Early Clinical Experience with Trifluridine/Tipiracil for Refractory Metastatic Colorectal Cancer: The ROS Study

Simple Summary Trifluridine/tipiracil is an oral combination therapy currently approved as a salvage-line treatment in patients with metastatic colorectal cancer refractory to, or not, candidates for available therapies. However, there is no consensus on the specific factors that should be considered to select patients who benefit the most from trifluridine/tipiracil in clinical practice. The aim of our retrospective cohort study was to assess the early clinical experience with trifluridine/tipiracil in Spain and identify potential survival markers. Our findings endorse the real-life efficacy and safety of trifluridine/tipiracil for refractory metastatic colorectal cancer, as well as revealing the presence of ≤2 metastatic sites, absence of liver metastasis, alkaline phosphatase levels < 300 IU, trifluridine/tipiracil dose reductions, and neutrophil/lymphocyte ratio < 5 as survival markers. Combinations of these markers may help physicians to identify subsets of patients with refractory metastatic colorectal cancer that may benefit the most from trifluridine/tipiracil in their daily practice. Abstract Trifluridine/tipiracil is currently approved for metastatic colorectal cancer (mCRC) refractory to available therapies. However, there is no consensus on factors that predict treatment outcomes in daily practice. We assessed the early clinical experience with trifluridine/tipiracil in Spain and potential survival markers. This was a retrospective cohort study of mCRC patients who participated in the trifluridine/tipiracil early clinical experience programme in Spain. The primary outcome was overall survival (OS). Associations between OS and patient characteristics were assessed using multivariate Cox regression analyses. A total of 379 patients were included in the study. Trifluridine/tipiracil was administered for a median of 3.0 cycles and discontinued mainly due to disease progression (79.2%). The median OS was 7.9 months, with a 12-month OS rate of 30.5%. Cox analyses revealed that the following variables independently enhanced OS: ≤2 metastatic sites, no liver metastasis, alkaline phosphatase < 300 IU, trifluridine/tipiracil dose reductions, and neutrophil/lymphocyte ratio < 5. Grade ≥ 3 toxicities were reported in 141 (37.2%) patients, including mainly afebrile neutropaenia (23.2%), anaemia (12.1%), and thrombocytopaenia (5.3%). This study supports the real-life efficacy and safety of trifluridine/tipiracil for refractory mCRC and identifies tumour burden, liver metastasis, alkaline phosphatase, dose reductions, and neutrophil/lymphocyte ratio as survival markers.


Introduction
Colorectal cancer is one of the three most commonly diagnosed cancers and the second leading cause of cancer death worldwide [1]. Only 39% of colorectal cancer patients are diagnosed with localised disease, and despite the improvements achieved in its management, the 5-year survival of patients with distant metastases drops to 14% [2]. Current therapies for metastatic colorectal cancer (mCRC) involve several active drugs administered either as monotherapy or in combination, including cytotoxic agents such as fluoropyrimidines, irinotecan or oxaliplatin, and targeted therapies against epidermal growth factor receptor (EGFR) or vascular endothelial growth factor (VEGF) [3,4]. However, the therapeutic management of mCRC refractory to these therapies remains challenging.
Trifluridine/tipiracil is an oral combination therapy consisting of an antineoplastic thymidine-based nucleoside analogue (trifluridine) and a thymidine phosphorylase inhibitor (tipiracil hydrochloride). Trifluridine is the active cytotoxic component, responsible for preventing tumour cell proliferation by interfering with DNA function [5,6], while tipiracil improves trifluridine bioavailability by inhibiting its catabolism [7,8]. Trifluridine/tipiracil administration is currently approved as a salvage-line treatment in patients with mCRC refractory to, or not candidates for, available therapies based on the pivotal phase III RECOURSE trial. Its findings revealed a median overall survival (OS) of 7.1 months with trifluridine/tipiracil versus 5.3 months with placebo [9]. Trifluridine/tipiracil improved OS irrespective of age, KRAS status, time from first metastasis or geographic region, and also enhanced progression-free survival (PFS), disease control, and performance status [9][10][11]. Subsequent post hoc analyses also pointed to neutropaenia as a surrogate marker of trifluridine/tipiracil efficacy [12], and supported low tumour burden, indolent disease, and absence of liver metastasis as prognostic factors [13].
Compassionate use programmes and cohort studies endorsed the use of trifluridine/tipiracil in non-trial conditions [14][15][16][17][18][19]. Many of them also addressed the unmet need of identifying prognostic or predictive factors, which included pre-treatment performance status [15,16,[18][19][20], KRAS status [18], time to metastasis [19], number of metastases [15], time from metastasis diagnosis [19] or first-line therapy [17], alkaline phosphatase levels [15,18], platelet and leucocyte counts [15,18], and neutrophil/leucocyte ratio [15,20], as well as on-treatment neutropaenia [15,17,19,21], and dose reductions [15,16]. However, there is no consensus on the specific factors that should be considered to select patients who benefit the most from trifluridine/tipiracil. In light of the above, this study aimed to provide further insights into the efficacy and safety of trifluridine/tipiracil for refractory mCRC in the early clinical experience in Spain, and to identify factors that may help physicians to predict better outcomes in daily practice.

Materials and Methods
ROS was a retrospective cohort study conducted in the Departments of Medical Oncology at 35 Spanish hospitals according to Good Pharmacoepidemiology Practices, the Declaration of Helsinki, and national regulations. The study was approved by the ethics committee of Hospital General Universitario Gregorio Marañón (Madrid, Spain) and alive patients gave their written informed consent; informed consent was waived in deceased and lost-to-follow-up patients.

Patient Population
The study included patients aged over 18 years, with histologically or cytologically confirmed colorectal adenocarcinoma, and who had participated in the early clinical experience programme for trifluridine/tipiracil. To participate in this program, patients had to be previously treated with, or not considered candidates for, available therapies (fluoropyrimidine-, oxaliplatin-, or irinotecan-based chemotherapies, anti-VEGF or anti-EGFR agents). Eligible patients must have completed the early clinical treatment with trifluridine/tipiracil at study enrolment. Patients who had received other investigational drugs or anticancer therapies (chemotherapy, immunotherapy, biological response modifiers, or endocrine therapy) while receiving trifluridine/tipiracil were excluded.

Study Treatment
Trifluridine/tipiracil (Lonsurf ® , Laboratoires Servier, Suresnes, France) was taken orally at a starting dose decided at the discretion of the treating oncologist and administered twice daily, after breakfast and dinner, on days 1 to 5 and 8 to 12 of each 28-day cycle. Patients were treated by the supervising physicians according to the recommendations included in the product information. Dose adjustments, delays, and discontinuations were performed according to clinical criteria.

Assessments
All study data were retrospectively collected from patient medical charts, progress reports submitted to the health authorities, and a data log owned by Laboratorios Servier S.L. containing previously collected information on compassionate use of trifluridine/tipiracil. These data included demographics and baseline clinical characteristics, including the Eastern Cooperative Oncology Group (ECOG) performance status, medical history of colorectal cancer, and antitumour therapies prior to trifluridine/tipiracil. Dates and doses of trifluridine/tipiracil were collected at the first treatment cycle, along with subsequent treatment modifications and the total number of administered cycles. Data available on follow-up ECOG performance status, radiologic progression and best response to trifluridine/tipiracil according to Response Evaluation Criteria in Solid Tumours (version 1.1) were also retrieved [22], as well as grade ≥ 3 adverse events attributed to trifluridine/tipiracil. These events were coded using the Medical Dictionary for Regulatory Activities (MedDRA) and their severity graded by National Cancer Institute Common Toxicity Criteria (version 4.03) [23]. Other therapies received after ending trifluridine/tipiracil and the survival status (alive, dead, or lost to follow-up) at data collection were also retrieved.

Outcomes
The primary study outcome was OS, which was defined as the time from trifluridine/tipiracil start to patient death from any cause. Secondary efficacy outcomes included PFS measured as the time from trifluridine/tipiracil start to radiologic disease progression or death from any cause, and PFS measured as the time from trifluridine/tipiracil start to clinical disease progression or death from any cause. Other secondary efficacy outcomes were OS and PFS rates at 2, 4, 6, 8, 10, and 12 months, overall response rate (percentage of patients with a complete or partial response), disease control rate (percentage of patients with complete or partial response or stable disease, with stable disease assessed at least 6 weeks after starting trifluridine/tipiracil), changes in ECOG performance status (improved, maintained, or worsened), and associations between OS and patient characteristics. Trifluridine/tipiracil exposure/management and grade ≥ 3 treatment-related adverse events were additional secondary outcomes.

Statistical Considerations
Sample size calculation was based on the median OS reported in the RECOURSE trial [9], a two-sided test, power of 99%, type I error of 0.01, loss rate ≤ 10%, and maximum follow-up of 12 months. The estimated sample size was 467 patients, for which the lower and higher critical values were 6.1 and 8.4 months, respectively.
Kaplan-Meier analyses were performed to assess OS and PFS, including survival curves, median estimates, and 95% confidence intervals (CIs). Descriptive statistics were used to determine survival rates, overall response rate, disease control rate, changes in ECOG performance status, trifluridine/tipiracil exposure/management, and treatmentrelated adverse events. Bivariate Cox regression analyses assessed associations between OS and the following characteristics: age ≥ 65 or <65 years, sex, ECOG performance status 0-1 or ≥2, time from metastasis diagnosis < 18 or ≥18 months, number of metastatic sites ≤2 or ≥3, presence or absence of liver metastatic lesions, synchronous or metachronous metastases, right or left colon primary tumour location, primary tumour molecular status (KRAS, RAS [KRAS + NRAS], BRAF, PI3K, HER2, and microsatellite instability), primary tumour surgery, previous antitumour lines < 3 or ≥3, and laboratory data prior to the first trifluridine/tipiracil cycle (platelet count ≥ 400 × 10 9 /L or <400 × 10 9 /L, leukocyte count ≥ 10 × 10 9 /L or <10 × 10 9 /L, alkaline phosphatase level ≥ 300 or <300 IU, and haemoglobin level ≥ 11 or <11 g/dL). A multivariate Cox regression model was built using characteristics with p < 0.20 in bivariate analyses, including the calculation of hazard ratios (HRs) and 95% CIs. Another multivariate Cox regression model was built adding neutropaenia, neutrophil/lymphocyte ratio, and dose reduction to the previously mentioned characteristics. OS in the prognostic subsets of patients with low tumour burden (≤2 metastatic sites), indolent disease (≥18 months from metastasis diagnosis), and absence of liver metastases [13] was also explored using log-rank tests and Cox regressions.
Missing data were not considered in the analyses, and a significance level of 0.05 was used for statistical testing. The statistical analyses were performed with IBM SPSS Statistics version 22.0 (IBM Corp., Armonk, NY, USA).

Patient Characteristics
A total of 402 patients were screened between June and November 2019, 23 of whom were not eligible (Supplementary Figure S1). Thus, 379 patients were finally included in this study, whose baseline characteristics and tumour molecular status are described in Tables 1 and 2, respectively.

Study Treatment
The planned dose of trifluridine/tipiracil at the first cycle ranged from 25.  Table S1). A total of 116 patients needed at least one of the 145 dose reductions and 191 patients needed at least one of the 294 dose delays reported during the trifluridine/tipiracil treatment. The median (IQR) number of administered cycles was 3.0 (2.0-4.0), and end of treatment was due mainly to disease progression (79.2%), followed by general state impairment (12.7%), toxicity (4.5%), patient decision (1.8%), and other reasons (1.8%).
The median PFS measured until radiologic disease progression or death was 3.2 months (95% CI 3.0-3.4; Figure 3a), with PFS rates at 2, 4, 6, 8, 10, and 12 months of 82.9% Figure 2. Kaplan-Meier plots for overall survival in the exploratory analysis of prognostic subsets. They included the comparison of patients with ≤2 metastatic sites and ≥18 months from metastasis diagnosis versus those with ≥3 metastatic sites and/or <18 months (a), patients with ≤2 metastatic sites, ≥18 months from metastasis diagnosis and absence of liver metastasis versus those with ≤2 metastatic sites, ≥18 months from metastasis diagnosis and liver metastasis (b), and patients with ≥3 metastatic sites and/or <18 months from metastasis diagnosis and absence of liver metastasis versus those with ≥3 metastatic sites and/or <18 months from metastasis diagnosis and liver metastasis (c). CI: confidence interval, HR: hazard ratio, OS: overall survival.

Discussion
The results from the ROS cohort study support the efficacy and safety of trifluridine/tipiracil for refractory mCRC in the early clinical experience in Spain. More than half of patients maintained or even improved their performance status, nearly one-third achieved disease control, and median PFS and OS reached 3 months and almost 8 months, respectively. These results are in line with the benefits observed with trifluridine/tipiracil in the RECOURSE trial [9][10][11], along with other compassionate use programmes and observational studies that confirm the use of trifluridine/tipiracil as a feasible treatment alternative for refractory mCRC in a real-life setting [14,15,[18][19][20].
PFS is a commonly used outcome for refractory mCRC that is usually reported until radiologic disease progression or death. However, the real disease progression usually occurs between two radiologic assessments [14,24], and the data available on PFS measured until clinical progression or death is still limited. This study shed some light on this issue, supporting the achievement of similar findings when considering clinical progression. Nonetheless, it is noteworthy that PFS seems to be poorly correlated with OS after secondline treatment and OS remains the most robust outcome estimate when assessing treatment efficacy for mCRC [25]. In this regard, our findings also revealed that the median OS might even reach 12.4 months in the subset of patients with ≤2 metastatic sites, ≥18 months from metastasis diagnosis, and absence of liver metastasis. These data support the exploratory analysis of Tabernero et al. [13], who defined a subgroup of patients with good prognostic characteristics including low tumour burden (<3 metastatic sites) and indolent disease (≥18 months from metastasis) and another with poor prognostic characteristics that included high tumour burden and/or aggressive disease. The reported median OS in these subgroups were 9.3 and 5.3 months, respectively. Our findings are in line with these data, showing that the 231 patients with ≤2 metastatic sites and ≥18 months from metastasis showed a median OS of 9.1 months versus 6.0 months in the 148 patients with ≥3 metastatic sites and/or <18 months. However, as in our study, the best prognosis could be identified within the subset of patients with good prognostic factors and absence of liver metastasis, with a median OS that reached 16.4 months [13].
In addition to the tumour burden and liver metastasis, our study supports the role of alkaline phosphatase levels, dose reductions, and neutrophil/lymphocyte ratios as survival markers. These findings agree with recently published observational studies that also revealed longer OS in patients with a single metastatic site [15], liver metastases [26], alkaline phosphatase levels < 200 or ≤500 IU/L [15,18], trifluridine/tipiracil dose reductions [15,16], and neutrophil/lymphocyte ratios < 5 [15,20]. However, they also found the potential influence of other factors such as ECOG performance status [15,16,[18][19][20]26], platelet count ≤ 350 × 10 9 /L [15], KRAS status [18], time to synchronous or metachronous metastasis [19], leukocyte count < 8 × 10 9 /L [18], or neutropaenia as an adverse event [15,17,19,21], which did not independently affect survival in our study. Based on these findings, some attempts have been made to achieve nomograms to screen for the patients who can benefit the most from trifluridine/tipiracil. These include the nomogram reported by Fernández-Montes et al. (2020) in mCRC patients treated with trifluridine/tipiracil, which considered the ECOG performance status, presence of multiple metastatic sites, carcinoembryonic antigen > 10 ng/mL, platelet counts > 350 × 10 9 /L, and phosphatase > 500 IU/L [15]. Another was the Colon Life nomogram, based on refractory mCRC patients treated with trifluridine/tipiracil, regorafenib, or other treatments, that included the ECOG performance status, primary tumour resection, lactate dehydrogenase, and peritoneal involvement [27]. Likewise, the REGOTAS study developed a scoring system to predict OS after trifluridine/tipiracil, including the ECOG performance status, aspartate transaminase > 40 IU/L, C-reactive protein ≥ 1.0 mg/dL, and cancer antigen 19-9 > 37.0 U/mL, as well as another for regorafenib including aspartate transaminase > 40 IU/L, C-reactive protein ≥ 1.0 mg/dL, number of metastatic sites ≥ 3, and <18 months from first-line chemotherapy [26]. A common scoring model for trifluridine/tipiracil and regorafenib was also reported based on the ECOG performance status, ≤18 months form metastasis diagnosis, and prior chemotherapy for ≥2 months beyond the progressive disease [28]. Furthermore, the REBECCA study proposed another score for regorafenib based on the ECOG performance status, time from metastasis diagnosis, initial regorafenib dose, number of metastatic sites, liver metastases, and KRAS mutations [29]. However, no specific nomogram or scoring system has been validated for refractory mCRC, and further assessments of prognostic/predictive factors are still needed to make the best therapeutic decisions.
The safety profile of trifluridine/tipiracil was consistent with that previously reported in clinical trials [30], compassionate use programmes [17][18][19], and other observational studies [15,16,20,21]. No new safety concern arose and the main ≥3 grade treatmentrelated adverse events included afebrile neutropaenia, anaemia, and thrombocytopaenia. In addition, most of these adverse events could be managed with dose reductions or delays, and only led to treatment discontinuation in 4.5% of patients.
Although the prescribing conditions of compassionate use programmes are less strict than clinical trials and more clearly reflect routine clinical practice, the retrospective data extraction from medical charts entailed limitations due to data availability. We cannot therefore exclude the potential underestimation of treatment-related adverse events, or the influence of missing data in the lack of association observed between tumour molecular markers and OS. Likewise, we cannot discern how the low prevalence of variables such as the mutated KRAS might have affected the study findings, and the absence of association between ECOG performance status and OS is likely derived from the limited number of patients with ECOG 2. Another limitation is the absence of a central review of scans performed at each participating site. Furthermore, the study design included no comparator group, which would have enabled the magnitude of improvement derived from trifluridine/tipiracil to be quantified. Likewise, a comparator group would have been needed to assess whether the identified survival markers might also affect patients not receiving trifluridine/tipiracil and further clarify their role in colorectal cancer. However, it is noteworthy that this study assessed a wide range of patient characteristics as potential markers of survival and provides additional insight into real-life use of trifluridine/tipiracil in 35 Spanish hospitals, enhancing the generalisability of its findings and expanding the information to consider when treating mCRC in daily practice.

Conclusions
This study endorses the real-life efficacy and safety of trifluridine/tipiracil for refractory mCRC, as well as supporting the role of specific survival markers such as tumour burden, liver metastasis, alkaline phosphatase, dose reductions, and neutrophil/lymphocyte ratio. Combinations of these markers may help physicians to identify subsets of patients that may benefit the most from trifluridine/tipiracil. Further studies are still needed to confirm our findings, support the most appropriate combination of prognostic/predictive factors to optimise trifluridine/tipiracil treatment, and clarify the role of the best supportive care in patients supposed to have poor response to chemotherapy and worse survival.

Informed Consent Statement:
Alive patients gave their written informed consent; informed consent was waived in deceased and lost-to-follow-up patients.

Data Availability Statement:
All data relevant to the study are included in the article or uploaded as supplementary information. Further data are available from the authors upon reasonable request and with permission of the Spanish Cooperative Group for the Treatment of Digestive Tumours (TTD).