Abstract
Pancreatic ductal adenocarcinoma (PDAC) is a lethal disease. The majority of patients are diagnosed with locally advanced or metastatic disease with a prognosis of short months. Therapeutic options are limited and until recently, there was no standard second-line chemotherapy option. Liposomal constructs have been engineered to encapsulate chemotherapy thereby preventing premature metabolism, improving distribution and minimizing toxicity. Favourable preclinical data on liposomal irinotecan and early phase trials, led to a recently published phase III trial of liposomal irinotecan in combination with fluorouracil and folinic acid in patients with metastatic PDAC, who progressed after gemcitabine-based chemotherapy. As a direct result, the United States Food and Drug Administration (FDA) and European Medicines Agency (EMA) have approved the use of liposomal irinotecan in this setting. However, first-line treatment options for this disease now include the combination regimen, FOLFIRINOX, in patients with good performance status, and the role of second-line combination treatment with liposomal irinotecan in this setting is unclear. Recent advances have changed the therapeutic landscape, as clinicians are now able to choose a sequential approach to treatment tailored to the individual patient characteristics. This article reviews current treatment options for metastatic PDAC and focuses on the efficacy, safety and place in therapy of liposomal irinotecan.
Keywords
Introduction to pancreatic cancer
Despite decades of research, pancreatic ductal adenocarcinoma (PDAC) remains one of the most lethal malignancies. Although it is a relatively uncommon cancer, due to its high mortality rate it represents a significant global health burden. Worldwide in 2012 there were 338,000 new cases of pancreatic cancer and 331,000 associated deaths [Ferlay et al. 2015]. The striking similarity between incidence and deaths highlights the dismal prognosis of this disease. Further, by 2030 it has been predicted that pancreatic cancer will surpass breast, prostate and colorectal cancer to become the second leading cause of cancer related deaths in the United States [Rahib et al. 2014].
More than 95% of those diagnosed with pancreatic cancer will succumb to it and across Europe the 5-year overall survival (OS) ranges from 2–9% [Cancer Research UK]. This poor survival is multifactorial, attributed to the systemic and aggressive nature of PDAC, its complex mutational landscape [Waddell et al. 2015], desmoplastic stroma [Kleeff et al. 2007] and the current lack of effective therapies.
Surgery is the one curative treatment for PDAC but only a minority of patients present with potentially operable disease (10–20%) [Willett et al. 2005; Gandy et al. 2016]. Further, as the disease is characterized by early micrometastatic spread [Sohal et al. 2014] long-term survival following surgery is poor with 80% of patients experiencing a local or distant recurrence within 2 years [Geer et al. 1993].
The majority of patients present with locally advanced inoperable (40%) or metastatic disease (40–45%) where the outlook is even bleaker, with a median OS of 6–11 months and 2–6 months respectively [Willett et al. 2005; Sclafani et al. 2015; Cancer Research UK].
Current treatment for advanced disease
First-line treatment
Chemotherapy remains the cornerstone of treatment for advanced PDAC. Following the Burris and colleagues study in 1997, single-agent gemcitabine became the gold standard first-line treatment for patients with advanced disease [Burris et al. 1997]. This study demonstrated a modest improvement in survival of 1.24 months (5.65 versus 4.41 months, p = 0.0025) with gemcitabine versus 5-fluoropyrimidine (5-FU) but became a standard treatment due to the associated increment in clinical benefit rate from 4.2% with 5-FU to 23.8% with gemcitabine.
Despite a myriad of trials investigating the use of more intensive combination regimens, often with a gemcitabine backbone, no chemotherapy regimen was found to be superior to single-agent gemcitabine until more recently. In light of nonstatistically significant trends in improved survival seen with gemcitabine combinations in a number of large phase III studies Sultana and colleagues conducted a meta-analysis of 51 trials, including almost 10,000 patients. This demonstrated a survival benefit with gemcitabine combination chemotherapy versus gemcitabine alone [hazard ratio (HR) = 0.91, 95% confidence interval (CI) 0.85–0.97] [Sultana et al. 2007]. The same group also assessed different gemcitabine combinations and found a trend favouring gemcitabine plus capecitabine or a platinum [Sultana et al. 2008]. Based on such analyses the National Comprehensive Cancer Network (NCCN) guidelines and recent American Society of Clinical Oncology (ASCO) guidelines include these gemcitabine combinations in their recommendations [Sohal et al. 2016].
With the elucidation of the molecular nature of pancreatic cancer, a wide variety of targeted agents have also been investigated, again often with a gemcitabine backbone. The targets studied have included EGFR, KRAS signalling, MEK, mTOR, HER2, VEGF and the Hedgehog pathway, all with disappointing clinical results, despite sometimes promising preclinical and early phase studies [Sclafani et al. 2015]. This is not entirely surprising considering the significant inter-tumoural heterogeneity seen in PDAC. In addition to common alterations in KRAS, TP53, SMAD4 and CDKN2A [Bryant et al. 2014], there are also a multitude of infrequently mutated genes in pancreatic cancer [Waddell et al. 2015] and the majority of the trials to date have been in unselected populations.
The one targeted agent to have demonstrated a survival benefit in a phase III study is erlotinib in combination with gemcitabine versus gemcitabine alone (HR 0.82, 95% CI 0.69–0.99, p = 0.038). However, in the results published in 2007 the numerical increase in median OS was only 12 days [Moore et al. 2007] with a higher incidence of adverse events, including rash, interstitial lung disease and diarrhoea. The group of patients who developed a rash of grade ≥2 did have a more clinically meaningful benefit (median OS 10.5 months with grade ≥2 rash versus 5.8 months with grade 1 rash versus 5.3 months for those who did not develop a rash) but this combination is rarely used in clinical practice despite being US FDA and EMA-approved.
The first combination chemotherapy regimen to demonstrate a statistically significant survival benefit versus gemcitabine in a phase III study was the ACCORD-11 study in 2011. This study demonstrated that the triplet FOLFIRINOX (5-FU, irinotecan and oxaliplatin) improved progression-free survival (PFS), OS and response rate (RR) when compared with single-agent gemcitabine (OS 11.1 versus 6.8 months, HR 0.57, p < 0.001, PFS 6.4 versus 3.3 months, HR 0.47, p < 0.001, RR 31.6% versus 9.4%, p < 0.001) [Conroy et al. 2011]. Unsurprisingly this increased efficacy came at the cost of increased toxicity and this is a regimen for patients with excellent performance status (PS). However, despite this increased toxicity the FOLFIRINOX patients had a significantly longer time to deterioration in quality of life than the gemcitabine patients [Gourgou-Bourgade et al. 2013]. Furthermore, since the adoption of this regimen into the clinic various institutional series have been published suggesting modifications to the regimen to increase its tolerability without reducing efficacy [Mahaseth et al. 2013; Mantripragada et al. 2016].
Following ACCORD-11 the MPACT study of the addition of nab-paclitaxel to gemcitabine reported a median OS of 8.5 months in the combination arm versus 6.7 months in the gemcitabine alone arm (HR 0.72, p < 0.0001). Again PFS and RR were also improved (PFS 5.5 versus 3.7 months, HR 0.69, p < 0.0001, RR 23% versus 7%, p < 0.001) [von Hoff et al. 2013]. Again toxicity including sensory neuropathy was increased in the combination arm, although quality of life was not assessed, and this is a regimen for good PS patients.
The lack of direct comparisons of these regimens leaves some uncertainty as to the best first-line treatment for advanced PDAC and as yet there are no established biomarkers for selecting patients for a particular therapy. For patients with a PS of 2, single-agent gemcitabine remains a reasonable option. For those patients who are fitter, PS 0–1, either FOLFIRINOX or gemcitabine-nab-paclitaxel are preferred and the decision depends on a combination of factors including patient comorbidities, toxicity profile, consideration of central venous access and patient preference and reimbursement. Clinical trials in the first-line setting are also recommended and there are currently numerous ongoing trials globally of targeted drugs with a chemotherapy backbone and increasing studies of immune-directed approaches.
Second and subsequent line treatment
Historically there has been very little use of second-line chemotherapy for PDAC. However, with improvements in first-line therapy for advanced disease, this is now changing. Unfortunately, there is no good quality evidence to support a particular regimen following first-line treatment with FOLFIRINOX or gemcitabine and nab-paclitaxel as the few trials of second-line treatment conducted were following progression on gemcitabine monotherapy.
The randomized phase III CONKO-01 study sought to establish if the combination of oxaliplatin, folinic acid and 5-FU (OFF) was superior to best supportive care (BSC). Despite terminating early due to recruitment difficulties (n = 46 patients) this study demonstrated a survival benefit of 2.52 months with chemotherapy (mOS 4.82 OFF versus 2.30 BSC, HR 0.45, CI: 0.24–0.83, p = 0.008) [Pelzer et al. 2011]. CONKO-003, another phase III study, then evaluated whether oxaliplatin was required, randomizing 168 patients to OFF or 5-FU and folinic acid (FF). Both median OS and PFS favoured the addition of oxaliplatin (mOS 5.9 months OFF versus 3.3 months FF, HR 0.66, 95% CI 0.48–0.91, log-rank p = 0.01) with similar toxicity other than peripheral neuropathy which was, as would be expected, higher in the oxaliplatin arm (grade 1/2 neurotoxicity: 38.3% versus 7.1%, p < 0.001) [Oettle et al. 2014]. However, the results from the randomized phase III PANCREOX study were not aligned with CONO-003, with patients receiving FOLFOX having an inferior OS than those receiving FF (6.1 versus 9.9 months, HR 1.78, p = 0.02). It has been suggested that this disparity is due to a higher level of treatment discontinuation in the FOLFOX arm than the FF arm (20.4% versus 1.9% respectively) and an imbalance in the post-discontinuation treatments received in each arm (6.8% in the FOLFOX arm and 25% in the FF arm) [Gill et al. 2014].
The most recent phase III study is the NAPOLI-1 trial of nanoliposomal irinotecan, which is discussed in much greater detail below, and provides another second-line regimen for consideration [Wang-Gillam et al. 2016].
Additional regimens have been investigated in the second or subsequent line setting in a number of small phase II studies and case series, including capecitabine and oxaliplatin (CAPOX), taxanes, 5-FU and irinotecan (FOLFIRI) as well as FOLFIRINOX and gemcitabine plus nab-paclitaxel. These studies have reported mixed results and such regimens require further assessment in larger studies [Oettle et al. 2000; Xiong et al. 2008; Yoo et al. 2009; Hosein et al. 2013; Lee et al. 2013].
Extrapolating from the data available it is not unreasonable to consider that second-line chemotherapy may provide a clinical benefit. A systemic analysis of 34 second-line studies, including over 1500 patients who had progressed on gemcitabine, reported a median OS of 2.8 months for patients who received BSC and 6 months for patients who received second-line treatment [Rahma et al. 2013]. The choice of second-line regimen then depends upon the first-line regimen used, the patients’ PS, residual toxicities and comorbidities and in this very palliative setting it must be remembered that toxicities and quality of life are of paramount importance. There are even less data to support third or subsequent line treatment and here clinical trial entry is strongly encouraged for fit patients.
Overall therefore, despite the small improvements in the treatment options for pancreatic cancer in the last decade described above, prognosis remains poor and there is an urgent need to develop novel therapies, consider new combinations and appropriately select those patients who may benefit.
Introduction to irinotecan
Camptothecin is a naturally occurring cytotoxic alkaloid that targets topoisomerase I, a nuclear enzyme that reduces the torsional stress of supercoiled DNA during the replication, recombination, transcription, and repair of DNA [Garcia-Carbonero et al. 2002]. Irinotecan is a synthetic derivative of camptothecin, with functional groups to enhance solubility and was first approved by the US FDA in 1996 [Hsiang et al. 1985]. As a prodrug, it is converted in to the active metabolite SN38, predominantly in the liver, by carboxylesterases. SN-38 is subsequently conjugated, with significant pharmacogenetic variability, into inactive, nontoxic SN38-glucuronide (SN-38G). Both irinotecan and SN-38 have a labile α-hydroxy-δ-lactone ring that undergoes pH-dependent reversible hydrolysis [Swami et al. 2013]. In acidic conditions, the more active, potent lactone form predominates, while in more basic conditions, the inactive, less toxic carboxylate form is favoured. These properties contribute to the marked heterogeneity of the main dose-limiting toxicities observed in patients, that of neutropaenia and late onset diarrhoea. The direct effect of SN-38 on intestinal epithelium is thought to be responsible for irinotecan-induced diarrhoea [Hecht, 1998], which can be severe, resulting in dose reductions or omissions leading to ineffective treatment administration.
There is therefore a rationale for the development of formulations that can improve the distribution and protect the premature metabolism of irinotecan to achieve maximum efficacy while minimizing toxicity. Liposomal constructs have been engineered to improve the circulation time and intra-tumoural drug concentration of anthracyclines, for example, PEGylated liposomal doxorubicin [Gabizon, 2001], and nab-paclitaxel [Gradishar, 2006], but it has been difficult to replicate the technology in other classes of chemotherapy drugs.
Liposomal irinotecan (MM-398, PEP02, nal-Iri)
In 2006, a novel liposomal construct containing irinotecan was developed [Drummond et al. 2006]. Unlike previous liposomal irinotecan preparations [Messerer et al. 2004], a reduced toxicity profile was reported, as well as increased tumour efficacy when compared with irinotecan in animal models [Drummond et al. 2006]. The nanoparticle consists of a lipid bilayer scaffold, which encapsulates the drug complex and facilitates in vivo drug retention. A transmembrane gradient of triethylammonium cations was used to drive the irinotecan into the liposome where upon sucrose octasulphate, a highly charged anion, trapped the irinotecan within the liposome (see Figure 1).
A schematic of encapsulation of irinotecan inside the liposome. Irinotecan is exchanged with triethylammonium cations across the bilipid layer.
Pharmacokinetics, pharmacodynamics and pharmacogenetics
Preclinical in vivo pharmacokinetic studies have shown liposomal encapsulation of irinotecan was associated with significantly longer circulation times (t½ = 10.7 h) compared with unencapsulated irinotecan (t½ = 0.27 h) [Drummond et al. 2006]. At 24 h, 23.3% of injected liposomal irinotecan still remained, with no detectable conversion to SN-38 or the carboxylate form. Conversely, only 2% of the injected dose of unencapsulated irinotecan remained at 30 min, 35% of which had been converted to SN-38 and subsequently conjugated to the in the inactive carboxylate form. A similar pharmacokinetic profile was reported in patients by Chang and colleagues [Chang et al. 2015]. In this phase I trial, treatment with liposomal irinotecan at 120 mg/m2 was characterized by slow clearance (mean = 0.0591 l/m2/h), small volume of distribution (mean = 1.8 l/m2 ≅ plasma volume) and prolonged terminal half-life of total irinotecan in circulation (mean = 29.5 h) when compared with published pharmacokinetic data for unencapsulated irinotecan [Chabot, 1997; Chang et al. 2015]. The plasma concentration profile of liposomal irinotecan matched that of the total irinotecan, suggesting the release of irinotecan from the liposome occurs slowly over time.
The mechanism of irinotecan release is not fully understood. Preclinical models have shown liposomal irinotecan increases tumoural drug retention, resulting in local release and conversion into irinotecan and SN-38 [Drummond et al. 2006]. However, the correlation between the Cmax and AUC0-∞ of liposomal irinotecan and the active metabolite SN-38, reported by Chang and colleagues, was weak [Chang et al. 2015]. This variability may be attributable to the small sample size or the pharmacogenetic variability of irinotecan metabolism.
Patients with a genetic polymorphism in the gene encoding uridine diphosphate glucuronosyltransferase (UGT) 1A1, have a lower than normal capacity to metabolize SN-38, the active metabolite of irinotecan. UGT conjugates SN-38 into the inactive SN-38 glucuronide (SN-38G). The presence of the UGT1A1*28 allele, present in approximately 17% of Whites [Lampe et al. 1999], has been shown to cause a 70% reduction in expression of UGT, leading to an increased exposure of patients to the cytotoxic metabolite, SN-38. Patients homozygous for the UGT1A1*28 allele are 3.5 times more likely to suffer from grade 3/4 toxicities [Palomaki et al. 2009] and may benefit from a dose reduction of up to 40% [Innocenti et al. 2014].
In a phase I trial, a woman with heterozygosity of UGT1A1*6/*28 died from grade 4 diarrhoea and was found to have Cmax and AUC0-∞ of SN-38 that were 3-times higher than in other patients treated at the same dose [Chang et al. 2015]. Moreover, in a recent phase III study, the dose of liposomal irinotecan was reduced by 25% in 14 patients homozygous for the UGT1A1*28 allele, 5 of whom were able to escalate to the standard dose, 2 of whom needed a further dose reduction and 1 discontinued secondary to grade 3 vomiting [Wang-Gillam et al. 2016].
Although early phase studies have demonstrated the different pharmacokinetic and toxicity profiles of patients homozygous and heterozygous for the UGT1A1*28 allele, the available data are not conclusive for defining a precise genotype-based dosage [Toffoli et al. 2006; Biason et al. 2008; Palomaki et al. 2009]. There are therefore no current guidelines on how UGT1A1 testing could be used to effect treatment with clinical practice.
Clinical Efficacy
Preclinical in vivo efficacy data have shown the enhanced anti-tumour effect of liposomal irinotecan compared with equivalent dose of free irinotecan in human pancreatic cancer cell line xenograft mouse models [Hann et al. 2007]. Chang and colleagues reported the results of a phase I dose-escalation study of liposomal irinotecan, in 11 patients with treatment refractory advanced solid tumours [Chang et al. 2015]. Liposomal irinotecan was delivered as a 90-min intravenous infusion every 21 days. Overall, two patients, including one patient with pancreatic cancer who was treated with the higher dose of 180 mg/m2, achieved a partial response to treatment. The disease control rate was 45% for the intention-to-treat population.
The favourable pharmacokinetics described in this study, led to a nonrandomized phase II trial that sought to establish the efficacy and toxicity of single-agent liposomal irinotecan, at 120 mg/m2 3-weekly, in patients with metastatic pancreatic cancer after progression following first-line gemcitabine-based chemotherapy [Ko et al. 2013]. A total of 40 patients were enrolled, 60% of whom had a Karnofsky score of 90–100. The protocol was amended during the second stage of the study to permit a starting dose of 100 mg/m2. However, 27 of the 40 patients remained on 120 mg/m2 for the duration of their treatment. The mean and median number of cycles received was, 5.88 and 2.5, respectively. Half of the patients achieved disease control defined as objective response (7.5% of patients) or stable disease for more than two cycles (42.5% of patients). The median PFS and OS were, 2.4 and 5.2 months, respectively.
This phase II study provided evidence of antitumour activity and led to the pivotal NAPOLI-1 trial of liposomal irinotecan in advanced pancreatic cancer [Wang-Gillam et al. 2016]. Wang-Gillam and colleagues published the results from this global, randomized, open-label, phase III study in which 417 patients previously treated with gemcitabine-based chemotherapy, were assigned to receive either, liposomal irinotecan monotherapy (120 mg/m2) every 3 weeks or folinic acid and fluorouracil (2000 mg/m2 over 24 h every week for the first 4 weeks of every 6-week cycle, based on the CONKO-003 trial [Oettle et al. 2014]. A third arm, consisting of 2-weekly combination treatment of liposomal irinotecan (80 mg/m2) with folinic acid and fluorouracil (2400 mg/m2 infusion over 46 h) was added later in a protocol amendment. The rationale for protocol amendment was based on preliminary data from the PEPCOL study; a randomized phase II trial which showed liposomal irinotecan in combination with 5-fluorouracil/leucovorin (5-FU/LV) to have activity and an acceptable safety profile in patients with colorectal cancer [Chibaudel et al. 2016]. In vitro data suggest liposomal irinotecan may improve vascular function thereby increasing the delivery and accumulation of a second drug when used in combination [Baker et al. 2008].
In NAPOLI-1, OS was the primary endpoint and patients were stratified by baseline albumin level, PS and ethnicity. Although all patients had previously been treated with gemcitabine, 12% of patients had received this in the neoadjuvant, adjuvant or locally advanced setting. Interestingly, over half the patients enrolled had previously been treated with fluorouracil or irinotecan-based regimes; 56% and 15%, respectively.
Median OS was 6.1 months in the combination arm and 4.2 months in those assigned 5-FU/LV control (HR 0.67, 95% CI 0.49–0.92; p = 0.012). In patients who were allocated liposomal irinotecan monotherapy, median OS was 4.9 months, which was not significantly different to the control arm, (HR 0.99, 95% CI 0.77–1.28; p = 0.94) (see Table 1). Interestingly, quality of life scores and clinical benefit response did not differ significantly between treatment groups. The authors acknowledge that using a different 5-FU/LV regimen in the combination arm to the control arm is not standard design but argue it was unlikely to have created bias in favour of the investigational arm which delivered lower dose intensities of 5-FU.
Summary table of main efficacy results from the three arms of NAPOLI-1 trial [Wang-Gillam et al. 2016].
CI, 95% confidence interval; Ca19-9, carbohydrate antigen 19-9; 5-FU, 5-fluorouracil; LV, leucovorin; ORR, overall response rate; OS, overall survival; PFS, progression-free survival.
In a pre-planned analysis of each subgroup, OS was increased in patients with poorer prognostic features who were treated in the combination arm compared with those treated with 5-FU/LV; lower PS (Karnofsky score 70–80), albumin <40 g/l, Ca19-9 ≥40 U/ml, liver metastases, stage IV at diagnosis and patients whose time from diagnosis to enrolment was less than the median. However, it should be noted that patients with a Karnofsky score <70% or an albumin <30 g/l were not included in the study.
Safety
The dose-limiting toxicities for liposomal irinotecan are myelosuppression and diarrhoea. The two patients treated with 180 mg/m2 of liposomal irinotecan suffered from grade 4 toxicities. One patient had grade 4 neutropaenia, while the second patient developed grade 4 febrile neutropaenia, grade 4 thrombocytopaenia with a bleeding event and grade 4 diarrhoea. The dose escalation was therefore stopped and 120 mg/m2 was determined to be the maximum tolerated dose [Chang et al. 2015] (see Table 2). At a dose of 120 mg/m2, the phase II trial found the majority of patients experienced at least one grade 3/4 event. The three deaths that occurred within 30 days of study treatment were attributed to infection in the setting of neutropaenia [Ko et al. 2013]. Overall, one-third of patients treated with liposomal irinotecan, either as monotherapy (120 mg/m2) or in combination (80 mg/m2) with 5-FU/LV, were reported by Wang-Gillam and colleagues to suffer from adverse events necessitating a treatment dose reduction, compared with only 4% of patients treated with 5-FU/LV [Wang-Gillam et al. 2016]. In a predefined subgroup analysis recently reported, the safety profile of the combination arm was generally similar across patient subgroups, apart from an increased risk of grade ≥3 neutropaenia and reduced neutrophil counts in Asian patients [Chen et al. 2016]. Single-agent liposomal irinotecan was associated with more diarrhoea (any grade) than the combination arm as well as more grade 4 treatment-related adverse events, 16% versus 10%, respectively. Moreover, 4 of the 5 treatment-associated toxicities which resulted in death, occurred in patients being treated with single-agent liposomal irinotecan.
Summary table of trials investigating liposomal irinotecan in advanced pancreatic cancer.
5-FU, 5-fluorouracil; LV, leucovorin; NA, not applicable.
Liposomal irinotecan, at the lower dose of (80 mg/m2), in combination with 5-FU/LV appears to be associated with a more tolerable toxicity profile as well as improved efficacy compared with liposomal irinotecan monotherapy (120 mg/m2). However, in the context of second-line treatment for metastatic pancreatic cancer, all toxicity grades need to be considered and it is worth highlighting that over half of patients who received liposomal irinotecan, as monotherapy or in combination, reported diarrhoea, vomiting or nausea.
Place in therapy
Single-agent liposomal irinotecan does not have a role in the treatment of metastatic pancreatic cancer in patients who have progressed on gemcitabine-based treatment. Liposomal irinotecan monotherapy at 120 mg/m2 performs similarly to single-agent 5-FU/LV, but has a less favourable toxicity profile. First-line treatment with gemcitabine-based chemotherapy has been standard of care for patients with metastatic pancreatic cancer for over 10 years and in October 2015 the results of NAPOLI-1 led to US FDA and EMA approval of liposomal irinotecan in combination with 5-FU/LV for the treatment of patients with metastatic pancreatic cancer following disease progression on gemcitabine-based therapy. This is the first treatment regime to be approved for use in the second-line setting for pancreatic cancer. However, as previously discussed, FOLFIRINOX has more recently emerged as an alternative to gemcitabine in the first-line setting, in patients with a good PS after demonstrating superior survival outcome. It is therefore unclear how combination liposomal irinotecan and 5-FU/LV will be used in clinical practice relative to first-line FOLFIRINOX.
Despite a clear preclinical rationale for liposomal preparations, the clinical efficacy and toxicity of liposomal irinotecan has not been directly compared with unencapsulated irinotecan in the treatment of patients with pancreatic cancer either as a single agent or in combination. The efficacy of unencapsulated irinotecan as a single agent, following failure of gemcitabine-based chemotherapy, was studied in a phase II of 33 patients treated with 150 mg/m2 irinotecan every 2 weeks. They reported a PFS and OS of 2.0 months (95% CI, 0.7–3.3) and 6.6 months (95% CI, 5.8–7.4), respectively [Yi et al. 2009] (see Table 3). A similar PFS of 2.7 months (95% CI, 2.1–2.9) was found in patients treated with liposomal irinotecan in the NAPOLI-1 trial, although the OS was less, at 4.9 months (95% CI, 4.2–5.6). Toxicities following different irinotecan preparations were directly compared in a randomized phase II trial of patients with locally advanced or metastatic gastric or gastro-oesophageal junction adenocarcinoma. Grade 3/4 neutropaenia and diarrhoea was reported in 11.4% and 27.3% of patients treated with 120 mg/m2 of liposomal irinotecan. While 15.9% and 18.2% of patients suffered from grade 3/4 neutropaenia and diarrhoea, respectively, following 300 mg/m2 of unencapsulated irinotecan [Roy et al. 2013].
Trials of unencapsulated irinotecan as a single agent or in combination with 5-FU in pretreated patients with advanced pancreatic cancer.
N, number of patients; PFS, progression-free survival; OS, overall survival, D, Day; W, Week; LV, leucovorin; 5-FU, fluorouracil; CI, continuous infusion.
The combination of unencapsulated irinotecan with 5-FU/LV has also previously been explored. In a randomized phase II trial comparing modified FOLFOX (5-FU/LV and oxaliplatin) with FOLFIRI (5-FU/LV and irinotecan), the median OS of patients in the FOLFIRI arm was disappointing at 3.8 months [Yoo et al. 2009]. The two nonrandomized studies [Neuzillet et al. 2012; Zaniboni et al. 2012] with small sample sizes reported a more favourable median OS of 5 and 6.6 months, respectively, following treatment with encapsulated irinotecan in combination with 5-FU/LV using the schedules shown in Table 3. In one of these studies, Zaniboni and colleagues reported combination treatment resulted in grade 3/4 neutropaenia and diarrhoea in 20% and 12% of patients, respectively [Zaniboni et al. 2012]. Interestingly, a similar incidence of grade 3/4 neutropaenia and diarrhoea was reported in patients receiving combination liposomal irinotecan and 5-FU/LV within the NAPOLI-1 trial (Table 2).
Patients with metastatic pancreatic cancer, who are well for second-line treatment, may have a favourable biology and a longer survival independent of choice of therapy. Although combination treatment of liposomal irinotecan and 5FU/LV met criteria for a statistically significant increase in OS compared with the control, the overall clinical benefit to the individual patient must be considered. A 2-weekly regime, associated with grade 3/4 neutropaenia in 27% of patients and any grade diarrhoea, nausea and vomiting in over 50% of patients, is not insignificant in the palliative setting. Although, it must be noted that the quality of life of patients on combination treatment was not reported to be appreciably different from those allocated 5-FU/LV.
Despite preclinical studies suggesting otherwise, the toxicity profile of liposomal irinotecan does not seem to be significantly better than standard unencapsulated irinotecan in vivo. It would be interesting to compare the efficacy and toxicity profile of the two irinotecan preparations in FOLFIRINOX, which is likely to remain the treatment of choice in first-line metastatic pancreatic cancer for patients with a good PS.
Three trials involving liposomal irinotecan are currently recruiting patients with pancreatic cancer. A phase II comparative study to assess the safety, tolerability and efficacy of liposomal irinotecan and 5-FU/LV with and without oxaliplatin compared with nab-paclitaxel and gemcitabine, is currently recruiting patients with advanced pancreatic adenocarcinoma [ClinicalTrials.gov identifier: NCT02551991]. A randomized phase II trial is recruiting Japanese patients with metastatic pancreatic cancer to assess the safety, tolerability and the pharmacokinetics of liposomal irinotecan in combination with 5-FU/LV and to compare the efficacy of combination treatment with single-agent 5-FU/LV [ClinicalTrials.gov identifier: NCT02697058].
Patients are also currently being recruited to study the pharmacokinetic and pharmacodynamic profile of BBI608 (a first-in-class cancer stem cell inhibitor), when administered in combination with standard chemotherapies including combination liposomal irinotecan and 5-FU/LV in patients with metastatic pancreatic cancer [ClinicalTrials.gov identifier: NCT02231723].
There is a clear need for well-designed, randomized clinical trials in the second-line setting after FOLFIRNOX failure. The role of liposomal irinotecan and 5-FU/LV in this setting is unclear. A pre-planned subgroup analysis in NAPOLI-1, showed no survival benefit of combination therapy over control, in the 12 patients who had previously been treated with irinotecan; HR 1.25 (0.49–3.19). However, the results from NAPOLI-1 alter the therapeutic landscape for metastatic pancreatic cancer by enabling clinicians and patients to choose a sequential approach to treatment. For patients in whom first-line FOLFIRINOX is not the preferred option, gemcitabine-based chemotherapy followed by liposomal irinotecan and 5-FU/LV is a reasonable treatment paradigm for a subset of patients.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Conflict of interest statement
The authors declare that there is no conflict of interest.