Systematic review and network meta-analyses of third-line treatments for metastatic colorectal cancer

Abstract

Background

Limited treatment options are available in chemotherapy-refractory metastatic colorectal cancer (mCRC). The objective was to conduct a systematic literature review (SLR) and exploratory network meta-analysis (NMA) to compare the tolerability and effectiveness of SIRT with Y-90 resin microspheres, regorafenib, TAS-102 (trifluridine/tipiracil), and best supportive care (BSC) as third-line treatment in patients with mCRC.

Methods

An SLR was conducted to identify studies comparing two or more of the treatments and reporting overall survival (OS), progression-free survival, tumor response, or adverse event (AE) incidence. An exploratory NMA was conducted to compare hazard ratios (HRs) for OS using Markov chain Monte Carlo (MCMC) techniques.

Results

Seven studies were identified in the SLR: two double-blind randomized-controlled trials (RCT) for each drug, one open-label RCT, and two non-randomized comparative studies for SIRT. Patient selection criteria differed between studies, with SIRT studies including patients with liver-dominant colorectal metastases. Nausea and vomiting were more frequent with TAS-102 than regorafenib or SIRT; diarrhea was more common with TAS-102 and regorafenib than SIRT. The exploratory NMA suggested that all active treatments improved OS, with HRs of 0.48 (95% CrI 0.30–0.78) for SIRT with Y-90 resin microspheres, 0.63 (0.38–1.03) for TAS-102, and 0.67 (0.40–1.08) for regorafenib each compared to BSC.

Conclusions

Regorafenib, TAS-102 and SIRT using Y-90 resin microspheres are more effective than BSC in third-line treatment of mCRC; however, study heterogeneity made comparisons between active treatments challenging. SIRT is a viable treatment for third-line mCRC and its favorable AE profile should be considered in the therapeutic decision-making process.

Background

Colorectal cancer (CRC) is the third most common cancer in the world by incidence, but second in terms of mortality, with an estimated 881,000 deaths in 2018 (Bray et al. 2018). Furthermore, the annual incidence is estimated to rise to over 3 million cases by 2040 (Ferlay et al. 2019). Metastases are reported in at least half of all CRC cases, and population-based studies report that liver metastases develop in approximately 60–70% of all cases of metastatic CRC (mCRC), with metastases restricted to the liver in approximately 35–55% of patients with mCRC (van der Geest et al. 2015; Adam et al. 2012; van der Pool et al. 2012; Mekenkamp et al. 2010; Kumar et al. 2014; Sadahiro et al. 2013; Kumar et al. 2013; Oh et al. 2015; Kennecke et al. 2014). Many patients with mCRC eventually become insensitive or unresponsive to chemotherapy (chemorefractory) or cannot tolerate multiple cycles of chemotherapy (chemotherapy-intolerant). In clinical practice, of all mCRC patients receiving first-line chemotherapy, approximately 50% go on to receive second-line chemotherapy, and of these patients, approximately 25% go on to receive third-line chemotherapy (Abrams et al. 2014; Zafar et al. 2009).

Limited treatment options are available in chemorefractory mCRC. Regorafenib (STIVARGA®; Bayer, Leverkusen, Germany) received US Food and Drug Administration (FDA) approval in 2012, and EMA approval in 2013; trifluridine–tipiracil (TAS-102, LONSURF®; Servier, Suresnes, France, and Taiho, Princeton, NJ) received FDA approval in 2015 and EMA approval in 2016. Both treatments are approved in Europe for patients with mCRC “who have been previously treated with, or are not considered candidates for, available therapies” including fluoropyrimidine-, oxaliplatin- and irinotecan-based chemotherapies, anti-VEGF agents, and anti-epidermal growth factor receptor (EGFR) agents (Bayer 2018; Les Laboratoires Servier 2016). Standard chemotherapy regimens for these patients typically consist of two lines of therapy combining the above agents; therefore, regorafenib and TAS-102 are considered as third-line therapy. While anti-EGFR re-challenge strategies are emerging as a treatment option for patients with RAS wild-type mCRC, options for patients with RAS-mutated mCRC remain limited and additional third-line therapies are a strong medical need (Van Cutsem et al. 2015; Mauri et al. 2019).

For patients failing third-line therapy, or not eligible for current third-line treatments, no other systemic therapy options are currently available and disease management is restricted to best supportive care (BSC). In cases of disease progression after two lines of treatment in mCRC, BSC is associated with median survival times of 4–6 months (Foubert et al. 2014).

SIRT with SIR-Spheres Y-90 resin microspheres is a treatment option for patients with liver-only or liver-dominant mCRC, and is recommended by European Society for Medical Oncology (ESMO) 2016 Guidelines for patients who are refractory or intolerant to chemotherapy (category B recommendation for yttrium-90 resin microspheres) and National Comprehensive Cancer Network (NCCN) Guidelines v3.2020 (category 2A recommendation for “arterially directed catheter therapy, and in particular yttrium-90 microsphere selective internal radiation”) in this indication (Van Cutsem et al. 2016; Benson et al. 2020). The current guidelines do not, however, offer recommendations on the respective positions of SIRT with SIR-Spheres Y-90 resin microspheres, regorafenib, or TAS-102 in the therapeutic strategy, especially for patients with colorectal liver metastases, as the guideline committees did not identify evidence comparing these interventions.

The objective of the present study was to conduct a systematic literature review with a descriptive analysis of the adverse event and overall survival (OS) data, and to conduct an exploratory network meta-analysis (NMA) comparing the relative clinical effectiveness and tolerability of SIRT with SIR-Spheres Y-90 resin microspheres, regorafenib, TAS-102, and BSC as third-line treatment in patients with mCRC.

Methods

The study was conducted in three stages: a systematic literature review designed to identify studies comparing two or more of the following interventions: SIR-Spheres Y-90 resin microspheres, TAS-102, regorafenib, or BSC; a descriptive analysis of the reported adverse event rates and median OS; and an exploratory network meta-analysis (NMA) comparing hazard ratios (HR) for OS.

Systematic review

A priori inclusion and exclusion criteria were developed using the Population, Intervention, Comparator, Outcome, Study Design (PICOS) methodology (Table 1). The PICOS criteria were designed to identify any studies investigating the use of BSC, TAS-102, regorafenib, or SIRT with Y-90 resin microspheres in third-line treatment of mCRC reporting OS, progression-free survival, tumor response, or adverse events. The study design criterion was structured to include comparative non-randomized studies of SIRT in addition to randomized-controlled trials on the basis of opinion from clinical experts solicited during the design of the literature search strategy, noting that there was likely to be a paucity of RCT data on SIRT in third-line mCRC.

Table 1 Population, intervention, comparators, outcomes, and study design (PICOS) criteria employed in the systematic literature review

The PubMed, EMBASE, and Cochrane Collaboration databases, and clinicaltrials.gov were searched using standard filters to identify clinical trials (Supplementary Tables 1 and 2). After removal of duplicate studies, titles and abstracts of the unique articles were screened by two independent reviewers and any conflicts were resolved by a third reviewer. Screening of the full-text copies of studies included after title and abstract screening was then conducted, and data were extracted from the papers included after full-text screening. Data on adverse events were extracted in two aggregated categories: Common Terminology Criteria for Adverse Events (CTCAE) Grades 1 and 2, and CTCAE Grades 3+. Adverse event data were only presented for those events occurring in > 5% of patients with Grade 1 and 2 events or > 2% of patients with Grade 3+ events in two or more studies, and those events that were potentially related to SIRT: bile duct complications, gastritis, gastrointestinal (GI) ulcers, radiation pneumonitis, and radioembolization-induced liver disease (REILD). Full-text screening and data extraction results were validated by an independent reviewer.

Evidence synthesis

Adverse event rates and median survival were summarized, and an exploratory NMA was conducted to provide the estimates of comparative effectiveness for the specified outcomes. The network meta-analysis was conducted using published methodology (Hawkins et al. 2016). The model parameters were estimated using Markov chain Monte Carlo (MCMC) techniques as implemented in JAGS 4.3.0 (Plummer 2003). Three chains were run starting from different initial values. Models were initially run for 20,000 iterations as a burn-in period and a further 20,000 iterations for estimation and convergence. Further iterations were considered if the Monte Carlo sampling error and convergence were judged not to be adequate. Convergence was assessed using Brooks–Gelman–Rubin plots and by examining history plots (the informal ‘blue finger test’) (Brooks and Gelman 1998). Adequacy of Monte Carlo sampling error was judged using the Rhat statistic and model fit was compared using the Deviance Information Criterion (DIC) (Brooks and Gelman 1998).

The exploratory NMA was conducted using HRs for OS assuming that placebo, BSC, and 5-FU monotherapy could be treated as a common BSC comparator. Both fixed and random effects pairwise meta-analyses were also undertaken, with the random effects meta-analyses incorporating an informative prior based on a review of previous meta-analyses (Turner et al. 2012). The effect of including data from non-randomized studies of SIRT was investigated through a series of analyses in which the variance around the mean HRs from the non-randomized studies was inflated (and hence the precision decreased) using a factor 1/wj where 0 < wj < 1 (Efthimiou et al. 2017). In the primary analysis, wj was set to 1, assigning equal weight to randomized and non-randomized studies. A validation analysis was then conducted with wj ≈ 0, approximating an analysis of RCTs only, with sensitivity analyses conducted with wj = 0.2, 0. 25, \(0.3\dot{3}\), 0.5, and \(0.6\dot{6}\).

Results

Search results

The literature searches retrieved 1524 studies, of which 1334 were unique (Supplementary Fig. 1). Title and abstract screening resulted in the exclusion of 1294 studies, leaving 40 studies for full-text screening. After screening of the full-text articles, seven studies were ultimately deemed eligible to be included in the analyses. The studies consisted of: one randomized open-label trial and two non-randomized studies comparing SIRT with BSC (Hendlisz et al. 2010; Bester et al. 2012; Seidensticker et al. 2012), two randomized double-blind trials comparing TAS-102 with placebo (Yoshino et al. 2012; Mayer et al. 2015), and two randomized double-blind trials comparing regorafenib with placebo (the RECOURSE and CONCUR) (Grothey et al. 2013; Li et al. 2015). While one of the non-randomized studies of SIRT (Bester et al. 2012) included patients with both liver metastases secondary to colorectal cancer (66% of subjects) and secondary to other sites (34%), the study was ultimately included as the results from the colorectal primary subgroup were reported separately from the overall population (Bester et al. 2012).

Study characteristics

The key characteristics and designs of the included studies are summarized in Table 2. The SIRT studies included an open-label randomized trial and two non-randomized comparative studies. The TAS-102 and regorafenib studies were double-blind randomized trials. Differences in the nature of SIRT and BSC interventions received by patients prevented blinding.

Table 2 Study designs

The TAS-102 and regorafenib studies were placebo-controlled studies in which patients also received BSC. Bester 2012 and Seidensticker 2012 had control arms in which patients received BSC. Although Hendlisz 2010 was a randomized trial, it compared SIRT with concurrent 5-fluorouracil (5-FU) versus 5-FU alone, allowing cross-over from the 5-FU alone arm to SIRT on progression at the investigators’ discretion. Since patients enrolled in Hendlisz 2010 had already progressed on a prior treatment regimen including 5-FU and 5-FU is not typically prescribed as monotherapy for mCRC, the control arm was assumed to be similar to BSC, although the impact of 5-FU itself is uncertain. Notably, 70% of patients in the 5-FU arm did cross over, including ten patients who received SIRT, potentially confounding overall survival outcomes of the trial (Hendlisz et al. 2010).

The demographics and clinical characteristics of study subjects are summarized in Table 3. The study subjects had a similar mean age across studies. In general, there was a slight excess of males included in the studies. Performance status was generally preserved with most patients (95–100%) having Eastern Cooperative Oncology Group (ECOG) scores of 0 or 1.

Table 3 Summary of subject demographics

The proportions of patients’ exposure to previous chemotherapy is summarized in Table 4. Almost all patients had received previous systemic therapy and all patients included in the selected studies were deemed refractory or intolerant to other available systemic agents available at the time of enrolment. Prior bevacizumab use varied between the studies; all patients in the Mayer et al. and Grothey et al. studies had received previous treatment with bevacizumab, with approximately 50% and 80% of patients having received bevacizumab in the Seidensticker et al. and Yoshino et al. studies, respectively. Reporting of the number of treatment lines received by the enrolled patients was inconsistent across studies, making quantitative comparisons between the different study populations difficult. The proportion of patients with extrahepatic disease (aside from the primary tumor) was not consistently reported across studies, but Bester et al. reported extrahepatic disease in 38% of patients receiving SIRT and 33% of patients receiving BSC; Yoshino et al. reported a total of 78% of patients receiving TAS-102 with two or more metastatic organs, compared with 81% of patients receiving placebo; and Hendlisz et al. specified extrahepatic disease as an exclusion criterion (Hendlisz et al. 2010; Bester et al. 2012; Seidensticker et al. 2012; Yoshino et al. 2012).

Table 4 Summary of prior systemic therapies

Due to the limited data available on response and progression in the SIRT studies, only results for adverse event incidence and OS are presented.

Adverse events

The percentages of patients reporting symptoms or non-laboratory signs of a grade 1 or 2 adverse event (with a prevalence > 5% in at least two studies) is shown in Fig. 1, with the corresponding percentages for grade 3, 4, or 5 events (with a prevalence > 2% in at least two studies) presented in Fig. 2. There was some evidence of increased rates of diarrhea with TAS-102 and regorafenib compared with SIRT and increased rates of nausea and vomiting with TAS-102 compared with regorafenib and SIRT. Regorafenib was also associated with increased rates of hand–foot syndrome, hypertension, and rash or desquamation. Radioembolization-induced liver disease (REILD) was only reported in the Bester et al. and Seidensticker et al. studies in which the incidence rates at grade 3+ were 0.3% and 10.3%, respectively. The only other potentially SIRT-related adverse event (as defined in the present study) that occurred in > 5% of patients was grade 1–2 GI ulcer, which occurred in 10.3% of patients in the Seidensticker et al. study. Reporting of laboratory test abnormalities was highly inconsistent throughout the evidence base and events were not, therefore, aggregated or analyzed for comparison.

Fig. 1
figure1

Summary of grade 1 or 2 adverse events occurring in > 5% of patients in two or more studies. BSC best supportive care, GI gastrointestinal, REILD radioembolization-induced liver disease, SIRT selective internal radiation therapy. Note: areas of circles represent the proportion of study subjects reporting specific symptoms

Fig. 2
figure2

Summary of grade 3, 4, and 5 adverse events occurring in > 2% of patients in two or more studies

Overall survival

The reported median survival for each study arm is shown in Fig. 3. In all studies, median survival is less than 12 months and SIRT, TAS-102, or regorafenib is each associated with prolonged survival (median OS: 6.4–11.9 months) compared to non-active, BSC (median OS: 3.5–7.3 months). The longest survival on BSC was observed in the Hendlisz 2010 trial of SIRT, in which patients in the control group were allowed to cross-over to active treatment after progression for ethical reasons (Hendlisz et al. 2010).

Fig. 3
figure3

Median overall survival (months) in each study arm. BSC best supportive care, OS overall survival, SIRT selective internal radiation therapy

Brooks–Gelman–Rubin, trace, and marginal density plots showed that the NMA converged on a solution within the 20,000 iterations after the burn-in period (Supplementary Fig. 2 and 3). The results of the fixed effects NMA are shown in Fig. 4 and the results of the fixed and random effects pairwise meta-analyses, and the random effects NMA are shown in Fig. 5. In the pairwise fixed effects meta-analyses, all active treatments were shown to improve OS, with mortality hazard ratios of 0.47 (95% CrI 0.34–0.65) for SIRT with Y-90 resin microspheres, 0.66 (0.56–0.77) for TAS-102, and 0.71 (0.60–0.83) for regorafenib each compared to BSC. Hazard ratios were similar in the pairwise random effects meta-analyses, at 0.49 (95% CrI 0.25–0.93) for SIRT, 0.66 (0.56–0.77) for TAS-102, and 0.67 (0.48–0.93) for regorafenib, each compared to BSC. Rank probabilities from the random effects NMA showed that there was a 75% probability of SIRT with Y-90 resin microspheres being the most effective treatment, compared with 15% for TAS-102, and 10% for regorafenib (Fig. 6); there was a 0% probability that BSC would be the optimal treatment in terms of OS.

Fig. 4
figure4

Overall results from the fixed effects network meta-analysis showing the mortality hazard ratio for each treatment relative to a reference of 1 with best supportive care. BSC best supportive care, HR hazard ratio, SIRT selective internal radiation therapy

Fig. 5
figure5

Individual study results for the random effects network meta-analysis. BSC best supportive care, HR hazard ratio, OS overall survival, SIRT selective internal radiation therapy

Fig. 6
figure6

Probability of treatment rankings based on the random effects network meta-analysis. BSC best supportive care, SIRT selective internal radiation therapy

Sensitivity analyses in which the variance around the OS hazard ratios from the two non-randomized studies of SIRT was inflated showed that the 95% CrI for SIRT relative to BSC only crossed 1 when the variance was four times greater than in the base case analysis (wj = 0.25; Supplementary Fig. 4). With a fourfold increase in variance around the hazard ratios from non-randomized studies, the mean hazard ratio for OS with SIRT relative to BSC was 0.59 (95% CrI 0.32–1.06).

While SIRT was ranked first based on OS, with TAS-102 second and regorafenib third (Fig. 6), there was substantial uncertainty as to the ranking of treatments due to the heterogeneity in study designs and the variation in comparator treatments. Although results of the network meta-analysis must be considered with caution, they demonstrate that all three active treatments are associated with a statistically significant reduction in the mortality risk compared to BSC.

Discussion

The present study is the first meta-analysis of SIRT, regorafenib, and TAS-102 for patients with chemotherapy-refractory or -intolerant mCRC. Previous meta-analyses have observed similar results for TAS-102 and regorafenib, at a starting dose of 160 mg/day, with no statistically significant difference in OS between the two drug regimens (Abrahao et al. 2018). Although both drugs are associated with increased incidence of treatment-emergent grade 3–4 AEs, the types of AEs associated with each drug are different. Regorafenib is frequently associated with AEs commonly reported for tyrosine kinase inhibitors, such as diarrhea, hand–foot skin reaction, or hypertension, while TAS-102 is more frequently associated with hematological abnormalities, as also evidenced in the present review. Regorafenib was reportedly associated with higher toxicity than TAS-102 overall (Abrahao et al. 2018).

The recently published ReDOS trial of regorafenib, comparing standard administration at a starting dose of 160 mg/day versus dose escalation at a starting dose of 80 mg/day, reported increased OS for the dose escalation strategy with lower incidence of grade 3 adverse events commonly associated with regorafenib (Bekaii-Saab et al. 2019). Based on published meta-analyses and outcomes of the present NMA, it is uncertain whether this OS increase would match the magnitude of clinical benefit observed for SIRT with Y-90 resin microspheres (Sonbol et al. 2019).

A key limitation of the present review was that the included studies of SIRT with Y-90 resin microspheres recruited exclusively subjects with colorectal metastases to the liver (with or without other extrahepatic disease), consistent with the liver-directed nature of the intervention, while studies of systemic treatment with regorafenib or TAS-102 included patients with any metastatic location. This difference in study design warrants caution when comparing median OS outcomes across the studies. This exploratory NMA was undertaken comparing HRs for OS, which are expected to capture differences in baseline characteristics across studies, as patients in the control group of each study of SIRT also have liver-dominant or liver-only metastatic disease. Nevertheless, it is worth noting that the median OS for patients receiving BSC was comparable in all studies, regardless of these differences in selection criteria, and that the lowest median OS was observed in the BSC group of the Seidensticker et al. 2012 study which only included patients with liver-dominant metastases (Seidensticker et al. 2012). This may reflect the prognostic importance of the liver and of liver metastases for patients with chemotherapy-refractory or -intolerant mCRC (Hendlisz et al. 2010).

Although the potential for confounding due to sample selection bias in the observational studies of SIRT included in this review should be noted, OS outcomes were consistent with those reported in large single-arm observational studies of regorafenib, TAS-102 and SIRT using Y-90 resin microspheres (de Groot J 2018; Andersen et al. 2019; Golfieri et al. 2015; Cosimelli et al. 2010; Kennedy et al. 2015; Saxena et al. 2015; Tohme et al. 2014; Lahti et al. 2015; Fendler et al. 2013; Maleux et al. 2016; Sofocleous et al. 2015; Nace et al. 2011). In ten identified studies of SIRT using Y-90 resin microspheres published between 2010 and 2019 which enrolled 50 or more patients with chemotherapy-refractory or -intolerant, liver-dominant colorectal metastases, median OS was between 6.9 and 13.8 months (median 10.2 months, pooled mean 9.9 months, and N = 1476). Sensitivity analysis also showed the findings to be insensitive to substantial increases in the modeled variance around the results from the non-randomized studies, with a fivefold increase in variance still associated with an average OS hazard ratio of 0.62 with SIRT relative to BSC. Other, more sophisticated techniques could have been employed to investigate the effect of including non-randomized studies in the analysis such as third-level hierarchical Bayesian modeling or utilizing data from randomized studies to modify the priors; (Efthimiou et al. 2017; Cameron et al. 2015); however, given that the non-randomized study data were confined to the single network edge comparing SIRT versus BSC, these techniques would have been unlikely to provide further insights into any bias arising from the non-randomized study designs.

Regorafenib and TAS-102 are recommended in ESMO, NCCN, French Intergroup, and Spanish consensus clinical guidelines for patients failing fluoropyrimidine-, oxaliplatin-, and irinotecan-based first- and second-line chemotherapy regimens (Van Cutsem et al. 2016; Benson et al. 2020; Phelip et al. 2019; Vera et al. 2019). SIRT using Y-90 resin microspheres is recommended in the ESMO and French guidelines in a similar setting, for patients with liver-only or liver-dominant colorectal metastases who are refractory or intolerant to chemotherapy, while NCCN guidelines are less specific about the recommended position for SIRT in the therapeutic strategy. These guidelines and the OS outcomes of the present review support offering patients presenting with colorectal liver metastases refractory to first- and second-line chemotherapy treatment with either systemic therapy options or SIRT with Y-90 resin microspheres depending on clinician and patient preference. In considering the mCRC armamentarium, it is worth mentioning the options available to specific sub-groups of patients with mCRC. Patients with BRAF V600E mutant mCRC (approximately 10% of mCRC cases) may be suitable for BRAF targeted therapies, such as encorafenib in combination with cetuximab, which received FDA in this indication after prior therapy in April 2020 (US Food and Drug Administration 2020; Kopetz et al. 2019). Numerous other BRAF pathway inhibitors, including dabrafenib, vemurafenib, and trametinib, are currently undergoing trials either alone or in combination with EGFR and/or MEK inhibitors in patients with BRAF V600E mutant mCRC (Ducreux et al. 2019). Moreover, patients with mismatch repair deficient (dMMR) and/or microsatellite instability-high (MSI-H) (approximately 3–5% of mCRC cases) may benefit from immune checkpoint inhibitors such as pembrolizumab, or nivolumab either alone or in combination with ipilimumab, all of which were approved by the FDA in 2017–2018 for use in patients with dMMR/MSI-H mCRC having progressed following treatment with fluoropyrimidine, oxaliplatin, and irinotecan (Kamatham et al. 2019; Le et al. 2020; Overman et al. 2018). Enrollment of patients in trials of these and other agents may also be a viable option for patients with advanced mCRC refractory to multiple lines of chemotherapy, with the NCCN specifically noting in the CRC guidelines their belief that “the best management for any patient with cancer is in a clinical trial” (Benson et al. 2020).

With regard to the sequencing of locoregional therapy relative to systemic therapies, Jeyarajah et al. recently published the findings of a Delphi panel of experienced practitioners, including surgical oncologists, transplant surgeons, and hepatopancreatobiliary surgeons, concluding that SIRT with Y-90 microspheres may be effective at multiple points in the algorithm of liver-dominant mCRC management, including the complete treatment of small metastases, as first-line therapy for liver metastases either alone or in combination with chemotherapy, in combination with second- or third-line chemotherapy, and as salvage therapy for chemotherapy-refractory patients. The authors further recommended considering the various positions in therapy alongside the principle that SIRT should be introduced in the treatment algorithm to control liver tumor progression before the liver has been damaged severely by chemotherapy (Jeyarajah et al. 2020).

From a clinical perspective, the adverse event profiles of each agent, along with patient performance status, are likely to determine treatment choice (Argiles et al. 2019). While ECOG performance statuses 1–2 are considered negative prognostic factors for each intervention, the adverse event profile of SIRT using Y-90 resin microspheres is superior to both regorafenib and TAS-102, with a low incidence of all-grade or grade 3–4 AEs known to severely affect patient quality of life, such as diarrhea, hand–foot skin reaction, vomiting, or fatigue. The adverse event profile of SIRT is well established, and risk of complications from the procedure can be reduced with adequate treatment planning and patient selection (Sangro et al. 2017). SIRT may, therefore, be preferred in selected patients with colorectal liver metastases refractory to first- and second-line chemotherapy, especially when the AE profile of TAS-102 and regorafenib represents a disincentive for treatment with these agents. The incidence of the five potentially SIRT-related AEs recorded in the present study was low, with only grade 3+ REILD and grade 1–2 GI ulcer showing incidence rates above 5% in one study each.

Liver-related laboratory test abnormalities were not reported in a consistent manner across the included studies, and a meaningful analysis or presentation was not possible. A previous study of 606 patients with mCRC undergoing SIRT showed that a high proportion of patients had mild-to-moderate (mostly grade 1 or 2) baseline laboratory abnormalities prior to SIRT with Y-90 resin microspheres, including alkaline phosphatase, AST, albumin, and hemoglobin (Kennedy et al. 2015). While the study showed clinically significant increases in severe (grade 3 and 4) laboratory test values for total bilirubin, albumin, alkaline phosphatase, and aspartate aminotransferase after SIRT with Y-90 microspheres, all incidence rates were below 10% 90 days after treatment (Kennedy et al. 2015).

In indications such as mCRC in which patients are facing a poor prognosis with limited therapeutic options, the availability of locoregional options such as SIRT in addition to regorafenib and TAS-102 addresses a great unmet clinical need, especially in RAS mutant patients for whom anti-EGFR agents such as cetuximab or panitumumab are ineffective (Van Cutsem et al. 2015). Effective physician–patient communication is an essential element in ensuring appropriate treatment selection, and for setting expectations for treatment outcomes and adverse event incidence. As both regorafenib and TAS-102 are orally administered, patient adherence to the choice of treatment is also essential (Argiles et al. 2019). From the patient perspective, SIRT using Y-90 resin microspheres represents a relevant alternative with lower administration burden for the patient and no degradation of quality of life (Cosimelli et al. 2010). The adverse event profile of SIRT using Y-90 resin microspheres has also been established across the non-comparative evidence base in this indication for patients intolerant to chemotherapy and/or patients aged ≥ 70 years (Golfieri et al. 2015; Cosimelli et al. 2010; Kennedy et al. 2015; Saxena et al. 2015; Tohme et al. 2014; Lahti et al. 2015; Fendler et al. 2013; Maleux et al. 2016; Sofocleous et al. 2015; Nace et al. 2011).

Conclusions

This systematic review and exploratory NMA of regorafenib, TAS-102, and SIRT using Y-90 resin microspheres in third-line mCRC demonstrate that all interventions are effective compared to BSC. Although SIRT was shown to have a greater OS benefit than systemic therapy for selected patients with liver-dominant colorectal metastases, this particular result needs to be interpreted with caution due to the differences in study design across interventions.

The improved adverse event profile of SIRT using Y-90 resin microspheres over systemic therapy should be considered in the physician and patient therapeutic decision-making process. Further research is needed to establish the relative costs of these interventions, and to determine the cost-effectiveness and budget impact of SIRT, regorafenib, and TAS-102 versus BSC.

References

  1. Abrahao ABK, Ko Y-J, Berry S, Chan KKW (2018) A comparison of regorafenib and TAS-102 for metastatic colorectal cancer: a systematic review and network meta-analysis. Clin Colorectal Cancer 17(2):113–120

    PubMed  Google Scholar 

  2. Abrams TA, Meyer G, Schrag D et al (2014) Chemotherapy usage patterns in a US-wide cohort of patients with metastatic colorectal cancer. J Natl Cancer Inst 106(2):371

    Google Scholar 

  3. Adam R, De Gramont A, Figueras J et al (2012) The oncosurgery approach to managing liver metastases from colorectal cancer: a multidisciplinary international consensus. Oncologist 17(10):1225–1239

    CAS  PubMed  PubMed Central  Google Scholar 

  4. Andersen SE, Andersen IB, Jensen BV, Pfeiffer P, Ota T, Larsen JS (2019) A systematic review of observational studies of trifluridine/tipiracil (TAS-102) for metastatic colorectal cancer. Acta Oncol 58(8):1149–1157

    CAS  PubMed  Google Scholar 

  5. Argiles G, Arnold D, Prager G, Sobrero AF, Van Cutsem E (2019) Maximising clinical benefit with adequate patient management beyond the second line in mCRC. ESMO Open 4(2):e000495

    PubMed  PubMed Central  Google Scholar 

  6. Bayer AG (2018) STIVARGA® summary of product characteristics. European Medicines Agency, Amsterdam. https://www.ema.europa.eu/en/documents/product-information/stivarga-epar-product-information_en.pdf. Accessed 28 Oct 2019

  7. Bekaii-Saab TS, Ou F-S, Ahn DH, Boland PM, Ciombor KK, Heying EN et al (2019) Regorafenib dose-optimisation in patients with refractory metastatic colorectal cancer (ReDOS): a randomised, multicentre, open-label, phase 2 study. Lancet Oncol 20(8):1070–1082

    CAS  PubMed  Google Scholar 

  8. Benson AB, Venook AP, Al-Hawary MM, Mustafa A, Chen YJ, Ciombor KK et al (2020) NCCN clinical practice guidelines in oncology: colon cancer, version 3. https://www.nccn.org/professionals/physician_gls/pdf/colon.pdf. Accessed 3 June 2020

  9. Bester L, Meteling B, Pocock N, Pavlakis N, Chua TC, Saxena A et al (2012) Radioembolization versus standard care of hepatic metastases: comparative retrospective cohort study of survival outcomes and adverse events in salvage patients. J Vasc Interv Radiol 23(1):96–105

    PubMed  Google Scholar 

  10. Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A (2018) Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 68(6):394–424

    PubMed  PubMed Central  Google Scholar 

  11. Brooks SP, Gelman A (1998) General methods for monitoring convergence of iterative simulations. J Comput Graph Stat. https://www.tandfonline.com/doi/abs/10.1080/10618600.1998.10474787. Accessed 17 June 2019

  12. Cameron C, Fireman B, Hutton B, Clifford T, Coyle D, Wells G, Dormuth CR, Platt R, Toh S (2015) Network meta-analysis incorporating randomized controlled trials and non-randomized comparative cohort studies for assessing the safety and effectiveness of medical treatments: challenges and opportunities. Syst Rev 4:147

    PubMed  PubMed Central  Google Scholar 

  13. Cosimelli M, Golfieri R, Cagol PP, Carpanese L, Sciuto R, Maini CL et al (2010) Multi-centre phase II clinical trial of yttrium-90 resin microspheres alone in unresectable, chemotherapy refractory colorectal liver metastases. Br J Cancer 103(3):324–331

    CAS  PubMed  PubMed Central  Google Scholar 

  14. Ducreux M, Petersen L, Öhler L, Bergamo F, Metges J, de Groot J et al (2018) O-012: safety and effectiveness of regorafenib in patients with metastatic colorectal cancer (mCRC) in routine clinical practice: final analysis from the prospective, observational CORRELATE study. Ann Oncol 29(s5):104

    Google Scholar 

  15. Ducreux M, Chamseddine A, Laurent-Puig P, Smolenschi C, Hollebecque A, Dartigues P, Samallin E, Boige V, Malka D, Gelli M (2019) Molecular targeted therapy of BRAF-mutant colorectal cancer. Ther Adv Med Oncol 11:1758835919856494

    CAS  PubMed  PubMed Central  Google Scholar 

  16. Efthimiou O, Mavridis D, Debray TP, Samara M, Belger M, Siontis GC, Leucht S, Salanti G (2017) GetReal work package 4. Combining randomized and non-randomized evidence in network meta-analysis. Stat Med 36(8):1210–1226

    PubMed  Google Scholar 

  17. Fendler WP, Tiega DBP, Ilhan H, Paprottka PM, Heinemann V, Jakobs TF et al (2013) Validation of several SUV-based parameters derived from 18F-FDG PET for prediction of survival after SIRT of hepatic metastases from colorectal cancer. J Nucl Med 54(8):1202–1208

    CAS  PubMed  Google Scholar 

  18. Ferlay J, Ervik M, Lam F, Colombet M, Mery L, Piñeros M, Znaor A, Soerjomataram I, Bray F (2019) Global cancer observatory: cancer tomorrow. International Agency for Research on Cancer, Lyon. https://gco.iarc.fr/tomorrow. Accessed 18 Oct 2019

  19. Foubert F, Matysiak-Budnik T, Touchefeu Y (2014) Options for metastatic colorectal cancer beyond the second line of treatment. Dig Liver Dis 46(2):105–112

    PubMed  Google Scholar 

  20. Golfieri R, Mosconi C, Giampalma E, Cappelli A, Galaverni MC, Pettinato C et al (2015) Selective transarterial radioembolisation of unresectable liver-dominant colorectal cancer refractory to chemotherapy. Radiol Med (Torino) 120(8):767–776

    Google Scholar 

  21. Grothey A, Cutsem EV, Sobrero A, Siena S, Falcone A, Ychou M et al (2013) Regorafenib monotherapy for previously treated metastatic colorectal cancer (CORRECT): an international, multicentre, randomised, placebo-controlled, phase 3 trial. Lancet 381(9863):303–312

    CAS  PubMed  Google Scholar 

  22. Hawkins N, Scott DA, Woods B (2016) ‘Arm-based’ parameterization for network meta-analysis. Res Synth Methods 7(3):306–313

    PubMed  Google Scholar 

  23. Hendlisz A, den Eynde MV, Peeters M, Maleux G, Lambert B, Vannoote J et al (2010) Phase III trial comparing protracted intravenous fluorouracil infusion alone or with Yttrium-90 resin microspheres radioembolization for liver-limited metastatic colorectal cancer refractory to standard chemotherapy. J Clin Oncol 28(23):3687–3694

    CAS  PubMed  Google Scholar 

  24. Jeyarajah DR, Doyle MBM, Espat NJ, Hansen PD, Iannitti DA, Kim J, Thambi-Pillai T, Visser BC (2020) Role of yttrium-90 selective internal radiation therapy in the treatment of liver-dominant metastatic colorectal cancer: an evidence-based expert consensus algorithm. J Gastrointest Oncol 11(2):443–460

    PubMed  PubMed Central  Google Scholar 

  25. Kamatham S, Shahjehan F, Kasi PM (2019) Immune checkpoint inhibitors in metastatic colorectal cancer: current status, recent advances, and future directions. Curr Colorectal Cancer Rep 15:112–121

    Google Scholar 

  26. Kennecke H, Yu J, Gill S et al (2014) Effect of M1a and M1b category in metastatic colorectal cancer. Oncologist 19(7):720–726

    PubMed  PubMed Central  Google Scholar 

  27. Kennedy AS, Ball D, Cohen SJ, Cohn M, Coldwell DM, Drooz A et al (2015) Multicenter evaluation of the safety and efficacy of radioembolization in patients with unresectable colorectal liver metastases selected as candidates for 90Y resin microspheres. J Gastrointest Oncol 6(2):134–142

    PubMed  PubMed Central  Google Scholar 

  28. Kopetz S, Grothey A, Yaeger R, Van Cutsem E, Desai J, Yoshino T, Wasan H, Ciardiello F, Loupakis F, Hong YS, Steeghs N, Guren TK, Arkenau HT, Garcia-Alfonso P, Pfeiffer P, Orlov S, Lonardi S, Elez E, Kim TW, Schellens JHM, Guo C, Krishnan A, Dekervel J, Morris V, Calvo Ferrandiz A, Tarpgaard LS, Braun M, Gollerkeri A, Keir C, Maharry K, Pickard M, Christy-Bittel J, Anderson L, Sandor V, Tabernero J (2019) Encorafenib, binimetinib, and cetuximab in BRAF V600E-mutated colorectal cancer. N Engl J Med 381(17):1632–1643

    CAS  PubMed  Google Scholar 

  29. Kumar R, Jain K, Beeke C et al (2013) A population-based study of metastatic colorectal cancer in individuals aged ≥ 80 years: findings from the South Australian Clinical Registry for Metastatic Colorectal Cancer. Cancer 119(4):722–728

    PubMed  Google Scholar 

  30. Kumar R, Price TJ, Beeke C et al (2014) Colorectal cancer survival: an analysis of patients with metastatic disease synchronous and metachronous with the primary tumor. Clin Colorectal Cancer 13(2):87–93

    PubMed  Google Scholar 

  31. Lahti SJ, Xing M, Zhang D, Lee JJ, Magnetta MJ, Kim HS (2015) KRAS status as an independent prognostic factor for survival after Yttrium-90 radioembolization therapy for unresectable colorectal cancer liver metastases. J Vasc Interv Radiol 26(8):1102–1111

    PubMed  Google Scholar 

  32. Le DT, Kim TW, Van Cutsem E, Geva R, Jäger D, Hara H, Burge M, O’Neil B, Kavan P, Yoshino T, Guimbaud R, Taniguchi H, Elez E, Al-Batran SE, Boland PM, Crocenzi T, Atreya CE, Cui Y, Dai T, Marinello P, Diaz LA Jr, André T (2020) Phase II open-label study of pembrolizumab in treatment-refractory, microsatellite instability-high/mismatch repair-deficient metastatic colorectal cancer: KEYNOTE-164. J Clin Oncol 38(1):11–19

    PubMed  Google Scholar 

  33. Les Laboratoires Servier (2016) LONSURF® summary of product characteristics. European Medicines Agency, Amsterdam. https://www.ema.europa.eu/en/documents/product-information/lonsurf-epar-product-information_en.pdf. Accessed 28 Oct 2019

  34. Li J, Qin S, Xu R, Yau TCC, Ma B, Pan H et al (2015) Regorafenib plus best supportive care versus placebo plus best supportive care in Asian patients with previously treated metastatic colorectal cancer (CONCUR): a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet Oncol 16(6):619–629

    CAS  PubMed  Google Scholar 

  35. Maleux G, Deroose C, Laenen A, Verslype C, Heye S, Haustermans K et al (2016) Yttrium-90 radioembolization for the treatment of chemorefractory colorectal liver metastases: technical results, clinical outcome and factors potentially influencing survival. Acta Oncol Stockh Swed 55(4):486–495

    CAS  Google Scholar 

  36. Mauri G, Pizzutilo EG, Amatu A, Bencardino K, Palmeri L, Bonazzina EF, Tosi F, Carlo Stella G, Burrafato G, Scaglione F, Marsoni S, Siravegna G, Bardelli A, Siena S, Sartore-Bianchi A (2019) Retreatment with anti-EGFR monoclonal antibodies in metastatic colorectal cancer: systematic review of different strategies. Cancer Treat Rev 73:41–53

    CAS  PubMed  Google Scholar 

  37. Mayer RJ, Van Cutsem E, Falcone A, Yoshino T, Garcia-Carbonero R, Mizunuma N et al (2015) Randomized trial of TAS-102 for refractory metastatic colorectal cancer. N Engl J Med 372(20):1909–1919

    PubMed  Google Scholar 

  38. Mekenkamp LJM, Koopman M, Teerenstra S et al (2010) Clinicopathological features and outcome in advanced colorectal cancer patients with synchronous vs metachronous metastases. Br J Cancer 103(2):159–164

    CAS  PubMed  PubMed Central  Google Scholar 

  39. Nace GW, Steel JL, Amesur N, Zajko A, Nastasi BE, Joyce J et al (2011) Yttrium-90 radioembolization for colorectal cancer liver metastases: a single institution experience. Int J Surg Oncol 2011:e571261

    Google Scholar 

  40. Oh SY, Kim DY, Suh KW (2015) Oncologic outcomes following metastasectomy in colorectal cancer patients developing distant metastases after initial treatment. Ann Surg Treat Res 88(5):253–259

    PubMed  PubMed Central  Google Scholar 

  41. Overman MJ, Lonardi S, Wong KYM, Lenz HJ, Gelsomino F, Aglietta M, Morse MA, Van Cutsem E, McDermott R, Hill A, Sawyer MB, Hendlisz A, Neyns B, Svrcek M, Moss RA, Ledeine JM, Cao ZA, Kamble S, Kopetz S, André T (2018) Durable clinical benefit with nivolumab plus ipilimumab in DNA mismatch repair-deficient/microsatellite instability-high metastatic colorectal cancer. J Clin Oncol 36(8):773–779

    CAS  PubMed  Google Scholar 

  42. Phelip JM, Tougeron D, Léonard D, Benhaim L, Desolneux G, Dupré A et al (2019) Metastatic colorectal cancer (mCRC): french intergroup clinical practice guidelines for diagnosis, treatments and follow-up (SNFGE, FFCD, GERCOR, UNICANCER, SFCD, SFED, SFRO, SFR). Dig Liver Dis 51(10):1357–1363

    PubMed  Google Scholar 

  43. Plummer M (2003) JAGS: a program for analysis of Bayesian graphical models using Gibbs sampling. In: Proceedings of the 3rd International Workshop on Distributed Statistical Computing (DSC 2003), March 20-22, Vienna, Austria. ISSN 1609-395X

  44. Sadahiro S, Suzuki T, Tanaka A et al (2013) Hematogenous metastatic patterns of curatively resected colon cancer were different from those of stage IV and autopsy cases. Jpn J Clin Oncol 43(4):444–447

    PubMed  Google Scholar 

  45. Sangro B, Martínez-Urbistondo D, Bester L, Bilbao JI, Coldwell DM, Flamen P et al (2017) Prevention and treatment of complications of selective internal radiation therapy: expert guidance and systematic review. Hepatol Baltim Md. 66(3):969–982

    Google Scholar 

  46. Saxena A, Meteling B, Kapoor J, Golani S, Morris DL, Bester L (2015) Is yttrium-90 radioembolization a viable treatment option for unresectable, chemorefractory colorectal cancer liver metastases? A large single-center experience of 302 patients. Ann Surg Oncol 22(3):794–802

    PubMed  Google Scholar 

  47. Seidensticker R, Denecke T, Kraus P, Seidensticker M, Mohnike K, Fahlke J et al (2012) Matched-pair comparison of radioembolization plus best supportive care versus best supportive care alone for chemotherapy refractory liver-dominant colorectal metastases. Cardiovasc Interv Radiol 35(5):1066–1073

    Google Scholar 

  48. Sofocleous CT, Violari EG, Sotirchos VS, Shady W, Gonen M, Pandit-Taskar N et al (2015) Radioembolization as a salvage therapy for heavily pretreated patients with colorectal cancer liver metastases: factors that affect outcomes. Clin Colorectal Cancer 14(4):296–305

    PubMed  PubMed Central  Google Scholar 

  49. Sonbol MB, Benkhadra R, Wang Z, Firwana B, Walden DJ, Mody K et al (2019) A systematic review and network meta-analysis of regorafenib and TAS-102 in refractory metastatic colorectal cancer. Oncologist 24(9):1174–1179

    CAS  PubMed  PubMed Central  Google Scholar 

  50. Tohme S, Sukato D, Nace GW, Zajko A, Amesur N, Orons P et al (2014) Survival and tolerability of liver radioembolization: a comparison of elderly and younger patients with metastatic colorectal cancer. HPB 16(12):1110–1116

    PubMed  PubMed Central  Google Scholar 

  51. Turner RM, Davey J, Clarke MJ, Thompson SG, Higgins JPT (2012) Predicting the extent of heterogeneity in meta-analysis, using empirical data from the Cochrane Database of Systematic Reviews. Int J Epidemiol 41(3):818–827

    PubMed  PubMed Central  Google Scholar 

  52. US Food and Drug Administration (2020) FDA approves encorafenib in combination with cetuximab for metastatic colorectal cancer with a BRAF V600E mutation. https://www.fda.gov/drugs/resources-information-approved-drugs/fda-approves-encorafenib-combination-cetuximab-metastatic-colorectal-cancer-braf-v600e-mutation. Accessed 3 June 2020

  53. Van Cutsem E, Lenz HJ, Köhne CH, Heinemann V, Tejpar S, Melezínek I, Beier F, Stroh C, Rougier P, van Krieken JH, Ciardiello F (2015) Fluorouracil, leucovorin, and irinotecan plus cetuximab treatment and RAS mutations in colorectal cancer. J Clin Oncol 33:692–700

    PubMed  Google Scholar 

  54. Van Cutsem E, Cervantes A, Adam R, Sobrero A, Van Krieken JH, Aderka D et al (2016) ESMO consensus guidelines for the management of patients with metastatic colorectal cancer. Ann Oncol 27(8):1386–1422

    PubMed  Google Scholar 

  55. van der Geest LG, Lam-Boer J, Koopman M et al (2015) Nationwide trends in incidence, treatment and survival of colorectal cancer patients with synchronous metastases. Clin Exp Metastasis 32(5):457–465

    PubMed  Google Scholar 

  56. van der Pool AE, Damhuis RA, Ijzermans JN et al (2012) Trends in incidence, treatment and survival of patients with stage IV colorectal cancer: a population-based series. Colorectal Dis 14(1):56–61

    PubMed  Google Scholar 

  57. Vera R, González-Flores E, Rubio C, Urbano J, Valero Camps M, Ciampi-Dopazo JJ, Orcajo Rincón J, Morillo Macías V, Gomez Braco MA, Suarez-Artacho G (2019) Multidisciplinary management of liver metastases in patients with colorectal cancer: a consensus of SEOM, AEC, SEOR, SERVEI, and SEMNIM. Clin Transl Oncol. https://doi.org/10.1007/s12094-019-02182-z

    Article  PubMed  Google Scholar 

  58. Yoshino T, Mizunuma N, Yamazaki K, Nishina T, Komatsu Y, Baba H et al (2012) TAS-102 monotherapy for pretreated metastatic colorectal cancer: a double-blind, randomised, placebo-controlled phase 2 trial. Lancet Oncol 13(10):993–1001

    CAS  PubMed  Google Scholar 

  59. Zafar SY, Marcello JE, Wheeler JL et al (2009) Longitudinal patterns of chemotherapy use in metastatic colorectal cancer. J Oncol Pract 5(5):228–233

    PubMed  PubMed Central  Google Scholar 

Download references

Funding

This study was funded by Sirtex Medical United Kingdom Ltd.

Author information

Affiliations

Authors

Corresponding author

Correspondence to Richard F. Pollock.

Ethics declarations

Conflict of interest

TW has no conflicts of interest to declare; NH is a partner of Visible Analytics Ltd, which received consultancy fees, research grant, and expenses from Sirtex Medical United Kingdom Ltd (a subsidiary of Sirtex Medical Ltd. the manufacturer of SIR-Spheres Y-90 resin microspheres) to conduct the systematic literature review and network meta-analysis; RFP is a director and shareholder of Covalence Research Ltd, which received consultancy fees from Sirtex Medical United Kingdom Ltd, to conduct sensitivity analyses, generate tables and figures, and write the manuscript; FC is a full-time employee of Sirtex Medical United Kingdom Ltd; SS is a director and full-time employee of Sirtex Medical United Kingdom Ltd; PR reports honoraria from Amgen, Servier, Shire, Sirtex, BMS, Pierre Fabre, and Roche, participation in advisory board meetings for Shire, Celgene, BMS, Pierre Fabre, receipt of travel support from Amgen, BMS, Bayer, Roche, and Servier, and his institution has received research funding from Sanofi.

Ethical approval

This article does not contain any studies with human participants or animals performed by any of the authors.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary material 1 (DOCX 1036 kb)

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Walter, T., Hawkins, N.S., Pollock, R.F. et al. Systematic review and network meta-analyses of third-line treatments for metastatic colorectal cancer. J Cancer Res Clin Oncol 146, 2575–2587 (2020). https://doi.org/10.1007/s00432-020-03315-6

Download citation

Keywords

  • Colorectal cancer
  • Neoplasm metastasis
  • Meta-analysis