Locally Advanced Pancreatic Cancer: Percutaneous Management Using Ablation, Brachytherapy, Intra-arterial Chemotherapy, and Intra-tumoral Immunotherapy

Within the management of pancreatic cancer, endoscopic ultrasound (EUS)–guided RFA has been successfully used for pain palliation by targeting the celiac ganglion [36]. However, the additive value of cytoreductive RFA in the context of LAPC remains controversial, regardless of the utilized method (i.e., open [37‐43], percutaneous [16], endoscopic [44‐46]). Literature on minimally invasive (percutaneous or endoscopic) RFA is scarce. Three articles employed EUS and one utilized a percutaneous approach. With a median overall survival (mOS) of 6 months from cytoreductive RFA, outcome was poor. Compared to the open approach, these methods demonstrated significantly reduced morbidity (0% vs. 53%) and mortality (0% vs. 6%). D’Onofrio et al. reported on the only percutaneous series of patients with unresectable PDAC (n = 18), all of whom were pre-treated with chemotherapy [16]. Technical success was achieved in 16 patients (93%), mOS was 6 months, and none of the patients experienced major complications. Most frequently reported severe complications for pancreatic RFA include pancreatitis, biliary injury, portal vein thromboses, pancreatic fistulas, hemorrhages, duodenal perforations, and gastric ulcers or fistulas [47]. The positioning of the pancreas, entwined with delicate structures, complicates the utilization of thermal ablative methods. For example, the heat-sink effect of adjacent large blood vessels can hinder successful RF ablation. Also, to ensure minimal collateral thermal damage, a safe distance to delicate surrounding parenchyma (i.e., blood vessels and duodenum) must be maintained [48]. This can be minimized by constant intra-procedural cooling of such structures. Ambiguity regarding procedure regimens (temperature, power, duration, minimum distance to vulnerable structures) and timing prior to or after chemo(radio)therapy remains [40, 41]. The currently recruiting randomized controlled PELICAN trial aims to determine the survival benefit of chemotherapy and cytoreductive RFA compared to sole chemotherapy (NCT03690323).

The immunomodulation presented after RFA is a result of ICD through the release of antigens and DAMPS such as interleukin (IL)-1, IL-6, IL-8, and tumor necrosis factor (TNF)-α [35]. Specific for thermal tissue damage is the release of a DAMP called heat shock protein (HSP)-70, involved in proper folding and transport of proteins, and believed to play a key role in the immunological response [49]. HSP-70, mainly exposed in the peripheral non-coagulative ablation zone, is elevated in the serum of patients after RFA, which can lead to an immunological anticancer response through activation and maturation of DCs [50, 51]. However, in this peripheral zone, created by diffusion of heat outwards, IL-6, HSP-70, and hypoxia-related pathways have also been implicated to stimulate outgrowth of tumor cells in this area, thus potentially causing (early) recurrences [52‐55]. In the coagulative central zone, protein denaturation and destruction of the blood and lymph vessels impedes proper immune infiltration as well as antigen diffusion or transport and subsequent presentation in draining lymph nodes. Clinical evidence of an RFA-induced immune response in PDAC is currently limited to one publication. In this article, Giardino et al. analyzed pre- and post-operative peripheral blood of 10 LAPC patients and found RFA to activate adaptive immune subsets (CD4⁺ and CD8⁺ T-cells) and myeloid DCs, while maintaining stable numbers of immune inhibiting Tregs [56•].

Three groups previously described results following percutaneous MWA in LAPC [17‐19]. They included a total of 32 patients treated under US or computed tomography (CT) guidance, mostly focusing on the technique’s feasibility and safety. All reported 100% technical success rates, without any 30-day mortality. Carrafielo et al. included 10 patients (5 percutaneous, 5 laparotomic) with LAPC unresponsive to chemotherapy and were the only group reporting severe morbidity (n = 2, 20%) and survival outcomes (1-year OS of 80%) [17]. Major complications included one (early) pancreatitis and one (late) pseudoaneurysm of the gastroduodenal artery. All articles were uniform to conclude the technique’s ability to temporarily improve quality of life, its feasibility, and safety. However, comprehensive and longer term survival results are lacking, thus conclusions on the efficacy of MWA in LAPC are premature. Nonetheless, similar to RFA, a major drawback of this technique remains the use of thermal energy in a highly delicate organ surrounded by vulnerable structures.

MWA induces a similar immune response as described for RFA, with upregulation of serum HSP-70, although apparently to a lesser extent [59]. (Pre-)clinical literature on immune modulation of MWA in the context of pancreatic cancer is lacking.

To date, four articles have been published on the use of cryoablation in the context of LAPC, two in an open [61, 62] and two in a percutaneous setting [12, 63]. As monotherapy, cryoablation demonstrated a mOS of 12.6 months [63], and in combination with internal radiotherapy a mOS of 16.2 months [12] was noted (neither article specified whether OS from diagnosis or treatment). Niu et al. were the only group reporting on sole percutaneous cryoablation and included patients with stage II (n = 3), III (n = 11), and IV (n = 18) PDAC whose tumor was deemed unresectable [63]. Clinical benefit response (based on pain scores and analgesic consumption) was 84.4%, and the mOS was 12.6 months (incl. all stages). In order to overcome potentially incomplete destruction at the ablation border, Xu et al. described percutaneous cryoablation in combination with brachytherapy in LAPC (n = 49), demonstrating a mOS of 16.2 months with a 6% severe complication rate [12]. The freezing process can injure delicate parenchyma including the duodenum, bile ducts, or blood vessels, resulting in inflammation, fistulas, abscesses, or bleeding. Division of surrounding tissue or application of warm pads may reduce the risk of these freezing-related injuries [59].

Cryoablation has been described as a potent immunostimulatory inducer among thermal ablative therapies, able to release most non-denatured proteins and induce profound DC antigen loading [64, 65]. It has even resulted in extreme cases of inflammatory responses owing to the induced cytokine storm which increases vascular permeability, resulting in abundant tissue edema and in some cases cryoshock [66]. On the other hand, cryoablation has also been described as an immunosuppressant modality [67]. Increased levels of IL-10, an immunosuppressive cytokine that promotes Treg differentiation, have been proposed as part of the explanation for this [68, 69]. Furthermore, portions of the ablation zone may not exhibit ICD, thus impeding generation of a proper anti-tumor immune response [49]. In this regard, the rate of tissue freezing has been suggested as an important determinant of immunologic response [70]. Literature on the immune response after cryoablation in pancreatic cancer patients is currently limited to one pre-clinical study. White et al. compared early immunological responses after cryoablation and IRE in a rodent model of pancreatic cancer. They noted no significant changes in any immune cell type after cryoablation, compared to a robust intra-tumoral infiltration of macrophages and CD3⁺ T-cells after IRE [71].

Studies on percutaneous IRE demonstrate survival outcomes after diagnosis (mOSd) between 13.3 and 27 months (vs. 15.3–24.9 months in open IRE), and survival outcomes after treatment (mOSt) between 7 and 27 months (vs. 6.4–12 months in open IRE). A large database study has established that IRE in combination with chemotherapy (mOS 16 months) prolongs survival compared to sole chemotherapy (mOS 8 months) after propensity score matching [83]. Leen et al. reported on the largest LAPC cohort (n = 75) with the longest survival outcomes (mOSt 27 months). These results may in part be explained by the retrospective nature of the research, inherently leading to selection bias, in combination with favorable patient selection criteria. The largest prospective cohort was reported by Ruarus et al., who treated 40 LAPC and 10 recurrent patients using (ce)CT-guided percutaneous IRE. They reported mOSd and mOSt of 17 months and 9.6 months, respectively. Major complication rates varied substantially among studies on percutaneous IRE (0–40%) but were generally lower compared to open procedures (8–53%). The most frequently encountered major complications during and after IRE procedures include pancreatitis, fistulas, ileus, or delayed gastric emptying, intra-abdominal hemorrhages, ulcers, portal vein thromboses, and biliary-related issues (leakage and/or cholangitis). To reduce the chances of IRE-related infections, prophylactic antibiotics are a must. Reported mortality rates after treatment varied between 0 and 6% (vs. 0–13% in an open setting). Notably, Xu et al. have recently published on the importance of adjuvant chemo- and/or radiotherapy after IRE, concluding that additive adjuvant therapy significantly improves survival [84].

Immunomodulation in IRE has been described as the most potent compared to thermal ablative techniques, specifically in terms of protein release and T-cell activation [64]. With IRE specifically, hypoxia is relieved by increasing blood vessel permeability and microvessel density, counteracting hypoxia-induced immunosuppression [85••]. In addition, the preservation of lymph vessels allows DCs to transport released antigens from the ablated area to the draining lymph nodes. Thereafter, the activated effector T-cells are able to infiltrate the residual ablated mass owing to intact blood vessel structures. Scheffer et al. have demonstrated the clinical immune modulatory effects after percutaneous IRE in the peripheral blood of 10 LAPC patients [86•]. Two weeks post-IRE, Tregs were decreased, PD-1⁺ T-cells were elevated, and boosting or de novo induction of a Wilms Tumor-1 (pancreatic TAA)-specific T-cell response was observed. These findings are consistent with a systemic and tumor-specific immune stimulatory effect. Pandit et al. also evaluated the post-IRE immune response in LAPC patients (n = 11) [87•], in which three Treg subsets displayed an inverse correlation with time, demonstrating attenuation of immunosuppression.

As stand-alone local therapy, median survival outcomes in LAPC patients percutaneously treated ranges between 7.3 and 11 months. Additional increase in survival can be achieved if ¹²⁵I seed implantation is combined with other interventional modalities such as cryoablation (mOS 16.2 months) [12] or TACE (mOS 17.6 months) [8]. Reported severe morbidity varies between 0 and17% (vs. 0–11% open setting), whereas mortality among studies was 0% (vs. 1.3% open setting). The high rate of severe morbidity reported by Yang et al. (17%) can be explained by patient selection, including only LAPC patients whose disease was complicated by obstructive jaundice [28]. Most common (major) complications for percutaneous ¹²⁵I seed implantation include pancreatitis, fistulas, seed migration, ulcers, infections, leakage, and intestinal perforations [29].

Apoptosis is the predominant form of cell death in radiation-based therapies, but high doses of radiation can also lead to necrosis. Similar to other ablative therapies, paradoxical immunomodulatory effects have been reported after radiotherapy treatment [96]. Immunopermissive effects may be initiated by a combinatory release of antigens and DAMPs, followed by activation of the adaptive system. However, radiotherapy has been suggested to promote inactivation of DCs and NK cells, consequently leading to recruitment of immunosuppressive MDSCs and Tregs. There is currently no literature published on the immunomodulatory effects of internal radiation therapy in pancreatic cancer, but external radiation effects were demonstrated by Fujiwara et al. in a subcutaneous PDAC mouse model [97]. Radiotherapy appeared to induce innate immune permissive responses by activating Toll-like receptors (TLRs) and pro-inflammatory chemokines. However, effects on the adaptive system seemed double-edged as immune activating (TNF receptors) and immune suppressive (transforming growth factor (TGF)-β) pathway genes were upregulated. Furthermore, intra-tumoral CD8⁺ T-cells remained scarce.

A meta-analysis compared systemic administration of chemotherapy (n = 143) with IAIC (n = 155) [9]. The analysis comprised 6 randomized controlled trials (RCTs) including 298 patients with advanced pancreatic disease (stages III/IV). They concluded that IAIC resulted in increased median survival (5–21 months vs. 2.7–14 months), superior clinical benefits (78% vs. 29%), and fewer overall complications (49% vs. 71%) and hematological side effects (61% vs. 86%). Although chemotherapeutic toxicity (e.g., neutropenia, thrombocytopenia) rates are lower, they can still arise in IAIC. In addition, specifically with IAIC, (post-procedural) infections or vascular dissections may occur. Although IAIC has shown to be of value, the expansion into common clinical practice is limited by the difficulty of the procedure, with i.v. chemotherapy being more readily available and cheaper. However, especially for patients resistant to standardized systemic chemotherapy, this may be a suitable neo-adjuvant or palliative treatment option.

(Locoregional) Chemotherapy has been implicated to elicit tumor-specific immune responses [14, 99]. The apoptotic ICD leads to extracellular accumulation of nucleic acids (DAMPs), promoting release of type I interferon (IFN), followed by maturation of DCs and activation of effector T-cells [100]. However, to date no literature has been published specifically describing the immune response after local administration of chemotherapy in pancreatic cancer. Michelakos et al. describe clinical results in PDAC patients (n = 248) of whom a portion received systemic FOLFIRINOX (± radiation) and were compared to untreated controls [101]. The treated group exhibited dense CD8⁺ T-cell infiltration, high CD4⁺ T-cell numbers, and low Treg cell density.