Background
Over 55,000 people will develop pancreatic cancer this year in the United States, with 95% of cases being fatal, making it the fourth leading cause of cancer-associated mortality [
1‐
3]. As a lethal malignancy with a dismal 5-year survival rate of 8%, it has the lowest survival rate of any major cancer and is projected to surpass colorectal cancer as the second leading cause of cancer-related death by 2030 without breakthroughs in prevention and therapy as we have seen in colon cancer [
1‐
3]. While the majority of pancreatic cancers are ductal adenocarcinomas, approximately 20% of pancreatic cancer develops from mucinous-type pancreatic cysts, and this percentage may be underreported [
2,
4]. Pancreatic cystic lesions are typically discovered incidentally on cross-sectional imaging, occurring in approximately 2% of American adults, with a 37% prevalence in individuals older than 80 years old [
5]. The prevalence of pancreatic cysts in the United States has grown dramatically over the last two decades due to an aging population and advances in imaging techniques, leading to a public health dilemma. Although certain types of pancreatic cysts carry little to no malignant potential, the majority are neoplastic and includes mucinous cystic neoplasm (MCN) and intraductal papillary mucinous neoplasm (IPMN), which carry a clinically significant potential for malignant transformation. The natural history of a mucinous pancreatic cyst is variable, with the overall risk of progression to pancreatic cancer progressing to invasive cancer generally linked to its number of high-risk features [
6,
7•]. The overall risk of progression to malignancy for a mucinous cyst without high-risk features is reported to range from 1 to 25%, and it is difficult with current imaging techniques and cyst fluid analysis to discern which cyst will undergo malignant transformation [
2,
8‐
10]. Molecular testing of cyst fluid may improve this risk stratification but is very expensive and requires further validation [
7•,
11,
12]. Identification of a mucinous pancreatic cyst requires the clinician and patient to choose between indefinite radiographic surveillance (MRI or CT) or surgical resection, both of which have considerable limitations. Surveillance for malignancy carries significant economic and psychological burdens, and CT imaging includes some degree of radiation exposure and possible nephrotoxicity. On the other hand, surgical resection possesses a substantial risk for serious adverse events (20–40%) and mortality (1–5%) and still requires post-operative surveillance [
13‐
15]. This clinical dilemma highlights an pressing need to develop effective, less expensive, and more minimally invasive approaches for this patient population.
In this respect, endoscopic ultrasound (EUS)-guided pancreatic cyst ablation has emerged as an innovative, promising minimally invasive approach for the treatment of neoplastic pancreatic cysts [
16•,
17]. EUS-guided cyst ablation was first demonstrated by Gan et al. where 80% alcohol was infused into all types of pancreatic cysts (after cyst fluid aspiration) and lavaged for 3–5 min. Patients were then followed for 6 months. Overall, a complete response rate of 35% rate was found with a 0% risk of adverse events [
18]. This study was followed by the prospective randomized EPIC trial, which demonstrated a 33% rate of complete cyst ablation after ethanol lavage with a serious adverse event rate of 4–5% (pancreatitis) [
19]. To date, ten published studies have investigated the safety and efficacy of EUS-guided pancreatic cyst ablation using ethanol, only two of which used a randomized design. These studies are summarized in Table
1. Although ethanol ablation is feasible, the use of alcohol alone for pancreatic cyst ablation has consistently resulted in disappointing efficacy. These poor results were illustrated most recently in a prospective trial by Gomez et al. which found a dismal 9% rate of complete ablation and a 4% risk of pancreatitis [
22]. Overall, the clinical value of pancreatic cyst ablation with ethanol alone is unfavorable due to the suboptimal response and significant adverse event rates and its use as a single agent should be abandoned.
Table 1
Summary of EUS-guided ablation trials: ethanol, ethanol + paclitaxel, alcohol vs. gemcitabine + paclitaxel. Italics = EtOH only. Bold = EtOH + paclitaxel. Bold italics = EtOH + paclitaxel–gemcitabine vs. saline + paclitaxel–gemcitabine
|
Prospective (pilot)
|
5–80% EtOH (25)
|
35% CR
7% PR
|
0%
|
|
Prospective (RCT)
|
80% EtOH (25)
Saline (17)
|
33% CR
2
0% CR
|
24% (4% Pancreatitis, 20% Ab pain)
12% (Ab pain)
|
|
Retrospective
|
80% EtOH (13)
|
38% CR
|
8% (Ab pain)
|
|
Retrospective
|
99% EtOH (13)
|
85% CR
|
–
|
|
Prospective
(pilot)
|
80% EtOH (23)
|
9% CR
44% PR
|
8% (4% Pancreatitis, 4% Ab pain)
|
|
Prospective
|
88–99% EtOH + paclitaxel (14)
|
79% CR
14% PR
|
21% (7% Pancreatitis, 14% Ab pain)
|
|
Prospective
|
99% EtOH + paclitaxel (10)
|
60% CR
20% PR
|
10% (Ab pain)
|
|
Prospective
|
99% EtOH + paclitaxel (47)
|
62% CR
13% PR
|
4% (2% Pancreatitis, 2% Ab pain)
|
|
Prospective
|
100% EtOH + paclitaxel (22)
|
50% CR
25% PR
|
23% (10% Pancreatitis, 13% Ab pain)
|
|
Prospective
(RCT)
|
80% EtOH paclitaxel + gemcitabine (18)
Saline + paclitaxel + gemcitabine (21)
|
61% CR
67% CR
|
28% (6% Pancreatitis, 22% Ab pain)
0%
|
Oh et al. added the innovative step of infusing and leaving paclitaxel within the pancreatic cyst after ethanol lavage [
23]. The addition of paclitaxel (a chemotherapeutic agent that arrests cellular microtubule assembly and mitosis) has been shown in separate trials to raise complete ablation rates to 50–79% (Table
1). The significant increase in efficacy of ethanol lavage plus paclitaxel infusion compared to ethanol alone has resulted in this combination therapy becoming the preferred approach for pancreatic cyst ablation today and offering complete ablation rates similar to that seen in other ablative strategies in gastroenterology.
Despite this increase of efficacy, a significant limitation of alcohol ablation with or without paclitaxel for pancreatic cyst ablation has been the associated serious adverse events of pancreatitis, peritonitis, and venous thrombosis in 0–10% of patients [
18‐
26,
27••]. Importantly, the mechanism for these reported complications is believed to be secondary to the potent inflammatory and toxic effects of alcohol on the surrounding pancreatic parenchyma and/or neighboring vessels. The recently published prospective, randomized, double-blind ChARM trial evaluated a completely alcohol-free chemoablation approach [
27••]. In this study, 39 patients were randomized to saline cyst lavage or conventional alcohol lavage with both arms then treated with a chemotherapeutic cocktail tailored specifically for pancreatic neoplasia (38 mg gemcitabine + 6 mg/mL paclitaxel). At one year post-treatment, 61% of patients in the alcohol arm achieved complete ablation compared with 67% in the alcohol-free arm. These findings suggest that alcohol is not required for effective pancreatic cyst ablation when a chemotherapy cocktail specifically designed for pancreatic neoplasia is used. More importantly, the rate of adverse events was significantly lower in the alcohol-free arm (
p = .01) as all minor and serious adverse events occurred in the alcohol arm. Minor adverse events were comprised entirely of abdominal pain, occurring in four patients in the alcohol arm (22%) and in zero patients in the alcohol-free arm. Collectively, these findings demonstrate that removal of alcohol from pancreatic cyst ablation reduces adverse events to that comparable to EUS-FNA. This is an important development, since significantly improving the risk profile of the procedure, while preserving its efficacy, increases the attractiveness of this therapeutic option for patients with pancreatic cysts. This trial has led to a larger prospective, randomized, NIH-funded, multicenter trial using a larger sample size and several technical improvements to the procedure with the results expected in four years [
28].
An important metric of the ablative approach is the treatment durability over time. Two trials to date have addressed this, with DeWitt et al. demonstrating the long-term durability of pancreatic cyst ablation at two years, showing that the majority of patients who have an effective ablation will also have elimination of detectable KRAS mutations [
26]. In a recent large, prospective, follow-up study of 164 patients undergoing EUS-guided ablation with alcohol followed by paclitaxel, Choi et al. demonstrated that when patients achieved complete EUS-guided pancreatic cyst ablation, 98.3% remained in remission at six-year follow-up, demonstrating an excellent durable response following ablative therapy [
29].
Multiple areas of uncertainty remain with this approach, including demonstration of clinical reduction in incidence rates of pancreatic adenocarcinoma in appropriately matched groups, a better understanding of which patients are best treated with this approach, and an evidence-based demonstration of the financial profile of this approach compared with a surveillance and/or surgical strategy. An international white paper addressing some of these issues is expected to be published in the coming year.
Patient Selection
If EUS-guided cyst ablation is to add significant clinical value; it should not be performed on certain high-risk lesions which are unlikely to fully respond to treatment. It should also be avoided in small, low-risk pancreatic cysts with little chance of malignant progression as they are more appropriate for routine surveillance. Ablative strategies should be focused on cystic tumors which are technically amendable to cyst ablation and do not show overt signs of malignancy.
Only cysts with a likelihood of progression to malignancy are currently candidates for ablation. For that reason, the first step is to use clinical, radiographic, cytologic, and chemical analysis by EUS-FNA to diagnose and risk stratify a mucinous cyst [
30]. There remains a level of uncertainty in the accurate diagnosis of pancreatic cyst type in 2019, and further investigation is required to improve this accuracy. Although 2% of the US population will have a pancreatic cyst, the estimated prevalence of cysts > 2 cm is estimated at only 0.8% [
31]. In a recent large study evaluating the long-term risk of pancreatic malignancy in patients with mucinous pancreatic cysts, 577 patients with presumed branch duct IPMN under surveillance at Massachusetts General Hospital were evaluated for a primary outcome of risk for malignancy over five years, compared to the US population. The overall rate of development of malignancy was 5.5%; however, of patients with cysts 1.5 cm or smaller, the overall risk of malignant transformation was 0.9%, demonstrating a very low risk for progression to malignancy for mucinous cysts < 1.5cm without other high-risk features [
32]. Additionally, it is technically challenging to perform EUS-guided pancreatic cyst ablation on cysts < 1.5 cm, as there is 0.8cc of volume within the FNA needle itself, which prohibits meaningful exchange of chemotherapeutic injectate. Consequently, cyst size ranging from 1.5 to 5 cm is the typical range of pancreatic cysts treated in most trials to date [
16•,
17]. Cyst morphology is an important consideration. Multiseptated cysts present a unique challenge to ablation since treatment must ensure that each loculation of the cyst is injected with the ablation agent. Although cysts can appear multiseptated on MRI-MRCP, EUS-FNA may collapse the entire cyst, indicating communication between these individual compartments which may be not appreciated on cross-sectional imaging. Most trials to date have favored unilocular to oligolocular cyst morphology and have avoided cyst ablation in lesions with more than 4–5 discrete cell chambers [
16•].
When reviewing previous trials and the experience of our centers, the recommended ideal indications and contraindications for pancreatic cyst ablation can be used to guide therapy and are summarized as the following:
Conclusions
EUS-guided pancreatic cyst ablation is a promising and emerging minimally invasive technique for the elimination of a known precursor lesion to a lethal malignancy which currently has poor therapeutic options limited to either a morbid surgery or indefinite surveillance. Careful patient selection is required, as is careful attention to the technical aspects of the procedure, post-procedure evaluation, and follow-up surveillance. In our experience, patients are almost universally in favor of eliminating an appropriate neoplastic-type pancreatic cyst and overwhelmingly favor a minimally invasive approach to do so. An emerging body of literature demonstrates the efficacy of appropriately performed pancreatic cyst ablation, a durable effect when a complete response is achieved, and a significantly improved safety profile with an alcohol-free protocol. Where cyst ablation fits into current treatment algorithms should be an ongoing discussion based on efficacy, safety, and the risk-benefit ratio unique to that patient. Additionally, EUS-guided ablation may offer a favorable cost profile when compared to EUS-FNA with molecular testing, Whipple surgery, or surveillance MRI-MRCP.
There are several areas of limitations and uncertainty with this technique. Most notably, there are no trials yet which demonstrate that this approach objectively reduces a patient’s progression to pancreatic adenocarcinoma. This emerging technique represents an exciting treatment option for appropriately selected patients; however, further studies are required to further develop the efficacy, safety, and treatment indications and to define which patients are best offered this emerging treatment option.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.