Effectiveness and cost analysis of liposomal irinotecan in patients with pancreatic adenocarcinoma: a multicenter real-world study

Abstract

Background:

Systemic chemotherapy is crucial for pancreatic adenocarcinoma (PAC). It serves as a palliative chemotherapy for unresectable PAC and as a (neo-)adjuvant chemotherapy for (borderline) resectable PAC. Survival benefits were found with the use of liposomal irinotecan combined with 5-fluorouracil and leucovorin (nal-IRI/5-FU/LV) in the global phase III NAPOLI-1 trial. However, limited evidence supports the cost-effectiveness of liposomal irinotecan throughout the PAC treatment course.

Objectives:

We investigated the real-world efficacy and cost implications of nal-IRI/5-FU/LV in PAC management.

Design:

A multicenter real-world study.

Methods:

We analyzed real-world data from electronic health records to evaluate outcomes of nal-IRI/5-FU/LV exposure in this retrospective study. Effectiveness was assessed using overall survival (OS). Medical costs were those incurred from imaging or laboratory tests, chemotherapy, surgery, radiotherapy, and supportive care during inpatient, outpatient, and emergency department visits. Propensity score matching was used to adjust for potential differences in patient characteristics between the nal-IRI and non-liposomal irinotecan groups. The incremental cost-effectiveness ratio (ICER) was calculated by dividing cost differences by survival gains.

Results:

Overall, 1734 patients diagnosed with PAC were screened, and 705 who received at least one cycle of chemotherapy were included in the analysis. The median OS was significantly longer in the nal-IRI group than in the non-liposomal irinotecan group (28.0 vs 18.3 months, p = 0.011). Patients in the nal-IRI group had more hospital visits and admissions during follow-up and incurred higher costs ($61,430 vs $39,129 per patient), which resulted in an ICER of $2787 per month of survival gained ($33,285 per life-year gained).

Conclusion:

Despite its higher cost, nal-IRI/5-FU/LV significantly improved the survival of patients with PAC compared to the conventional chemotherapies. Our findings suggest that it may be a cost-effective treatment option for gemcitabine failure despite the high cost of liposomal irinotecan, considering the poor prognosis of PAC.

Introduction

Systemic chemotherapy is crucial for managing unresectable pancreatic adenocarcinoma (PAC), which has a poor prognosis with a 5-year survival rate of approximately 10%. FOLFIRINOX (FFX; a combination of oxaliplatin, irinotecan, leucovorin, and 5-fluorouracil) and gemcitabine combined with nab-paclitaxel (GN) are standard first-line regimens for unresectable PAC, with subsequent lines historically showing limited success. ¹ NAPOLI-1 trial revealed that liposomal irinotecan combined with 5-fluorouracil and leucovorin (nal-IRI/5-FU/LV) is an effective subsequent therapy after gemcitabine-based treatment failure.^1,2 Recently, NALIRIFOX (liposomal irinotecan combined with 5-FU/LV and oxaliplatin) showed superior outcomes to GN in the NAPOLI-3 trial, which was expected to expand the use of the nal-IRI combination chemotherapy for PAC treatment.

The efficacy of nal-IRI/5-FU/LV as a second or later-line therapy has been investigated in various real-world studies. The results were comparable to those of a pivotal study.^{3 –5} However, limited studies have evaluated the effectiveness and cost of liposomal irinotecan over the entire treatment course of pancreatic cancer. Gourzoulidis et al. ⁶ estimated the total cost per patient to be €21,468 for nal-IRI/5-FU/LV compared with €4758 for 5-FU/LV. Despite the financial burden, liposomal irinotecan is undoubtedly a promising agent for the treatment of pancreatic cancer because of its potential clinical benefits.

We previously launched a multicenter web-based registry for PAC, called the K-PaC registry, which began as a national project in which clinical data of patients with histologically proven PAC were included. Data were obtained from two tertiary hospitals (National Cancer Center (NCC) and Seoul National University Bundang Hospital (SNUBH)) in Korea. ⁷ In this study, we aimed to evaluate the effectiveness and medical cost of nal-IRI/5-FU/LV in the management of PAC, considering exposure to liposomal irinotecan.

Methods

Study population

Patients initially diagnosed with PAC, who received at least one cycle of chemotherapy, were eligible for this retrospective study. The study was performed between January 2014 and February 2020 at two referral institutions in Korea. Because nal-IRI/5-FU/LV is approved for patients with PAC after gemcitabine failure in Korea, those who did not receive this second-line chemotherapy after first-line therapy were excluded. We also excluded patients who died within 6 months of diagnosis. Patient data, including demographics, diagnoses, treatment, and follow-ups, were obtained from the clinical data warehouse and K-PaC registry, whereas cost data were collected from medical claims records. To confirm the reliability of the data extracted from the participating institutions, we compared our study population with that included in a recently published study using representative claims data from the National Health Insurance Service (NHIS) in Korea. ⁸ This study was approved by the Institutional Review Board of the NCC (Approval Number, NCC 2021-0211) and SNUBH (Approval Number, B-2204-748-101), Korea, in compliance with the Declaration of Helsinki and was registered at ClinicalTrials.gov (identifier: NCT04984174). The requirement for informed consent was waived owing to the retrospective nature of this study. The reporting of this study conforms to the STROBE statement (see Supplemental File). ⁹

Study endpoints

In this study, we divided all enrolled patients into a nal-IRI group and a non-liposomal irinotecan group based on nal-IRI exposure. We analyzed the effectiveness and medical costs during the follow-up. Effectiveness was assessed using overall survival (OS) and was defined as the difference between the time of diagnosis and any cause of death or the last follow-up. We calculated the total costs (1 USD = 1300 KRW) associated with imaging or laboratory tests, administering chemotherapy, surgery, or radiation therapy, supportive care, and managing adverse events during inpatient, outpatient, and emergency department visits that occurred after the initial diagnosis. We assumed the incremental cost-effectiveness ratio (ICER) to be the difference in mean total medical costs (per patient) divided by the difference in the median OS between the two groups.^10,11

Statistical analysis

Propensity score matching (PSM) was used to adjust for potential differences in patient characteristics between the nal-IRI and non-liposomal irinotecan groups. Propensity scores were calculated based on potential confounding variables (age, sex, year of diagnosis, cancer stage, and initial chemotherapy regimen). We matched the two groups of patients using a 1:2 nearest-neighbor matching protocol with a caliper width of 0.1. Chi-square and t-tests were used to compare categorical and continuous variables, respectively. OS was estimated using Kaplan–Meier survival analysis, and group differences were assessed via log-rank tests. To address potential bias due to unequal follow-up durations, we also calculated restricted mean survival time (RMST, τ = 36 months), truncated at 36 months, and incorporated it into the cost-effectiveness analysis. Cox proportional hazards regression was used to estimate the effects of independent variables on patient survival, including age, sex, year of diagnosis, initial chemotherapy regimen, and exposure to nal-IRI/5-FU/LV. Significance was set at a two-sided p-value of <0.05. All analyses were performed using the SAS statistical software (version 9.4; SAS Institute, Cary, NC, USA).

Results

Comparison of patients in the study cohort with the NHIS database

First, we compared the clinical characteristics of our study population with those of nationwide patients identified in the NHIS database. The patient characteristics are presented in Table S1. The number of patients diagnosed with PAC gradually increased over time in both groups. The study population comprised more patients in their 50s who received a higher proportion of radiotherapy and chemotherapy. They were more likely to receive palliative chemotherapy with aggressive combination regimens such as FFX or GN. Median OS was 10.4 months (95% confidence interval (CI), 9.9–10.9) for those in the study population and 10.1 months (95% CI, 9.9–10.3) for those from the NHIS database (Figure S1). These findings suggest that our study population comprised patients with PAC in real-world clinical settings.

Patient characteristics

Of the 1734 patients diagnosed with PAC who were screened, 705 were included in the analysis after excluding 985 who did not receive second-line chemotherapy and 44 who died within 6 months of diagnosis (Figure 1).

Figure 1.

Patient flow of the study.

Table S2 shows the patient characteristics before and after PSM. Most patients were diagnosed with PAC in their 60s and 70s, with the diagnosis rate increasing with age. Both treatment groups were well-balanced, except for the year of the first confirmed diagnosis and chemotherapy regimen. Most patients underwent the first chemotherapy session for palliative purposes. Nal-IRI/5-FU/LV use has increased over the years, with more patients receiving GN as their first therapy than those not exposed to liposomal irinotecan. After PSM, the baseline characteristics were well-balanced between the two groups. The treatments other than chemotherapy for both groups after PSM are summarized in Table S3. More patients in the non-liposomal irinotecan group underwent surgery (26.3% vs 34.4%); however, the difference was not significant. Radiotherapy was similarly administered in both groups (29.2% vs 33.6%).

Effectiveness of nal-IRI treatment

The median follow-up durations were 27.6 and 18.5 months in the nal-IRI group and non-liposomal irinotecan groups, respectively. We compared the OS based on liposomal irinotecan exposure. Patients who received nal-IRI/5-FU/LV had significantly longer OS than those who did not (Figure 2(a)). Median OS was 28.0 months (95% CI, 24.4–31.4) in the nal-IRI group and 18.3 months (95% CI, 16.2–20.2) in the non-liposomal irinotecan group (p = 0.011).

Figure 2.

Overall survival of patients according to the liposomal irinotecan exposure in the matched cohort. (a) Overall study population. (b) Patients in metastatic disease at time of diagnosis.

The 1-year and 2-year OS rates were 91% and 61% in the nal-IRI group and 75% and 39% in the non-liposomal irinotecan group, respectively (Table 1). For analysis of only patients with metastatic PAC at initial diagnosis, those treated with nal-IRI/5-FU/LV also had a significantly longer median OS (23.3 vs 14.4 months, p = 0.002; Figure 2(b)).

Table 1.

Treatment outcomes in the matched cohort.

Variables	nal-IRI (+) (n = 137)	nal-IRI (−) (n = 256)
Median follow-up duration, months (Q1, Q3)	27.6 (17.5, 36.7)	18.5 (12.2, 30.4)
Median OS, months (95% CI)	28.0 (24.4–31.4)	18.3 (16.2–20.2)
Median OS, months (95% CI) from nal-IRI/5-FU/LV treatment	6.7 (4.4–9.3)	–
Survival rate, % (95% CI)
1-Year	90.5 (84.2–94.4)	75.00 (69.2–79.9)
2-Year	61.3 (52.6–68.9)	38.67 (32.7–44.6)
3-Year	34.5 (26.3–42.9)	27.51 (21.8–33.5)
4-Year	20.2 (12.3–29.4)	21.01 (15.4–27.3)
5-Year	7.7 (1.84–19.23)	18.0 (12.1–24.8)

5-FU, 5-fluorouracil; CI, confidence interval; LV, leucovorin; Nal-IRI, liposomal irinotecan; OS, overall survival; Q1, 25th percentile; Q3, 75th percentile.

The treatment outcomes of patients treated with nal-IRI/5-FU/LV are summarized in Table S4. Of the 137 patients, 74 (54%) were treated with nal-IRI/5-FU/LV as a third-line therapy. The median number of chemotherapy cycles was 4, and the median OS from the time nal-IRI/5-FU/LV began was 6.7 months (95% CI, 4.4–9.3). Regarding safety, common adverse events observed were nausea (28.5%), anorexia (22.6%), and fatigue (22.6%).

Factors associated with OS

Univariate and multivariate analyses were performed using the Cox proportional hazards model for survival, including the independent variables (Table 2). Metastatic cancer and old age (⩾80 years) were associated with a poorer prognosis, whereas nal-IRI/5-FU/LV exposure was associated with an improved prognosis in both univariate (hazard ratio (HR), 0.74; 95% CI, 0.58–0.94) and multivariate-adjusted analysis (HR, 0.65; 95% CI, 0.51–0.83).

Table 2.

Univariate and multivariate analysis by Cox proportional hazard model on survival.

Variables	Univariate		Multivariate
	HR	95% CI	Adjusted HR	95% CI
Sex
Female	1	(Reference)
Male	1.286	1.018–1.623	1.267	0.987–1.627
Age
Less than 50	1	(Reference)	1	(Reference)
50–59	0.581	0.368–0.916	0.636	0.391–1.034
60–69	0.555	0.360–0.856	0.585	0.371–0.922
70–79	0.675	0.426–1.068	0.763	0.465–1.253
80 and over	1.613	0.750–3.469	2.247	1.005–5.026
Cancer stage
Resectable	1	(Reference)	1	(Reference)
Borderline or locally advanced	1.022	0.720–1.451	1.115	0.736–1.690
Metastatic	2.147	1.603–2.876	2.387	1.628–3.499
Year of initial diagnosis
2014	1	(Reference)	1	(Reference)
2015	1.473	0.188–11.532	2.614	0.296–23.063
2016	2.564	0.346–15.932	3.158	0.397–25.090
2017	2.193	0.289–15.078	2.717	0.348–21.234
2018	2.086	0.289–15.078	2.137	0.276–16.574
2019	1.549	0.213–11.263	1.870	0.238–14.683
2020	1.591	0.220–11.527	1.723	0.219–13.554
Regimen of first chemotherapy
Gemcitabine plus nab-paclitaxel	1	(Reference)	1	(Reference)
Folfirinox	0.714	0.545–0.937	1.020	0.755–1.377
Gem based	0.974	0.505–1.878	1.103	0.494–2.467
Gem mono	0.448	0.278–0.721	0.838	0.460–1.527
5-FU based	0.773	0.312–1.916	1.135	0.385–3.351
Others	0.805	0.530–1.224	1.333	0.788–2.256
Nal-IRI exposure
Non-exposure	1	(Reference)	1	(Reference)
Exposure	0.74	0.58–0.941	0.651	0.51–0.83

5-FU, 5-fluorouracil; CI, confidence interval; HR, hazard ratio; Nal-IRI, liposomal irinotecan.

Cost analysis

Medical cost data were available only from the NCC, and the total cost of 329 patients at a single institution was analyzed. Baseline characteristics of patients before and after PSM in the NCC are summarized in Table S5. The difference in the median OS was 8 months, which was longer in the nal-IRI group than in the non-liposomal irinotecan group (Figure S2). The patterns of medical use and total costs per patient or visit incurred after diagnosis in both groups are summarized in Table 3.

Table 3.

Medical use and costs in the matched cohort of a single institution.

Variables	Nal-IRI (+)	Nal-IRI (−)
Medical use (per patient)
Average number of hospital admissions	16.5	7.5
Average admission days	68.3	48.2
Number of outpatient visits	80.9	63.9
Number of emergency department visits	4.0	2.6
Cost (per visit), $ a
Inpatient	501.9	448.8
Outpatient	194.8	221.1
Emergency department	226.1	248.6
Cost (per patient), $
Total	61,430.3	39,129.1
Inpatient	34,258.2	21,518.1
Outpatient	16,209.9	14,335.4
Emergency department	869.5	581.6
Chemotherapy drugs	23,791.0	8302.6
Nal-IRI/5-FU/LV	6920.9	n/a
Monthly cost (per patient per month), $
Total	2457.7	2182.0
Inpatient	1422.7	1293.0
Outpatient	675.4	895.9
Emergency department	34.4	38.6
Chemotherapy drugs	949.0	454.0

^a1 USD = 1300 KRW.

5-FU, 5-fluorouracil; LV, leucovorin; n/a, not applicable; Nal-IRI, nanoliposomal irinotecan.

Patients in the nal-IRI group had more hospital visits and admissions during follow-up and incurred higher average costs than those in the non-liposomal irinotecan group ($61,430 vs $39,129). The average cost per visit showed that the nal-IRI group had higher inpatient costs ($502 vs $449), lower outpatient costs ($195 vs $221), and emergency department visit costs ($226 vs $249).

The total cost per patient was higher ($61,430 vs $39,129) in the nal-IRI group than in the non-liposomal irinotecan group. To account for the difference in follow-up periods (27.3 vs 18.3 months), we compared monthly costs (per patient per month) and found that the difference was attenuated ($2458 vs $2182). The average monthly cost of chemotherapy was higher ($949 vs $454) in the nal-IRI group than in the non-liposomal irinotecan group.

The ICER for the nal-IRI group was estimated at $2787 per additional month of median survival, equivalent to $33,285 per life-year gained. To complement the median-based estimate and reduce bias due to unequal follow-up durations, we also conducted an ICER analysis using RMST, standardized to a 36-month time horizon. After truncating or prorating patient-level costs accordingly, the incremental cost was $9936, and the incremental RMST gain was 4.58 months (95% CI, 2.53–6.63). This yielded an ICER of $2170 per life-month gained, or $26,040 per life-year gained.

Discussion

In this study, we evaluated the effectiveness and cost implications of nal-IRI/5-FU/LV throughout the treatment course of PAC using real-world data. Our findings revealed that exposure to nal-IRI/5-FU/LV was associated with improved OS compared with the non-liposomal irinotecan group, with a median OS of 28.0 versus 18.3 months. This survival advantage underscores that nal-IRI/5-FU/LV is a valuable therapeutic option for PAC despite its high costs. The calculated ICER of $2787/month of survival suggests that nal-IRI/5-FU/LV treatment is within an acceptable cost-effectiveness range, considering the poor prognosis of PAC.^12,13

In terms of effectiveness, our results align with those of previous studies, demonstrating the clinical benefit of nal-IRI/5-FU/LV in second or later treatment settings. Findings from the NAPOLI-1 trial revealed significant survival improvements with nal-IRI/5-FU/LV compared with 5-FU alone, with similar trends observed in real-world studies.^2,14,15 The studies reported a wide range of OS (range, 4.3–9.4 months) and PFS (range, 2.0–4.1 months), which is comparable to our OS result of 6.7 months from the start of nal-IRI/5-FU/LV. The wide range of OS may be due to different previous lines of therapy. Patients usually receive nal-IRI/5-FU/LV as a second- or third-line treatment, and those using earlier lines of therapy have improved survival compared with those using later lines.^5,14,16 Considering that in our previous study, the OS was 4.9 months in patients treated with third- or later-line therapy, the results of this study appear to be reasonable real-world evidence. ⁵

In some comparative studies, the effectiveness of nal-IRI/5-FU/LV compared with conventional regimens was evaluated.^{17 –21} Tossey et al. ²¹ conducted a single-center retrospective analysis comparing nal-IRI/5-FU/LV with FOLFIRI (irinotecan, LV, and 5-FU) and reported a trend of favorable OS (7.1 vs 6.7 months) and PFS (4.1 vs 3.1 months) in the nal-IRI/5-FU/LV group. FFX regimen was compared with nal-IRI/5-FU/LV as second-line therapy in recent retrospective studies. Tezuka et al. reported a favorable OS trend with FFX (7.4 vs 11.8 months). However, Otsu et al. ¹⁷ reported the opposite trend (11.2 vs 9.1 months). ¹⁸ nal-IRI/5-FU/LV has shown superior survival (9.1 vs 4.8 months) to S-1, which is another conventional treatment option in PAC, in a more recent retrospective study. ¹⁹ The liposomal encapsulation of irinotecan enables prolonged systemic circulation and enhances antitumor efficacy while modulating toxicity profiles. ⁴ This may explain the modestly superior outcomes reported in retrospective comparisons with conventional chemotherapeutic agents. Given that the safety profile of the nal-IRI/5-FU/LV regimen was manageable and did not adversely affect health-related quality of life (QoL), it can be considered clinically meaningful. ²⁰ In real-world practice, the sequence of chemotherapy regimens is determined by the physician, patient, and contextual factors and can therefore vary widely. Hence, we evaluated the efficacy of nal-IRI/5-FU/LV across the entire treatment course and demonstrated that nal-IRI exposure was significantly associated with a survival benefit, which was maintained in the subgroup analysis of patients with metastatic cancer.

The clinical implications of these findings are significant for the management of PAC. The ability of nal-IRI/5-FU/LV to prolong survival represents a critical advancement in PAC treatment, particularly in patients with disease progression after gemcitabine-based treatment. However, the economic impact of nal-IRI, given its significantly higher cost than conventional regimens, complicates its routine use in resource-constrained settings. Notably, cost-analysis studies on nal-IRI/5-FU/LV for PAC are lacking. In a previous study, the nearly 30-fold higher cost is considered a limitation, considering the modest survival benefit compared with FOLFIRI. ²¹ In a Greek report, the cost difference was estimated to be €21,468 and €79,799 for nal-IRI/5-FU/LV and 5-FU/LV based on a cost per life-year gained versus 5-FU/LV alone. ⁶ Our results show that the incremental cost of $33,285 for the prolonged survival in the nal-IRI/5-FU/LV group supports its use within acceptable willingness-to-pay (WTP) thresholds of $100,000–200,000/quality-adjusted life years (QALY) based on previous studies.^12,13,22

A nal-IRI/5-FU/LV regimen is widely used as a second or later-line therapy for metastatic PAC, which limits the ability to evaluate cost-effectiveness. Recently, nal-IRI was included in NALIRIFOX, which comprised nal-IRI, oxaliplatin, LV, and 5-FU, showing utility as a new first-line treatment in metastatic PAC. ²³ Shao et al. ¹³ constructed a partitioned survival model from the perspective of United States public payers based on clinical data from the NAPOLI-3 trial, and reported $206,340/QALY for NALIRIFOX compared with GN, suggesting that it is not cost-effective at a WTP threshold of $150,000/QALY. From China’s perspective, NALIRIFOX yielded the highest life-years, QALYs, and cost. It resulted in an ICER of $346,330/QALY compared with GN, indicating that it was not cost-effective as a first-line treatment at the WTP threshold of $38,223/QALY. ²⁴

This study assessed the real-world cost-effectiveness of liposomal irinotecan across the entire treatment course of PAC using electronic health records and insurance claims, beyond isolated regimen comparisons seen in clinical trials. Real-world data based studies offer complementary perspectives to clinical trials, particularly for evaluating real-world effectiveness, cost implications, and toxicity in representative patient populations rather than the controlled population of a clinical trials. ²⁵ Importantly, the calculated ICER ($33,285 per life-year gained) aligns with the WHO-recommended threshold of 1–3 times gross domestic product per capita, supporting the economic viability of nal-IRI/5-FU/LV in Korea. ²⁶

Our cost analysis utilized longitudinal patient-level data and may provide more realistic estimations of total cost and survival in the public payer setting, reflecting cumulative burden across various lines of treatment. In terms of drug tolerability, common treatment-related adverse events were nausea, vomiting, diarrhea, fatigue, anorexia, anemia, and neutropenia, which were consistent with previous reports.^2,4,5,27 Furthermore, nal-IRI/5-FU/LV has been covered by National Health Insurance in Korea since August 2021, greatly reducing the financial burden on patients.

This study had some limitations despite its strengths. Most notably, the cost-effectiveness analysis was not conducted as a head-to-head comparison of two different regimens; however, it was based on differences in exposure to liposomal irinotecan over the entire course of treatment, which did not account for quality-adjusted components in life-years gained, a major limitation to the interpretation of the cost-effectiveness of the nal-IRI/5-FU/LV. Cost analyses with different possible regimens (FFX, FOLFIRI, S-1, or others) after gemcitabine failure are required in the future; however, given the similar effectiveness and fewer side effects, it can be cost-effective as a second- or later-line therapy compared with FFX.^17,27 In addition, although we reported life-years gained as an outcome, the absence of patient-reported QoL data (e.g., EQ-5D or SF-6D) limits the calculation of QALYs. Future cost-effectiveness studies incorporating utility weights and disaggregated supportive care costs (e.g., antiemetics and G-CSFs) are warranted. The retrospective nature of the study introduces potential biases related to patient and treatment selection and outcome measurement. Notably, the benefit of liposomal irinotecan exposure may have been overestimated because of the lack of performance status, an important clinical factor in estimating survival outcomes, and the inclusion of the pre-treatment period. To address immortal time bias, we separately analyzed OS from the initiation of nal-IRI, which was 6.7 months, consistent with prior literature.^2,5,19 We also used PSM to strengthen the validity of our findings by mitigating confounding factors inherent in real-world data.

Conclusion

In this real-world study, the combination with liposomal irinotecan (nal-IRI/5-FU/LV) demonstrated a significant survival benefit compared to non-liposomal irinotecan regimens in patients with PAC. Although total medical costs were higher in the nal-IRI/5-FU/LV group, the ICER remained within the accepted WTP thresholds in Korea, supporting its clinical and economic value. Given the poor prognosis and limited treatment options for most patients with unresectable PAC, nal-IRI/5-FU/LV represents a justifiable therapeutic option despite its high cost. Further research incorporating QALYs is warranted to more precisely assess its cost-effectiveness, particularly in light of newer treatment regimens such as NALIRIFOX, to ensure sustainable adoption in real-world clinical practice.