Review of Key Evidence
We reviewed data from the main analyses of the phase III studies of the three drugs [16,17,18]; their efficacy is summarized in Table 1. The patient characteristics varied between the studies. Regarding the main histologic subtype, the proportions of patients with liposarcoma, leiomyosarcoma, and other histologic subtypes were 0%, 44.7%, and 55.3%, respectively, in the phase III study of pazopanib; 27%, 73%, and 0%, respectively, in the phase III study of trabectedin; and 33.8%, 65.7%, and less than 1%, respectively, in the phase III study of eribulin. Regarding the number of previous chemotherapy regimens for advanced STS defined in the inclusion criteria, for pazopanib the requirement was at least one regimen containing anthracycline and a maximum of four previous lines of systemic therapy for metastatic disease (no more than two lines of combination regimens); for trabectedin, the requirement was at least either a combination of an anthracycline and ifosfamide or an anthracycline plus one or more additional cytotoxic chemotherapy regimen(s); for eribulin, the requirement was at least two standard systemic regimens including an anthracycline. All studies included patients who had received the standard treatment for STS, namely an anthracycline (e.g., doxorubicin)-containing regimen, and the median number of previous chemotherapy regimens in the three studies was two.
Impact of Prolonged Overall Survival in Cancer Therapy
The aims of treatment for advanced STS are to reduce symptoms and prolong survival. Therefore, improvements in OS have a large clinical impact, particularly in cases of recurrent STS with a poor prognosis. Recent phase III studies for cancer therapy tend to have OS as the primary end point. In a phase III study designed to evaluate the new drug olaratumab as a first-line treatment option for advanced STS, the primary end point is OS .
The PALETTE (PAzopanib expLorEd in sofT-Tissue sarcoma—a phasE III study) trial was an international, multicenter, randomized, double-blind, placebo-controlled phase III study designed to evaluate the clinical benefits of pazopanib in patients with advanced STS . Patients with metastatic STS progressing despite previous standard chemotherapy (n = 369) were randomized to either the pazopanib arm (800 mg orally once daily; n = 246) or the placebo arm (n = 123) of the study in a ratio of 2:1. The primary end point was PFS, and the secondary end points included OS. Median PFS was 4.6 months in the pazopanib arm and 1.6 months in the placebo arm, showing a significant improvement in PFS in the pazopanib arm (hazard ratio, HR 0.31; 95% confidence interval, CI 0.24–0.40; P < 0.0001); however, there was no significant improvement in OS (HR 0.86; 95% CI 0.67–1.11; P = 0.25) (Table 1). Patients with liposarcoma were excluded from this phase III study, because the phase II study showed insufficient evidence for the efficacy of pazopanib in the treatment of liposarcoma .
The ET743-SAR-3007 trial was a multicenter, randomized, open-label, active-controlled, parallel-group, phase III study designed to evaluate the clinical benefits of trabectedin in patients with advanced STS . Patients with locally advanced or metastatic STS after previous therapy with an anthracycline and at least one additional systemic regimen (n = 518) were randomized to either the trabectedin arm (n = 345) or the dacarbazine arm (n = 173) of the study in a ratio of 2:1. In the trabectedin arm, trabectedin (1.5 mg/m2) was administered as a 24-h intravenous infusion, whereas in the dacarbazine arm, dacarbazine (1 g/m2) was administered by intravenous infusion over 20–120 min. In each arm, the study drug was administered on day 1 of every 21-day cycle. OS was the primary end point, and PFS was a secondary end point. PFS was significantly improved in the trabectedin arm compared with the dacarbazine arm (median PFS, 4.2 versus 1.5 months; HR 0.55; 95% CI 0.44–0.70; P < 0.001); however, interim analysis showed no improvement in OS, the primary end point (HR 0.87; P = 0.37) (Table 1). One reason for the non-significant result for OS may be the effect of the active comparator, dacarbazine .
It should be noted that the approved dose of trabectedin in Japan is different from that in the USA and Europe; the approved dose in Japan, 1.2 mg/m2, which was used in the ET743-SAR-3007 phase III study, is lower than the 1.5 mg/m2 dose used in other countries. The approved dose in Japan is based on the results of a phase I study in Japanese advanced STS patients who had received treatment with at least one anthracycline-based regimen . Dose-limiting toxicity was experienced by two of three patients who received trabectedin at a dose of 1.5 mg/m2; one had grade 3 increased creatine phosphokinase and grade 3 anorexia, and the other had grade 4 decreased platelet count.
The phase III E7389-G000-309 trial was conducted as a multicenter, open-label, active-controlled, parallel-group study to evaluate the clinical benefits of eribulin in patients with advanced STS . Patients with locally recurrent, locally advanced, or metastatic STS who had received treatment with systemic regimens including an anthracycline (n = 452) were randomized to either the eribulin arm (n = 228) or the dacarbazine arm (n = 224) in a ratio of 1:1. In the eribulin arm, eribulin mesylate (1.4 mg/m2) was administered as an intravenous infusion over 2–5 min on days 1 and 8 of every 21-day cycle, whereas in the dacarbazine arm, dacarbazine (850, 1000, or 1200 mg/m2) was administered as an intravenous infusion over 15–60 min on day 1 of every 21-day cycle. The primary end point was OS, and the secondary end points included PFS. OS was significantly improved in the eribulin arm compared with the dacarbazine arm (median OS, 13.5 versus 11.5 months; HR 0.77; 95% CI 0.62–0.95; P = 0.0169); however, there was no significant difference between the two arms in the secondary end point of PFS (HR 0.88; 95% CI 0.71–1.09; P = 0.23) (Table 1).
Review of Histologic Subtypes
In the treatment of advanced STS, it is necessary to choose an anticancer drug that is effective against the specific histologic subtype. However, because STS is a rare tumor, it is difficult to prospectively accumulate clinical data from patients with different histologic subtypes to evaluate the efficacy of particular drugs against each subtype. Therefore, the findings of subgroup analyses by histologic subtype should be taken into account when deciding which drug to use in clinical practice. The chemosensitivity of each histologic subtype to anticancer drugs must also be considered. Therefore, in this section, we review subgroup analyses of data from the phase III studies [16,17,18] and phase II studies [20, 22] of pazopanib, trabectedin, and eribulin, and propose treatment strategies for advanced STS.
Chemosensitivity of Different Histologic Subtypes of STS
The chemosensitivity of each histologic subtype to anticancer drugs is described in UK guidelines for the management of STS  and in an article summarizing the opinions of international experts regarding patient-tailored management of STS . On the basis of this information and our combined clinical experience, we examined the treatment for each histologic subtype, and list the primary drugs of choice and substitution drugs of choice in Table 2.
In the international multicenter phase III study of pazopanib , patients with liposarcoma were not included in the study. This is because in the phase II study , the primary end point, namely progression-free rate at 12 weeks (PFR12 weeks), was 26%, which was lower than the 44% value for leiomyosarcoma, 49% for synovial sarcoma, and 39% for other types of STS, as well as the benchmark value of PFR12 weeks (40%); this last value was used as the minimum at which an anticancer drug can be considered effective as a second-line treatment for STS in phase II studies, based on information from a database of clinical studies including data from patients with various subtypes of STS . Therefore, current evidence suggests that eribulin and trabectedin are preferable to pazopanib for the treatment of liposarcoma.
Subgroup analysis by histologic subtype, using data from a phase III study, has shown trabectedin to be effective in terms of improvement in PFS for patients with myxoid/round-cell liposarcoma (HR 0.41; 95% CI 0.17–0.98) . The effect of trabectedin on these subtypes was also shown in a randomized, phase II study in Japanese patients with translocation-related sarcoma (HR 0.03; 95% CI <0.01–0.22) .
Subgroup analysis by histologic subtype, using data from a phase III study (the E7389-G000-309 trial), has shown eribulin to be effective in terms of improvement in OS for patients with liposarcoma (HR 0.51; 95% CI 0.35–0.75) . Therefore, eribulin can be expected to improve OS in patients with liposarcoma. Further analyses showed eribulin to be effective for patients with dedifferentiated liposarcoma (HR 0.45; 95% CI 0.25–0.82) and pleomorphic liposarcoma (HR 0.21; 95% CI 0.06–0.76) .
On the basis of the above information, we propose the treatment strategy for liposarcoma shown in Fig. 1. Eribulin may be the optimal treatment option for liposarcoma, particularly dedifferentiated liposarcoma and pleomorphic liposarcoma after doxorubicin treatment. Trabectedin may be the first option for myxoid/round-cell liposarcoma after doxorubicin treatment. Also, either eribulin or trabectedin can be used for each liposarcoma, depending on conditions such as safety issues and method of administration (for inpatients or outpatients).
Patients with leiomyosarcoma made up the greatest proportion (44.7%) of the study population in the international phase III study of pazopanib , in which the primary end point of PFS was met. Subgroup analysis by histologic subtype, as well as predictive analysis for histology subtype using Cox models, showed pazopanib to be effective for leiomyosarcoma in terms of PFS .
In the phase III study of trabectedin, patients with leiomyosarcoma made up the greatest proportion (73%) of the study population, and the primary end point of OS was not met ; however, trabectedin treatment resulted in similar OS compared with dacarbazine (the active comparator). On the other hand, subgroup analysis of data for the secondary end point (PFS) showed an improvement with trabectedin compared with dacarbazine (HR 0.55; 95% CI 0.42–0.73). This finding suggests that trabectedin may be useful in improving PFS.
In the phase III study of eribulin , patients with leiomyosarcoma made up the greatest proportion (65.7%) of the study population, and the primary end point of OS was met for the overall population. Subgroup analysis by histologic subtype of the data for OS showed the effect of eribulin on leiomyosarcoma to be similar to that of dacarbazine (HR 0.93; 95% CI 0.71–1.20). This is thought to be because dacarbazine, known to be effective for patients with leiomyosarcoma , was the active comparator. Therefore, eribulin is deemed to be similarly effective to dacarbazine for patients with leiomyosarcoma. Our proposed strategy for the treatment of leiomyosarcoma based on the above information is shown in Fig. 1.
The selection of patients for treatment with any anticancer drug is based not only on considerations of efficacy but also on considerations of safety of the drug. Safety information from preapproval clinical studies is based on safety data from the limited patients selected in accordance with the inclusion criteria. However, in clinical practice, STS patients have more varied baseline characteristics, and some have poor performance status and decreased organ function. Therefore, the information obtained from clinical studies should be supplemented by expert opinions from specialists with appropriate clinical experience.
In this section, we review the results of the safety analyses from the phase III studies of pazopanib, trabectedin, and eribulin [16,17,18]; a subgroup analysis of Japanese patients in the phase III study of pazopanib ; Japanese phase II studies of trabectedin and eribulin [22, 29]; and the Japanese package insert for each drug [30,31,32]. We summarize the safety precautions that should be considered when selecting patients for treatment with these drugs in clinical practice, taking into account the administration methods for each drug and our combined clinical experience.
The main safety issues to be considered when selecting patients for treatment with pazopanib, trabectedin, and eribulin are summarized in Table 3. The safety profile of each drug should be considered in addition to its efficacy, and appropriate follow-up should be carried out after administration. Moreover, apart from the adverse events listed here, any potential safety issues should be addressed carefully, and treatment should be individualized according to the clinical condition of each patient.
In both the international phase III study of pazopanib  and the associated subgroup analysis of data from Japanese patients , liver dysfunction was reported as an adverse event. In the pazopanib arm of the overall population (n = 240) and in the Japanese subgroup (n = 31), the incidence of increased alanine aminotransferase (ALT) (all grades) was 46% and 52%, respectively; the incidence of increased aspartate aminotransferase (AST) was 51% and 65%, respectively; and the incidence of increased total bilirubin was 29% and 42%, respectively . Therefore, liver function should be monitored before and during treatment.
In the phase III study, pneumothorax occurred in 3% of patients (8/240) in the pazopanib arm , but there were no occurrences in the Japanese subgroup . However, in a post-marketing report based on data from Japanese patients, the prevalence of pneumothorax was reported to be 10.3% (6/58 patients) ; therefore, physicians should be aware of the possible onset or worsening of pneumothorax, particularly in patients with lung metastases.
Because the mechanism of action of pazopanib includes antiangiogenic effects , hemorrhage is possible in patients with brain metastases. Therefore, safety information should be provided and explained to patients before administration of this drug.
The Japanese package insert for pazopanib states that patients with risk factors for cardiac dysfunction may have exacerbation of symptoms . Anthracyclines such as doxorubicin, which have cardiotoxic effects, may be used as first-line treatment for advanced STS; thus, caution is required when pazopanib is used as the second-line treatment.
In the phase III study, hypertension (grade 3 or higher) occurred in the pazopanib arm, in both the overall population (grade 3, 7%; grade 4, 0%) and the Japanese subgroup (grade 3, 16%; grade 4, 0%) . However, protocols for the management of hypertension have been established; thus, control of associated symptoms should be possible in patients who experience hypertension as an adverse effect of pazopanib. Safety information should be provided to patients before administration of the drug, and appropriate monitoring of blood pressure during treatment is essential.
In the pazopanib arm of the phase III study, grade 3 or higher fatigue was reported in less than 14% and in 3% of patients in the overall population and the Japanese subgroup, respectively; and grade 3 or higher diarrhea in 5% and 6%, respectively . These adverse effects are readily recognized by patients, but the symptoms are difficult to control. No management of fatigue has been established, and it may become necessary to reduce the dosage of pazopanib in patients experiencing this adverse effect. Regarding diarrhea, if symptoms fail to improve with supportive care, interruption or reduction of dosage is necessary.
In regard to the administration of pazopanib, the oral route is convenient for patients.
Abnormal changes in laboratory test values associated with bone marrow suppression were recorded in the clinical studies of trabectedin. In the trabectedin arm (n = 340) of the phase III study, neutropenia (grade 3, 21%; grade 4, 16%) and anemia (grade 3, 14%; grade 4, 0%) were reported . In the trabectedin arm (n = 36) of the Japanese phase II study, neutropenia (grade 3, 31%; grade 4, 36%), leukopenia (grade 3, 31%; grade 4, 25%), thrombocytopenia (grade 3, 11%; grade 4, 6%), lymphopenia (grade 3, 19%; grade 4, 3%), and anemia (grade 3, 19%; grade 4, 0%) were reported . Therefore, blood tests and careful examination of patients for the presence of infection should be carried out before administration of trabectedin; regular blood tests are also required during treatment.
Febrile neutropenia (grade 3, 11%; grade 4, 3%) was reported in the trabectedin arm of the Japanese phase II study .
Abnormal changes in laboratory test values associated with liver dysfunction were also recorded in the clinical studies of trabectedin. In the trabectedin arm of the phase III study, increased ALT (grade 3, 25%; grade 4, 1%), increased AST (grade 3, 12%; grade 4, 1%), and increased alkaline phosphatase (ALP) (grade 3, 1%; grade 4, 0%) were reported . In the trabectedin arm of the Japanese phase II study, increased ALT (grade 3, 47%; grade 4, 14%), increased AST (grade 3, 33%; grade 4, 8%), increased gamma-glutamyltransferase (grade 3, 22%; grade 4, 3%), and increased ALP (grade 3, 6%; grade 4, 0%) were reported . In the same study, grade 1 or 2 increased blood bilirubin (11%) was also reported in the trabectedin arm . In clinical practice, we have encountered serious cases of liver dysfunction in patients treated with trabectedin. Therefore, monitoring of liver function is required in these patients, both before and during treatment.
As with pazopanib, trabectedin may also cause the onset or worsening of cardiac dysfunction when administered to patients with a history of treatment with anthracycline drugs and those with cardiac dysfunction. Before administration of trabectedin, patients are likely to have a history of treatment with doxorubicin, which has cardiotoxic effects; therefore particular caution is required.
Rhabdomyolysis (grade 4, 3%) and increased creatine phosphokinase (grade 3, 6%; grade 4, 0%) were reported in the trabectedin arm of the Japanese phase II study , and we have encountered serious cases in trabectedin-treated patients in clinical practice. In addition to vigilance for the initial symptoms (e.g., muscular pain and weakness), monitoring is required for increased creatine phosphokinase or increased blood or urinary myoglobin concentration. Appropriate management, including discontinuation of treatment, should be taken if patients had abnormal changes in these values, and physicians and medical staff should be aware of this before the start of treatment.
Fatigue of grade 3 or higher was reported in 6% and 3% of patients in the trabectedin arm of the phase III study and the Japanese phase II study, respectively.
Because trabectedin is administered by 24-h intravenous infusion using a central venous catheter or port, hospitalization is required. Tissue necrosis may occur, and extravasation is a concern. If extravasation occurs, the administration of trabectedin should be interrupted immediately; medical staff should be aware of this measure.
Abnormal changes in laboratory test values associated with bone marrow suppression were recorded in the clinical studies of eribulin. In the phase III study (eribulin arm, n = 226), neutropenia (grade 3, 20%; grade 4, 15%), leukopenia (grade 3, 8%; grade 4, 2%), and thrombocytopenia (grade 3, <1%; grade 4, 0%) were reported . In the Japanese phase II study (n = 51), neutropenia (grade ≥3, 86%), leukopenia (grade ≥3, 75%), lymphopenia (grade ≥3, 33%), anemia (grade ≥3, 14%), and febrile neutropenia (grade ≥3, 8%) were reported . The incidence of adverse events associated with bone marrow suppression reported in the clinical studies of eribulin was lower than in those of trabectedin [17, 22], although direct comparison of safety data is difficult because of the different background characteristics of the patients in these studies. However, close examination for signs of bone marrow suppression is still required before administration of eribulin, as well as appropriate monitoring after administration, because bone marrow suppression can affect vital functions. The incidence of febrile neutropenia was higher in the Japanese phase II study  than the phase III study . However, the overall incidence of febrile neutropenia was less than 10% in both studies.
Abnormal changes in laboratory test values associated with liver dysfunction were recorded in the Japanese phase II study, namely ALT increase (grade ≥3, 6%) and AST increase (grade ≥3, 4%) . In clinical practice, we have not encountered any serious cases of liver dysfunction in patients treated with eribulin; however, it is essential to assess liver function before its administration.
In the phase III study of eribulin, grade 3 or higher peripheral sensory neuropathy (grade 3, 2%; grade 4, 0%) was recorded in the eribulin arm . In contrast, in the Japanese phase II study, the severity of peripheral sensory neuropathy was grade 1 and 2 . The likelihood of the onset of peripheral sensory neuropathy depends largely on previous treatment; therefore, caution is required in patients previously treated with drugs such as vincristine that cause peripheral sensory neuropathy. However, vincristine is generally administered to young children with rhabdomyosarcoma; therefore, its target patients differ from those of eribulin. However, as peripheral sensory neuropathy may occur as a result of cumulative toxicity, caution is required in the long-term administration of eribulin.
Overall, we consider that the toxicities of eribulin are manageable, and that management of adverse effects of eribulin is relatively straightforward on the basis of our experience.
Eribulin needs to be administered by intravenous infusion. However, as this takes only 2–5 min, it is possible to administer the drug in an outpatient visit.