Incidence and risk of hematologic toxicities in cancer patients treated with regorafenib

Regorafenib, an oral vascular endothelial growth factor receptor tyrosine-kinase inhibitor, has been approved for the treatment of several malignancies. As a non-traditional cytotoxic chemotherapeutic agent, regorafenib is often associated with hematologic toxicities. Here we searched PubMed and Embase up to June 2017 for relevant clinical trials. Eligible studies include trials in which subjects treated with 160 mg of regorafenib daily during the first 21 days of each 28-day cycle, and adequate safety data profile reporting thrombocytopenia, anemia, neutropenia and leucopenia. Statistical analyses were conducted to calculate the overall incidences, relative risks (RRs) and their 95% confidence intervals (CIs). A total of 2,341 subjects from 16 trials were included in the present studies. The incidences of regorafenib associated all-grade and high-grade hematologic toxicities were: thrombocytopenia, 22% and 3%; anemia, 20% and 3%; neutropenia, 10% and 2%, and leucopenia, 13% and 2%, respectively. Regorafenib-treated subjects had a significant increased risk of all-grade (RR=6.35; 95% CI, 3.19-12.64) and high-grade (RR=6.27; 95% CI, 1.69-23.26) thrombocytopenia, all-grade (RR=2.76; 95% CI, 1.63-4.68) and high-grade (RR=5.38; 95% CI, 1.60-18.06) anemia. Our results suggested that regorafenib therapy was associated with significantly increased risks of hematological toxicities, and hematologic monitoring at regular intervals should be advised to clinician.


INTRODUCTION
Tyrosine kinase inhibitors (TKIs) are small molecules that bind to the activation domain of tyrosine kinase receptors, and have emerged as an important kind of anti-cancer agents. Regorafenib (also referred as Stivarga, BAY 73-4506) can inhibit the activity of angiogenic, stromal and oncogenic tyrosine kinases by targeting vascular endothelial growth factor receptors 1, 2, 3 (VEGFR1, VEGFR2 and VEGFR3), tyrosine protein kinase receptor Ret, tyrosine-protein kinase TIE-2, basic fibroblast growth factor receptor-1, platelet-derived growth factor beta, proto-oncogene RAF-1, c-KIT, BRAF and p38 MAP kinase [1,2]. Currently, it has been approved by the United States Food and Drug Administration (FDA) for the treatment of metastatic colorectal cancer (CRC) [3], advanced gastrointestinal stromal tumor (GIST) [4] and recently, advanced hepatocellular carcinoma (HCC) [5].
Currently, regorafenib is being investigated in several types of tumors and an increase in the application of regorafenib could be expected in the near future. However, although hematologic toxicities associated with regorafenib have been reported in numerous studies, there has been no systematic attempt to synthesize these data and the overall risk of hematologic toxicities induced by regorafenib has yet to be assessed. Accordingly, here we conducted a systematic review and meta-analysis of available clinical studies to determine the overall incidence and risk of developing hematologic toxicities in subjects treated with regorafenib.

Population characteristics
A total of 2,341 subjects were included in this study (regorafenib: 1,735; control: 606). 1,466 of these patients had CRC (regorafenib: 1,145; control: 321) from 10 trials. 603 patients had HCC (regorafenib: 410; control: 193) from 2 trials. 75 subjects had GIST (regorafenib: 75; control: 0) from 2 trials. 182 patients had soft tissue sarcoma (STS; regorafenib: 90; control: 92) from 1 trial. The schedule and dose of regorafenib for all trials were 160 mg once daily orally for the first 21 days of each 28-day cycle, the current FDA-recommended dose until disease progression or unacceptable toxicity. The median treatment ranged from 1.6 months [38] to 10.0 months [34]. The clinic-pathological characteristics of eligible studies were summarized in Table 1. The numbers of allgrade and high-grade events for each trial were presented in Table 2. It should be noted that not all trials consistently reported the four hematologic adverse events of our interest.

Publication bias
We found no evidence of publication bias for RRs of both all-grade and high-grade thrombocytopenia, anemia, neutropenia and leucopenia by either Egger or the Begg test (p > 0.05).

DISCUSSION
To our knowledge, this is the first meta-analysis focusing specifically on hematologic toxicities associated with regorafenib. Our results revealed that the overall incidences of regorafenib-associated all-grade and highgrade (grade 3 and 4) hematologic toxicities, respectively: thrombocytopenia: 22% and 3%; anemia: 20% and 3%; neutropenia: 10% and 2% and leucopenia: 13% and 2%. Furthermore, our analysis from randomized controlled trials demonstrated a significantly increased risk of all- grade and high-grade thrombocytopenia and anemia with regorafenib treatment compared with control. Although not statistically significant, the risks of both all-grade and high-grade neutropenia and leucopenia inclined to increase in regorafenib-treated patients.
Traditional VEGF pathway targeted agents had been linked to a number of mechanism-driven toxicities such as hypertension [6], hepatic toxicity [7], handfoot skin reaction [8] and arterial thromboembolism [9]. Recently, plenty of data also implied the clinical association of hematological toxicities with these agents such as sunitinib [10], sorafenib [11], bevacizumab [12], although the frequency and severity vary among the different agents (Table 3). It had been demonstrated that a significantly increased risk of both high-grade and allgrade thrombocytopenia and neutropenia was discovered in sunitinib-treated subjects [10] and sorafenib-treated subjects [11]. Bevacizumab, despite not generally treated as a drug prone to cause hematologic toxicities, was also associated with an increased risk of all-grade thrombocytopenia and neutropenia and high-grade neutropenia in a meta-analysis [12]. However, Schutz et al. revealed that sorafenib was associated with a lower risk of high-grade anemia [11] and bevacizumab showed protective effects of both all-grade and high-grade anemia [12]. Our results failed to show a statistically increased risk of all-grade and high-grade neutropenia and leucopenia in regorafenib treated patients. These discrepancies were partly due to the differences in the mechanisms of action among these VEGF-targeted agents, the type of underlying malignancies, insufficient follow-up data and use of blood transfusions among different trials. It was noted that regorafenib had a biochemical structure similar to sorafenib differing only in the fluorine on the phenyl ring [1,41]. As showed in Table  3, regorafenib and sorafenib had similar high-grade incidence rates of hematologic toxicities, which were relatively lower compared with other anti-angiogenic agents, such as sunitinib. This was consistent with in vitro studies revealing that sunitinib had more activity against both c-KIT and FLT-3 kinases than other inhibitors [42]. However, based on our meta-analysis, regorafenib appeared to have a lower incidence rate of all-grade thrombocytopenia, neutropenia and anemia compared with sorafenib. Although the mechanisms underlying this difference had not been completely explained, it could not rule out that the structural dissimilarity between regorafenib and sorafenib resulted in the different inhibitory effects on angiogenesis related receptors such as VEGFR2 and fibroblast growth factor receptor 1 [1].
The observed hematologic toxicities associated with regorafenib treatment could be explained by the   [43]. FLT-3 was mainly expressed on committed myeloid, lymphoid precursors as well as the more mature monocytic lineage [44], and its activation played an important role in normal hematopoiesis and cellular growth [45]. FLT and c-KIT ligands, in combination with interleukin-3 (IL-3), had been discovered to regulate the proliferation of hematopoietic progenitor cells [46]. Previous studies had demonstrated that animals with FLT-3 knockout cells displayed a global disruption of hematopoiesis [43]. Our results were consistent with these pre-clinical results and supported the hypothesis that VEGF blockade could increase the risk of myelosuppression. The hematologic toxicities, especially thrombocytopenia and neutropenia, were one of the most common drug-related adverse events leading to treatment adjustment and discontinuation in clinical trials [28,30,31,34,38]. High-grade hematologic toxicities were usually clinically significant and required medical intervention. Our study showed that regorafenib-treated subjects had a higher risk of high-grade thrombocytopenia and neutropenia. Although not statistically significant because of limited trials involved in the present study, the relative risk of neutropenia and leucopenia intend to be higher in patients treated with regorafenib. These adverse events could potentially lead to overwhelming sepsis and hemorrhage in patients. In fact, some studies demonstrated fatal adverse events because of bleeding during regorafenib-based therapy [37,38]. Moreover, it was essential to point out that VEGFR inhibition could cause impairment in the coagulation and endothelial cell function without dysregulating the platelet count [47]. Since we did not have access to any individual patient data, we could not correlate the risks of infection and bleeding with thrombocytopenia and neutropenia. In fact, there were currently no methods to predict subjects at high risk and therefore regular monitoring of complete blood counts was needed. In clinic, because there were no established guidelines in the follow-up and treatment of regorafenib induced high-grade neutropenia and thrombocytopenia yet, temporary dose interruption or dose reduction to 50%-75% of original contents were conducted based on the severity and individual toleration [28,30,31,34].
Here we conducted a comprehensive review using the most up-to-date published data, which made our results more extensive and valid. In addition, with accumulating evidence and enlarged sample size, we had enhanced the statistical power to provide more precise and reliable estimates. However, our study was restricted by some limitations. First, this was a meta-analysis conducted at the trial level and no clinicopathological variables at the patient level could be analyzed. Second, pooled incidence rates had significant heterogeneities. It might be due to the different types of underlying malignancies, sample size, insufficient follow-up data among the included trials. Third, we could not determine the risk of regorafenibinduced hematologic toxicities in different regimens because of the small number of studies available for each regimen. Forth, we could not correlate our data with dose delays and/or interruptions or with hematologic support measures applied. Fifth, different version of Common Terminology Criteria for Adverse Events (CTCAEs) were applied in different trials. However, as far as we know, there were no differences among these versions in term of the definition of hematologic toxicities.
In conclusion, our meta-analysis revealed that regorafenib was associated with an increased risk of hematological toxicities. Clinical doctors should be acknowledged of these potential adverse events and hematologic monitoring at regular intervals might be advised.

MATERIALS AND METHODS
The present study was conducted in compliance with the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions and was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [48].
Other publications on the topic, including conference abstract, review articles, pre-clinical papers, editorials, early versions of data later published, articles not dealing with regorafenib were not included ( Figure 1).
Considering that recent studies with regorafenib might be unpublished, electronic searches were also conducted using the major international congresses' proceedings (American Society of Clinical Oncology Annual Meeting and European Society of Medical Oncology). Moreover, the reference lists of all studies fulfilling the eligibility criteria were further examined for any relevant studies missed by the electronic searches. When multiple publications of the same clinical trial appeared or if there was a case mix between different publications, only the most recent and/or most complete reporting study was included. Any discrepancies were settled by discussion and consensus.

Data extraction
Identified abstracts were collected and full texts of potentially relevant studies were reviewed for the trial design and reporting of hematologic toxicities. The following items were extracted from every study: full name of the first author, region, year of publication, underlying malignancy, median follow-up, number of patients for analysis, median age, gender, European cooperative oncology group performance status (ECOG PS), median treatment duration, median overall survival, median progression-free survival ( Table 1), number of events of the following adverse events (both all-grade and high-grade): thrombocytopenia, anemia, neutropenia and leucopenia ( Table 2). All the reviewers discussed and resolved any discrepancies in the extracted information. The number of subjects evaluated for toxicity was used as the number analyzed for each study, unless it was indicated otherwise. When studies using crossover designs were described, only data available from before the crossover were applied. In cases where this was not available, those trials were not included.

Statistical analysis
The primary analysis investigated the incidence, relative risk and corresponding 95% CI of all-grade and high-grade hematologic toxicities in cancer patients treated by regorafenib. To calculate the incidence, the number of subjects receiving regorafenib alone and the number of subjects with hematologic toxicities (both all-grade and high-grade) were extracted from the eligible single-arm and randomized controlled trials. The proportion of patients with thrombocytopenia, anemia, neutropenia and leucopenia and 95% CIs were derived from every study. We calculated both RRs and CIs with data extracted only from randomized controlled trials, comparing the incidence of each adverse event in subjects assigned to regorafenib with subjects assigned to control treatment. To calculate 95% CIs, the variance of a logtransformed study specific RR was derived by the delta method. Statistical heterogeneity between different trials and subgroups was assessed by Cochrane's Q statistic. The I 2 statistic was calculated to assess the extent of inconsistency contributable to the heterogeneity across different studies [49]. The assumption of homogeneity was considered invalid for I 2 > 25% or p<0.10. Summary RRs and incidences were calculated using fixed-effects or random-effects models depending on the heterogeneity of included trials. Potential publication bias was assessed by visual inspection of a funnel plot, and also evaluated using the tests of Egger et al. [50] and Begg et al. [51]. Twosided p <0.05 were considered statistically significant. All analysis was performed using Stata version 12.0 (StataCorp, USA).