Cost-effectiveness of systematic chemotherapy for metastatic pancreatic cancer: a retrospective study using Japanese clinical data

doi:10.21203/rs.3.rs-4360252/v1

Background: Gemcitabine plus nab-paclitaxel (GnP)and modified FOLFIRINOX (mFFX)are recommended as standard first-line treatments for metastatic pancreatic cancer in Japan. Considering the emergence and associated costs of adverse events in clinical practice, it is unclear which treatment is more cost-effective: mFFX or GnP. This study aimed to evaluate the cost-effectiveness of mFFX and GnPin patients with metastaticpancreatic cancer in clinical practice in Jaan.

Methods: This retrospective cohort study compared patients with metastatic pancreatic cancer treated with mFFX or GnP between December 2013 and February 2017 at the National Cancer Center Hospital East. A partitioned survival model was developed featuring five mutually exclusive health states, namely, 1st line stable disease (1stSD), 1st progressive disease (1stPD), 2nd line stable disease (2ndSD), 2nd progressive disease (2ndPD), and death. A cost-effectiveness analysis was performed to obtain the total costs and quality-adjusted life-years (QALYs). A 2% annual discount rate was applied to costs and outcomes. A one-way sensitivity analysis(OWSA) was performed to assess the effect of parameter uncertainty on the results. mFFX treatment involved the intravenous infusion of oxaliplatin, levofolinate, and irinotecan, while GnP treatment involved weekly intravenous administration of nab-paclitaxel and gemcitabine. Dosages were adjusted based on observed adverse events.

Results: The total costs of the mFFX and GnP arms were JPY 3,193,597 (USD 31,006) and JPY 3,913,171 (USD 37,992), respectively. Incremental total costs of JPY 719,574 (USD 6,986) were associated with GnP compared to mFFX. Total QALYs in the mFFX and the GnP arms were 0.427 and 0.435, respectively.

Conclusion: GnP was associated with slightly higher QALYs than mFFX in Japanese clinical practice and should be the preferred first-line treatment for metastatic pancreatic cancer in Japan.

Gemcitabine plus nab-paclitaxel

modified FOLFIRINOX

pancreatic cancer

chemotherapy

cost-effectiveness

Pancreatic cancer (PC) is a leading cause of cancer-related mortality in Japan [1]. The majority of PC patients are diagnosed with metastatic disease with a remarkably poor prognosis [2]. Gemcitabine (GEM) has long been the standard of care for the treatment of metastatic pancreatic cancer (MPC). Several randomized phase III studies have compared several promising chemotherapies with GEM alone [3–5]. Recently, FOLFIRINOX (FFX) and gemcitabine plus nab-paclitaxel (GnP) have shown superior survival rates compared to GEM alone as first-line treatments for patients with MPC [3, 4]. FFX causes adverse events (AEs) including neutropenia, thrombocytopenia, febrile neutropenia, diarrhea, and neuropathy. To decrease the AEs and increase tolerability, modified FOLFIRINOX (mFFX) has been used as the first-line option for patients with advanced PC. Phase II trials in Japan have shown that mFFX improves the safety profile while maintaining efficacy [5]. Therefore, mFFX is also recommended as a first-line treatment for MPC patients in Japan [6, 7]. However, as no randomized comparative head-to-head trial between these GnP, FFX, and mFFX chemotherapy regimens has been conducted yet, the optimal first-choice treatment for patients with MPC remains unclear. Additionally, the daily regimen price per body surface area in Japan is more expensive for GnP than for FFX and mFFX because nab-paclitaxel in the GnP regimen is the branded drug. Therefore, cost-effectiveness should be considered, among other factors, when selecting first-line treatment for patients with MPC.

Recently, the cost-effectiveness of the first-line treatment of MPC was evaluated in the UK [8]. A study demonstrated that the total cost of FFX (GBP 18,456) was lower than that of GnP (GBP 20,806), while the quality-adjusted life-years (QALYs) for FFX (0.687) was slightly higher than that for GnP (0.608). Thus, FFX is the most effective first-line treatment for patients with MPC. However, these studies did not consider the emergence of low-grade AEs and the associated medical costs of clinical practice in Japan.

Although grade 3 or higher toxicities occur frequently in patients treated with mFFX and GnP, grade 1 or 2 toxicities induced by chemotherapy in cancer patients have not been considered, even though they have been shown to impair patients' quality of life (QoL) [4–5, 9–10]. Furthermore, the cost-effectiveness of mFFX, which is frequently used in Japanese clinical practice, has not been clarified in previous studies. Therefore, we evaluated the cost-effectiveness of mFFX and GnP in Japanese MPC patients from a healthcare provider's perspective using actual clinical record data in a cancer center in Japan.

This study was conducted with the approval of the ethics committee of the National Cancer Center Hospital East (NCCHE) and in accordance with epidemiological research guidelines (2019 − 179).

Patient population and data collection

All participants in this study were patients with MPC treated with mFFX or GnP between December 1, 2013 and February 28, 2017, at the NCCHE (Fig. 1). The patients were followed up until March 31, 2021. All patients survived the NCCHE or out-of-hospital outcomes (survival or death at home and transfer). Patients who died at the NCCHE were defined as the analysis population because transfer patients are difficult to track and their data cannot be collected. This study was based on the medical record data from a single hospital. Therefore, to utilize as much patient data as possible from the collected data, the overall survival (OS) and progression-free survival (PFS) data for each arm were extracted from the entire population, and safety data such as direct medical costs and frequency of AEs were extracted.

mFFX consisted of oxaliplatin at a dose of 85 mg/m², administered as a 2-hour intravenous infusion, immediately followed by levofolinate at a dose of 200 mg/m², as a 2-hour intravenous infusion, with the addition of irinotecan after 30 minutes at a dose of 150 mg/m², as a 90-minute intravenous infusion. Immediately following this treatment, 5-FU was administered every 2 weeks at 2400 mg/m2 by continuous IV infusion over 46 hours. Nab-paclitaxel was administered intravenously at a dose of 125 mg/m² over 30 minutes, and gemcitabine was administered intravenously at a dose of 1,000 mg/m² over 30 min, once weekly for 3 consecutive weeks, with each treatment cycle lasting 4 weeks. mFFX and GnP doses were adjusted based on the observed AEs. Treatment was continued until disease progression became evident or until unacceptable toxicity occurred.

Model structure and design

We created a partitioned survival model to compare the healthcare costs and clinical outcomes associated with mFFX and GnP in the 1st line treatment MPC patients [11.12]. As shown in Fig. 2, five mutually exclusive health states were included in the model: 1st line stable disease (1stSD), 1st progressive disease (1stPD), 2nd line stable disease (2ndSD), 2nd progressive disease (2ndPD), and death. The primary endpoints of the model were used to calculate the incremental cost, incremental QALY, and incremental cost-effectiveness ratio (ICER) of the GnP arm against the mFFX arm. Transition state cycles were one month, and a lifetime horizon was used. Our cost-effectiveness analysis was conducted from the healthcare provider’s perspective using a standard rate of 2% annually to discount future costs and QALYs, according to the guidelines for cost-effectiveness evaluation in Japan (Guideline for cost-effectiveness evaluation in Japan). The model was constructed using Microsoft Excel for Microsoft 365, and additional statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA) and R version 3.6.3.

Efficacy

We used the individual patient data (IPD) of patients treated with mFFX or GnP between December 1, 2013 and February 28, 2017, at the NCCHE to calculate the transition probabilities at 1st PFS, 1stPD, 2nd PFS, and 2ndPD for each arm as efficacy parameters. Specifically, we estimated survival curves using the number of days in the four states for each patient and a functional model. We then used the estimated survival curves to calculate the transition probabilities. The specific definitions of the four periods are as follows:

1st PFS: Number of days from the first dose of first-line chemotherapy to the day before the confirmation of PD of first-line chemotherapy.
1stPD: Number of days from PD of the first chemotherapy to death or the day before the first dose of the second chemotherapy.
2nd PFS: Number of days from PD of the second chemotherapy to the day before the date of PD of the second chemotherapy.
2ndPD: Number of days from PD of the first chemotherapy to death.

The six model functions selected for each period were exponential, Weibull, Gompertz, log-normal, log-logistic, and generalized gamma distributions. We calculated the Akaike information criterion, and Bayesian information criterion, and visually compared the Kaplan–Meier curves with the model curves to select the most appropriate model function from these three factors [12, 13]. Curve visualization was also checked to assess the consistency between OS and PFS curves and natural history in Japan [14, 15]. Only the probability of transition from 1stPD to 2ndSD and death was calculated by multiplying the probability of transition per unit month in 1stPD by the proportion of second-line treatment in each arm.

Health-related QoL scores

We calculated the QoL scores (U_Calc) for each state in the analytical model by setting the value of the disease state without AEs of chemotherapy (U_{SD base}) and the value of disutility due to the AEs of chemotherapy (Table 1). The formula is expressed as U_Calc = U_{SD base} - disutility. No QoL investigation regarding U_SD base for 1stSD and 2ndSD in MPC patients has been conducted before in Japan; a previous study using the Vignette-based method evaluated QoL in the general population in Japan for MPC [15]. Therefore, we established the QoL scores from this study for "Stable disease + no adverse event" as the U_{SD base} for 1stSD and 2ndSD.

As 1stSD and 2ndSD assume the use of chemotherapy to treat MPC, patients are affected by AEs. As chemotherapy also impacts a patient's QoL, in this study, we considered the disutility of the AEs associated with PC treatment for our cost-effectiveness assessment (Table S1). Specifically, the type, severity, and number of days of onset of specific AEs during the 1stSD and 2ndSD were collected from IPD for each arm collected from medical records, and the total disutility per year was calculated as well as the point estimate of Disutility in the GnP and mFFX arms. The estimates were calculated as median values. Six medical professionals specializing in oncology selected specific adverse effects. These included febrile neutropenia, neutropenia, diarrhea, nausea and vomiting, peripheral neuropathy, fatigue, alopecia, and dysgeusia. Parameters of disutility were considered for each severity level, which was either grade 1/2 or grade 3/4 in the common terminology criteria for adverse events ver 4.0 and 5.0 [16, 17].

Finally, QoL scores for 1stPD and 2ndPD were based on the results of a QoL survey administered to healthcare professionals (Table 1). As chemotherapy was discontinued in 1stPD and 2ndPD, we assumed that no disutility due to AEs occurred.

Costs

Direct medical costs were calculated as cost parameters for the population analyzed in this study. Specifically, the type and amount of medical resources used for patients were extracted from medical records and multiplied by the Japanese drug price as of 2020, based on the information on medical fees [19].

Direct medical costs were defined as "all costs due to resource use that are completely attributable to medical intervention or disease use" as listed in the electronic medical record data, including inpatient and outpatient hospitalization, drug costs, surgery, laboratory tests, and all other costs that can be billed for receipts. Cost parameters were calculated for each month in 1stSD, 1stPD, 2ndSD, and 2ndPD, up to 12 months from the start date of each state (Table 1). The costs after 13 months were assumed to continue to be the costs up to the 12th month. Costs were converted as 1 USD = 103 JPY (Exchange rate as of April 2020). To determine the impact of costs on the ICER, a breakdown of monthly costs was calculated for both arms (Fig. S1-2).

Sensitivity Analysis

A one-way sensitivity analysis (OWSA) was performed using the upper and the lower confidence interval values for the selected parameters between mFFX and GnP, with a maximum variation of plus or minus 10%. In addition, we performed a scenario analysis in OWSA wherein the QoL scores and transition probabilities for the GnP and mFFX arms in the first-line PFS were increased or decreased from the base case to 1%, 3%, 5%, and 10%. This is because the survival rate of MPC is short, the difference in QALYs between regimens is very small, and the impact of QoL scores and transition probability on the ICER is expected to be large.

Patient characteristics

The baseline characteristics of the study participants are summarized in Table 2. There were no significant differences in baseline characteristics between the mFFX and GnP arms. The proportion of patients treated with second-line chemotherapy was higher in the mFFX than in the GnP arm (79% vs. 56%, respectively). Specifically, in the mFFX arm, GEM or GnP was administered as second-line chemotherapy, whereas in the GnP arm S-1 was administered as second-line chemotherapy. Most patients received treatment for MPC within these three regimens.

Efficacy

The dose intensities of gemcitabine and nab-paclitaxel for GnP treatment were 90.9% and 78.3%, respectively, whereas those of irinotecan, oxaliplatin, and 5-fluorouracil for mFFX treatment were 87.3%, 72.7%, and 98.6%, respectively (Table 2). 1st PFS tended to be longer in the mFFX than in the GnP arm, however, the Kaplan–Meier curves for the two were similar (Fig. S4, Table S3). There were no significant differences in the 2nd PFS of the mFFX and GnP arms (Fig. S4, Table S3).

Base case analysis

The results of the base case model are presented in Table 3. The total costs of the mFFX and GnP arms were JPY 3,193,597 (USD 31,006) and JPY 3,913,171 (USD 37,992), respectively. The incremental total cost of GnP compared with mFFX was JPY 719,574 (USD 6,986). Total QALYs or the mFFX and GnP arms were 0.427 and 0.435, respectively. The incremental total QALYs for the GnP compared to the mFFX arm was 0.008. Therefore, the ICER of GnP for mFFX was JPY 86,089,864 (USD 835,824).

One-way sensitivity analysis

The results of the OWSA for the ICER of the mFFX versus that of the GnP arm are shown in tornado diagrams (Fig. S3). Almost all parameters affected the ICER for the GnP arm compared to that for the mFFX arm. However, only when the QoL score during 1st PFS or the probability of maintaining the 1st PFS increased to 10% or more, the ICER for the mFFX arm was higher than that for the GnP arm (Table S4).

This is the first real-world cost-effectiveness study to evaluate and compare the cost-effectiveness of mFFX compared with GnP as a first-line treatment for MPC in Japan that also considers the costs of second-line chemotherapy and AEs associated with chemotherapy. We demonstrated the increased incremental costs of JPY 719,574 associated with GnP compared with mFFX, and the slightly increased QALYs of 0.008 associated with GnP compared with mFFX. These results suggest that GnP is a regimen with a slight increase in QoL value-weighted efficacy relative to the incremental cost of mFFX, which raises concerns about its economic impact.

Several studies have analyzed the cost-effectiveness of mFFX and GnP as first-line treatments for MPC [20, 21]. In the US, the cost-effectiveness of first-line treatment of MPC has been evaluated using a 3-state Markov mode and showed that the total cost of GnP ($49,007) was lower than that of FFX ($116,087), and QALYs for FFX was higher (0.725) than that for GnP (0.603) [20]. Although the two treatments were not compared, FFX was more cost-effective than GnP as first-line therapy for MPC. Zhou et al. evaluated the cost-effectiveness of FFX and GnP from a Chinese perspective and found that the QALY of FFX and GnP were 0.427 and 0.435, respectively [21]. The ICER of FFX versus GnP was 835,824 USD per QALY gained. Because of the differences in modeling approaches, costs, and healthcare systems among countries, conclusions drawn from one country cannot be applied to another. Additionally, these analyses only provided information from clinical trials. To decrease AEs and increase tolerability to FFX, mFFX has been used as a first-line option for advanced PC in clinical practice in Japan. Therefore, the results of these studies using FFX may not apply to patients in clinical practice in Japan.

Our analyses provided information on IPD and actual hospital costs. Although slight, we demonstrated that GnP was associated with a higher QALY than mFFX; therefore, our results contradict some of the previously reported results. Several factors may have contributed to this result. First, there were small differences in the median 1st PFS between the mFFX and the GnP arms (6.2 months vs 5.5 months). The results of our study were similar to those of previous clinical trials [3–5]. However, because of the similar PFS duration between arms, a slight change in the parameters extrapolated to the economic model could have reversed the results. Additionally, the QoL score, considering the occurrence of AEs during the 1stSD period, was higher in the GnP arm than in the mFFX arm (0.565 vs. 0.547). The lower incidence and severity of AEs for GnP than for mFFX may lead to an increase in QALYs in the GnP arm. Therefore, when the first-line treatment for patients with MPC is selected considering economic impact, it is important to factor not only the efficacy but also the incidence and severity of AEs.

We analyzed IPD and actual hospital costs, including not only those associated with grade 3 or higher but also with grade 1 or 2 toxicities. Grade 1 or 2 toxicities induced by chemotherapy in cancer patients impair the QoL of cancer patients [9, 10]. In particular, peripheral neuropathy, alopecia, and dysgeusia of even grade 1 or 2 reduce the QoL of cancer patients. Therefore, the consideration of toxicity data, including grade 1 or 2 AEs, added a new dynamic to this study. In this study, the incidences of grade 1 or 2 AEs were nausea and vomiting, fatigue, peripheral neuropathy, alopecia, and dysgeusia, which are known AEs that impair QoL in patients with cancer (Table S2). The occurrence rate of most AEs was higher in the mFFX arm than in the GnP arm. These results suggest that GnP is safer, given the lower grade AEs, than mFFX as the first-line treatment for MPC.

Our study has several limitations. First, patients who were analyzed for cost and QoL score in this study were not included among the patients transferred to another hospital. However, the results of the OWSA demonstrated that the parameter that most affected the ICER was the QoL score or the transition probability during 1st PFS. Because the transition probability during 1st PFS was analyzed for all patients, including those transferred to another hospital, we considered that the impact would be limited. Second, we did not consider the 3rd line of treatment and beyond for patients with MPC. However, the results of the OWSA demonstrated that the impact of the efficacy with 2nd line of treatment and beyond on the ICER was lower than that of other parameters; therefore, the impact of the third line of treatment and beyond would be limited.

Finally, the small sample size may contribute to the uncertainty of the results. Since this study is based on medical records from a single hospital, uncertainty is a concern, mainly in the efficacy of the regimen and the frequency of AEs. However, by directly using medical data from a representative cancer center hospital in Japan, this study was able to comprehensively consider the actual status of medical resource use. Further expansion of the number of hospitals and correction of uncertainties are desirable in the future.

Our results show that GnP treatment is associated with a slightly higher QALY than mFFX treatment in Japanese clinical practice. While concerns persist regarding the potential impact of GnP on healthcare costs, its favorable safety profile suggests its potential benefits in real-world clinical practice.

Ethics approval and consent to participate

This study was conducted with the approval of the ethics committee of the National Cancer Center Hospital East (NCCHE) and in accordance with epidemiological research guidelines (2019-179).

Consent for publication

Not applicable.

Availability of data and materials

All data are anonymized at NCCEE and the authors of this study alone have access to it.

Competing interests

The authors declare that they have no competing interests. Takumoto Yuki joined Boehringer Ingelheim in April 2023; this study was designed and conducted before this date.

Funding

This study was supported by a JSPS KAKENHI grant (Grant Number 19H03874).

Authors' contributions

SA had full access to all the data in this study and takes responsibility for the integrity of the data and the accuracy of the data analysis. SA headed the manuscript preparation. TY designed the study and conducted all statistical analyses. TJ and MK collected patient data and curated the database. KT and AM managed and supervised the research and reviewed and interpreted the research policy. All authors read and approved the final manuscript.

Foundation for promotion of cancer research. Cancer statistics in Japan-2022. Foundation for promotion of cancer research, 2022. https://ganjoho.jp/public/qa_links/report/statistics/pdf/cancer_statistics_2022.pdf. Accessed 12 Dec 2023.
Lau MK, Davila JA, Shaib YH. Incidence and survival of pancreatic head and body and tail cancers: A population-based study in the United States. Pancreas. 2010;39(4):458-62.
Conroy T, Desseigne F, Ychou M, Bouché O, Guimbaud R, Bécouarn Y, et al. FOLFIRINOX versus gemcitabine for metastatic pancreatic cancer. N Engl J Med. 2011;364(19):1817-25.
Von Hoff DD, Ervin T, Arena FP, Chiorean EG, Infante J, Moore M, et al. Increased survival in pancreatic cancer with nab-paclitaxel plus gemcitabine. N Engl J Med. 2013;369(18):1691-1703.
Ozaka M, Ishii H, Sato T, Ueno M, Ikeda M, Uesugi K, et al. A phase II study of modified FOLFIRINOX for chemotherapy-naïve patients with metastatic pancreatic cancer. Cancer Chemother Pharmacol. 2018;81(6):1017-23.
National Comprehensive Cancer Network. NCCN guideline. Pancreatic Adenocarcinoma. version 2.2022. https://www.nccn.org/professionals/physician_gls/pdf/pancreatic.pdf. Accessed 12 Dec 2023.
Okusaka T, Nakamura M, Yoshida M, Kitano M, Uesaka K, Ito Y, et al. Committee for revision of clinical guidelines for pancreatic cancer of the Japan pancreas society. Clinical practice guidelines for pancreatic cancer 2019 from the Japan pancreas society: A synopsis. Pancreas. 2020 Mar;49(3):326-35.
Gharaibeh M, McBride A, Alberts DS, Erstad B, Slack M, Alsaid N, et al. Economic evaluation for the UK of systemic chemotherapies as first-line treatment of metastatic pancreatic cancer. Pharmacoeconomics. 2018;36(11):1333-43.
de Mol M, Visser S, den Oudsten BL, Lodder P, van Walree N, Belderbos H, et al. Frequency of low-grade adverse events and quality of life during chemotherapy determine patients’ judgement about treatment in advanced-stage thoracic cancer. Support Care Cancer. 2019;27(9):3563-72.
Guérin A, Chen L, Ionescu-Ittu R, Marynchenko M, Nitulescu R, Hiscock R, et al. Impact of low-grade adverse events on health-related quality of life in adult patients receiving imatinib or nilotinib for newly diagnosed Philadelphia chromosome positive chronic myelogenous leukemia in chronic phase. Curr Med Res Opin. 2014;30(11):2317-28.
Schaefer AJ, Bailey MD, Shechter SM, Roberts MS. Modeling medical treatment using Markov decision processes. In: Brandeau ML, Sainfort F, Pierskalla WP, editors. Operations research and health care. Berlin:Springer; 2005. p. 593-612.
Woods B, Sideris E, Latimer PS, et al. DSU technical support:document 19. Partitioned survival analysis for decision modelling in healthcare: A critical review. NICE Decision Support Unit. 2017. https://www.sheffield.ac.uk/sites/default/files/2022-02/TSD19-Partitioned-Survival-Analysis-final-report.pdf. Accessed 28 Nov 2023.
National institute for health and care excellence. Guide to the methods of technology appraisal 2013 [Internet]. London: national institute for health and care excellence (NICE); 2013. Process and Methods Guides No. 9.
Statistics Bureau Japan. Statistical Handbook of Japan 2021. 2021. http://www.stat.go.jp/data/nihon/02.html. Data is available by 5 Years of Age, Tables 2-20. Accessed 12 Dec 2023.
Ministry of health, labour and welfare. Director-general for statistics and information policy, Vital statistics of Japan 2019. https://www.mhlw.go.jp/toukei/saikin/hw/jinkou/houkoku19/dl/all.pdf. Accessed 10 Dec 2023.
Takumoto Y, Sasahara Y, Narimatsu H, Murata T, Akazawa M. Correction: Health state utility values for metastatic pancreatic cancer using a composite time trade-off based on the vignette-based approach in Japan. Erratum for. Health Econ Rev. 2022;12(1):63.
Japan clinical oncology group. Common Terminology Criteria for Adverse Events (CTCAE). version 4.03; 2010. http://www.jcog.jp/doctor/tool/CTCAEv4J_20170912_v20_1.pdf. Accessed 12 Dec 2023.
Japan clinical oncology group. Common Terminology Criteria for Adverse Events (CTCAE). version 5.0; 2017. http://www.jcog.jp/doctor/tool/CTCAEv5J_20210901_v24_1.pdf. Accessed 12 Dec 2023.
Various Information of medical fee. The master record of medical service compensation in FY2021 revision. https://shinryohoshu.mhlw.go.jp/shinryohoshu/kaitei/doKaiteiR04/. Accessed 10 Nov 2022.
Gharaibeh M, McBride A, Bootman JL, Patel H, Abraham I. Economic evaluation for the US of nab-paclitaxel plus gemcitabine versus FOLFIRINOX versus gemcitabine in the treatment of metastatic pancreas cancer. J Med Econ. 2017;20(4):345-52. doi: 10.1080/13696998.2016.1269015.
Zhou J, Zhao R, Wen F, Zhang P, Wu Y, Tang R, et al. Cost-effectiveness analysis of treatments for metastatic pancreatic cancer based on PRODIGE and MPACT trials. Tumori. 2016;2016(3):294-300. doi: 10.5301/tj.5000499.

Table 1: Model parameters

State	mFFX/Both				GnP				Reference
Parameter for the QoL Score
State	N	U_Calu	U_{sd base}	Disutility	N	U_Calu	U_{sd base}	Disutility
1stSD	18	0.547	0.634	0.087	96	0.565	0.634	0.069	[16]
2ndSD	63	0.541	0.634	0.093	-	-	-	-	[16]
1stPD/2ndPD	114	0.150	0.15	0.000	-	-	-	-	Estimated from healthcare providers
Parameter for the Cost (USD)
State	Month	Median	Q1	Q3	Month	Median	Q1	Q3
1stSD	1	6,021	7,503	12,684	1	8,587	5,292	8,850	Calculated from medical records in NCCHE
	2	3,786	1,557	2,278	2	1,653	2,807	4,207
	3	3,784	967	2,160	3	1,459	2,832	4,930
	4	2,692	1,537	2,387	4	1,699	2,481	4,208
	5	2,953	1,082	1,347	5	1,279	2,700	4,055
	6	2,960	2,063	2,516	6	1,722	2,555	4,247
	7	2,677	531	1,493	7	1,785	2,608	4,075
	8	2,531	1,307	4,435	8	1,785	1,765	3,432
	9	2,536	900	2,103	9	1,785	2,321	3,605
	10	2,742	1,478	1,646	10	1,785	2,690	4,338
	11	2,433	1,323	1,790	11	1,785	2,296	3,820
	12	2,696	1,635	2,434	12	1,785	2,555	5,051
2ndSD	1	3,608	464	3,011	1	739	1,388	8,103
	2	1,640	1,415	3,725	2	2,570	1,652	2,112
	3	1,861	767	1,415	3	1,091	1,403	3,633
	4	1,669	585	1,007	4	796	1,403	2,544
	5	1,669	860	1,146	5	1,003	1,418	2,662
	6	2,071	611	1,919	6	1,265	1,963	2,457
1stPD	All period	7,313	-	-	All period	-	-	-
2ndPD	All period	2478.981	-	-	All period	-	-	-

mFFX, modified FOLFIRINOX; GnP, Gemcitabine plus nab-paclitaxel; 1stSD, 1st line stable disease; 2ndSD, 2nd line stable disease; 1stPD, 1st line progressive disease; 2ndPD, 2nd line progressive disease; U_Calc, quality of life value considering disutility; U_Base, quality of life value without considering disutility.

Table 2: Patient characteristics

Contents	mFFX	GnP
N	48	279
Age (median, min-max)	64 (34 - 78)	66 (26 - 86)
Sex
Male	27 (56%)	174 (62%)
Female	21 (44%)	105 (38%)
Progress
Death (Other hospital)	20 (42%)	103 (37%)
Death (NCCHE)	18 (38%)	96 (34%)
Survival	2 (4%)	24 (9%)
Unknown	8 (17%)	56 (20%)
2nd line chemotherapy
BSC	10 (21%)	122 (44%)
GEM	18 (38%)	0 (0%)
GnP	15 (31%)	0 (0%)
mFFX	0 (0%)	54 (19%)
S-1	2 (4%)	100 (36%)
GEM+Erlo	1 (2%)	0 (0%)
Other	2 (4%)	3 (1%)
Number of chemotherapies
1st line only	10 (21%)	122 (44%)
2nd line	25 (52%)	122 (44%)
3rd line	11 (23%)	31 (11%)
4th line	2 (4%)	4 (1%)
Relative Dose Intensity % (mean, SD)
CPT-11	87.3 (9.2)	-
LV	97.2 (2.5)	-
5-FU	98.6 (4.7)	-
OXA	72.7 (21.8)	-
GEM	-	90.7 (9.5)
NPTX	-	78.3 (23.5)

mFFX, modified FOLFIRINOX; GnP, Gemcitabine plus nab-paclitaxel; NCCHE, National cancer center hospital east; BSC, Best supportive care; GEM, Gemcitabine; Erlo, Erlotinib; CPT-11, Capecitabin; LV, Leucovorin; 5-FU, Fluorouracil; OXA, Oxaliplatin; NPTX, nab-paclitaxel

Table 3: Total cost, total QALYs, and ICER for each regimen (USD)

State	Total cost		Incremental Total Cost	Total QALY		Incremental total QALY
State	mFFX	GnP	Incremental Total Cost	mFFX	GnP	Incremental total QALY
All	31,006	37,992	6,986	0.427	0.435	0.008
1stSD	18,588	24,328	5,740	0.334	0.342	0.008
1stPD	7,185	7,144	-41	0.012	0.012	0.000
2ndSD	1,819	3,678	1,858	0.064	0.067	0.003
2ndPD	3,413	2,842	-571	0.017	0.014	-0.003

mFFX, modified FOLFIRINOX; GnP, Gemcitabine plus nab-paclitaxel; 1stSD, 1st line stable disease; 2ndSD, 2nd line stable disease; 1stPD, 1st line progressive disease; 2ndPD, 2nd line progressive disease

No competing interests reported.

SupplementalMaterial20240503.docx

Cost-effectiveness of systematic chemotherapy for metastatic pancreatic cancer: a retrospective study using Japanese clinical data

Status:

Version 1

Abstract

Figures

Introduction

MATERIALS AND METHODS

Patient population and data collection

Model structure and design

Efficacy

Health-related QoL scores

Costs

Sensitivity Analysis

RESULTS

Patient characteristics

Efficacy

Base case analysis

One-way sensitivity analysis

DISCUSSION

Conclusion

Declarations

References

Tables

Additional Declarations

Supplementary Files

Status:

Version 1