Original Study

Cost Effectiveness Analysis of

Pharmacokinetically-Guided 5-Fluorouracil in
FOLFOX Chemotherapy for Metastatic
Colorectal Cancer
Daniel A. Goldstein,1 Qiushi Chen,2 Turgay Ayer,2 David H. Howard,1,3
Joseph Lipscomb,1,3 R. Donald Harvey,1 Bassel F. El-Rayes,1
Christopher R. Flowers1
Abstract
Dosing chemotherapy based on pharmacokinetics (PK) instead of body surface area (BSA) has been shown to
decrease interindividual variability in drug exposure. PK-guided 5-fluorouracil (5-FU) instead of BSA-guided 5-
FU for patients with metastatic colorectal cancer (mCRC) leads to decreased toxicity and increased overall
survival (OS). In this article, we use Markov modeling to show that this is a cost-effective strategy, costing
$23,000 per quality-adjusted life-year (QALY) gained.
Background: Dosing chemotherapy based on BSA results in marked interindividual variability in drug exposure. A
randomized trial showed increased OS and decreased toxicity with PK-guided compared with BSA-based 5-FU
dosing in patients with mCRC. The objective of this study was to compare the cost effectiveness of PK-based 5-
FU dosing with BSA-based 5-FU dosing in patients with mCRC receiving FOLFOX (5-FU, leucovorin, and oxalipla-
tin). Materials and Methods: We developed a Markov model to evaluate the cost effectiveness of PK FOLFOX
compared with BSA FOLFOX. Progression risks and cause-specific mortality were extrapolated from the fitted survival
models. Costs for administration and management of adverse events were estimated based on 2013 Medicare
reimbursement rates and average sale prices. Results: PK FOLFOX provided 2.03 QALYs at a cost of $50,205
compared with BSA FOLFOX, which provided 1.46 QALYs at a cost of $37,173. The incremental cost-effectiveness
ratio (ICER) was $22,695 per QALY. The ICER remained < $50,000 per QALY in all univariate and multivariate
sensitivity analyses. Conclusion: At a $50,000 per QALY threshold, PK FOLFOX is cost effective for mCRC. Because
of the cost effectiveness profile and OS advantage with PK FOLFOX, it should be evaluated further in comparative
effectiveness studies.

Clinical Colorectal Cancer, Vol. 13, No. 4, 219-25 ª 2014 Elsevier Inc. All rights reserved.
Keywords: Body surface area, Colorectal neoplasms, Cost-benefit analysis, Fluorouracil, Quality-adjusted life years

Introduction
Colorectal cancer is the second most common cancer in women
and the third most common cancer in men, with an estimated
1
Department of Hematology and Medical Oncology, Winship Cancer Institute, Emory
University, Atlanta, GA
2
H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of
Technology, Atlanta, GA
3
Department of Health Policy and Management, Rollins School of Public Health,
Emory University, Atlanta, GA
Submitted: Jul 8, 2014; Accepted: Sep 10, 2014; Epub: Sep 21, 2014
Address for correspondence: Daniel A. Goldstein, MD, Winship Cancer Institute of
Emory University, 1365C Clifton Rd NE, Suite C5010, Atlanta, GA 30322
Fax: 404-778-3366; e-mail contact: dgolds8@emory.edu
worldwide incidence of more than 1.2 million in 2008, and mor-
tality > 600,000.1 In 2010, $14 billion was spent in the United
States on management of colorectal cancer.2 The current front-line
standard of care for metastatic colorectal cancer (mCRC) is
5-fluorouracil (5-FU)-based chemotherapy as a bolus and contin-
uous infusion over 46 hours, given with FOLFOX (5-FU, leuco-
vorin, and oxaliplatin). However, there is clear evidence that 5-FU
dosing based on body surface area (BSA) results in substantial
interindividual variability in drug exposure.3-6
Dose adjustments of 5-FU based on pharmacokinetics (PK) are
feasible using an immunoassay for 5-FU, to individualize subse-
quent doses to achieve an optimal target exposure.7 In a randomized
trial in mCRC, PK-guided 5-FU was shown to increase median

1533-0028/$ - see frontmatter ª 2014 Elsevier Inc. All rights reserved.

http://dx.doi.org/10.1016/j.clcc.2014.09.007 Clinical Colorectal Cancer December 2014 - 219
Downloaded for Anonymous User (n/a) at Clinical Hospital Center Rijeka from ClinicalKey.com by Elsevier on September
01, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
 Cost Effectiveness Analysis of 5-FU in FOLFOX Chemotherapy
overall survival (OS) and decrease toxicity compared with conven- Because the state transitions can always occur at any time within
tional BSA- and toxicity-based dosing.8 In a nonrandomized phase each period, we applied a half-cycle correction13 when calculating
II trial, PK-guided 5-FU in the FOLFOX regimen was compared life-years (LYs). The primary outputs of the model included total
with a BSA-based group.9 In this trial, 118 patients received direct costs, LYs, quality-adjusted LYs (QALYs), and the incre-
PK-adjusted 5-FU in first-line FOLFOX chemotherapy, and the mental cost-effectiveness ratio (ICER).
comparison arm consisted of 39 patients. When adjusting dose
using PK, 5-FU was monitored during the 46-hour infusion, and Patients and Treatment Regimens
subsequent dosing was based on targeting a specific area under the In the absence of randomized controlled trials with a head-
curve value of 20-24 mg/h/L. In the PK-adjusted arm, the median to-head comparison between PK and BSA FOLFOX, we used
OS and median progression-free survival (PFS) times were 28 and results from one study to estimate survival for patients treated
16 months, respectively. In the BSA-based group, OS and PFS were using a PK protocol and another to estimate survival for patients
22 and 10 months, respectively. Grade 3/4 toxicities included: for whom dosing was determined based on BSA. The PK arm
diarrhea in 2 patients, (1.7%), mucositis in 1 patient (0.8%), and was based on a phase II nonrandomized study by Capitain et al.,
neutropenia in 21 patients (18%) for patients in the PK group and who compared PK FOLFOX with BSA FOLFOX.9 The BSA arm
5 (12%), 6 (15%), and 10 (25%), for patients in the BSA group, was based on a phase III randomized study by Tournigand et al.14
respectively. Although the 2 studies did not have identical target populations,
Because of these differences in response and toxicity, and limited the survival outcomes of the baseline BSA arm of the 2
data regarding whether these benefits warrant the additional cost, we studies are comparable (Table 1).9,14,15 Moreover, the median
analyzed the cost-effectiveness of PK-guided FOLFOX. We devel- PFS and OS were also similar to the FOLFOX arm in a similar
oped a Markov model using a hypothetical cohort of patients with study.15
mCRC to compare the cost-effectiveness of PK-guided FOLFOX
with conventional BSA-guided FOLFOX. Progression Risk and Mortality
Progression risks and cause-specific mortalities for each arm
Materials and Methods were derived from the PFS and OS curves of the corresponding
We built Markov models to evaluate the lifetime costs and study. Engauge Digitizer16 was used to extract the data points
outcomes for each treatment strategy from the third-party payer’s from each PFS and OS plot, and then used to fit parametric
perspective. Health states defined for the clinical course included: survival models. Statistical analyses demonstrated that Weibull
(1) first-line treatment before progression; (2) second-line treatment and log-logistic models provided a good fit for all curves according
after progression; and (3) death (Figure 1). We used previously to the Akaike information criterion, Schwarz Bayesian criterion,17
published estimates of health utilities and costs associated with each and visual inspection (Supplemental Table 1 in the online
health state.10,11 Cost estimates were all valued in 2013 US dollars. version). Weibull models permit the hazard rate to increase over
All costs and outcomes were discounted by 3% annually. time and are thus more suitable for modeling events occurring
The Markov model was implemented in Cþþ language and early during follow-up periods. We therefore selected the Weibull
statistical analyses were performed in R.12 model for all survival curves in this analysis. Next, we estimated
The Markov model simulated the transition of health states the risk for each cycle based on the fitted survival models.
after diagnosis of mCRC with a cycle length of 2 weeks corre-
sponding to the time between chemotherapy administrations. Based
Table 1 Patient Characteristics
on the standard of care, we considered 2 lines of treatment with
different utilities in each line of treatment. After initiating the PK BSA BSA BSA
second-line treatment, no more progression events were considered. Regimen FOLFOX FOLFOX FOLFOX FOLFOX
Study Capitain Capitain Hochster Tournigand
et al, 20129 et al, 20129 et al, 200815 et al, 200414
Figure 1 Markov Model Patient n 118 39 49 111
Median Age, Years 65 63 62 65
Age Range, Years 35-81 32-80 35-79 40-75
Male Sex, % 59 62 57 72
Performance 78 77 100 94
Status 0-1, %
Performance 22 23 0 6
Status 2-3, %
Percentage With 10 10 47 30
Lung Metastases
Percentage With 56 60 76 80
Liver Metastases
Median PFS 16 months 10 months 8 months 8.7 months
Median OS 28 months 22 months 20.6 months 19.2 months
Abbreviations: BSA ¼ Body Surface Area; FOLFIRI ¼ 5-Fluorouracil, Leucovorin, and Irinotecan;
FOLFOX ¼ 5-Fluorouracil, Leucovorin, and Oxaliplatin; mCRC ¼ Metastatic Colorectal Cancer; Abbreviations: BSA ¼ body surface area; FOLFOX ¼ 5-fluorouracil, leucovorin, and oxaliplatin;
PK ¼ Pharmacokinetics. OS ¼ overall survival; PFS ¼ progression-free survival; PK ¼ pharmacokinetics.

220 - Clinical Colorectal Cancer December 2014

Downloaded for Anonymous User (n/a) at Clinical Hospital Center Rijeka from ClinicalKey.com by Elsevier on September
01, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
 Daniel A. Goldstein et al
In particular, because OS is defined as the period between the
Table 2 Drug Costs for 2-Week Cycle
start of treatment and death, we computed the cause-specific
mortality rates, p1(t), at cycle t based on the fitted Weibull Drug and Dose Cost Total Cost
OS model, OS(t), using the formula: First-Line FOLFOX

Pðt  T  t þ 1Þ ðOSðtÞ  OSðt  1ÞÞ Oxaliplatin 85 mg/m2 $0.535 per 0.5 mg $169.16
p1 ðtÞ ¼ ¼ : (1) Leucovorin 350 mg $4.501 per 50 mg $31.51
PðT  tÞ OSðtÞ
5-FU 400 mg/m2 bolus $1.995 per 500 mg $2.97
PFS is the period between the start and progression of 5-FU 2400 mg/m2 continuous $1.995 per 500 mg $17.81
disease or death. We computed the combined risks p2(t) based infusion
on the Weibull PFS model using Formula 1 and computed Palonosetron 250 mcg I.V. $19.384 per 25 mcg $193.84
p2(t)  p1(t) as the estimate of the progression risks. Estimates Dexamethasone 12 mg p.o. $0.64 per 4 mg $1.92
of mortality and progression risk beyond the follow-up time Prochlorperazine 30  10 mg $0.89 per 10 mg $26.70
in the clinical trials were extrapolated based on the fitted tablets
survival models. The overall mortality was defined according to Total $443.91
the maximum value of cause-specific mortality and the Second Line FOLFIRI
background mortality. We use the US life tables to estimate the Irinotecan 180 mg/m2 $4.46 per 20 mg $74.66
background mortality for each age group separately.18 In each Leucovorin 400 mg/m 2
$4.501 per 50 mg $66.97
run of the model simulation, the initial age was sampled from 5-FU 400 mg/m2 bolus $1.995 per 500 mg $2.97
the distribution of age at diagnosis for mCRC in the 5-FU 2400 mg/m2 continuous $1.995 per 500 mg $17.81
Surveillance, Epidemiology, and End Results database since infusion
the year 2000. Palonosetron 250 mcg I.V. $19.384 per 25 mcg $193.84
Dexamethasone 12 mg p.o. $0.64 per 4 mg $1.92
Utility Estimates Prochlorperazine 30  10 mg $0.89 per 10 mg $26.70
To compute the total QALYs in the Markov model, survival time tablets
was adjusted for quality of life using health utilities. We assumed Atropine 0.4 mg I.V. $1.73 $1.73
the health utility of patients during first-line therapy is 0.85 and Loperamide 2-4 mg q.4h., 30 $0.20  30 $6.00
0.65 during second-line therapy based on estimates by Ramsey tablets
et al.19 We established the probability of Grade 3/4 adverse events Total $392.60
(AEs) from the trials used for the models.15,20 AEs can reduce
Abbreviations: FOLFIRI ¼ 5-fluorouracil, leucovorin, and irinotecan; FOLFOX ¼ 5-fluorouracil,
patients’ utility so we applied disutility estimates for these temporary leucovorin, and oxaliplatin; 5-FU ¼ 5-fluorouracil; p.o. ¼ orally; q.4h. ¼ every 4 hours.
health states to each major AE based on data from Aballea at al.21
These included febrile neutropenia, vomiting, and diarrhea. outpatient physician service cost based on relative value units using
We assumed the duration of each AE was 5 days. For each AE, the method described in Tumeh et al.10 The costs for each proce-
the average disutility was weighted by the incidence reported in dure are listed in Table 3.
the clinical study. Because our primary end point was to compare the effectiveness
and costs for different treatments in first-line therapy, we assumed
Cost Estimates the subsequent treatment was FOLFIRI (5-FU, leucovorin, and
Direct costs included drug costs, administration costs, moni-
toring costs, and AE costs. The drug costs for FOLFOX were based
Table 3 Costs of Drug Administration and Routine
on the following doses: oxaliplatin 85 mg/m2 intravenous (I.V.), Management
leucovorin 350 mg I.V., 5-FU 400 mg/m2 I.V. bolus, and
2400 mg/m2 continuous infusion over 46 hours. Each cycle lasted CPT Median Minimum Maximum
2 weeks. All doses were calculated based on a US mean BSA of Description Code Cost, $ Range, $ Range, $
1.86 m2.22 We assigned 5-FU costs to the PK FOLFOX group Outpatient Follow-Up 99213 49.67 43.28 66.32
based on an average dose. Some patients would receive more and Visit
some less based on the results of PK testing. In the PK group, the Inpatient First Visit 99222 134.73 116.73 179.17
cost of the PK assay was added for 4 cycles. Because these are Inpatient Follow-Up 99232 70.09 61.58 94.16
Visit
commonly used drugs, we did not round drugs up to the full vial
Inpatient Discharge 99238 70.77 60.98 93.4
sizes. We used the 2013 Centers for Medicare and Medicaid services Visit
average sales price for the unit price of each drug.23 The costs of Chemotherapy 96413 143.24 99.45 190.62
drugs are listed in Table 2. Administration, I.V.,
First Hour
Administration costs and AE costs were calculated according to
Chemotherapy 96415 30.62 22.7 39.5
the Medicare physician fee schedule for 2013. Cost for chemo-
Administration, I.V.,
therapy administration was based on a 4-hour infusion every Each Additional Hour
2 weeks. Each medical service performed by a physician is associated
Assumptions: national payment amount (ie, GPCI) ¼ 1; conversion factor (for 2013) ¼ 34.023;
with a specific Health Care Procedure Coding System code and/or data of 2013 year.
the Current Procedure Terminology code.24 We computed the Abbreviations: CPT ¼ Current Procedure Terminology; GPCI ¼ Geographic Practice Cost Index.

Clinical Colorectal Cancer December 2014 - 221

Downloaded for Anonymous User (n/a) at Clinical Hospital Center Rijeka from ClinicalKey.com by Elsevier on September
01, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
 Cost Effectiveness Analysis of 5-FU in FOLFOX Chemotherapy
irinotecan) and identical in each arm based on recent standard
regimens.14 Treatment regimen and drug costs for the second-line
treatment are described in Table 2.
To estimate the cost of Grade 3/4 AEs, we assumed that patients
with febrile neutropenia would be hospitalized, and others would be
managed as outpatients. For hospitalization because of febrile
neutropenia, we computed the cost of inpatient service (physician
fee and hospital reimbursement) based on the diagnosis-related
group (DRG) code and length of stay.10 The costs for other AEs
included the costs of 1 clinic visit and corresponding medications.
For each treatment strategy, the costs of each AE were obtained by
multiplying the incidence reported in published studies and the unit
cost for each type of AE (Table 4).9,15,19-21 Because the OS in both
arms was short we made an assumption that the health care costs
not related to cancer were the same in both arms.
Sensitivity Analysis
A series of sensitivity analyses were performed to evaluate the
robustness of the model and to address the uncertainty in parameter
estimation. Ranges and distributions used in sensitivity analysis
are summarized in Table 4. Utilities were varied over their
95% confidence intervals. For each procedure, physician fees of
different carriers were usually adjusted by different geographic
pricing cost index adjustment factors, and the ranges of physician

Table 4 Model Parameters: Baseline Values, Ranges and Distributions for Sensitivity Analysis

Variable Value Minimum Maximum Reference Distribution

Cost (Per Cycle)
Administration costa 284.77 177.70 375.44 e g (28.295, 10.064)
FOLFOX drug cost 443.91 355.13 532.69 e g (100, 4.439)
PK test costb 400.00 300.00 500.00 e g (64, 6.25)
Second-line drug cost 392.60 314.08 471.12 e g (100, 3.926)
Adverse Event Costs
FN 11,565.60 7092.16 27,655.84 e g (2.067, 5596.247)
Diarrhea 71.50 57.20 85.80 e g (100, 0.715)
Vomiting 811.12 648.90 973.34 e g (100, 8.111)
Hypertension 51.89 41.51 62.27 e g (100, 0.519)
Adverse Event
Incidence
BSA arm
FN 0.040 0.011 0.136 Hochster et al, 200815 b (1.96, 47.04)
Diarrhea 0.310 0.198 0.449 Hochster et al, 200815 b (15.19, 33.81)
Vomiting 0.310 0.198 0.449 Hochster et al, 200815 b (15.19, 33.81)
Hypertension 0.000 0.000 0.073 Hochster et al, 200815 b (1, 48)
PK arm
FN 0.013 0.002 0.095 c
b (1.534, 116.466)
Diarrhea 0.017 0.003 0.102 Capitain et al, 20129 b (2.006, 115.994)
Vomiting 0.310 0.198 0.449 d
b (36.58, 81.42)
Hypertension 0.000 0.000 0.073 Capitain et al, 20129 b (1, 117)
Second-line
FN 0.030 0.017 0.051 Bennouna et al, 201320 b (12.3, 397.7)
Diarrhea 0.080 0.058 0.110 Bennouna et al, 201320 b (32.8, 377.2)
Vomiting 0.030 0.017 0.051 Bennouna et al, 201320 b (12.3, 397.7)
Hypertension 0.010 0.004 0.025 Bennouna et al, 201320 b (4.1, 405.9)
Utilities
Before progression 0.850 0.680 1.000 Ramsey et al, 200019 b (14.15, 2.497)
After progression 0.650 0.520 0.780 Ramsey et al, 200019 b (34.35, 18.496)
Disutility of FNe 45% 54% 36% Aballea at al, 200721 b (54.55, 66.672)
Disutility of vomiting 19% 23% 15% Aballea at al, 200721 b (80.81, 344.506)
Disutility of diarrhea 36% 43% 29% Aballea at al, 200721 b (63.64, 113.138)
Discount factor 0.030 0.000 0.050 Uniform (0, 0.05)

Abbreviations: BSA ¼ body surface area; FN ¼ febrile neutropenia; FOLFOX ¼ 5-fluorouracil, leucovorin, and oxaliplatin; PK ¼ pharmacokinetics.
a
Clinical visit and chemotherapy.
b
Authors’ personal contact.
c
Febrile neutropenia was not reported in Capitain et al, 20129, so this was extrapolated from Hochster based on neutropenia rates reported in Capitain et al, 20129.
d
Nausea and vomiting was not reported in Capitain et al, 20129. It is therefore assumed to be the same as for BSA FOLFOX in Hochster et al, 200815.
e
Adverse events were assumed to last for 5 days.

222 - Clinical Colorectal Cancer December 2014

Downloaded for Anonymous User (n/a) at Clinical Hospital Center Rijeka from ClinicalKey.com by Elsevier on September
01, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
 Daniel A. Goldstein et al
sampling distributions. We used g distribution for cost parameters,
Table 5 Base Case Results
b distribution for parameters bounded on the 0 to 1 interval
BSA PK (eg, AE incidence estimates and utilities), and the uniform distri-
FOLFOX FOLFOX Incremental bution for the discount factor. The range for the discount factor was
Life Years 1.990 2.597 0.607 0.00 to 0.05. In multivariate analysis, we ran 10,000 replications in
QALYs 1.458 2.031 0.573 the probabilistic sensitivity analysis. The baseline values, ranges, and
Cost, $ 37,173.10 50,177.42 13,004.32 distributions for model parameters are presented in Table 4.
QALYs Before Progression 0.709 1.465 0.756
LYs Before Progression 0.837 1.726 0.889 Results
Cost of Adverse Events, $ 1009.69 568.53 441.16 In the base case analysis, effectiveness and costs were compared
Cost of First-Line 15,852.31 34,266.75 18,414.44 between the PK and BSA group. PK FOLFOX provided
Chemotherapy
2.03 QALYs at a cost of $50,177 compared with BSA FOLFOX with
Cost of Second-Line 20,311.10 15,342.15 4968.96
Chemotherapy 1.46 QALYs at a cost of $37,173. The ICER for PK FOLFOX was
ICER (QALY) e e 22,695.15 $22,695 per QALY. In terms of LYs, PK FOLFOX provided
ICER (LY) e e 21,423.92 2.60 LYs compared with BSA FOLFOX with 1.99 LYs. The ICER of
PK 5-FU was $21,423 per LY. All numerical results are summarized
Abbreviations: BSA ¼ body surface area; FOLFOX ¼ 5-fluorouracil, leucovorin, and oxaliplatin; in Table 5.
ICER ¼ incremental cost-effectiveness ratio; Ly ¼ life-year; PK ¼ pharmacokinetics; QALY ¼
quality-adjusted life-year. In the univariate sensitivity analysis, parameters that influenced
ICERs the most differed between the model arms (Figure 2). As we
fees were defined by the lowest and highest cost among all carriers in adjusted the hazard ratios, the model results remained robust, with
the Centers for Medicare & Medicaid Services database. Similar for the ICER range being $20,000 to 25,000 per QALY. Variation in
DRG rates, we computed the lowest and highest DRG rates among other parameters had very limited effect (< $4000 per QALY) on
all providers in the United States using the methods previously the ICER of PK. One parameter of interest was the cost of the PK
described.10 Drug costs were varied within  20% of their baseline test, and the ICER ranged between $22,000 and 23,300 per QALY
values as previously done by Goulart and Ramsey.25 To define the when the PK test cost was between $300 and $500.
ranges of the mortalities and progression risks for the sensitivity The results of probabilistic sensitivity analysis are shown in the scatter
analysis, we first adjusted the PFS/OS curves S(t) in the base case by plot in Figure 3. Each point in the scatter plot corresponds to 1 sample
a hazard ratio adjustment g, as [S(t)g], and then reestimated of parameter values. Two reference lines indicating willingness to
the risks for every cycle based on the adjusted survival curves. pay (WTP) of $50,000 and $100,000 are presented in Figure 3.
The adjustment hazard ratio was chosen to be 1  0.2 (1 indicates All points were < the $100,000 per QALY WTP line, and most
that the base case value was used). points were < the $50,000 WTP line. In particular, 99% of samples
In univariate sensitivity analysis, we examined the effect of each in the probabilistic sensitivity analysis had the ICER < the WTP
parameter separately on the ICER. In probabilistic sensitivity of $41,800 per QALY, and 95% < $33,300 per QALY. Considering
analysis, parameters were varied simultaneously according to their the WTP of $50,000 per QALY and $100,000 per QALY, the

Figure 2 Tornado Diagram: ICER of BSA FOLFOX Compared With PK FOLFOX

Abbreviations: BSA ¼ Body Surface Area; FN ¼ Febrile Neutropenia; FOLFOX ¼ 5-Fluorouracil, Leucovorin, and Oxaliplatin; HR ¼ Hazard Ratio; ICER ¼ Incremental Cost-Effectiveness Ratio; OS ¼
Overall Survival; PFS ¼ Progression-Free Survival; PK ¼ Pharmacokinetics.

Clinical Colorectal Cancer December 2014 - 223

Downloaded for Anonymous User (n/a) at Clinical Hospital Center Rijeka from ClinicalKey.com by Elsevier on September
01, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
 Cost Effectiveness Analysis of 5-FU in FOLFOX Chemotherapy
Figure 3 Probabilistic Sensitivity Analysis of PK FOLFOX
Versus BSA FOLFOX
Abbreviations: BSA ¼ Body Surface Area; FOLFOX ¼ 5-Fluorouracil, Leucovorin, and Oxali-
platin; PK ¼ Pharmacokinetics; QALY ¼ Quality-Adjusted Life-Year; WTP ¼ Willingness to Pay.
percentage increased to 99.7% and 100%, respectively. The cost-
effectiveness acceptability curve is shown in Figure 4.
Discussion
We performed a cost-effectiveness analysis of PK-guided FOL-
FOX compared with standard BSA-guided FOLFOX in patients
with mCRC. In the base case, PK FOLFOX improved OS by
approximately 6 months, and had lifetime incremental costs of
$13,000 with an ICER of $22,700 per QALY. In all 1-way and
99.7% of probabilistic sensitivity analyses, the ICER of PK
remained < $50,000 per QALY.
For many years there has been significant interest in the meta-
bolism of 5-FU by dihydropyrimidine dehydrogenase (DPD)26 and
the use of pharmacogenomic strategies to guide 5-FUebased ther-
apy. Patients with DPD deficiency are susceptible to severe toxicity
after 5-FU exposure. Many strategies have been developed to
individualize 5-FU dosing, but so far none have reached standard
practice.27 The use of PK-guided dosing appears to be one of the
Figure 4 Cost Effectiveness (CE) Acceptability Curve
224 - Clinical Colorectal Cancer December 2014
Downloaded for Anonymous User (n/a) at Clinical Hospital Center Rijeka from ClinicalKey.com by Elsevier on September
01, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
most promising approaches to individualize 5-FUebased chemo-
therapy regimens.
In a time of increasing concerns regarding the cost of cancer care,
there are many examples of when personalized medicine might
produce cost savings. For example, in early-stage breast cancer,
using genomic testing to select the patients most likely to benefit
from adjuvant chemotherapy has been demonstrated to be a cost-
saving strategy.28-30 In mCRC, Kirsten rat sarcoma viral oncogene
homolog mutation testing helps to select the patients most likely to
benefit from anti-epidermal growth factor receptor therapy, and is
also a cost-saving strategy.31 In both of these examples, the cost of
the test is outweighed by the cost of expensive therapy and toxicity
that is avoided in patients who would not benefit. Using PK-guided
FOLFOX is an example of personalized cancer treatment that is
cost-effective.
Our study does have some limitations, in particular, the quality
of the trials that were used to build the model. A randomized trial
of PK-guided 5-FU demonstrates increased OS and PFS in
addition to decreased toxicity.8 However, in this trial, 5-FU was
given with leucovorin alone which is not the current standard of
care in the United States. The more standard regimen today is to
add oxaliplatin to 5-FU and leucovorin, forming the FOLFOX
regimen. PK FOLFOX was analyzed in a trial, however, it was
not a randomized design. Despite demonstrating improved
OS and PFS with decreased toxicity, patients in the BSA
comparison arm were recruited nonrandomly from a different
institution. This trial was used to extract the data for the model.
We did successfully combat this limitation by using data from
the studies by Hochster et al15 and Tournigand et al14 to validate
the BSA group. Furthermore, our sensitivity analyses showed
that the model is robust against the uncertainties in survival risks
for each arm.
In the United States, the standard of care in first-line therapy for
mCRC is to use bevacizumab with FOLFOX. However there are no
trials available using PK-adjusted FOLFOX with bevacizumab, thus
our study did not include bevacizumab. It is also noteworthy that
bevacizumab has a substantial cost, and would significantly inflate
the total costs, presumably equally, for both arms. Additionally, the
clinical use of PK-directed FOLFOX with or without bevacizumab
should not be different.
Conclusion
PK-adjusted 5-FU in the FOLFOX regimen for mCRC appears
to be cost effective in our model with wide variations in model
parameters in univariate and multiple variable probabilistic
sensitivity analyses. These data provide the baseline for develop-
ment of further comparative effectiveness studies using PK-guided
5-FU. In an era of technological advances and concerns regarding
costs for cancer care, this study provides evidence to support the
use of personalized medicine. Using new technology in a
personalized strategy is more cost effective than a “1 size fits all”
approach.
Clinical Practice Points
 Dosing chemotherapy based on PK instead of BSA has been
shown to decrease interindividual variability in drug exposure.

 Pharmacokinetics-guided 5-FU instead of BSA-guided 5-FU for
patients with mCRC leads to decreased toxicity and increased
OS.
 Markov modeling demonstrates that PK 5-FU is a cost effective
strategy, costing $23,000 per QALY gained.
Disclosure
The authors have stated that they have no conflicts of interest.
Supplemental Data
The supplemental table accompanying this article can be
found in the online version at http://dx.doi.org/10.1016./j.clcc.
2014.09.007.
References
1. Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of
worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer 2010;
127:2893-917.
2. Mariotto AB, Yabroff KR, Shao Y, Feuer EJ, Brown ML. Projections of the cost of
cancer care in the United States: 2010-2020. J Natl Cancer Inst 2011; 103:117-28.
3. Beumer JH, Chu E, Salamone SJ. Body-surface area-based chemotherapy dosing:
appropriate in the 21st century? J Clin Oncol 2012; 30:3896-7.
4. Saif MW, Choma A, Salamone SJ, Chu E. Pharmacokinetically guided dose
adjustment of 5-fluorouracil: a rational approach to improving therapeutic out-
comes. J Natl Cancer Inst 2009; 101:1543-52.
5. Walko CM, McLeod HL. Will we ever be ready for blood level-guided therapy?
J Clin Oncol 2008; 26:2078-9.
6. Saam J, Critchfield GC, Hamilton SA, Roa BB, Wenstrup RJ, Kaldate RR. Body
surface area-based dosing of 5-fluoruracil results in extensive interindividual vari-
ability in 5-fluorouracil exposure in colorectal cancer patients on FOLFOX regi-
mens. Clin Colorectal Cancer 2011; 10:203-6.
7. Beumer JH, Boisdron-Celle M, Clarke W, et al. Multicenter evaluation of a novel
nanoparticle immunoassay for 5-fluorouracil on the Olympus AU400 analyzer.
Ther Drug Monit 2009; 31:688-94.
8. Gamelin E, Delva R, Jacob J, et al. Individual fluorouracil dose adjustment based
on pharmacokinetic follow-up compared with conventional dosage: results of a
multicenter randomized trial of patients with metastatic colorectal cancer. J Clin
Oncol 2008; 26:2099-105.
9. Capitain O, Asevoaia A, Boisdron-Celle M, Poirier AL, Morel A, Gamelin E.
Individual fluorouracil dose adjustment in FOLFOX based on pharmacokinetic
follow-up compared with conventional body-area-surface dosing: a phase II, proof-
of-concept study. Clin Colorectal Cancer 2012; 11:263-7.
10. Tumeh JW, Moore SG, Shapiro R, Flowers CR. Practical approach for using
Medicare data to estimate costs for cost-effectiveness analysis. Expert Rev Phar-
macoecon Outcomes Res 2005; 5:153-62.
11. Tumeh JW, Shenoy PJ, Moore SG, Kauh J, Flowers C. A Markov model assessing
the effectiveness and cost-effectiveness of FOLFOX compared with FOLFIRI for
Downloaded for Anonymous User (n/a) at Clinical Hospital Center Rijeka from ClinicalKey.com by Elsevier on September
01, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
Daniel A. Goldstein et al
the initial treatment of metastatic colorectal cancer. Am J Clin Oncol 2009;
32:49-55.
12. Team RC. R: A Language and Environment for Statistical Computing 2013,
Available at: http://web.mit.edu/r_v3.0.1/fullrefman.pdf.
13. Sonnenberg FA, Beck JR. Markov models in medical decision making a practical
guide. Med Decis Making 1993; 13:322-38.
14. Tournigand C, Andre T, Achille E, et al. FOLFIRI followed by FOLFOX6 or the
reverse sequence in advanced colorectal cancer: a randomized GERCOR study.
J Clin Oncol 2004; 22:229-37.
15. Hochster HS, Hart LL, Ramanathan RK, et al. Safety and efficacy of oxaliplatin
and fluoropyrimidine regimens with or without bevacizumab as first-line treatment
of metastatic colorectal cancer: results of the TREE Study. J Clin Oncol 2008;
26:3523-9.
16. Engauge Digitizer - Digitizing software. Available at: http://digitizer.sourceforge.net.
Accessed June 15, 2014.
17. Auget JL, Balakrishnan N, Mesbah M, Molenberghs G. Advances in Statistical
Methods for the Health Sciences. Boston: Springer; 2007.
18. Arias E. United States life tables, 2007. Natl Vital Stat Rep 2011; 59:1-60.
19. Ramsey SD, Andersen MR, Etzioni R, et al. Quality of life in survivors of colo-
rectal carcinoma. Cancer 2000; 88:1294-303.
20. Bennouna J, Sastre J, Arnold D, et al. Continuation of bevacizumab after first
progression in metastatic colorectal cancer (ML18147): a randomised phase 3 trial.
Lancet Oncol 2013; 14:29-37.
21. Aballea S, Chancellor JV, Raikou M, et al. Cost-effectiveness analysis of oxaliplatin
compared with 5-fluorouracil/leucovorin in adjuvant treatment of stage III colon
cancer in the US. Cancer 2007; 109:1082-9.
22. Centers for Disease Control and Prevention. FastStats Homepage. Available at:
http://www.cdc.gov/nchs/fastats/. Accessed June 15, 2014.
23. CMS.gov. Centers for Medicare & Medicaid Services. 2013 ASP Drug Pricing
Files. Available at: http://www.cms.gov/Medicare/Medicare-Fee-for-Service-Part-
B-Drugs/McrPartBDrugAvgSalesPrice/2013ASPFiles.html. Accessed June 15,
2014.
24. CMS.gov. Centers for Medicare & Medicaid Services. Medicare physician fee
schedule. Available at: http://www.cms.gov/apps/physician-fee-schedule/. Accessed
September 1, 2013.
25. Goulart B, Ramsey S. A trial-based assessment of the cost-utility of bevacizumab
and chemotherapy versus chemotherapy alone for advanced non-small cell lung
cancer. Value Health 2011; 14:836-45.
26. van Staveren MC, Guchelaar HJ, van Kuilenburg AB, Gelderblom H, Maring JG.
Evaluation of predictive tests for screening for dihydropyrimidine dehydrogenase
deficiency. Pharmacogenomics J 2013; 13:389-95.
27. Yen JL, McLeod HL. Should DPD analysis be required prior to prescribing flu-
oropyrimidines? Eur J Cancer 2007; 43:1011-6.
28. Lamond NW, Skedgel C, Rayson D, Lethbridge L, Younis T. Cost-utility of the
21-gene recurrence score assay in node-negative and node-positive breast cancer.
Breast Cancer Res Treat 2012; 133:1115-23.
29. Holt S, Bertelli G, Humphreys I, et al. A decision impact, decision conflict and
economic assessment of routine Oncotype DX testing of 146 women with node-
negative or pNImi, ER-positive breast cancer in the U.K. Br J Cancer 2013;
108:2250-8.
30. Blohmer JU, Rezai M, Kummel S, et al. Using the 21-gene assay to guide adjuvant
chemotherapy decision-making in early-stage breast cancer: a cost-effectiveness
evaluation in the German setting. J Med Econ 2013; 16:30-40.
31. Behl AS, Goddard KA, Flottemesch TJ, et al. Cost-effectiveness analysis of
screening for KRAS and BRAF mutations in metastatic colorectal cancer. J Natl
Cancer Inst 2012; 104:1785-95.
Clinical Colorectal Cancer December 2014 - 225