1 BACKGROUND

Colorectal cancer is one of the most common malignancies, and its incidence has gradually increased. Metastases are already present in 15-25% at the first diagnosis of colorectal cancer. Moreover, the proportion of patients who eventually progress to metastasis is high (50%).1 For several decades, the main treatment for metastatic colorectal cancer (mCRC) has been fluorouracil (5-FU)-based chemotherapy as either the first- or second-line treatment.2 The efficacy of 5-FU can be markedly improved when combined with calcium folinate (CF).3 No other chemotherapeutic drugs were available for patients who experienced treatment failure with 5-FU, oxaliplatin, and irinotecan when the current clinical trial was conducted with raltitrexed and S-1 (RS regimen) in 2015. Although antiepidermal growth factor receptor monoclonal antibody (with or without irinotecan) and the multitarget inhibitor, regorafenib, can be used in mCRC as the third- or fourth-line treatment,4 their use is restricted to approximately 50% of patients, those with the wild-type Ras gene, and they are expensive for some patients in China. Furthermore, regorafenib was not available in China at the time.

In vivo and in vitro studies demonstrated that raltitrexed or S-1 had no complete cross-resistance with 5-FU, and can be used in patients with mCRC who have had failure with 5-FU. Previous studies have reported that in patients who experienced failure with first-line therapy of 5-FU/CF, the overall response rate (ORR) of capecitabine as second-line therapy was 0%, whereas that of raltitrexed was 16%.5 Deoxyuridylic acid is converted to deoxythymidine monophosphate in the cells by thymidylate synthase (TS). Approximately 85% of 5-FU is degraded and inactivated by dihydropyrimidine dehydrogenase (DPD). TS is a key enzyme in the metabolism of folic acid and is a target enzyme of 5-FU. Genetic variation in TS can affect the toxicity and efficacy of 5-FU, and inhibition of TS activity is an approach to overcome 5-FU resistance. Raltitrexed is a new-generation water-soluble TS inhibitor. In vitro studies have demonstrated a synergistic effect when raltitrexed was combined with 5-FU.6 S-1 is a 5-FU derivative, which contains the active component-tegafur (FT207) and two biological modifiers, gimeracil (CDHP) and oteracil. CDHP is a DPD inhibitor that can inhibit the degradation of 5-FU, which is released from FT207 and catalyzed by DPD, and thereby maintains the effective concentration of 5-FU in the blood and tumor tissue for a prolonged duration, thus obtaining a similar effect to that of sustained intravenous infusion of 5-FU. The increase in DPD activity is one of the mechanisms of 5-FU resistance, so the inhibition of DPD activity is a strategy to overcome 5-FU resistance. Theoretically, as S-1 contains DPD inhibitor it may partially overcome the resistance of cancer cells to 5-FU. Thus, we hypothesized that the combination of raltitrexed and S-1, which is another 5-FU derivative, can improve the outcome of patients with mCRC when implemented as the third-line treatment.

Although oral regorafenib and fruquintinib—potent and highly selective vascular endothelial growth factor receptor inhibitors—are available in China and TAS-102—developed in Japan and will soon be available in China—can be used as third-line options in mCRC, their effects are still limited with a median overall survival (mOS) benefit of less than 2.7 months compared with the best suppurative care.7, 8 Moreover, some patients cannot benefit from regorafenib and fruquintinib, either because of the toxicity or high price. Given these reasons, new effective drugs or regimens must be developed to improve the survival of patients without the suitable choice of drugs. In our retrospective study, in 18 patients who had experienced failure with 5-FU, oxaliplatin, and irinotecan were treated with the RS regimen, we found a median progression-free survival (mPFS) and mOS of 2.5 and 7.0 months, respectively. The patients were well tolerant to the RS regimen, and most of the adverse events were graded 1-2.9 Therefore, this prospective study was performed to further evaluate the efficacy and safety of the RS regimen in patients in whom conventional therapy had failed.

2 PATIENTS AND METHODS 2.1 Patient eligibility

This was a single-arm, two-center, prospective phase II trial. The main eligibility criteria were: ≥18 years of age; histologically confirmed adenocarcinoma of the colon or rectum; Eastern Cooperative Oncology Group (ECOG) 0-1; expected survival period ≥3 months; measurable objective tumor lesions according to the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1, ≥1; adequate blood test, hepatic and renal functions; and progression during or within 3 months following the last administration of approved standard therapies, which must include fluoropyrimidine, oxaliplatin, and irinotecan. All patients provided a written informed consent, and the study was approved by the Ethics Committee of the Fudan University Shanghai Cancer Center.

2.2 Treatment

Raltitrexed 3 mg/m2 intravenous infusion was administered once on the first day and was repeated every 3 weeks. The infusion time was about half an hour with antiemetics for premedication, including palonosetron and dexamethasone. S-1 was administered orally twice per day and was administered after breakfast and dinner. S-1 administration was continued for 2 weeks and discontinued for 1 week, and this was repeated every 3 weeks. The daily dose of S-1 was calculated according to the body surface area (BSA) <1.25 m2, 80 mg/day; ≥1.25 to <1.5 m2, 100 mg/day; ≥1.5 m2, 120 mg/day. Treatment was continued until documented progression of disease (PD), intolerable toxicity, or unwillingness to continue treatment.

2.3 Evaluation of efficacy and toxicity

Tumor evaluations were performed by computer tomography or magnetic resonance imaging every two cycles until disease progression or intolerable toxicity. Short-term efficacy measures were complete response (CR), partial response (PR), stable disease (SD), and PD according to the RECIST 1.1 criteria. Toxicity was graded according to the United States National Cancer Institute Common Toxicity Criteria version 4.0. Blood tests were conducted weekly, and side effects were recorded.

2.4 Endpoints and statistical analysis

Overall survival (OS) was defined as the time from inclusion to death due to any cause. Progression-free survival (PFS) was defined as the time from inclusion to PD. The ORR was defined as the rate of CR and PR, while the disease control rate (DCR) was defined as the rate of CR, PR, and SD. The primary endpoint was PFS, whereas the secondary endpoints were ORR, DCR, OS, and toxicity. Patients who provided the informed consent for the RS regimen were included in the intention-to-treat (ITT) population. The per-protocol (PP) population included patients who met the eligibility criteria, completed at least two chemotherapy cycles and one measurement according to the study protocol, while patients who stopped treatment or missed measurement because of adverse events or unwillingness were excluded. All of our statistical analyses were conducted using SPSS 23.0 (SPSS Inc., Chicago, IL) and R software 3.6.1 (https://www.r-project.org/). R packages survival and survminer were used for the survival analysis. The Kaplan-Meier curve and log-rank test were used to compare the PFS and OS between the subgroups. Pearson's chi-square test was used to compare the DCR between the subgroups. Two-sided P-values < .05 were considered statistically significant.

3 RESULTS 3.1 General information

In this study, 105 patients with advanced colorectal cancer between November 2015 and May 2019 in whom therapy with fluorouracil, oxaliplatin, or irinotecan had failed were included. All patients were pathologically confirmed to have colorectal adenocarcinoma and had complete imaging of at least one measurable lesion and laboratory test data. All patients had previously received at least two-line chemotherapy without raltitrexed or S-1 and had no contraindications for chemotherapy. Among them, six patients received the regimen but did not complete the first assessment due to adverse events or unwillingness. The characteristics of the patients are given in Table 1.

TABLE 1. Patient baseline (ITT group, n = 105; PP group, n = 99) ITT group (n = 105) PP group (n = 99) Gender Male: 67 (63.81%) Male: 62 (62.63%) Female: 38 (36.19%) Female: 37 (37.37%) Age Median (25-75%): 60 (51-66) Median (25-75%): 60 (51-66) Line of systemic chemotherapy 3 Line: 80 (76.19%) 3 Line:76 people (76.77%) >3 Line: 25 people (23.81%) >3 Line: 23 people (23.23%) Primary tumor site Right-sided colon: 25 (23.81%) Right-sided colon: 23 (23.23%) Left-sided colon: 29 (27.62%) Left-sided colon†: 27 (27.27%) Rectum: 50 (47.62%) Rectum: 48 (48.48%) Unknown: 1 (0.95%) Unknown: 1 (1.01%) Primary tumor lesion Exist: 28 (26.67%) Exist: 26 (26.26%) None: 77 (73.33%) None: 73 (73.74%) Histological grade Well (grade I): 0 (0.00%) Well (grade I): 0 (0.00%) Moderate (grade II): 49 (46.67%) Moderate (grade II): 46 (46.46%) Poor (grade III): 42 (40.00%) Poor (grade III): 39 (39.39%) Unknown: 14 (13.33%) Unknown: 14 (14.14%) Gene status Wild-type: 43 (40.95%) Wild-type: 39 (39.39%) K-ras/N-ras mutation: 47 (44.76%) K-ras/N-ras mutation: 45 (45.45%) B-raf mutation: 6 (5.71%) B-raf mutation: 6 (6.06%) Unknown: 9 (8.57%) Unknown: 9 (9.09%) Number of metastatic sites <3: 80 (76.19%) <3: 76 (76.77%) ≥3: 25 (23.81%) ≥3: 23 (23.23%) Previous targeted therapy history Yes: 62 (59.05%) Yes: 58 (58.59%) No: 43 (40.95%) No: 41 (41.41%) 5-FU or capecitabine history in second-line chemotherapy Yes: 60 (57.14%) Yes: 56 (56.57%) No: 45 (42.86%) No: 43 (43.43%) †In our study, left-sided colon only includes descending colon, sigmoid colon. The rectum was analyzed separately. 3.2 Effects

In the ITT population (105 patients), the ORR was 7.62%, and the DCR was 48.57%. The mPFS and mOS were 2.5 and 8.0 months, respectively. In the PP population (99 patients), the ORR was 8.08%, and the DCR was 51.52%. The mPFS and mOS were 2.5 and 8.0 months, respectively (Figure 1, Table S1).

Survival analysis of intention-to-treat (ITT) group and per-protocol (PP) group. A, Overall survival (OS) analysis of ITT group, medium progression-free survival (PFS) months (95% CI): 2.5 (1.5-3.0). B, PFS analysis of ITT group, medium OS months (95% CI): 8.0 (7.0-10.0). C, OS analysis of PP group, medium PFS months (95% CI): 2.5 (1.8-3.0). D, PFS analysis of PP group, medium OS months (95% CI): 8.0 (7.0-10.0) [Colour figure can be viewed at wileyonlinelibrary.com] 3.3 Adverse events

Adverse events (Table 2) included neutropenia, anemia, thrombocytopenia, abnormal liver enzymes, pyrexia, diarrhea, nausea, vomiting, anorexia, skin eruption, oral ulceration, and fatigue, and most of the adverse events were graded 1-2. The incidence of grade 3 hematological and nonhematological toxicities was 22.9%. The incidence of grade 4 hematological toxicity was less than 7% (7/105), comprising neutropenia, anemia, and thrombopenia.

TABLE 2. Adverse events Adverse eventa Grade Incidence (%) Neutropenia All grades 37.14 Grade I 15.24 Grade II 9.52 Grade III 8.57 Grade IV 3.81 Anemia All grades 45.71 Grade I 24.76 Grade II 13.33 Grade III 6.67 Grade IV 0.95 Thrombocytopenia All grades 20.95 Grade I 7.62 Grade II 7.62 Grade III 2.86 Grade IV 2.86 Hepatic function abnormal ALT All grades 5.71 Grade I 4.76 Grade II 0.95 AST Grade I 2.86 BIL All grades 1.90 Grade I 0.95 Grade II 0.95 Nausea

All grades

Grade I

Grade II

Grade III

21.90

14.29

5.71

1.90

Diarrhea

All grades

Grade I

Grade II

Grade III

8.57

5.71

1.91

0.95

Skin eruption All grades 8.57 Grade I 6.67 Grade II 0.95 Grade III 0.95 Vomiting Grade I 7.62 Pyrexia All grades 6.67 Grade I 4.76 Grade II 1.90 Fatigue

All grades

Grade I

Grade II

6.67

5.72

0.95

Oral ulceration

All grades

Grade I

Grade II

Grade III

4.87

1.91

0.95

1.91

Note. Rare adverse event was defined as incidence <3%. Other rare adverse events: hand-foot syndrome, baldness, headache, epistaxis, edema of lower extremity. aWe reported adverse events with all-grade incidence >5%, or grade 3-4 adverse events happened in detail. 3.4 Subgroup analysis

Subgroup analysis of the ITT population is given in Table 3. The effects of the regimen were similar or not significantly different for the different pathological grades, primary lesions, different number of metastasis sites, with or without oncotarget therapy in the past, different primary site (left- or right-sided colon cancer, divided by colonic splenic flexure), even different gene mutation status (K-Ras/N-Ras mutant, B-Raf mutant, or wild-type mCRC). Patients with colon cancer appeared to have better PFS than those with rectum cancer, irrespective of left- or right-sided colon, with mPFS of 3.00 and 1.65 months, respectively (Figure 2A, P = .044). In the six patients carrying the B-Raf mutation, DCR was 50%, and mPFS was 2.75 months (Figure 2B). Patients with a metastatic site number less than 3 had a better OS and no difference in PFS (Figure 2C).

TABLE 3. Subgroup analysis of disease control rate (DCR), progression-free survival (PFS), and overall survival (OS) in ITT group Numbers DCR Median PFS Median OS Overall 105 48.57% 2.50 8.00 Gender P = .490 P = .32 P = .12 Female 38 52.63% 2.25 10.00 Male 67 46.27% 2.50 7.00 Age P = .918 P = .77 P = .52 <60 49 49.98% 2.00 8.00 ≥60 56 48.21% 2.50 8.00 Histological grade P = .480 P = .13 P = .74 II 47 42.94% 2.0 7.00 III 42 52.38% 3.0 8.50 Unknown 14 42.86% 1.5 7.25 Gene P = .978 P = .46 P = .46 Wild-type 43 48.84% 2.50 7.50 K-ras/N-ras mutation 47 51.06% 2.00 8.00 B-raf mutation 6 50.00% 2.75 6.25 Primary tumor site P = .006 P = .044 P = .65 Right-sided colon 25 56.00% 3.00 6.00 Left-sided colon 29 68.97% 3.00 9.00 Rectum 50 34.00% 1.65 8.00 Primary tumor resection P = .465 P = .37 P = .21 No 28 53.57% 3.00 6.07 Yes 77 46.75% 2.00 8.00 Number of metastasis site P = .116 P = .81 P = .053 <3 77 53.25% 2.50 9.50 ≥3 27 37.04% 2.00 7.00 Baseline ALB P = .016 P = .0022 P = .014 <40 g/L 47 34.04% 1.50 7.00 ≥40 g/L 57 59.65% 3.00 10.00 Baseline HB P = .484 P = .14 P = .061 <115 g/L 36 41.67% 1.50 6.07 ≥115 g/L 68 51.47% 2.75 8.50 Targeted therapy history P = .935 P = .93 P = .69 Yes 62 48.39% 2.00 8.00 No 43 48.84% 2.50 7.50 5-FU or capecitabine history in second-line therapy P = .140 P = .05 P = .75 Yes 60 41.67% 1.65 8.00 No 45 57.78% 3.00 8.00 Number of systemic therapy P = .001 P < .001 P = .051 3 Line 80 57.50% 3.00 9.00 >3 Line 25 20.00% 1.50 7.00 Univariable survival analysis of primary site, gene status, metastatic site number baseline Alb, chemotherapy line and usage of 5-FU in second-line chemotherapy in intention-to-treat (ITT) group. A, Kaplan-Meier survival curve of progression-free survival (PFS) and overall survival (OS) according to primary site (P = .044, P = .65). B, Kaplan-Meier survival curve of PFS and OS according to gene status (P = .46, P = .46). C, Kaplan-Meier survival curve of PFS and OS according to number of metastatic sites (P = .81, P = .053). D, Kaplan-Meier survival curve of PFS and OS according to baseline Alb level (P = .0022, P = .014). E, Kaplan-Meier survival curve of PFS and OS according to line of chemotherapy (P P = .051). F, Kaplan-Meier survival curve of PFS and OS according to usage of 5-FU in second-line chemotherapy (P = .05, P = .75) [Colour figure can be viewed at wileyonlinelibrary.com]

We found that patients with baseline albumin (ALB) >40 g/L had better mPFS and mOS (Figure 2D). The effects of the RS regimen were better when implemented as the third-line therapy than after third-line treatment (Figure 2E). If 5-FU or capecitabine was not used as the second-line treatment, the patients had a mPFS of 3.0 months, which was significantly longer than the mPFS of 1.9 months in continuous 5-FU treatment patients (Figure 2F, P = .05). Additional