagnosed with locally advanced pancreatic cancer unresectable due to technical aspects. Patients with an artificial pacemaker, arrhythmia requiring antiarrhythmic therapy, and patients with tumor ingrowth in hollow organs with mucosal infiltration were not considered candidates for IRE. Morphological evidence was obtained in all cases (ductal adenocarcinoma). CT/MRI with intravenous contrast revealed that tumours were unresectable and had no distant metastases; in five unclear cases, positron emission tomography (PET) was performed. In most cases (n = 22) the tumor was unresectable due to the signs of arterial involvement, specifically superior mesenteric artery (SMA), celiac trunk (CT), and common hepatic artery (CHA). One female patient had a complete occlusion of the superior mesenteric and portal veins (SMV/PV) with infiltrate transfer to the intestinal veins; vascular reconstruction was not technically possible. All patients underwent induction chemotherapy. Two patients received stereotactic radiation therapy. The most common first-line chemotherapy was FOLFIRINOX. Adverse events with induction chemotherapy were observed in 16 patients (69.6%); grade III complications were reported in six (26.1%) patients. RECIST tumor response of stabilization was observed in 18 (78.3%) patients; a partial response was noted in five (21.7%) patients. Description of patient characteristics prior to induction chemotherapy is detailed in Table 1.

Table 1. Characteristics of patients before chemotherapy

The median CA-19-9 level prior to electroporation was 59 IU/ml (5–1123 IU/ml), the median tumor size was 35 mm. Note that tumors with CT involvement of more than 180o, but without involvement of the gastroduodenal artery, were considered technically resectable. In this case, the surgery included distal pancreatectomy and CT removal without arterial reconstruction.

The procedure was performed using AngioDynamics NanoKnife with mandatory cardiac synchronization. To do so, midline laparotomy was completed under combined anesthesia with obligatory muscle relaxation in 22 patients. Percutaneous IRE was performed in one case. IRE was a standalone procedure in 22 patients; one patient underwent palliative pancreaticoduodenectomy in combination with IRE of SMA-surrounding infiltrate. Four patients received lymph node dissection or lymphadenectomy, sero-muscular membrane of the stomach due to invasion was removed in two patients.

Ultrasonic guidance was used to place the probes and measure the spacing between them; the probes were placed in a caudal-to-cranial direction in two cases (Fig. 1a) and in the anterior-posterior direction in the other 21 cases (Fig. 1b).

Fig. 1. Probe placement in pancreatic tumor.

Caudal-cranial (a) and anterior-posterior probe placement (b). T — tumor.

Generally, 4 probes (3–5) were used. The active tip length was 10–20 mm, voltage was 1200–1500 V/cm, the pulse length was 90 μs. At an early stage, we routinely used probes with the 20 mm active part, voltage of 1200 V/cm. Currently (n = 16) we use probes with the 15 mm active tip and voltage of 1500 V/cm. In two cases after electroporation of the node central part, it was necessary to shorten the active tip to 10 mm to treat the tumour periphery. After a series of test pulses with specified parameters (10 pulses between each pair of probes), the current intensity was estimated. If values were satisfactory (20–40 A), electroporation was performed as planned (a total of 90 pulses between each pair of probes). If necessary, voltage or pulse duration was corrected. Five patients underwent IRE in one ablation zone. In other cases (n = 18), the probes were repositioned and/or spacing between them was changed under ultrasound guidance. One patient required local destruction in five zones. Ablation volume between each pair of probes was considered effective when current strength of 20 A was achieved at the initial stage followed by increase with subsequent pulses. In two cases there was no increase in current strength between any probe pair during the procedure, and ablation was repeated (between the probes without current strength increase).

Table 2. Perioperative characteristics

Note. PD — pancreatoduodenal resection

Post-surgery 19 patients (86.3%) received systemic adjuvant chemotherapy. The most common chemotherapy included gemcitabine ± capecitabine (73.9%, n = 17, in one case in combination with radiation therapy) and FOLFIRINOX (8,7%, n = 2). Three patients (13.4%) received no additional treatment until progression.

Routine follow-up examination included abdominal CT or MRI with intravenous contrast, chest X-ray, CA-19-9 test, hematology, biochemistry, and coagulation. If necessary, additional tests, PET-CT and chest CT, were carried out. A follow-up examination was carried out a month after the surgery to assess whether the entire tumour was ablated. Then the examination was repeated every three months to check for recurrence and metastasis.

GraphPad Prism 6 software was used for statistical data processing. Long-term outcomes were assessed both from the induction therapy initiation and electroporation. The progression date was defined as the date of progression signs detection based on objective examination (CT, MRI, PET) or as the last discontinuation visit.

Results

Both resectable tumors and patients with distant metastases were excluded. Resectable tumors were removed; when metastases were detected, the surgeon completed the intervention. The median surgery time was 250 minutes due to intraoperative revision and mobilization for final resectability assessment often complicated with scarring after prior surgeries (exploratory laparotomy, formation of biliodigestive anastomosis). The median electroporation time (probe placement and ablation) was 90 minutes. The rest of the procedure included laparotomy, adhesiolysis, and revision. In female patient, who underwent palliative pancreaticoduodenectomy with prior infiltrate electroporation around SMA, the surgery took 600 minutes. Median blood loss was 50 ml (50–1200 ml). Intraoperative data demonstrated effective treatment of the entire infiltrate in 21 cases (91.3%).

Complications were observed in five patients (21.7%), three patients (13.0%) had grade ≥III complications. Three patients had arrhythmia (13.0%) which led to hemodynamic disorder in one case. Two patients (8.7%) developed pancreatic necrosis, one was fatal (mortality rate 4.3%).

An autopsy of the patient who died after IRE did not reveal tumor cells in the treatment zone. However, adenocarcinoma clusters were found in the intestinal mesentery.

The first follow-up radiologic examination (CT/MRI) after IRE showed a complete response to local treatment in 21 patients. CT revealed tumor density decrease; there was no contrast accumulation in the infiltrate after treatment. MRI also found no contrast agent accumulation; in addition, there was no diffusion restriction (Fig. 2).

Fig. 2. Axial MR scans demonstrate cystic transformation of pancreatic tumor remaining avascular in arterial (Fig. 2a) and venous (Fig. 2b) phases of examination (thin arrows).

Fatty tissue thickening along the celiac axis (Fig. 2c, thick arrow). DT MRI (b-factor=800 sec/mm²) (Fig. 2d) reveals no tumor, IRE area is marked by thin arrows.

CT showed that 25% of the tumor was not ablated in one patient; she underwent stereotactic radiation therapy.

The median follow-up time was 19 months. Currently eight patients are alive, four of them have progression.

The median TTP was 7 months from the surgery and 14 months from the initiation of systemic induction chemotherapy (Fig. 3).

Fig. 3. Time to progression after IRE (a) and initiation of chemotherapy (b).

Isolated local recurrence was observed in seven patients, three of them underwent stereotactic radiation therapy, and one received repeated IRE (10 months after the first procedure). The last female patient died of pulmonary embolism 16 months after the second IRE, liver metastases were found one month before death, chemotherapy was initiated.

In 11 patients progression included distant metastases, systemic therapy was resumed.

Eight patients are currently alive; time after surgery ranges 7 to 67 months. Median OS was 16 months from electroporation and 25 months from induction chemotherapy (Fig. 4). The female patient, who underwent IRE in combination with palliative gastrapancreatoduodenal resection, has been alive for over 5 years.

Fig. 4. Overall survival after IRE (a) and initiation of chemotherapy (b).

Discussion

At diagnosis half of the patients with pancreatic ductal adenocarcinoma has distant metastases for which systemic therapy is initiated [1, 10]. In patients with ECOG 0-1, a three-component regimen is usually used; in elderly and debilitated patients, gemcitabine with nab-paclitaxel is an alternative [11].

In another 10-15 % of patients, a potentially curative surgery may be performed: pancreaticoduodenectomy, distal pancreatectomy, or total duodenopancreatectomy. Chemotherapy usually complements surgery. In the remaining 30–50 % of patients, the tumor is considered borderline resectable or locally advanced and unresectable [1, 10]. Involvement of the PV, superior mesenteric vessels, celiac trunk, hepatic artery challenges R0 resection or makes it obviously impossible. The standard of care for such cases is systemic therapy. If a positive response is observed and tumour becomes resectable, surgery is indicated [12, 13, 14].

FOLFIRINOX helped to achieve high results in patients with LAPC. In systematic review and meta-analysis including 11 studies to evaluate long-term outcomes (315 patients), the median OS was 24.2 months (10–32.7 months). Such high results might challenge feasibility of local effect which, as it will be shown below, cannot be more successful for sure with increase of side effects risk. On the other hand, half of the patients in the study received radiation therapy, and a quarter of the patients underwent resection. It should be noted that we did not find a direct correlation between the resection rate in different centers and long-term results [2].

Local control methods should be considered when the tumour remains unresectable with systemic therapy. They include thermal and non-thermal local destruction, regional intra-arterial infusion chemotherapy, and radiation therapy [15].

Radiation therapy is typically used for local control, though randomized study certainly showed no improvement in long-term outcomes [16]. Five sessions of stereotactic radiation therapy with 33 Gy in combination with gemcitabine chemotherapy resulted in median OS of 13.9 months in 49 patients with LAPC; 1-year local control was achieved in 78% of patients [17]. The prospective non-randomized study conducted by Reyngold M. et al. showed that radiation therapy in unresectable pancreatic ductal cancer (including patients with medical contraindications to pancreatic resection) resulted in the median OS of 26.8 months from the diagnosis. The median OS from radiation therapy was 18.4 months [18]. In case of recurrence, stereotaxic radiation therapy may be used [19]. Although non-randomized trials show incoherent results and randomized trials show negative results, more than half of the patients with LAPC receive concomitant radiation therapy, as already mentioned in systematic review by M. Suker et al [2].

Other methods of local control are used much less frequently due to their limited capabilities and low efficiency [4]. Recently there have been more and more reports of a relatively new method of soft tissue ablation IRE. The method is based on non-thermal energy of ultrashort (80–90 microseconds) high-voltage electric fields (1500–3000 V) between the probes to create nanopores in the cell wall which disrupts cell homeostasis and initiates apoptosis [3]. Absence of a significant thermal effect, selective cell damage with no impact on the stroma, confident control of the ablation zone with probe positioning, and a clear boundary between the ablation zone and the intact tissue allows manipulations near hollow organs and directly on the vessels and ducts affected by the tumour. The technique is applicable to laparotomy, laparoscopy, and percutaneous procedures. It can also be used in combination with palliative resection [20].

In 2015 Martin RC et al. published an article showing safety of the procedure in a large number of patients both as a standalone surgery (n = 150) and in combination with palliative pancreatic resection (n = 50). Complication and mortality rates were 37% and 1.5%, respectively. The median OS was 24.9 months [5].

Ruarus et al. reported interesting IRE results. The study included 50 patients (40 of them with LAPC and 10 with local recurrence) who underwent percutaneous IRE. In the LAPC group, 18 and 22 patients received preoperative chemotherapy with gemcitabine and FOLFIRINOX, respectively. Complications were reported in 52% of patients, two patients died within 90 days (one of them from duodenal perforation). Although the authors reached the primary endpoint of 17 months median OS from diagnosis in the LAPC group, these modest long-term outcomes are confusing. It should be noted that survival did not depend on the induction chemotherapy regimen. The authors were able to identify three unfavorable prognostic factors: tumour size of ≥ 37 cm³, high (≥ 2000 IU/ml) baseline CA 19-9 with no decrease by ≥ 50% 3 months after IRE [7].

Multicenter Asian study by Yang PC et al. showed that tumor marker levels had no independent effect on prognosis. The study included 74 patients with a non-metastatic cancer less than 4 cm in diameter involving the celiac trunk or SMA, with no progression during preoperative chemotherapy. Laparotomy (93.8%) and a caudal-to-cranial direction of the probes (83.8%) were most common. Complications were reported in 17 patients, nine of them were Clavien–Dindo grade III. Fluid accumulation (n = 8) and ileus (n = 11) were the most frequent complications. Seven patients had gastrointestinal haemorrhage. Involvement of hollow organs (p = 0.002) and anterior-posterior probe spacing during the procedure (p = 0.004) were independent risk factors for complications. Five-year TTP was achieved in 28.8% of patients, 5-year OS was 31.2%. Adjuvant therapy with TS-1 (tegafur/gimeracil/otaracil) were preferred over gemcitabine regimens. As above mentioned, CA-19-9 levels had no significant impact on the prognosis in multivariate analysis, as well as tumour size and volume. However, these parameters affected the long-term outcomes in one-way analysis [6].

The IMPALA results which explored the IRE role in LAPC were published. The study included 59 patients with no progression by RECIST within three months from the disease onset. Fourteen patients were successfully resected, the median OS in this group was 34 months. Electroporation was performed in 15 patients; the median survival reached 16 months. The results in this group did not differ from those in patients neither without resection or IRE (n = 30, median OS was 15 months). Clavien–Dindo grade ≥III complications after IRE were observed in 53% of patients. Bile leakage (n = 3), gastric stasis (n = 3), PV/SMV thrombosis, pancreatic fistula (n = 1), gastric bleeding (n = 1), bile duct obstruction (n = 1) were reported. The authors explain such a high complication rate after electroporation by the fact that in this group, a metal stent had to be removed in two patients, and two more patients underwent excessive dissection before the procedure when trying to perform a resection. Within 90 days of IRE, two patients died (from bleeding due to an acute gastric ulcer and from liver failure). The authors conclude that, based on their own experience, electroporation cannot be indicated for patients with LAPC in routine practice [8]. Meta-analysis by Lafrancescina S shows that complications after IRE develop in 30.5% (0–59%) of patients, mortality is 3.4% (0–13%). Median OS ranges 15 to 27 months. The authors cautiously conclude that IRE may be recommended in LAPC when performed in an expert center after successful chemotherapy or chemoradiation therapy and careful consideration of possible risks [21]. Based on the systematic review data, Moris D et al. believe that IRE is not superior to the standard LAPC treatment and poses a risk of complications, including fatal. It is too early to talk about its use in routine practice [9]. The authors of the last two papers have high expectations for the ongoing large randomized trials.

In our paper we studied IRE results in 23 patients with LAPC. Surgeries were performed following successful induction therapy with disease stabilization (n = 18) or partial response (n = 5). CA-199 marker decrease (median) from 121 to 59 IU/ml was noted during preoperative therapy. Intraoperative data indicate that the entire infiltrate could not be included in the electroporation zone (8.6%) in two patients. Grade ≥III complications after surgery were observed in three patients (13.0%). Mortality was 4.3% (n=1). Examination results one month after the surgery were positive: CT/MRI found a significant vascularization decrease in the entire infiltrate almost in all patients (n = 21). Median CA 199 level decreased to 24.5 IU/ml.

In addition, good long-term outcomes were noted: the median TTP was 14 months after the treatment initiation and 7 months after electroporation. For local recurrence, repeated electroporation should be considered. This case was successful and the patient lived for another 16 months. Other three patients with local recurrence underwent radiation therapy.

One patient has been alive for more than five years. The median OS reached two years after the start of treatment (25 months). The long-term results are similar to those of the Suker M et al. [2]. Given the similar results in our work with the results of systemic therapy according to the FOLFIRINOX regimen, it seems that there is no expediency in the use of IRE in LAPC. However, it should be noted that half of the patients in the study of Suker M received radiation therapy, and a quarter of the patients underwent resection.

In addition, both approach and technique of IRE are being improved. Narayanan et al. believe that it is necessary to evaluate safety and efficacy of high-frequency IRE which may not require deep muscle relaxation, cardiac synchronization, or even general anesthesia [4]. It is also worth considering the impact of the learning curve on both immediate and long-term results. Interim results were shown in a prospective study AHPBA that included 152 patients from six centers. The centers involved have overcome the learning curve associated with IRE. Mortality was 2%, the complication rate reached 18%. The long-term results are impressive — the median overall survival was 30.7 months from the date of diagnosis [22].

The results of prospective works (NCT02674100, NCT02791503, NCT03257150) of IRE feasibility for patients with LAPC are expected. The results of the first phase works are also expected, for example, the work investigating the combination of IRE with immunotherapy [23].

Conclusions

Despite the inconsistent results, many authors find IRE in LAPC to be promising. Often IRE is the only effective option for local tumour treatment in LAPC. Palliative pancreatic resection may be combined with residual tumor IRE. Based on our experience in LAPC treatment using IRE, the median OS was 25 months from the induction therapy start and 16 months from electroporation. One patient is still alive for more than five years. Tumour IRE should be combined with perioperative treatment to improve long-term outcomes. For local recurrence after IRE, options for local disease control, including repeated electroporation, should be considered. In such cases, acceptable treatment outcomes may be expected.

However, possible complications should be considered, including fatal. In our study, one patient (4.3%) died. In literature data mortality was 3.4% (0-13%). Given the possibility of complications, including fatal, impact on long-term outcomes and feasibility of routine use should be determined in randomized trials.

No funding.

Compliance with ethical rights. The article does not include personal information. An informed consent was signed for the use of the off-label method.

The authors declare no conflicts of interest.