Accepted Manuscript Relationship between thymidine kinase 1 expression and trifluridine/tipiracil therapy in refractory metastatic colorectal cancer: A pooled analysis of two randomized clinical trials Takayuki Yoshino, Kentaro Yamazaki, Eiji Shinozaki, Yoshito Komatsu, Tomohiro Nishina, Hideo Baba, Akihito Tsuji, Yasushi Tsuji, Kensei Yamaguchi, Naotoshi Sugimoto, Tadamichi Denda, Kei Muro, Tetsuji Takayama, Taito Esaki, Yasuo Hamamoto, Toshikazu Moriwaki, Yasuhiro Shimada, Masahiro Goto, Norisuke Nakayama, Hirofumi Fujii, Takanori Tanase, Atsushi Ohtsu PII:
S1533-0028(18)30261-5
DOI:
10.1016/j.clcc.2018.07.009
Reference:
CLCC 489
To appear in:
Clinical Colorectal Cancer
Received Date: 28 May 2018 Revised Date:
10 July 2018
Accepted Date: 18 July 2018
Please cite this article as: Yoshino T, Yamazaki K, Shinozaki E, Komatsu Y, Nishina T, Baba H, Tsuji A, Tsuji Y, Yamaguchi K, Sugimoto N, Denda T, Muro K, Takayama T, Esaki T, Hamamoto Y, Moriwaki T, Shimada Y, Goto M, Nakayama N, Fujii H, Tanase T, Ohtsu A, Relationship between thymidine kinase 1 expression and trifluridine/tipiracil therapy in refractory metastatic colorectal cancer: A pooled analysis of two randomized clinical trials, Clinical Colorectal Cancer (2018), doi: 10.1016/j.clcc.2018.07.009. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1
ACCEPTED MANUSCRIPT Relationship between thymidine kinase 1 expression and trifluridine/tipiracil therapy in refractory metastatic colorectal cancer: A pooled analysis of two randomized clinical trials
RI PT
Takayuki Yoshinoa,*, Kentaro Yamazakib,*, Eiji Shinozakic, Yoshito Komatsud, Tomohiro Nishinae, Hideo Babaf, Akihito Tsujig,1, Yasushi Tsujih, Kensei
Yamaguchii,2, Naotoshi Sugimotoj, Tadamichi Dendak, Kei Murol, Tetsuji Takayamam,
SC
Taito Esakin, Yasuo Hamamotoo,3, Toshikazu Moriwakip, Yasuhiro Shimadaq,4,
Masahiro Gotor, Norisuke Nakayamas, Hirofumi Fujiit, Takanori Tanaseu and Atsushi
a
M AN U
Ohtsua
Department of Gastroenterology and Gastrointestinal Oncology, National Cancer
Center Hospital East, 6-5-1 Kashiwanoha, Kashiwa, Chiba 277-8577, Japan; b
Division of Gastrointestinal Oncology, Shizuoka Cancer Center, 1007
c
TE D
Shimonagakubo, Nagaizumi-cho, Sunto, Shizuoka 411-8777, Japan; Department of Gastroenterology Cancer, Institute Hospital of Japanese Foundation
for Cancer Research, 3-8-31 Ariake, Koto, Tokyo 135-8550, Japan; Department of Cancer Chemotherapy, Hokkaido University Hospital, Kita 14, Nishi
EP
d
5, Kita-ku, Sapporo, Hokkaido 060-8648, Japan;
AC C
e
Department of Gastrointestinal Medical Oncology, National Hospital Organization
Shikoku Cancer Center, 160 Kou, Minamiumemoto-machi, Matsuyama, Ehime 7910280, Japan; f
Department of Gastroenterological Surgery, Graduate School of Medical Sciences,
Kumamoto University, 1-1-1 Honjo, Kumamoto, Kumamoto 860-8556, Japan; g
Department of Clinical Oncology, Kochi Health Sciences Center, 2125-1 Ike, Kochi,
Kochi 781-8555, Japan; and Department of Clinical Oncology, Kobe City Medical
2
ACCEPTED MANUSCRIPT Center General Hospital, 2-1-1 Minatojima, Minamimachi, Chuo, Kobe, Hyōgo 6500047 Japan; h
Department of Medical Oncology, Tonan Hospital, Kita 4, Nishi 7-3-8, Chuo-ku,
Sapporo, Hokkaido, Japan; Department of Gastroenterology, Saitama Cancer Center, 780 Komuro, Inamachi,
RI PT
i
Kitaadachi District, Saitama 362-0806, Japan; j
Department of Medical Oncology, Osaka International Cancer Institute, 3-1-69
Division of Gastroenterology, Chiba Cancer Center, 666-2 Nitona-cho, Chuo-ku
Chiba, Chiba 260-8717, Japan; l
M AN U
k
SC
Otemae, Chuo-ku, Osaka, Osaka 541-8567, Japan;
Department of Clinical Oncology, Aichi Cancer Center Hospital, 1-1 Kanokoden,
Chikusa-ku, Nagoya, Aichi 464-8681, Japan; m
Department of Gastroenterology and Oncology, University of Tokushima, 2-50-1
n
TE D
Kuramoto-cho, Tokushima, Tokushima 770-8503, Japan;
Department of Gastrointestinal and Medical Oncology, National Hospital
Organization Kyushu Cancer Center, 3-1-1 Notame, Minami-ku, Fukuoka, Fukuoka
o
EP
811-1395, Japan;
AC C
Department of Chemotherapy, Tochigi Cancer Center, 4-9-13 Younan, Utsunomiya-
shi, Tochigi 320-0834, Japan; and Department of Internal Medicine, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan; p
Division of Gastroenterology, University of Tsukuba, 2-1-1 Amakubo, Tsukuba,
Ibaraki 305-8576, Japan; q
Gastrointestinal Oncology Division, National Cancer Center Hospital, 5-1-1, Tsukiji,
Chuo-ku, Tokyo 104-0045, Japan; r
Cancer Chemotherapy Center, Osaka Medical College Hospital, 2-7 Daigakumachi,
3
ACCEPTED MANUSCRIPT Takatsuki, Osaka 569-8686, Japan; s
Department of Gastroenterology, Kanagawa Cancer Center, 2-3-2 Nakao, Asahi-ku,
Yokohama, Kanagawa 241-8515, Japan; t
Department of Clinical Oncology, Jichi Medical University Hospital, 3311-1
u
RI PT
Yakushiji, Shimotsuke, Tochigi 329-0498, Japan;
Taiho Pharmaceutical Co., Ltd., 1-2-4 Uchikanda, Chiyoda, Tokyo 101-0047, Japan
1
SC
*These authors contributed equally to this work
Present address: Department of Clinical Oncology, Kagawa University Faculty of
2
M AN U
Medicine, 1750-1 Ikenobe, Miki-cho, Kita, Kagawa 761-0793, Japan
Present address: Department of Gastroenterology Cancer, Institute Hospital of
Japanese Foundation for Cancer Research, 3-8-31 Ariake, Koto, Tokyo 135-8550, Japan;
Present address: Department of Internal Medicine, Keio University School of
TE D
3
Medicine, 35 Shinanomachi, Shinjuku-ku 160-8582, Tokyo, Japan; 4
Present address: Department of Clinical Oncology, Kochi Health Sciences Center,
AC C
EP
2125-1 Ike, Kochi, Kochi 781-8555, Japan.
Corresponding author Professor Takayuki Yoshino, Department of Gastroenterology and Gastrointestinal Oncology, National Cancer Center Hospital East, 6-5-1 Kashiwanoha, Kashiwa-shi, Chiba 277-8577, Japan Tel/Fax: +81-4-7134-6920 Email:
[email protected]
4
ACCEPTED MANUSCRIPT
Email addresses of all authors: T. Yoshino
[email protected]
[email protected]
T. Nishina
[email protected]
H. Baba
[email protected]
A. Tsuji
[email protected]
Y. Tsuji
[email protected]
M AN U
Y. Komatsu
K. Yamaguchi
[email protected]
N. Sugimoto
[email protected] [email protected]
K. Muro
[email protected]
TE D
T. Denda
T. Takayama
[email protected] T. Esaki
[email protected]
EP
Y. Hamamoto
[email protected]
[email protected]
AC C
T. Moriwaki
SC
E. Shinozaki
[email protected]
RI PT
K. Yamazaki
[email protected]
Y. Shimada
[email protected]
M. Goto
[email protected]
N. Nakayama
[email protected] H. Fujii
[email protected]
T. Tanase
[email protected]
A. Ohtsu
[email protected]
5
ACCEPTED MANUSCRIPT Short title: TK1 expression and FTD/TPI therapy in mCRC
Word count, abstract: 225 (limit: 250) Word count, text: 2696 (limit: 8000)
RI PT
Number of tables/figures: 1/4 (limit: 7) Supplemental tables/figures: 3/4 Trial registration:
SC
Phase II (J003) JapicCTI: 090880
AC C
EP
TE D
Phase III PGx JapicCTI: 121918
M AN U
Phase III (RECOURSE) ClinicalTrials.gov: NCT01607957
6
ACCEPTED MANUSCRIPT CONFLICTS OF INTEREST T. Yoshino has received grants from GlaxoSmithKline K. K. and Boehringer Ingelheim GmbH. K. Yamazaki has received personal fees and non-financial support from Taiho Pharmaceutical; and personal fees from Chugai Pharma, Takeda
RI PT
Pharmaceutical, Yakult Honsha, Eli Lilly, Sanofi, Merck Serono, and Bayer. E.
Shinozaki has received personal fees from Taiho Pharmaceutical, Chugai Pharma,
Merck Serono, Takeda Pharmaceutical, Eli Lilly, Yakult Honsha, Ono, and Novartis.
SC
T. Nishina has received personal fees from Taiho Pharmaceutical. H. Baba has
received grants from Taiho Pharmaceutical. A. Tsuji has received personal fees from
M AN U
Daiichi Sankyo, Taiho Pharmaceutical, Chugai Pharma, Merck Serono, Takeda Pharmaceutical, and Bristol-Myers Squibb Japan. Y. Tsuji has received grants from Merck Serono, Eli Lilly Japan, Chugai Pharma, Taiho Pharmaceutical, Ono, Takeda Pharmaceutical, Daiichi Sankyo, Kyowa Kirin, Yakult Honsha, Nippon Kayaku, and
TE D
Medicon. K. Yamaguchi has received personal fees and grants from Taiho Pharmaceutical; grants from Eli Lilly, Ono, MSD, Daiichi Sankyo, Merck Serono, Dainippon Sumitomo, Boehringer Ingelheim, and Yakult Honsha; and personal fees
EP
from Eli Lilly, Merck Serono, Takeda Pharmaceutical, Chugai Pharma, and Yakult
AC C
Honsha. T. Denda has received grants from Sanofi, MSD, and Boehringer Ingelheim. K. Muro has received grants from Ono, MSD, Daiichi Sankyo, Kyowa Hakko Kirin, Shionogi Pharmaceutical, and Gilead Sciences; and personal fees from Chugai Pharma, Taiho Pharmaceutical, Takeda Pharmaceutical, Merck Serono, Eli Lilly, and Yakult Honsha. T. Takayama has received a grant from Taiho Pharmaceutical. T. Esaki has received grants and personal fees from Eli Lilly, Taiho Pharmaceutical, Ono, and Merck Serono; grants from Novartis, Daiichi Sankyo, DS Pharma, AstraZeneca, Boehringer Ingelheim and MSD; and personal fees from Chugai Pharma, Nihon
7
ACCEPTED MANUSCRIPT Kayaku, and Eisai. Y. Hamamoto has received personal fees from Taiho Pharmaceutical. T. Moriwaki has received grants and honoraria from Taiho Pharmaceutical, Takeda Pharmaceutical, Yakult Honsha, Sanofi-Aventis, and Chugai Pharma; grants from Boehringer Ingelheim and MSD Oncology; and honoraria from
RI PT
Bayer, Merck Serono, and Novelpharma. Y. Shimada has received grants and personal fees from Taiho Pharmaceutical, Eli Lilly, and Merck Serono; grants from MSD; and personal fees from Chugai Pharma, Ono, Novartis, Bayer Yakuhin, Yakult Honsha,
SC
Daiichi Sankyo, and Takeda Pharmaceutical. M. Goto has received grants, personal fees and non-financial support from Taiho Pharmaceutical, Yakult Honsha, Chugai
M AN U
Pharma, Ono, Takeda Pharmaceutical, Kyowa Hakko Kirin, Nippon Kayuku, Sumitomo Dainippon Pharma, and Novartis Pharmaceuticals. T. Tanase is employed by Taiho Pharmaceutical and owns Taiho Pharmaceutical stock. A. Ohtsu has received grants from Bristol-Myers Squibb. Y. Komatsu, N. Sugimoto, N. Nakayama, and H.
AC C
EP
TE D
Fujii have declared no conflicts of interest.
8
ACCEPTED MANUSCRIPT MICROABSTRACT No predictive biomarker to indicate the clinical benefit of trifluridine/tipiracil (FTD/TPI) has been identified. Individual patient data from 329 patients from two randomized placebo-controlled trials were analyzed to determine the relationship
RI PT
between thymidine kinase 1 (TK1) expression and FTD/TPI efficacy in refractory metastatic colorectal cancer. Patients with high TK1 expression showed an
AC C
EP
TE D
M AN U
SC
improvement in overall survival when treated with FTD/TPI.
9
ACCEPTED MANUSCRIPT ABSTRACT Background: High thymidine kinase 1 (TK1) activity increases the incorporation of trifluridine (FTD) into DNA; thus, FTD antitumor activity is likely to increase in patients with high tumoral TK1 activity. To date, no established predictive biomarker
RI PT
to indicate the clinical benefit of FTD/tipiracil (TPI) has been identified. We aimed to determine the relationship between TK1 expression and FTD/TPI efficacy in refractory metastatic colorectal cancer.
SC
Materials and Methods: Individual patient data from two randomized placebo-
controlled trials were analyzed. We measured TK1 protein expression in tumor tissue
M AN U
samples and its relationship with FTD/TPI clinical efficacy using overall survival (OS), progression-free survival (PFS), and disease control rate (DCR). Results: This study comprised 329 patients (FTD/TPI, 224; placebo, 105). FTD/TPI significantly improved OS versus placebo in the high-expression (cut-off ≥15%) TK1
TE D
group (median OS, 7.8 versus 6.8 months; hazard ratio [HR]=0.65; 95% confidence interval [CI] 0.46–0.93; P=0.018). The low-expression (cut-off <15%) TK1 group experienced a smaller OS benefit (9.3 versus 7.4 months; HR=0.88; 95% CI 0.63–
EP
1.23; P=0.45). For patients who received placebo, the high-expression TK1 group had
AC C
a slightly worse prognosis than the low-expression TK1 group. The tendency of FTD/TPI efficacy concerning PFS and DCR was not similar to that concerning OS between groups.
Conclusion: Patients with high TK1 expression showed an improvement in OS when treated with FTD/TPI. Further investigations are warranted to confirm this relationship.
KEYWORDS: biomarker, disease control rate, next-generation sequencing, overall
10
ACCEPTED MANUSCRIPT survival, progression-free survival
ABBREVIATIONS 5-FU - 5-fluorouracil
RI PT
CI - confidence interval DCR - disease control rate
ECOG PS - Eastern Cooperative Oncology Group performance status
FOLFIRI - leucovorin with irinotecan
M AN U
FOLFOX - leucovorin with oxaliplatin
SC
FFPE - formalin-fixed paraffin-embedded
FOLFOXIRI - leucovorin with oxaliplatin and irinotecan FTD/TPI - trifluridine/tipiracil HR - hazard ratio
TE D
ITT - intent-to-treat m - median
mCRC - metastatic colorectal cancer
EP
MSI - microsatellite instability
AC C
MSI-H - microsatellite instability-high NGS - next-generation sequencing OS - overall survival
PCR - polymerase chain reaction PFS - progression-free survival TK1 - thymidine kinase 1
11
ACCEPTED MANUSCRIPT INTRODUCTION Metastatic colorectal cancer (mCRC) treatment has evolved significantly in the past two decades. Building on the backbone of chemotherapy with 5-fluorouracil (5FU)/leucovorin with oxaliplatin (FOLFOX), irinotecan (FOLFIRI), or both
RI PT
oxaliplatin and irinotecan (FOLFOXIRI), targeted agents have shown clinical benefit as first-line and subsequent therapies, including angiogenesis inhibitors
(bevacizumab, aflibercept, and ramucirumab) and anti-epidermal growth factor
SC
receptor antibodies for wild-type RAS (cetuximab and panitumumab). The multityrosine kinase inhibitor regorafenib showed clinical benefit as a third-line and
M AN U
subsequent therapy.1–9 More recently, the anti-programmed cell death 1 monoclonal antibodies pembrolizumab and nivolumab were approved by the US Food and Drug Administration for the treatment of patients with unresectable or metastatic microsatellite instability-high (MSI-H)/mismatch repair-deficient CRC. Although
TE D
considerable efforts and resources have been dedicated to finding novel predictive biomarkers against mCRC, extended RAS/BRAF and MSI are the only validated biomarkers to date.10–13
EP
The clinical benefits of trifluridine/tipiracil (FTD/TPI) include significant
AC C
improvements in overall survival (OS) and progression-free survival (PFS) as confirmed in randomized, placebo-controlled studies in mCRC patients in Japan (J003 study), multiple nations (Japan, US, Australia, and EU countries; RECOURSE study), and Asia (TERRA study).14–16 Thus, FTD/TPI represents a standard treatment option as third-line and subsequent therapies for mCRC patients. FTD/TPI is a combination drug comprising the antineoplastic thymidine-based nucleoside analog FTD plus the thymidine phosphorylase inhibitor TPI at a molar ratio of 1:0.5 (weight ratio, 1:0.471).17,18 FTD is incorporated into DNA after
12
ACCEPTED MANUSCRIPT phosphorylation by thymidine kinase 1 (TK1), causing DNA dysfunction.18,19 The mechanism of action of FTD/TPI is distinct from that of 5-FU, an analog of uracil and 5-fluoro-2'-deoxyuridine, which inhibits thymidylate synthase.20,21 Thymidine phosphorylase degrades nucleosides such as thymidine. TPI inhibits thymidine
living organisms has been confirmed.18,20
RI PT
phosphorylase activity strongly and specifically, and inhibition of FTD degradation in
TK1 has a major role in the mechanism of action of FTD/TPI owing to FTD
SC
phosphorylation. High TK1 activity increases the incorporation of FTD into DNA;
thus, FTD antitumor activity is likely to increase in patients with high tumoral TK1
M AN U
activity.22 To date, no established predictive biomarker to indicate the clinical benefit of FTD/TPI has been identified. The primary objective of the present study was to retrospectively determine the level of TK1 expression in tumor tissue samples and any relationship with FTD/TPI clinical efficacy over placebo using individual patient
TE D
data from two randomized placebo-controlled clinical trials. The exploratory objectives were to investigate genomic and genetic factors predicting responsiveness
EP
to FTD/TPI.
AC C
MATERIALS AND METHODS Trial design and participants Data from two randomized, placebo-controlled trials were analyzed.14,15 In both studies, patients were randomized in a 2:1 ratio to receive either FTD/TPI plus best supportive care or placebo plus best supportive care. In the current study, mCRC patients refractory to standard chemotherapies and treated with either FTD/TPI or placebo were assessed for treatment outcomes in relation to TK1 expression as a prespecified analysis.
13
ACCEPTED MANUSCRIPT Treatments FTD/TPI (35 mg/m2/dose) was orally administered twice daily over 5 days for 2 weeks with 2 days of rest between weeks, followed by a 14-day resting period (one
RI PT
treatment cycle). This treatment cycle was repeated every 4 weeks until disease progression or unacceptable toxicity.
SC
Determination of TK1 expression
Archival formalin-fixed paraffin-embedded (FFPE) tumor tissue specimens collected
M AN U
during surgery or by biopsy were used. Immunohistochemistry was performed in the central laboratory to evaluate the TK1 expression. All evaluations were blinded. Detection was performed using anti-human TK1 rabbit polyclonal antibody (Taiho Pharmaceutical Co., Ltd.) and anti-human TK1 mouse monoclonal antibody (Immuno-Biological Laboratories Co., Ltd.) in phase II and III studies, respectively.
TE D
The TK1 antibodies were validated via multiple tissue microarray samples carried out by independent immunohistochemistry experts. The area stained in tumor cells was
EP
classified according to intensity by 5% intervals and scored as 0, 1+, 2+, or 3+ by independent pathologists. Positively-stained cells were defined as those with a
AC C
staining intensity of 2+ or 3+. Occupancy rates of the areas scored as 2+ and 3+ in tumor cells were calculated (low-scoring cells with 0/1+ expression were excluded) and divided into two groups (high or low TK1 expression) (Supplemental Figure 1).
Relationship between TK1 expression and clinical efficacy The OS, PFS, and disease control rate (DCR) were analyzed to evaluate differences in treatment effect by TK1-defined subgroup (high or low TK1 expression).
14
ACCEPTED MANUSCRIPT Exploratory endpoints For the exploratory endpoints, next-generation sequencing (NGS) and MSI analyses were centrally conducted. Using the QIAamp DNA FFPE Tissue Kit (Qiagen Inc.), genomic DNA was extracted from the FFPE slides of tumor tissue samples containing
RI PT
at least 70% of tumor cells (verified microscopically). Samples were checked for quality using polymerase chain reaction (PCR). The TruSight™ Tumor Panel
(Illumina Inc.) is an enrichment system targeting 174 amplicons within 26 well-
SC
known genes related to various types of cancer (Supplemental Table 1). The
resulting Variant Call Format files generated by MiSeq Reporter (Illumina Inc.) were
M AN U
analyzed using VariantStudio Data Analysis v.2.2. Software (Illumina Inc.). The OS, PFS, and DCR were calculated for each treatment group and each non-synonymous gene mutation subgroup (mutant or wild-type). For MSI status, DNA samples were PCR-amplified using primers fluorescently labeled for the mononucleotide repeat
TE D
markers of five loci (NR21, NR24, BAT25, BAT26, and MONO27), and then subjected to capillary electrophoresis using the MSI Analysis System (Promega).
EP
Statistical methods
AC C
This study was not specifically powered to assess the efficacy of FTD/TPI in any TK1-defined subgroups, and no sample size calculations were performed for our analyses. Patients whose TK1 expression was assessed were included in the current study. The analysis set for OS and PFS comprised the TK1-defined population. The analysis set for DCR comprised patients who were evaluable for both TK1 expression and tumor response. For OS and PFS, hazard ratios (HR) for study treatment were estimated using multivariate Cox proportional hazards modeling, with the covariates of TK1 protein expression level and the other baseline factors selected using a
15
ACCEPTED MANUSCRIPT stepwise method. Interactions between study treatment and TK1 protein expression were assessed in the same manner. The DCR odds ratio was estimated using a multivariate logistic regression model. NGS to identify genetic mutations was conducted only on the evaluable population of the phase III study.
RI PT
Sensitivity analysis was conducted to determine whether differences in
characteristics between the study populations influenced the OS and PFS. The HR of treatment efficacy was estimated by Cox proportional hazard modeling with the
SC
following five factors used as covariates: J-003 or RECOURSE study, number of
previous treatments, previous treatment with bevacizumab, previous treatment with
M AN U
anti-EGFR antibody, and presence of KRAS mutation.
To validate the predictive role of TK1 in FTD/TPI treatment in an explorative manner, the datasets of the TK1-defined population from the J-003 and RECOURSE studies were randomly divided into two equal datasets: a training cohort (n=166) and
TE D
a validation cohort (n=163). The HRs for OS and PFS in each TK1-defined subgroup (high/low) were estimated in each cohort. Simulations were also performed in the same manner.
EP
A two-sided P-value was calculated for each statistical test. No adjustments
AC C
were made for multiple comparisons. All statistical analyses were performed using SAS Version 9.2 (SAS Institute Inc.).
Ethical considerations
The studies were carried out in accordance with Good Clinical Practice guidelines and the Declaration of Helsinki and were approved by the institutional review boards of participating institutions. All patients provided written informed consent for the optional pharmacogenetics studies, in addition to that provided for the main phase II
16
ACCEPTED MANUSCRIPT and phase III studies.
RESULTS TK1 expression was evaluable in 329 patients (Supplemental Figure 2). The phase II
RI PT
and III studies comprised 150 and 179 participants, respectively. Overall, 224 patients received FTD/TPI and 105 received placebo. Gene mutation analysis by NGS was
evaluable in 159 patients (109, FTD/TPI; 50, placebo), all of whom were participants
SC
in the phase III study. Table 1 shows the baseline characteristics of participants in the FTD/TPI and placebo groups in the combined intent-to-treat (ITT) population, TK1-
M AN U
defined population, and gene mutation analysis evaluable population. The characteristics in the FTD/TPI and placebo groups in the TK1-defined and gene mutation analysis evaluable populations were generally balanced. Baseline patient characteristics were also generally similar among the three populations. The HRs of
TE D
OS and PFS were similar between the combined ITT and TK1-defined populations, although those in the gene mutation analysis evaluable population were slightly different. In the TK1-defined populations, the OS and PFS significantly improved in
EP
the FTD/TPI group compared with the placebo group (Figure 1A and B). The results
AC C
of the sensitivity analysis showed that there was almost no difference in HR values for OS and PFS before adjusting compared with after adjusting for patient characteristics (data not shown). In the immunohistochemistry studies, the mean and median percentages of
positively stained cells were 15.8% and 10%, respectively. Given that the median to mean TK1 expression ranged from 10% to 15.8%, 15% was chosen as the cut-off point for high or low TK1 expression (high: ≥15%; 160 patients, low: <15%; 169 patients) (Supplemental Table 2). There was no bias in patients’ characteristics
17
ACCEPTED MANUSCRIPT between the FTD/TPI and placebo groups as a result of dividing the patient populations based on TK1 expression levels.
Predictive utility of TK1
RI PT
As shown in Figure 2A, at the 15% cut-off, FTD/TPI significantly improved OS compared with placebo (median OS of FTD/TPI versus placebo, 7.8 versus 6.8
months; HR=0.65; 95% confidence interval [CI] 0.46–0.93; P=0.018) in the high
SC
TK1 expression group, and showed a greater OS benefit than the low TK1 expression group (9.3 versus 7.4 months; HR=0.88; 95% CI 0.63–1.23; P=0.45).
M AN U
A significant improvement in PFS was shown with FTD/TPI compared with placebo in both the high TK1 expression (median PFS of FTD/TPI versus placebo, 2.1 versus 1.1 months; HR=0.37, 95% CI 0.26–0.54; P<0.0001) and low TK1 expression (3.5 versus 1.7 months; HR=0.39, 95% CI 0.28–0.55, P<0.0001) groups
TE D
(Figure 2B). The PFS and DCR benefit was similar between the high and low TK1 expression groups, contrary to the findings for OS. Figure 3A and B show median time and HR for OS and PFS according to TK1-defined subgroups. Regarding the
EP
potential utility of TK1 as a biomarker for FTD/TPI efficacy, while OS was improved
AC C
in patients with high TK1 expression receiving FTD/TPI in every subgroup, a different trend was observed in terms of PFS, with patients with high TK1 expression performing less well. The DCR of FTD/TPI was significantly higher than that of placebo, regardless of TK1 expression at the 15% cut-off point (Supplemental Figure 3). A validation analysis, in which the patients were divided into a training cohort and a validation cohort, showed similar results (Supplemental Table 3).
Prognostic utility of TK1
18
ACCEPTED MANUSCRIPT Patients with high TK1 expression who received placebo had a slightly poorer prognosis than those with low TK1 expression in terms of OS by multivariate analysis (Figure 4). The numbers of patients with RAS and BRAF mutations, some of the
(data not shown).
Clinical significance of NGS profile
RI PT
generally known prognostic factors, were well-balanced between treatment groups
SC
Evaluable patients for MSI and NGS were analyzed for the exploratory objectives.
Ten gene mutations were identified overall (Supplemental Figure 4). Of these, five
M AN U
occurred at a frequency greater than 10% in the FTD/TPI group [TP53, APC, RAS (KRAS/NRAS), PI3CA, and FBXW7]. Although a trend toward better OS was observed in patients with the RAS wild-type gene, there was no clear relationship between gene status and clinical efficacy (data not shown). Gene mutations [TP53,
TE D
APC, RAS (KRAS/NRAS), PI3CA, and FBXW7] were added as covariates in a multivariate analysis to evaluate prognosis; however, all gene mutations were excluded from the model by stepwise selection. MSI analysis was conducted in 155
EP
DNA samples from the PGx study in the phase III study. However, the only patient
AC C
with MSI-H was in the placebo group, so a relationship with efficacy was not analyzed.
DISCUSSION
The purpose of this study was to determine whether there was any relationship between TK1 expression and FTD/TPI clinical efficacy in mCRC patients using data from two clinical trials. OS data in the patients receiving placebo indicated that high TK1 expression may be a marker of poor prognosis. Although similar findings have
19
ACCEPTED MANUSCRIPT previously been reported in several cancers, there are no reports in CRC.23–25 Regarding the potential clinical utility of TK1 as a biomarker for FTD/TPI efficacy, a greater OS benefit was observed in patients with high TK1 expression receiving FTD/TPI at the 15% cut-off point, while PFS and DCR benefits in patients receiving
RI PT
FTD/TPI were similar between high and low TK1 expression subgroups. The results of sensitivity analysis showed that the effect of differences in patient characteristics between the study populations of two randomized clinical trials on the OS and PFS
SC
was small (data not shown), and thus, do not explain the differences in OS and PFS benefits between the patients in the present studies. The results from the validation
M AN U
analysis supported the results of the multivariate analysis at the 15% cut-off point. A potential reason why the PFS and DCR results differed from the OS results is that tumors with low TK1 expression are more indolent than tumors with high TK1 expression. As the incorporation of FTD/TPI in tumors with low TK1 expression is
TE D
possibly lower, the pharmaceutical effect of FTD/TPI might also be lower; however, tumors with low TK1 expression might present slow growth because of their indolence. Therefore, tumors with low TK1 expression might show a similar PFS and
EP
DCR benefit compared with tumors with high TK1 expression.
AC C
Regarding mutation analysis, no particular tendency was observed for the distribution of genetic mutations in the FTD/TPI and placebo groups (Supplemental Figure 4). The distribution of mutations between high and low TK1 expression subgroups did not show a particular tendency. No differences were observed in the ratio or proportion of mutations between the FTD/TPI and placebo groups (data not shown). The exploratory analyses in this study did not identify any specific gene correlating with the clinical efficacy of FTD/TPI. Previous studies of FTD/TPI with mCRC observed that neutropenia, after
20
ACCEPTED MANUSCRIPT initiating FTD/TPI treatment, was associated with better prognosis in refractory mCRC patients26,27; thus, neutropenia might be a surrogate marker for adequate FTD/TPI dosing. As it is currently not possible to predict the efficacy of FTD/TPI before the administration of treatment, further study of biomarkers such as TK1 is
RI PT
warranted.
Limitations and generalizability
SC
The study is limited by its relatively small sample size due to TK1-defined subgroup analyses, and the fact that it was not prospectively designed to evaluate any
M AN U
relationship between TK1 expression and efficacy. Furthermore, it was unknown whether the tissue samples used to measure TK1 expression were collected from the primary tumor or from metastases in the pooled studies; this may have caused potential bias. In the J-003 study, assessment of MSI could not be conducted as no
TE D
tissue samples were collected. Finally, the study comprised only Japanese patients, thus limiting ethnic generalizability. Larger studies in patient populations comprising
EP
various ethnicities would, therefore, be of value.
CONCLUSION
AC C
Overall, the findings indicate that high TK1 expression could be a marker of poor prognosis in terms of OS in refractory mCRC patients. Based on OS data comparing FTD/TPI with placebo, high TK1 expression in patients treated with FTD/TPI might be suggestive of better outcomes. However, further investigations are needed to clarify whether there is any relationship between TK1 expression and FTD/TPI efficacy.
22
ACCEPTED MANUSCRIPT CLINICAL PRACTICE POINTS Several studies have shown that trifluridine/tipiracil (FTD/TPI) provides clinical benefits in terms of significant improvements in overall survival (OS) and progression-free survival in metastatic colorectal cancer (mCRC) patients. However,
RI PT
to date, no established predictive biomarker has been identified that can indicate the clinical benefit of FTD/TPI. Thymidine kinase 1 (TK1) plays a major role in the
mechanism of action of FTD/TPI owing to FTD phosphorylation. High TK1 activity
SC
increases the incorporation of FTD into DNA; thus, FTD antitumor activity is likely
to increase in patients with high tumoral TK1 activity. Our findings showed that high
M AN U
TK1 expression could be a marker of poor prognosis in terms of OS in refractory mCRC patients and that FTD/TPI treatment may be more effective in improving OS among patients with high TK1 expression versus those with low TK1 expression. The assessment of TK1 expression as a predictive biomarker of the clinical benefit of
TE D
FTD/TPI may be useful in clinical practice to determine which colorectal cancer
AC C
EP
patients are more likely to benefit from FTD/TPI treatment.
23
ACCEPTED MANUSCRIPT ACKNOWLEDGMENTS This investigation was supported by Taiho Pharmaceutical. The authors wish to thank the participating patients and their families; all the investigators, site staff, and operations staff who participated in the study; Toshihiko Doi, M.D., at the National
RI PT
Cancer Hospital East, Japan, for his contribution to the development of the studies; Katsuya Tsuchihara, M.D., at the National Cancer Hospital East, Japan, for his contribution to data interpretation; Professor Chikuma Hamada, at the Tokyo
SC
University of Science, Japan, for his contribution to the advice of statistical analysis; Hikari Chiba, Helen Roberton, and Dr Sarah Williams of Edanz Medical Writing for
M AN U
providing medical writing services, which were supported by Taiho Pharmaceutical.
FUNDING
This work was supported by Taiho Pharmaceutical, Tokyo, Japan. The funding source
TE D
participated in the study design; in the collection, analysis, and interpretation of the data; in the writing of the report; and in the decision to submit the article for
AC C
EP
publication.
24
ACCEPTED MANUSCRIPT REFERENCES 1. Saltz L. Colorectal cancer treatment: what’s next? (or: Is there life after EGFR and VEGF?). Gastrointest Cancer Res. 2008;2:S20–22.
RI PT
2. Van Cutsem E, Tabernero J, Lakomy R, et al. Addition of aflibercept to fluorouracil, leucovorin, and irinotecan improves survival in a phase III randomized trial in patients with metastatic colorectal cancer previously treated with an
SC
oxaliplatin-based regimen. J Clin Oncol. 2012;30:3499–3506.
3. Tabernero J. Ramucirumab versus placebo in combination with second-line
M AN U
FOLFIRI in patients with metastatic colorectal carcinoma that progressed during or after first-line therapy with bevacizumab, oxaliplatin, and a fluoropyrimidine (RAISE): a randomised, double-blind, multicentre, phase 3 study. Lancet Oncol. 2015;16:499–508.
TE D
4. Van Cutsem E, Oliveira J. Advanced colorectal cancer: ESMO clinical recommendations for diagnosis, treatment and follow-up. Ann Oncol. 2009;20:61–63.
EP
5. Douillard JY, Siena S, Cassidy J, et al. Randomized, phase III trial of panitumumab with infusional fluorouracil, leucovorin, and oxaliplatin (FOLFOX4) versus
AC C
FOLFOX4 alone as first-line treatment in patients with previously untreated metastatic colorectal cancer: the PRIME study. J Clin Oncol. 2010;28:4697–4705. 6. Grothey A, Van Cutsem E, Sobrero A, et al. Regorafenib monotherapy for previously treated metastatic colorectal cancer (CORRECT): an international, multicentre, randomised, placebo-controlled, phase 3 trial. Lancet. 2013;381:303–312. 7. Venook AP, Niedzwiecki D, Lenz H-J, et al. CALGB/SWOG 80405: Phase III trial of irinotecan/5-FU/leucovorin (FOLFIRI) or oxaliplatin/5-FU/leucovorin
25
ACCEPTED MANUSCRIPT (mFOLFOX6) with bevacizumab (BV) or cetuximab (CET) for patients (pts) with KRAS wild-type (wt) untreated metastatic adenocarcinoma of the colon or rectum (MCRC). J Clin Oncol. 2014;32:5s (suppl; abstr LBA3).
RI PT
8. Saltz LB, Clarke S, Díaz-Rubio E, et al. Bevacizumab in combination with oxaliplatin-based chemotherapy as first-line therapy in metastatic colorectal cancer: a randomized phase III study. J Clin Oncol. 2008;26:2013–2019.
SC
9. Van Cutsem E, Köhne CH, Hitre E, et al. Cetuximab and chemotherapy as initial
M AN U
treatment for metastatic colorectal cancer. N Engl J Med. 2009;360:1408–1417. 10. Douillard JY, Oliner KS, Siena S, et al. Panitumumab-FOLFOX4 treatment and RAS mutations in colorectal cancer. N Engl J Med. 2013;369:1023–1034. 11. Van Custem, Lenz HJ, Köhne CH, et al. Fluorouracil, leucovorin, and irinotecan
2015;33:692–700.
TE D
plus cetuximab treatment and RAS mutations in colorectal cancer. J Clin Oncol.
12. Le DT, Uram JN, Wang H, et al. PD-1 blockade in tumors with mismatch-repair
EP
deficiency. N Engl J Med. 2015;372:2509–2520.
AC C
13. Overman MJ, McDermott R, Leach JL, et al. Nivolumab in patients with metastatic DNA mismatch repair-deficient or microsatellite instability-high colorectal cancer (CheckMate 142): an open-label, multicentre, phase 2 study. Lancet Oncol. 2017;18:1182–1191.
14. Yoshino T, Mizunuma N, Yamazaki K, et al. TAS-102 monotherapy for pretreated metastatic colorectal cancer: a double-blind, randomised, placebocontrolled phase 2 trial. Lancet Oncol. 2012;13:993–1001.
26
ACCEPTED MANUSCRIPT 15. Mayer RJ, Van Cutsem E, Falcone A, et al. Randomized trial of TAS-102 for refractory metastatic colorectal cancer. N Engl J Med. 2015;372:1909–1919. 16. Xu J, Kim TW, Shen L, et al. Results of a randomized, double-blind, placebo-
RI PT
controlled, phase III trial of trifluridine/tipiracil (TAS-102) monotherapy in Asian patients with previously treated metastatic colorectal cancer: the TERRA study. J Clin Oncol. 2018;36:350–358.
SC
17. Emura T, Suzuki N, Fujioka A, Ohshimo H, Fukushima M. Potentiation of the
antitumor activity of alpha, alpha, alpha-trifluorothymidine by the co-administration
Oncol. 2005;27:449–455.
M AN U
of an inhibitor of thymidine phosphorylase at a suitable molar ratio in vivo. Int J
18. Temmink OH, Emura T, de Bruin M, Fukushima M, Peters GJ. Therapeutic potential of the dual-targeted TAS-102 formulation in the treatment of gastrointestinal
TE D
malignancies. Cancer Sci. 2007;98:779–789. 19. Tanaka N, Sakamoto K, Okabe H, et al. Repeated oral dosing of TAS-102 confers
EP
high trifluridine incorporation into DNA and sustained antitumor activity in mouse
AC C
models. Oncol Rep. 2014;32:2319–2326. 20. Emura T, Suzuki N, Yamaguchi M, Ohshimo H, Fukushima M. A novel combination antimetabolite, TAS-102, exhibits antitumor activity in FU-resistant human cancer cells through a mechanism involving FTD incorporation in DNA. Int J Oncol. 2004;25:571–578. 21. Emura T, Nakagawa F, Fujioka A, et al. An optimal dosing schedule for a novel combination antimetabolite, TAS-102, based on its intracellular metabolism and its incorporation into DNA. Int J Mol Med. 2004;13:249–255.
27
ACCEPTED MANUSCRIPT 22. Shintani M, Urano M, Takakuwa Y, Kuroda M, Kamoshida S. Immunohistochemical characterization of pyrimidine synthetic enzymes, thymidine kinase-1 and thymidylate synthase, in various types of cancer. Oncol Rep.
RI PT
2010;23:1345–1350. 23. Konoplev SN, Fritsche HA, O’Brien S, et al. High serum thymidine kinase 1 level predicts poorer survival in patients with chronic lymphocytic leukemia. Am J Clin
SC
Pathol. 2010;134:472–477.
24. Romain S, Spyratos F, Descotes F, et al. Prognostic of DNA-synthesizing enzyme
M AN U
activities (thymidine kinase and thymidylate synthase) in 908 T1–T2, N0–N1, M0 breast cancers: a retrospective multicenter study. Int J Cancer. 2000;87:860–868. 25. Nisman B, Yutkin V, Nechushtan H, et al. Circulating tumor M2 pyruvate kinase and thymidine kinase 1 are potential predictors for disease recurrence in renal cell
TE D
carcinoma after nephrectomy. Urology. 2010;76:513.e1–6. 26. Kasi PM, Kotani D, Cecchini M, et al. Chemotherapy induced neutropenia at 1-
EP
month mark is a predictor of overall survival in patients receiving TAS-102 for
AC C
refractory metastatic colorectal cancer: a cohort study. BMC Cancer. 2016;16:467. 27. Hamauchi S, Yamazaki K, Masuishi T, et al. Neutropenia as a predictive factor in metastatic colorectal cancer treated with TAS-102. Clin Colorectal Cancer. 2016;16:51–57.
ACCEPTED MANUSCRIPT
28
Table
RI PT
Table 1. Patient baseline characteristics and efficacy summary for each analysis population
ECOG PS
Number of
FTD/TPI
(N=646)
(N=323)
(N=969)
(N=224)
Male
390 (60)
193 (60)
583 (60)
Female
256 (40)
130 (40)
386 (40)
<65 y
360 (56)
182 (56)
542 (56)
≥65 y
286 (44)
141 (44)
Colon
401 (62)
197 (61)
Rectum
245 (38)
126 (39)
0
373 (58)
182 (56)
1
270 (42)
2
3 (<1)
<3
392 (61)
population
Placebo
Total
FTD/TPI
Placebo
Total
(N=105)
(N=329)
(N=109)
(N=50)
(N=159)
134 (60)
60 (57)
194 (59)
64 (59)
32 (64)
96 (60)
90 (40)
45 (43)
135 (41)
45 (41)
18 (36)
63 (40)
115 (51)
59 (56)
174 (53)
56 (51)
29 (58)
85 (53)
M AN U
Total
TE D
Primary lesion
Placebo
427 (44)
109 (49)
46 (44)
155 (47)
53 (49)
21 (42)
74 (47)
598 (62)
127 (57)
58 (55)
185 (56)
59 (54)
23 (46)
82 (52)
371 (38)
97 (43)
47 (45)
144 (44)
50 (46)
27 (54)
77 (48)
555 (57)
158 (71)
72 (69)
230 (70)
81 (74)
38 (76)
119 (75)
140 (43)
410 (42)
64 (29)
32 (30)
96 (29)
28 (26)
12 (24)
40 (25)
1 (<1)
4 (<1)
2 (1)
1 (1)
3 (1)
0 (0)
0 (0)
0 (0)
184 (57)
576 (59)
136 (61)
64 (61)
200 (61)
63 (58)
34 (68)
97 (61)
EP
Age
FTD/TPI
AC C
Sex
Gene mutation analysis evaluable
TK1-defined population
SC
Combined ITT population
ACCEPTED MANUSCRIPT
29
metastatic ≥3
254 (39)
139 (43)
393 (41)
88 (39)
41 (39)
Mutant
323 (50)
163 (51)
486 (50)
102 (46)
55 (52)
Wild-type
320 (50)
159 (49)
479 (50)
121 (54)
50 (48)
3 (<1)
1 (<1)
4 (<1)
1 (<1) 47 (21)
16 (32)
62 (39)
157 (48)
52 (48)
27 (54)
79 (50)
171 (52)
57 (52)
23 (46)
80 (50)
0 (0)
1 (<1)
0 (0)
0 (0)
0 (0)
2
112 (17)
58 (18)
170 (18)
27 (26)
74 (22)
26 (24)
15 (30)
41 (26)
chemotherapy
3
165 (26)
70 (22)
235 (24)
69 (31)
30 (29)
99 (30)
27 (25)
13 (26)
40 (25)
regimens
≥4
369 (57)
195 (60)
564 (58)
108 (48)
48 (46)
156 (47)
56 (51)
22 (44)
78 (49)
0.68 [0.59–0.79]
HR for PFS [95% CI]a
0.45 [0.39–0.52]
a
FTD/TPI compared with placebo.
AC C
Data in the table are presented as n (%).
0.78 [0.61–0.99]
0.86 [0.60–1.25]
0.39 [0.30–0.50]
0.47 [0.33–0.67]
EP
HR for OS [95% CI]a
TE D
Number of
SC
Missing
M AN U
KRAS status
129 (39)
RI PT
46 (42)
tumors
Abbreviations: CI, confidence interval; ECOG PS, Eastern Cooperative Oncology Group performance status; FTD/TPI, trifluridine/tipiracil; HR, hazard ratio; ITT, intent-to-treat; OS, overall survival; PFS, progression-free survival; TK1, thymidine kinase 1
31
ACCEPTED MANUSCRIPT Figure legends
Figure 1. Kaplan–Meier curves for (A) OS and (B) PFS in the TK1-defined population (FTD/TPI vs. placebo). The horizontal axes of the OS curve and PFS
RI PT
curve were truncated at 24 months and 12 months, respectively, because only a few patients were censored over the boundary. Abbreviations: CI, confidence interval; FTD/TPI, trifluridine/tipiracil; HR, hazard ratio; m, median; OS, overall survival;
SC
PFS, progression-free survival; TK1, thymidine kinase 1.
M AN U
Figure 2. Kaplan–Meier curves for (A) OS and (B) PFS stratified by (left) high and (right) low TK1 expression. A 15% cut-off point was used to divide the high and low TK1 expression groups. The horizontal axes of the OS curve and PFS curve were truncated at 24 months and 12 months, respectively, because only a few patients were
TE D
censored over the boundary. Abbreviations: CI, confidence interval; FTD/TPI, trifluridine/tipiracil; HR, hazard ratio; m, median; OS, overall survival; PFS,
EP
progression-free survival; TK1, thymidine kinase 1.
AC C
Figure 3. Forest plot of HRs for (A) overall survival (Reprinted with permission from Yoshino Takayuki et al. J Clin Oncol. 2017:35(4 Suppl);529. © 2017 American Society of Clinical Oncology. All rights reserved.) and (B) progression-free survival in the TK1-defined population. Abbreviations: FTD/TPI, trifluridine/tipiracil; HR, hazard ratio; CI, confidence interval; TK1, thymidine kinase 1.
Figure 4. OS in patients in placebo groups by high and low TK1 expression at the 15% cut-off. The horizontal axis of the OS curve was truncated at 24 months because
32
ACCEPTED MANUSCRIPT only a few patients were censored over the boundary. Abbreviations: CI, confidence
AC C
EP
TE D
M AN U
SC
RI PT
interval; HR, hazard ratio; m, median; OS, overall survival; TK1, thymidine kinase 1.
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT Supplemental Materials
Supplemental Table 1. Genes related to solid tumors Genes related to solid tumors APC
BRAF
CTNNB1
EGFR
ERBB2
FBXW7
FOXL2
GNAQ
GNAS
KIT
MAP2K1
MET
MSH6
NRAS
PIK3CA
PTEN
SMAD4
SRC
AC C
EP
TE D
M AN U
TP53
CDH1
RI PT
ALK
SC
AKT1
FGFR2 KRAS
PDGFRA STK11
ACCEPTED MANUSCRIPT Supplemental Table 2. Distribution of TK1 expression (2+ and 3+) FTD/TPI (%)
Placebo (%)
Total (%)
(N=224)
(N=105)
(N=329)
High ≥5%
196 (87.5)
87 (82.9)
283 (86.0)
Low <5%
28 (12.5)
18 (17.1)
46 (14.0)
High ≥10%
154 (68.8)
71 (67.6)
225 (68.4)
Low <10%
70 (31.3)
34 (32.4)
104 (31.6)
High ≥15%
111 (49.6)
49 (46.7)
160 (48.6)
Low <15%
113 (50.4)
56 (53.3)
169 (51.4)
High ≥20%
71 (31.7)
33 (31.4)
104 (31.6)
153 (68.3)
72 (68.6)
225 (68.4)
47 (21.0)
26 (24.8)
73 (22.2)
177 (79.0)
79 (75.2)
256 (77.8)
38 (17.0)
19 (18.1)
57 (17.3)
86 (81.9)
272 (82.7)
High ≥25% Low <25%
EP
Median
TE D
High ≥30%
Mean
SC
M AN U
Low <20%
Low <30%
RI PT
TK1-defined subgroups
186 (83.0)
15.8% 10.0%
AC C
Abbreviations: TK1, thymidine kinase 1; FTD/TPI, trifluridine/tipiracil
ACCEPTED MANUSCRIPT Supplemental Table 3. Hold-out validation analysis of HRs for OS and PFS at the 15% cut-off point Training cohort Validation cohort HR (95% CI) or (2.5th HR (95% CI) or (2.5th and 97.5th percentiles)* and 97.5th percentiles)* OS Hold-out (n=166) Hold-out (n=163) 0.659 0.578 (0.395–1.100) (0.336–0.995) Simulation Simulation 0.646 0.646 (0.396–0.961) (0.392–0.969) Low <15% Hold-out (n=166) Hold-out (n=163) 0.889 0.784 (0.523–1.514) (0.482–1.274) Simulation Simulation 0.859 0.858 (0.512–1.280) (0.512–1.285) PFS High ≥15% Hold-out (n=166) Hold-out (n=163) 0.433 0.267 (0.257–0.730) (0.135–0.530) Simulation Simulation 0.382 0.384 (0.227–0.569) (0.228–0.569) Low <15% Hold-out (n=166) Hold-out (n=163) 0.387 0.439 (0.219–0.684) (0.259–0.744) Simulation Simulation 0.371 0.370 (0.193–0.587) (0.192–0.575) Abbreviations: HR, hazard ratio; OS, overall survival; PFS, progression-free survival;
EP
TE D
M AN U
SC
RI PT
TK1 expression subgroup High ≥15%
AC C
TK1, thymidine kinase 1.
*HR and 95% CI at the 15% cut-off point were estimated by hold-out analysis. Mean of HR, 2.5th and 97.5th percentiles at the 15% cut-off point were estimated by simulation. The proportion of misspecification in the TK1 cut-off point was 35.07% for OS and 50.49% for PFS between two cohorts.
ACCEPTED MANUSCRIPT Supplemental Figure 1. TK1 protein expression level by IHC. TK1 expression was determined by IHC performed in the central laboratory. Cells categorized as 2+ and 3+ were evaluated (i.e., cells scoring 1+ and 0 were excluded). Examples of cells with different TK1 scoring are shown with arrows of different colors. In this case, the percentage of positive cells is 15%. If the cut-off is 5%–15%, this
RI PT
a
case would be classified into the high TK1 expression group. If the cut-off is 20%, this case would be classified into the low TK1 expression group.
AC C
EP
TE D
M AN U
SC
Abbreviations: IHC, immunohistochemistry; TK1, thymidine kinase 1
ACCEPTED MANUSCRIPT Supplemental Figure 2. Patient disposition a
Patients who could provide tissue samples obtained in biopsy or surgery and who
provided additional consent were enrolled in each study. b
One patient was excluded from the TK1-defined population because of an incomplete
RI PT
c
One patient was excluded because of inadequate tissue samples.
informed consent form and three were excluded because of inadequate tissue samples. d
One patient was excluded from the NGS-evaluable population because of an
SC
incomplete consent form, three patients because their tissue samples were not
obtained from the study site, and 20 patients because of inadequate tissue samples.
M AN U
Abbreviations: FTD/TPI, trifluridine/tipiracil; ITT, intent-to-treat; NGS, next-
AC C
EP
TE D
generation sequencing; PGx, pharmacogenomic; TK1, thymidine kinase 1
ACCEPTED MANUSCRIPT Supplemental Figure 3. Forest plot of odds ratios for disease control rate by TK1 and tumor response-evaluable patients in the TK1-defined population Five patients were excluded for not having target lesions. Abbreviations: CI, confidence interval; FTD/TPI, trifluridine/tipiracil; OR, odds ratio;
AC C
EP
TE D
M AN U
SC
RI PT
TK1, thymidine kinase 1
ACCEPTED MANUSCRIPT Supplemental Figure 4. Mutation frequency in the FTD/TPI and placebo groups Ten genes were identified as having mutations. Genes with >10% mutation frequency in the FTD/TPI group were TP53, APC, RAS, PIK3CA, and FBXW7.
AC C
PIK3CA
FBXW7
CDH1
PT EN
MSH6
4%
1.8 %
3.7 %
2%
10 %
4.6 %
12 %
SMAD4
3.7 %
6%
7.3 %
12.8 %
12.8 %
RAS
EP
APC
8%
TE D
20 10 0 T P53
6%
47.7 %
49.5 %
60 50 30
40
54 %
64 %
M AN U
74.3 %
90
84 %
TAS-102 Placebo
70
80
SC
100
RI PT
Abbreviation: FTD/TPI, trifluridine/tipiracil
BRAF