TAS-102 Independent and Combined Therapy in Metastatic Colorectal Cancer

Objective To evaluate the effectiveness and safety of TAS-102 in the treatment of metastatic colorectal cancer. Methods The pubmed, web of science, medline, cochrane library databases were searched for the literature on TAS-102 treatment of metastatic colorectal cancer. Extract data such as median Overall Survival (mOS), median Progression-Free Survival (mPFS) and the incidence of adverse events for meta-analysis. Results Our study found that the mOS of patients treated with TAS-102 was 7.74 (95%CI: 6.09–9.85) and the mPFS was 2.91 (95%CI: 2.38–3.57). The mOS in patients treated by TAS-102 Combined with bevacizumab is 10.41 (95%CI: 8.40-12.89) and the mPFS is 4.35 (95%CI: 3.05– 6.20). In the control experiment, the patients' mOS and mPFS were improved. TAS-102 + B VS. TAS- 102 (OR = 0.41, 95% CI: 0.18–0.93; OR = 0.72, 95% CI: 0.63–0.83). TAS-102 VS. Placebo(OR = 0.44, 95% CI: 0.29–0.67; OR = 0.51, 95% CI: 0.42–0.62). The incidence of adverse events in combination with bevacizumab will increase. is a humanized monoclonal antibody against vascular endothelial growth factor (VEGF), which plays an anti-tumor effect by blocking the formation of tumor blood vessels and regulating the immune function of patients 37 . In 2004, the FDA approved bevacizumab combined with chemotherapy drugs as the rst-line treatment for mCRC. A study showed that bevacizumab combined with rst-line chemotherapy for metastatic colorectal cancer can signicantly prolong the survival and PFS of patients with mCRC, improve the quality of life, increase the resectable rate of metastases, and improve the survival outcome of patients with mCRC 38,39 . The number of adverse events has also been signicantly reduced.


Introduction
By 2020, it is estimated that colorectal cancer is the cause of 935,000 cancer-related deaths worldwide, accounting for 9% of all cancer deaths 1 . In the initial diagnosis, approximately 25% of colorectal cancer patients have concurrent metastatic disease, and more than half of the patients are diagnosed as metastases 2,3 . Despite advances in the treatment of metastatic CRC (metastatic colorectal cancer), the survival rate is still poor. And the expected survival period without effective drug treatment is about 6 months 4-6 .
TAS-102 (tri uridine/tipiracil) is an oral anticancer drug containing a thymidine analogue (tri uridine). It is composed of active cytotoxic component FTD and effective thymidine phosphorylase inhibitor TPI hydrochloride. The molar ratio is 1:0.5 5 . FTD is the active cytotoxic component of the drug. TPI can prevent thymidine phosphorylase from rapidly degrading FTD into Inactive form 7,8 . FTD/TPI is established as the third-line treatment for metastatic colorectal cancer. According to the results of the international phase III RECOURSE study, the study reported the signi cant bene ts of FTD/TPI compared with placebo in terms of overall survival (OS), and acceptable security conditions 9,10 .
The e cacy and safety of FTD/TPI monotherapy in adults with refractory mCRC was rst demonstrated in a Japanese phase II trial by Yoshino et al. 5 and later in the pivotal phase III RECOURSE trial 4 . In these two studies, TAS-102 showed good effectiveness, signi cantly improving overall survival (mOS) and median progression-free survival (mPFS). TAS-102 combined with bevacizumab have a good effectiveness in the treatment of metastatic colorectal cancer, while reducing the incidence of adverse events.
Regorafenib and TAS-102 are both considered as new treatment options for salvage line therapy. A meta-analysis showed similar effectiveness of the two drugs, but the occurrence of adverse events may be different 12 . The main goal of clinical trials is to establish the effectiveness and safety of the drug in a carefully selected group of patients. However, there are still differences from real-world applications. The actual application of TAS-102 needs more attention. This study conducted a meta-analysis of clinical trials in the practical application of TAS-102, and compared the safety and effectiveness of drugs in controlled trials and uncontrolled trials.

Search Strategy
PubMed, Medline, Web of Science, Cochrane databases were searched for eligiity publications. The following keywords were used: "metastatic colorectal cancer" AND "TAS-102" OR "FTD/TPI". There is no time limit for searching until the nal search date is May 31, 2021. In addition, the reference list of applicable studies was manually checked for inclusion in other articles. Two researchers jointly complete this search process.

Inclusion and Exclusion
These articles are manually screened for relevance, rst by removing duplicate publications, and then checking the abstract if the title is not enough to exclude. If a publication appears to meet the inclusion criteria after reading the abstract, it will be marked for full text inspection. Review the design, intervention methods, and results of the study, and evaluate compliance with the requirements. Animal experiments, reviews, reports, non-refractory colorectal cancer, inappropriate outcome indicators, non-compliant intervention are the main exclusion principles.

Data Extraction and Quality Assessment
Two researchers independently extracted relevant information from each study: rst author, year of publication, demographic characteristics of participants include age, gender, ECOG performance status, (K) RAS status, grouping scheme, sample size, median os, median pfs, HR, the incidence of grade ≥ 3 AEs. We downloaded the full text. If in doubt, ask the original author for help. The NOS scale is used to evaluate the quality of the included controlled trials. The total score is 9 points, and the scores above 5 are included in the metaanalysis. However, for the included one-arm experiment, the rst 8 items of the MINORS item were selected for quality evaluation. Each item is 2 points and the total score is 16 points, 10 points or more enter our research.

Statistical Analysis
Based on the recommendations of the Cochrane collaboration, quantitative synthesis of the indicators included in the study. In metaanalysis, I 2 > 50% is considered heterogeneous. In the absence of signi cant heterogeneity, a xed-effect model is used; otherwise, a random-effect model is used 13 . Performing sensitivity analyze and subgroup analyze for the included studies to nd the reason of heterogeneity. Funnel plots were used to detect whether there is a small research effect 14 .

The characteristics of the included studies
855 studies were retrieved. Two investigators were screened and included 26 studies 4,5,9-11,15−35 . Including 14 controlled experiments and 12 single-arm experiments. The average age of 3780 participants was over 50 years old. The intervention methods are TAS-102 alone or combined with bevacizumab, and the control is (REG)regorafenib or placebo. The search and screening process is described in Fig. 1. All studies included in this study were based on moderate-to high-quality evidence. Table 1 provides a brief description of these 26 studies. In the included studies, the score of the controlled experiment was above 5, and the score of the uncontrolled experiment was above 10. The quality of the literature can support the meta-analysis. The appendix summarizes the literature quality evaluation situation.

Effectiveness of uncontrolled clinical trials
Pooling the PFS data from 12 uncontrolled clinical trials revealed that the mOS of patients with metastatic colorectal cancer was 8.11 (95% CI: 6.57-10.01). A random effect model was used, I 2 = 79.7% (Fig. 2). Subgroup analysis showed that the mOS of TAS-102 combined with bevacizumab treatment may be higher. Similarly, the mPFS of patients was 3.06 (95% CI: 2.52-3.70). A random effect model was used, I 2 = 91.8% (Fig. 3). Subgroup analysis did not nd a difference between TAS-102 + B and TAS-102 alone.  (Fig. 4). Similarly, the mOS death hazard ratio in the second scheme was 0.72 (95% CI: 0.63-0.83). A random effect model was used, I 2 = 58.7% (Fig. 4). A sensitivity analysis that eliminated studies one by one did not detect abnormalities. The funnel chart and Begg's test (Egger's test) show that there is no publication bias.
Under the rst scheme, compared with the control group, the mPFS was improved, and the risk ratio of death was 0.44 (95% CI: 0.29-0.67).
A random effect model was used, I 2 = 60.9% (Fig. 5). Similarly, the mPFS death hazard ratio in the second scheme was 0.51 (95% CI: 0.42-0.62). A random effect model was used, I 2 = 52.8% (Fig. 5). A sensitivity analysis that eliminated studies one by one did not detect abnormalities. The funnel chart and Begg's test (Egger's test) show that there is no publication bias.
9 studies showed the incidence of grade ≥ 3 AEs in the experimental group and the control group. Compared with the placebo group, there is no difference in safety under the treatment of TAS-102 alone. OR = 3.10 (95%CI: 0.88-10.87). Random effects model was used, I 2 = 84.7% (Fig. 6). Compared with the application of TAS-102 alone, the combination of bevacizumab may increase the incidence of adverse events. OR = 2.19 (95%CI: 1. 40-3.44). The xed effects model is applied, I 2 = 0.0% (Fig. 6). A sensitivity analysis that eliminated studies one by one did not detect abnormalities. The funnel chart and Begg's test (Egger's test) show that there is no publication bias.

Discussion
Almost 55% of colorectal cancer cases worldwide occur in more developed countries. Its incidence continues to rise in developing countries 36 . As with most cancer types, surgery is the main treatment method. For metastatic cancer, cytotoxic methods, such as neoadjuvant therapy and adjuvant therapy, are used before or after it. The main treatment options include uoropyrimidine, oxaliplatin, and irinotecan. TAS-102 is an anti-cancer drug that has entered people's eld of vision in recent years. Because of its excellent clinical e cacy and safety, it is often added to the treatment of colorectal cancer, gastric cancer in the middle and late stages and anti-cancer treatment programs for metastatic tumors.
Our study found that the mOS of patients treated with TAS-102 was 7.74 (95%CI: 6.09-9.85) and the mPFS was 2.91 (95%CI: 2.38-3.57). The mOS in patients treated by TAS-102 combined with bevacizumab is 10.41 (95%CI: 8. 40-12.89) and the mPFS is 4.35 (95%CI: 3.05-6.20). Combination therapy may have better effectiveness. As the current targeted drug for the treatment of metastatic colorectal cancer, it is a humanized monoclonal antibody against vascular endothelial growth factor (VEGF), which plays an anti-tumor effect by blocking the formation of tumor blood vessels and regulating the immune function of patients 37 . In 2004, the FDA approved bevacizumab combined with chemotherapy drugs as the rst-line treatment for mCRC. A study showed that bevacizumab combined with rst-line chemotherapy for metastatic colorectal cancer can signi cantly prolong the survival and PFS of patients with mCRC, improve the quality of life, increase the resectable rate of metastases, and improve the survival outcome of patients with mCRC 38,39 . The number of adverse events has also been signi cantly reduced.
Although uncontrolled trials can observe the survival of patients, they cannot specify the improvement in survival. We included 16 studies that included two controlled protocols (TAS-102 + B VS. TAS-102 AND TAS-102 VS. Placebo). In either scenario, we found a signi cant increase in mOS and mPFS. Surprisingly, we found that TAS-102 combined with bevacizumab will increase the incidence of grade ≥ 3 AEs (OR = 2.19, 95%CI: 1.40-3.44) compared to TAS-102 alone. The safety of bevacizumab is worthy of further consideration. This indicates that clinicians need to make careful decisions when making treatment options for patients with metastatic colorectal cancer, considering the patient's tolerance to anticancer drugs.
It is necessary to optimize the design plan when evaluating the e cacy of new drugs. Randomized controlled trials such as resource and TERRA are conducted in homogeneous populations, which can minimize the risk of bias 40 . In the current study, we have included real observational studies aimed at evaluating the effectiveness of a relatively small homogeneous population. These studies have the shortcomings of non-randomized controlled studies. The studies we included included controlled and uncontrolled experiments. And the demographic characteristics and disease manifestations of the participants in the experiment are also quite different. This will actually affect the accuracy of our nal results. Therefore, more rigorous and appropriate randomized controlled experiments need to be proposed.
The published meta-analysis of TAS-102 involves comparison of the effectiveness and safety of multiple therapeutic drugs [40][41][42][43][44][45] . Regorafenib, TAS-102, fruquintinib, panitumumab and cetuximab are recommended single-agent chemotherapy regimens for patients exhibiting disease progression. The safety of these drugs is di cult to assess. But the safety of the drug does affect the con dence of patients in the treatment plan. The most important thing is the improvement of symptoms and the management of side effects 47,48 .
In a retrospective study, potential predictive or prognostic biomarkers for the e cacy of regorafenib were evaluated 46 . The results indicate that the mutation status of KRAS and PIK3CA may be a predictor of the clinical bene t of regorafenib treatment. However, no biomarkers have been found to predict the e cacy and safety of TAS-102. The frequency of KRAS mutations should be worth exploring, because mOS and mPFS are independent of KRAS status in resource and TERRA, but are highly dependent on KRAS mutations in the second phase of trials in Japan 5 .
Although this study gives the survival status of TAS-102 as a single treatment, it is worth looking forward to the application of the drug combination program may further improve the e cacy of FTD/TPI, and some clinical trials are currently underway.