Quantity of Research Available
A total of 207 citations were identified in the literature search. Following screening of titles and abstracts, 170 citations were excluded and 37 potentially relevant reports from the electronic search were retrieved for full-text review. Three potentially relevant publications were retrieved from the grey literature search. Of these potentially relevant articles, 24 publications were excluded for various reasons, while 16 publications (two of which reported on the same study) met the inclusion criteria and were included in this report. Appendix 1 shows the PRISMA flowchart of the study selection.
Summary of Study Characteristics
Study Design
Pancreatic Tumors
Two of the included studies9, 10 were systematic reviews. The Rombouts SR10 reviewed four non-randomized studies that included patients with pancreatic tumors and the Scheffer SR9 reviewed non-randomized studies of patients with pancreatic cancer and patients with liver cancer; four of the studies were in patients with pancreatic cancer.
Five of the included primary studies examined patients with pancreatic tumors.11–15 Three of the studies11–13 were prospective before and after studies without a control group and two of the studies were retrospective before and after studies without a control group.14, 15
No relevant economic studies were identified.
Liver Tumors
One of the included SRs reviewed studies of patients with either liver or pancreatic cancer,9 eight of which included patients with liver cancer.
Nine of the included citations describing eight studies examined patients with liver cancer.11, 16–23 Six citations describing five studies were prospective before and after studies without a control group11, 16–20 and three21–23 were retrospective before and after studies without a control group.
Country of Origin
Pancreatic Tumors
Both of the systematic reviews that examined studies of patients with pancreatic cancer were conducted in the Netherlands.9, 10 Two of the primary studies were conducted in Italy11, 13 and three in the United States.12, 14, 15
Liver Tumors
The systematic review that included studies of patients with liver cancer (along with studies in patients with pancreatic cancer) was conducted in the Netherlands.9 Six of the included primary studies were conducted in Germany,16, 17, 19, 22–24 one in Japan,20 one in the Netherlands,18 and one in the United States.21
Patient Population
Pancreatic Tumors
All of the studies included patients with locally advanced pancreatic cancer (LAPC).10–15 The SRs examined the results of 14110 and 699 patients included in the various studies. The number of included patients in the primary studies ranged from ten13 to 200.14 Mean and median ages ranged from 6214 to 69.2 years11 and the percentage of male participants ranged from 40%13 to 58.5%.12 Patients in all studies either received chemotherapy and/or radiation therapy prior to or in conjunction with IRE.
Liver Tumors
The systematic review included 129 patients with hepatocellular carcinoma (HCC), colorectal liver metastases (CRLM), or other liver tumors,9 and the median age of the patients included in the SR studies ranged from 51 to 65. The number of included patients in the primary studies ranged from five20 to 52.22 Mean and median ages ranged from 6122 to 66.620 and the percentage of male participants ranged from 4018 to 82%23 The percentage of male participants ranged from 40%18 to 82%,15 and one study did not report the percentages of male or female participants.21
Interventions and Comparators
Pancreatic Tumors
All of the studies performed IRE using the NanoKnife. The systematic reviews included studies that used open, percutaneous, and laparoscopic approaches to IRE.9, 10 One of the included primary studies used a percutaneous approach,11 one used a laparoscopic approach,13 and two used an open surgical approach.14, 15 The approach was unclear in one study, but as most of the IRE procedures were done in conjunction with an operative procedure, it is likely the open approach was used.12
Liver Tumors
All of the included studies performed IRE using the NanoKnife. The systematic review included studies that used open, percutaneous, and laparoscopic approaches to IRE.9 Six of the included primary studies used a percutaneous approach,16, 20–24 and one used a laparoscopic approach.18 One study included patients who underwent open, laparoscopic, or percutaneous IRE.19
Outcomes
Pancreatic Tumors
The included studies reported survival outcomes, complications, and outcomes related to disease response and progression. More specifically:
seven studies reported general adverse events or complications,
10–15, 25one reported IRE related adverse events or complications,
10four reported overall mortality,
9–12one reported IRE related mortality,
10one reported disease progression,
9two reported disease recurrence following IRE,
12, 14two reported tumor response to IRE
13, 15one reported progression-free survival,
14one reported change in tumor volume.
11
Liver Tumors
Similarly, the studies examining liver tumors also reported survival outcomes, complications, and outcomes related to disease response and progression. More specifically:
eight studies reported general adverse events or complications
9, 18–24four reported IRE related complications or adverse events
9, 16, 18, 23two studies reported mortality
9, 21one studies reported disease progression
9two reported disease recurrence following IRE
19, 24two reported tumor response to IRE.
16, 20
Additional information regarding study and patient characteristics is reported in Appendix 2.
No relevant economic studies were identified, however, four studies13–15, 19 reported length of hospitalization. This information is reported in Appendix 4.
Summary of Critical Appraisal
The authors of the two included systematic reviews9, 10 both reported that they conducted the review according to PRISMA guidelines, and while it is assumed that this was followed, neither reported an a priori design in the text. Rombouts et al. reported that two authors performed selection of articles, but they did not report two authors performing data extraction or critical appraisal. There was no mention of critical appraisal at all, so it is unclear if it was done or if study quality was considered in the conclusions.10 Scheffer et al.9 did not mention double screening, extraction, or quality appraisal. Although the database searches were comprehensive for both SRs, it was unclear if the Rombouts review included a grey literature search.9 Although that may normally pose a potential risk for publication bias, the Scheffer review9 was conducted in a similar time period, searched the grey literature, and did not find any relevant inclusions from the grey literature. The authors of both SRs did not perform a meta-analysis, however complication rates were combined to aggregate the results. It is unclear if this was the least biased approach, however the authors did discuss the limitations of their reviews. None of the included studies in the SRs used a design that included a comparator group, thus the results are only descriptive and not comparative.9, 10 Additionally, Scheffer et al.9 considered most of the results of the included studies together, as opposed to separating the results based on the surgical approach, it was therefore difficult to determine whether the type of surgery (open, laparoscopic, or percutaneous) was taken into account. As open surgeries are generally more lengthy and are often associated with more adverse events when compared with less invasive approaches, this may not have been an appropriate choice. Rombouts et al.10 considered outcomes together as well as separated by surgical approach and did differentiate between open and percutaneous approaches in their conclusions.
None of the included primary studies randomized patients to interventions, had a comparator group, or attempted blinding of study participants or staff. Thus, the results of all of the included studies are descriptive and are not indicative of comparative results. Additionally, due to the lack of blinding, it is possible that patients or staff could be biased toward new technology.
Most of the included studies clearly reported the interventions, methods, and characteristics of patients,11–16, 18, 19, 22–24 though Sugimoto20 and Hosein21 did not report a description of the 5 and 29 patients who participated in their studies. Although a patient was recruited to replace the patient who could not receive IRE in the Eisele study,19 the characteristics of the patient who did not receive IRE were not well reported. Compliance with IRE was reliable in all studies.11
The Scheffer et al. primary study18 included a population that was 60% female. As liver cancer is more prevalent in men than women (the Canadian Cancer society statistics suggest that 75% of those diagnosed with liver cancer in 2016 will be men), this sample was likely not representative of the general population with liver cancer. Eller et al. included a population that was 79% male – while liver cancer is indeed more prevalent in men, women were likely underrepresented in the sample. While pancreatic cancer tends to be more prevalent in men in other countries (American men are 30% more likely to get pancreatic cancer than their female counterparts26), Canadian statistics suggest a more even distribution between the sexes2 and thus the 50/50 sex representation in the Belfiore study11 may be representative of the Canadian population with pancreatic cancer.
The included studies for both liver and pancreatic cancer examined patients with either unresectable primary cancers or unresectable metastatic cancer (one study also included patients with borderline resectable cancers)15 The patient populations were generally similar and the results are likely generalizable to the general population of patients with unresectable liver and pancreatic cancers.
Further detail regarding critical appraisal is available in Appendix 3.
Summary of Findings
What is the clinical effectiveness and safety of irreversible electroporation (IRE) in patients with tumors of the pancreas or liver?
Pancreatic Tumors
IRE Efficacy, Disease Progression, Survival
In the Rombouts SR,10 one study of patients receiving open surgery IRE (n = 54 patients) reported that IRE increased survival from an average of 11 months to an average of 20 months. These patients received IRE following unresectability and received IRE in conjunction with palliative bypass (where indicated) and chemotherapy. In the study using percutaneous IRE, the 6-month overall survival was 70% though patients with metastatic disease showed no improvement in survival following IRE and died “soon after the IRE procedure.”10 In the Scheffer SR,9 one study compared IRE patients to propensity score matched chemo-radiation patients. IRE was associated with improved local (14 months versus 6 months, P = 0.01), distant (15 months versus 9 months, P = 0.02), and overall survival (20 months versus 13 months, P = 0.03). After 20 months, there was no difference in survival between the groups.
Mean survival ranged from 7.5 months (range: 2.9 to 15.9) in the Paiella study (N = 10)13 to 12.95 months (95% Confidence Interval [CI] 11.57 to 14.33) the Belfiore study (N = 20).11 Median survival from the day of IRE was 12.03 months (95 CI, 7.71 to 23.12 months) in the Kluger study12 22 months (95% CI, 17.9 to 24.9) in the Kwon study15 and 24.9 months (range: 12.4 to 85 months) in the Martin study.14 Historical data on chemotherapy and chemoradiotherapy options reported by Martin et al.14 show a median survival of approximately 12 months.
Two studies reported median progression free survival; 11 months (95% CI, 3 to 10) in the Kwon study,15 12.4 months (range: 4.4 to 8.9) in the Martin study.14
Overall recurrence rate ranged from 29% in the Martin study14 58% in the Kluger study12 The Local disease control at the last available follow-up (range 6 to 14 months) was reported in all living patients (18/20) in the Belfiore study.11
Complications and Adverse Events
The Rombouts systematic review10 reported an overall complication rate of 48% (in 4 studies), with open surgeries having a complication rate of 51% and percutaneous approaches having a rate of 27%. Two studies reported complications that were specifically judged to be IRE-related; the IRE-related complication rate was 13% (15% in open, 9% in percutaneous surgery) and the morbidities included duodenal leakage, pancreatic leakage, bile leakage, and (progression of) portal vein thrombosis.10 The Scheffer SR9 reported an IRE-related complication rate of 19%, with 7% of the IRE-related complications being classified as “major”. The authors noted that one study reported general adverse events that were not considered IRE-related but did not describe them in the review.9
No major complications were reported in four of the primary studies.11, 13–15 The overall rates of minor complications (mostly gastrointestinal) ranged from 10%11 to 40%.14 Major morbidities (e.g. gastrointestinal bleeding, deep surgical site infection, need for stent placement) were reported in one study;12 the rate of major morbidities was 19%, 44% of which were thought to be IRE-related. Minor IRE-related complication rates were reported in one study; 11% of patients (n = 5) had a complication thought to be IRE-related two of the evets were bleeding events.15
Mortality
The overall mortality rate (based on the 4 included studies) reported in the Rombouts SR10 was 3% (2% in open, 9% in percutaneous surgery) and the IRE-related mortality (reported in 4 studies) was 2% (3% in open, 0% in percutaneous surgery). The Scheffer SR reported no periprocedural deaths, one possible IRE-related death (a rate of 2.3%) and three deaths (N = 43) in the three months following IRE for pancreatic tumors. The cause of these deaths was not reported but they were not attributed to IRE.
None of the included primary studies reported procedure-related mortality. In the Belfiore study (N = 20),11 two patients died 3- and 4-months after IRE due to disease progression and all patients died in the Paiella study13 (nine from advanced disease and one from septic shock that was related to ulcerative colitis and not IRE). Median postoperative mortality in the Kluger study12 was 26 days (interquartile range 8 to 42 days). No patients had died at the 90-day follow up in the Kwon study15 and the Martin study14 reported progression free survival, which was included in the section on disease progression.
Liver Tumors
IRE Efficacy, Disease Progression, Survival
The Shaffer SR reported six to 18 month efficacy data for 106 patients in five studies.9 Primary efficacy (defined as percentage of tumors successfully eradicated after the initial procedure based on follow-up imaging after 3 months) ranged from 67% to 100% and secondary efficacy (defined as successful tumor eradication 6 months after the first treatment) ranged from 55% to 93%.
Tumor control or response rate (generally defined as no tumor recurrence or tumor growth) at the final follow-up was reported in two studies and ranged from 71.4%24 to 83.3%.20 Complete response ranged from 18%21 to 91.7%16 and partial response 8.3%16 to 18%21 in the two studies that reported it. Overall recurrence rate was reported as 38% in the Eisele study.19
Overall and 2-year progression free survival was reported as 4.0 months (95%CI 1.4 to 6.6) and 18% (95% CI, 0% to 35%) in the Hosein study.21
Complications and Adverse Events
No serious IRE-related adverse events were reported in the Scheffer SR.9 Both the total and IRE-related complication rates were reported as 16%, 6% of which were stenosis or occlusion of portal vessels or bile ducts.
IRE-related adverse events were reported in 2 of 24 patients in the Granata study,16, 17 neither of which were considered major and one IRE-related event (ventricular extrasystoles without haemodynamic changes) was reported in the Scheffer study.18 In two patients in two studies, the IRE procedure either had to be prematurely halted (due to bleeding)24 or could not be performed due to pre-operative complications.21 Overall complication rate was reported in one study; 18.8% in the Dollinger study.22 Major adverse events were reported in two studies and ranged from 8%22 to 28.5%.24
Mortality
Mortality was not reported for liver cancer patients in the Schaffer SR.9 It was not clear if there were no mortalities in the included studies or if the authors did not report them. No treatment related mortalities16, 18–24 occurred in the primary studies.
Further detail regarding the clinical effectiveness and safety reported in the included studies is available in Appendix 4.
What is the cost-effectiveness of irreversible electroporation (IRE) in patients with tumors of the pancreas or liver?
No relevant cost-effectiveness or economic studies were identified.
Limitations
Although 15 studies were included in this review, the majority of the data is based on small, uncontrolled studies that were designed to demonstrate feasibility and early safety of the IRE procedure. The two systematic reviews included case series and case studies and the included primary studies did not include a comparator group. The results of this review are therefore descriptive and no conclusions can be made regarding the comparative efficacy of the IRE procedure versus other procedures or standard care.
The majority of the studies of patients with liver tumors included patients with tumors that were ≤ 3 centimeters (cm), thus the results are likely not generalizable to patients with larger tumors. Tumor sizes in the studies of patients with pancreatic cancer ranged from just under 3 cm to ≤ 6 cm, thus the results may be more generalizable to patients with various sizes of tumors.
In almost all of the included studies, the patients had either previously or were concurrently receiving chemotherapy, radiation therapy, or chemoradiation in order to treat their cancers. It is therefore difficult to determine whether tumor response or absence of disease progression was due to the IRE procedure alone. Additionally most studies examined the use of IRE in unresectable tumors, thus it is unclear whether results could generalize to patients with resectable tumors or if IRE is a treatment option for patients with resectable tumors.
