After approval from the local medical committee, we carried out a pilot study with satisfactory results: 50 patients corresponding to the inclusion criteria (Table 3.1) were treated in a 20-month period before or after a combination of chemo and chemoradiotherapy. Morbidity and mortality rate was 26 % and 2 %, respectively.

After an interim analysis of the first 25 patients, because of the pilot nature of the study and the lack of specific parameters of application for pancreatic cancer, we decided to decrease the RFA temperature from 105 to 90 °C in the following 25 patients. The overall complication rate significantly decreased to 8 % [43].

After these preliminary results, we felt confident to propose the procedure to our patients as an alternative treatment to stage III PDAC in a multimodal setting. We slightly changed our target and started to consider RFA as a treatment option that could modify the natural history of the disease. RFA should be considered a cytoreductive treatment, which provides direct and rapid necrosis of the mass. This debulking effect on tumor volume may induce the acceptance of a higher perioperative risk when compared to bypass operations alone [44]. We prolonged follow-up of the initial group of patients and found that most of them attended at planned controls. Therefore, we decided to review our results when 100 patients had been treated. In this group of patients treated with a new multimodal therapy, RFA associated to chemo and radiotherapy, the median OS and DSS were 20 and 23 months, respectively, and confirmed the preliminary data of the previous pilot study [45]

Between February 2007 and December 2014, 200 patients were treated with RFA in our department. All patients had preoperative diagnosis of stage III pancreatic carcinoma confirmed by contrast US, abdominal CT scan, and/or MRCP, all histologically proven by fine-needle aspiration. All demographics data are listed in Table 3.2. The male/female ratio was 112/88 with a mean age of 64 years. The tumor was located in the pancreatic head/uncinate process in 145 cases and in the body/tail in 55 cases. Median tumor size was 35 mm (IQR 34–48). Medium serum CA 19.9 at admission was 110 (IQR 28–479), with mean hospital stay of 10.7 days.

Thirty-nine percent of patients received RFA as up-front treatment when surgical palliation was needed or due to the lack of diagnosis or under staging. The remaining 61 % underwent neoadjuvant chemo or radiochemotherapy.

During RFA operation, associated surgery was performed in 45 % of patients: 49 % single bypass (biliary or gastric), 43 % double bypass, one cholecystectomy, and one pseudocyst-jejunostomy.

It has been proposed that tumor destruction due to radiofrequency ablation takes two steps: firstly, a direct damage proportional to the applied energy, tumor biology, and its microenvironment and, secondly, an indirect damage after the energy application. The latter represents the progression of tissue damage, which can arise from immunity stimulation due to the ablation [46].

As a matter of fact, the evidence of spontaneous regression of secondary untreated lesions after the ablation of the primary tumor site can lead to the evidence of immunity stimulation by the treatment [47–50].

It is indeed well known that, when a temperature higher than 60 °C is applied, in the treated area, multiple changes occur: enzymes inactivation, protein denaturation, and coagulative necrosis. The events mentioned above can modify the membrane permeability with cytolysis and metabolite accumulation. All the modifications can actively extend for days, also weeks, after the thermal application [34, 35].

Ablation site can be divided into three zones: the central zone where the highest temperature produces coagulative necrosis, the transition zone where sublethal hyperthermia produces cell apoptosis, and the peripheral zone which is excluded from direct temperature damage [51].

Many papers have studied the transition zone, considered as the primary site where inflammatory processes stimulate the immunity system [52, 53]. Murine models and in vivo studies showed the increase of HSP-70 (heat shock protein) after thermal ablation in the transition zone up to 5 days after the treatment [54–57].

HSP-70 is a heterogeneous family of extracellular proteins induced by stress. Calderwood et al. described their role in the intracellular signaling processes and immunity stimulation [75].

Thermal stress increases tissue HSPs and facilitates its permeability through cell membrane into interstitial space [58–62]. Stress-induced proteins can be also produced by neurons, monocytes, macrophages, B-lymphocytes, and tumor cells [63–69]. The main activity is to trigger the innate immunity response by releasing multiple cytokines and stimulation of adaptive immunity by its ability to enhance the peptide ligands. Moreover, the complex HSP peptide, when released by tumor apoptotic cell, links with antigen-presenting cells (APCs), represented by dendritic cells population, promoting the antigen presentation [70–78].

Furthermore, it has been extensively demonstrated a cytotoxic activity after thermal damage.

Multiple studies clearly show the evidence of immunity cells stimulation (like dendritic cells, B- and T-subsets) [79].

Other evidences attribute a power to increase the production of multiple pro-inflammatory cytokines like TNF-α, IL-6, IL-8, IL-10, HGF, and VEGF [80–86].

Fietta et al. underlined the importance of another factor: the lymphocytes T regulators (Tregs) [87]. T regulators are crucial to modulate the immunity response in order to control inflammatory reactions, compared to normal immunity response. Fietta et al. proved that, after RFA ablation, Tregs remain inactivated up to 30 days [87].

In conclusion, RFA stimulates immune activity through different mechanisms: enhancement of antigen presentation by increasing the activity of dendritic cells, stimulation of antitumor response by increasing cytotoxic activity, and suppression of negative modulation decreasing the Treg subset.

The immune stimulation might be particularly effective in pancreatic cancer where the lack of dendritic cells, the limited number of natural apoptosis, and the chemoresistance significantly restrict the immunogenic potential of the tumor. These peculiarities make the pancreatic cancer the “ideal candidate” for thermal ablation [88].

When downstaging is not achieved, preliminary data show good control of the disease with an encouraging progression-free survival (13 months).

The right timing of RFA, up-front or after neoadjuvant CHT, has been a matter of debate. We retrospectively analyzed 57 patients affected by stage III PDC who received short-term chemotherapy between February 2007 and June 2010 [90]. Among 57 treated cases, 29 patients had a minimum follow-up of 12 months. We compared their survival data with those from a similar group of patients affected by LAC, observed in the same period, who underwent RFA as up-front therapy, and we saw that there was no difference in survival associated with the timing of RFA (Fig. 3.13).We can therefore propose RFA as an up-front treatment with no fear of neglecting a neoadjuvant treatment.

We will keep offering RFA to our patients as one of the possible treatment modalities for stage III disease. The retrospective, preliminary results achieved in our series need to be validated, and therefore a prospective, randomized, controlled trial is ongoing. At the same time, we are carrying out an immunology study protocol with the aim to find out specific immunomodulatory effects after thermal ablation and their role in disease control. Immunity effects of RFA in pancreatic cancer have never been investigated.

As mentioned above, the complication rate is the “Achilles heel” of the surgical ablation. Therefore, on the way to decrease the risks of the procedure, the choice of a minimally invasive approach is strongly encouraged.

Endoscopic ultrasound (EUS) is a well-known procedure that was introduced in clinical practice more than 30 years ago. In the early 1990s, linear-array echoendoscopes equipped with a working channel safely guiding a needle into a target lesion have been introduced. By means of this technique, therapeutic interventions such as drainage of fluid collections and of obstructed pancreatic and biliary ducts, celiac plexus neurolysis, ablation of cyst neoplasms of the pancreas, biliary and gastroenteric anastomoses, injection of antitumor agents inside the tumor, and radiofrequency and cryothermal ablation of solid neoplasms are currently performed.

Radiofrequency thermal ablation (RFA) of solid tumors is widely applied for unresectable tumors, with surgical or percutaneous approaches, in several organs. However, in pancreatic neoplasms the procedure is not yet commonly adopted. RFA for locally advanced pancreatic cancer has been largely applied during laparotomy in retrospective reports demonstrating feasibility, safety, and benefits on overall survival.

Potential advantages of RFA of pancreatic neoplasms under EUS guidance are a less invasive approach, a more accurate identification of the target lesion due to the high-resolution images, and the short distance between the device and the neoplasm as compared to percutaneous approach.

The RFA with EUS approach (EUS-RFA) of pancreatic lesions in humans was carried out after several attempts in animal models. In 1999 Goldberg et al. [91] reported EUS-RFA in the pancreas of 13 pigs using a modified EUS 19-gauge needle electrode with a 1–1.5 cm tip. No major complications occurred but the ablated area in the pancreas was less than 1 cm size. Only one pig developed pancreatitis. In 2008 Carrara et al. [92] found a good correlation between the size of the ablated area and the duration of ablation in the pancreas of 14 pigs using a flexible bipolar ablation probe combining RFA and cryogenic cooling with carbon dioxide. Two pigs developed necrotizing pancreatitis. It was shown that the same device produced similar effects in the liver and spleen of 19 pigs [93]. A different device was tested in 2009 by Varadarajulu et al. [94]. An umbrella-shaped electrode array introduced through the lumen of a 19-gauge EUS-FNA needle was used to ablate the liver of 5 pigs. No major complications after the procedure have been reported with a mean area of coagulation necrosis at histopathology of 2.6 cm. This method presented the limit of a very large necrotic area for the application in the pancreas and the difficulty in determining the exact position of all the hooks of the umbrella-shaped electrode that lie on a different plane of the probe. In 2012, Gaidhane et al. [95] reported results of RFA of five porcine pancreas using the Habib EUS-RFA 1 Fr wire monopolar probe introduced through the lumen of a 19-gauge EUS-FNA needle: no major complications and moderate pancreatitis in one case were observed. In 2014, Sethi et al. [96] also tested the same device in mediastinal lymph nodes of 18 pigs: no complications occurred and a mean of 17.6 ± 10.3 % of the respected lymph node areas was ablated. Moreover, in 2012 Kim et al. [97] presented a 18-gauge monopolar endoscopic RFA probe tested in the body and tail of 10 porcine pancreas: administration of 50 W for 5 min produced a mean diameter of ablated pancreatic tissue of 23 ± 6.9 mm. In this report, no major complications were described. All these animal model studies demonstrated the feasibility and safety of EUS-RFA in pancreatic tissue with all the electrodes used; the further step was to perform this method in patients with pancreatic neoplasms unfit for surgery.

In 2012, Arcidiacono et al. [98] reported a prospective study on 22 patients affected by locally advanced pancreatic cancer using the abovementioned flexible bipolar ablation probe combining RFA and cryogenic cooling with carbon dioxide (ERBE^®) with an electric active part of 1.8 mm of diameter and 20 mm length.. Ablation parameters were 18 W for RF power, cooling pressure of 650 psi, and the application time depended on the size of the lesion (mean time 107 ± 86 s). The procedure was feasible in 16 out of 22 patients (72.8 %), and unsuccessful placement of the probe inside the tumor occurred in 6 cases and was due to stiffness of the gastrointestinal wall and of the tumor due to desmoplastic reaction, tumor infiltration, or fibrosis in patients who had already undergone radiation therapy. No early severe complications have been reported (mild abdominal pain in three cases, one minor duodenal bleeding, and increase in serum amylase in three cases). Four patients experienced late complications. In one case, hemobilia and jaundice were effectively treated by ERCP and biliary stent. A patient presented with jaundice and was successfully treated with ERCP and biliary stent. Another patient presented with duodenal stricture 1 month after the procedure, and the last patient developed an asymptomatic peripancreatic fluid collection. In conclusion, EUS-guided cryothermal ablation was feasible and safe but not applicable in 6 out of 22 patients, probably because the probe was not sharp enough to penetrate in such a stiff tissue as many pancreatic masses are.

The monopolar Habib EUS-RFA catheter (Emcision) with RITA 1500X or ERBE ICC200 RF generators has been used by Pai et al.[99] to treat eight patients. Six of them had mucinous cystic neoplasms of the pancreas and two had pancreatic neuroendocrine tumors (PNETs). No adverse events have been reported. Mild abdominal pain, resolved within 3 days, was the only adverse event. In two cases a complete resolution of the cyst was observed; in the other four, reduction in cyst size was obtained. With regard to the two patients with PNETs, a central necrosis with a change in vascularity was described after treatment. Despite the positive results of this method, the efficacy of RFA of cystic neoplasm and neuroendocrine tumors of the pancreas has not previously been demonstrated, differently from pancreatic cancer, and then, larger series are needed to propose RFA for the treatment of these neoplasms.

The last abovementioned RFA probe is a monopolar 18-gauge RFA electrode 140 cm long (STARmed^®), with a sharp conical 1 cm tip for energy delivery and an internal cooling system connected via a pump to an external cold (0 °C) saline solution source (Fig. 3.14). The VIVA RF generator (STARmed^®) has variable wattage settings. This system has been tested in two reports. In the first one, Lakhtakia et al. (2015) [100] ablated three patients affected by insulinomas who refused surgery or were considered unfit for surgery. No post-procedure complications were reported. In all the three cases, a rapid relief of insulinoma-related symptoms and a biochemical improvement were obtained, and the patients remained symptom-free at 11–12-month follow-up. In the second report by Song et al. (2015) [101], six patients with stage III or IV pancreatic cancer were treated. No post-procedure complication was described. At short-term follow-up, in only one case, CT scan showed necrosis with air bubbles in the tumor without infection or perforation; long-term follow-up was not reported. In our opinion, considering the reported studies, this last device resulted the most effective and reliable because of the similarity to a normal FNA needle and because no failures in insertion of the probe into the tumor were described. We recently started a feasibility and safety study of EUS-RFA using this device in five patients with stage III pancreatic adenocarcinoma; the first two cases have been performed with no postoperative complications or adverse events. No follow-up data are available because the procedures have been done very recently.