Cardiac safety of trabectedin monotherapy or in combination with pegylated liposomal doxorubicin in patients with sarcomas and ovarian cancer

Abstract Background As with other alkylating agents, cardiac dysfunction can occur with trabectedin therapy for advanced soft tissue sarcomas (STS) or recurrent ovarian cancer (ROC) where treatment options for advanced disease are still limited. Cardiac safety for trabectedin monotherapy (T) for STS or in combination with pegylated liposomal doxorubicin (T+PLD) for ROC was evaluated in this retrospective postmarketing regulatory commitment. Methods Patient data for multiple cardiac‐related treatment‐emergent adverse events (cTEAEs) were evaluated in pooled analyses of ten phase 2 trials, one phase 3 trial in STS (n = 982), and two phase 3 trials in ROC (n = 1231). Results Multivariate analyses on pooled trabectedin data revealed that cardiovascular medical history (risk ratio [RR (95% CI)]: 1.90 [1.24‐2.91]; p = 0.003) and age ≥65 years (RR [95% CI]: 1.78 [1.12‐2.83]; p = 0.014) were associated with increased risk for cTEAEs. Multivariate analyses showed increased risk of experiencing cTEAEs with T+PLD compared to PLD monotherapy (RR [95% CI]: 2.70 [1.75‐4.17]; p < 0.0001) and with history of prior cardiac medication (RR [95% CI]: 1.88 [1.16‐3.05]; p = 0.010). Conclusions For patients with STS or ROC who still have limited treatment options, trabectedin may be initiated after carefully considering benefit versus risk. Trial Registration (ClinicalTrials.gov): NCT01343277; NCT00113607; NCT01846611.


| INTRODUCTION
Trabectedin is a DNA-binding agent with a unique antitumor mechanism of action (MOA) targeting the transcription-coupled nucleotide excision repair (NER) system. Trabectedin was developed for the treatment of soft tissue sarcomas (STS) and epithelial ovarian cancer based on its novel cytotoxic activity. These cancers still have limited treatment options, particularly where advanced disease has progressed with other therapies. 1 Preclinical studies with trabectedin showed no toxicity in cultured rat myocytes in vitro, while single and repeated doses in Cynomolgus monkeys did not induce any relevant cardiac, vascular, or respiratory effects. 2 Further, a low incidence of cardiac-related treatment-emergent adverse events (cTEAE) was reported in previous analyses from earlier phase 1-2 clinical trials and one phase 3 (OVA-301), pharmacovigilance databases, and spontaneously reported cases; tachycardia or palpitations were the most common cTEAEs reported. 2 No clinically relevant left ventricular ejection fraction (LVEF) changes occurred in phase 1 combination trials, while LVEF decreases from baseline were similar [9% of patients (pegylated liposomal doxorubicin [PLD]) and 7% (trabected-in+PLD)] with no relevant symptoms in one phase 3 trial. 2 Trabectedin is now approved for STS in 80 countries and for ovarian cancer in combination with PLD in 71 countries. European Union (EU) approval for trabected-in+PLD for relapsed platinum-sensitive ovarian cancer was granted in October 2009. In the United States (US), trabectedin was approved for STS treatment following the failure of anthracycline-based chemotherapy in October 2015. Approval in the United States was contingent upon undertaking post-marketing requirements to characterize risk of cardiotoxicity and its sequelae with trabectedin to identify risk factors including previous treatments known to be cardiotoxic (e.g., anthracyclines).
As an extension to the cTEAE analysis reported in 2011 2 , we now report the findings of this retrospective pooled analysis of key cTEAEs for all patients enrolled in ten phase 2 trials and one phase 3 trial involving trabectedin monotherapy (T) for STS and other solid tumors and two phase 3 trials in combination with PLD for recurrent ovarian cancer (ROC).

| Overall safety evaluation plan and description of safety studies
Safety analysis sets incorporated two pooled analyses: Cardiac safety with T was evaluated using data from ten phase 2 and one phase 3 trial (SAR-3007 [NCT01343277]) in STS and other solid tumors at a dose and regimen of 1.5 mg/ m 2 every 3 weeks (q3wk), 24 h. Cardiac safety with combination trabectedin+PLD was derived from two phase 3 ovarian cancer trials (OVA-301 [NCT00113607] and OVC-3006 [NCT01846611]) where trabectedin (1.1 mg/m 2 q3wk; 3 h) was co-administered with PLD (30 mg/m 2 q3wk; 90 min). Phase 3 trial designs are described in Table 1. Key inclusion/exclusion criteria for enrollment are presented for each phase 3 trial (Table S1), cardiac safety evaluations by individual trial (Table S2), and exposure and cancer diagnoses for pooled data from phase 2 and 3 studies (Table S3). Study protocols and amendments were reviewed by an Independent Ethics Committee or an Institutional Review Board. proportional hazards models. Anthracycline exposure data are summarized for subjects who received anthracyclines prior to the study (i.e., prior anthracycline) and for subjects who received anthracyclines prior to and during the study (i.e., cumulative anthracycline).

|
Two parameters were used for the cardiac safety analysis: LVEF significant decline where available and cardiac-related AEs of special interest (cardiac-related AEs). Cross tabulation of ECG data were included when available.

| Cardiac-related adverse events
cTEAEs are summarized from time of first administration to 30 days after the last dose and graded using Common Terminology Criteria for Adverse Events (CTCAE; version 4.0). Incidences of cTEAEs are defined by eight Medical Dictionary for Regulatory Activities (MedDRA) high-level group terms (HLGTs) and associated preferred terms (PTs) and two Standardized MedDRA Queries (narrow SMQ) ( was defined as either return to baseline values or <Grade 2 ejection fraction decreased toxicity (CTCAE v4.0). In all three phase 3 studies (SAR-3007, OVA-301, and OVC-3006), LVEF assessments were performed at baseline and end of treatment. Additionally, OVC-3006 was amended to provide comprehensive cardiac evaluations of patients while on treatment. Collection time points for LVEF in each study are described in Table S2. Table 3 shows the number of patients exposed to study drug by individual trial and by pooled safety analysis sets for T and trabectedin+PLD. Nine hundred and eighty-two patients were exposed to T, and 619 patients were treated with trabectedin+PLD.

| Trabectedin in combination with PLD
cTEAEs were reported for 78 (12.6%) patients in the trabectedin+PLD group and 34 (5.6%) patients in the PLD monotherapy group; most commonly reported cTEAE was LVEF decrease (7.8% vs. 4.2%, respectively). Within these SMQ/HLGTs, palpitation was the only cTEAE reported with at least a 2% greater incidence in the trabectedin+PLD group compared with the PLD monotherapy group (3.2% vs. 1.0%) (Table S10). Kaplan-Meier analyses showed an increased risk of cTE-AEs with trabectedin+PLD compared with PLD monotherapy (Figure 1). Cumulative incident rate curves separated early and remained separated throughout treatment. Median time from first study dose to the onset of first occurrence of cTEAE was shorter with trabectedin+PLD (57 days) compared with PLD monotherapy (98 days), while most patients in both groups had similar resolutions of cTEAEs (57.1% and 55.9%) and time to resolution (8 days

| Unique study features
In OVA-301, fewer patients with a significant decrease from baseline in LVEF in the trabectedin+PLD group had a cardiovascular medical history compared with the PLD monotherapy group (23.8% vs. 52.6%). In OVC-3006, median cumulative PLD dose for patients with a cTEAE was lower with trabectedin+PLD treatment compared with PLD monotherapy (180.78 vs. 329.67 mg/m 2 ). Median cumulative anthracycline dose of ≥300 mg/m 2 was reported in 11/43 F I G U R E 1 Cumulative Incidence of Cardiac-Related Adverse Events Over Treatment Duration for Treated Patients (Pooled Studies ET743-OVC-3006 and ET743-OVA-301). PLD, pegylated liposomal doxorubicin (25.6%) in trabectedin+PLD patients with a cTEAE compared to 17/23 (73.9%) PLD monotherapy patients.
In OVC-3006, the median time from first dose of drug to the onset of first occurrence of a cTEAE was shorter with trabectedin+PLD compared with PLD monotherapy (68 days vs. 169 days); however, the median time to resolution was longer with PLD monotherapy group compared with trabec-tedin+PLD (29 days vs. 16 days). Median cumulative PLD dose for patients having significant decreases from baseline in LVEF was lower in the trabectedin+PLD group compared with the PLD monotherapy group (149.39 vs. 251.25 mg/ m 2 ). In addition, median cumulative anthracycline doses of ≥300 mg/m 2 were associated with a significant decrease from baseline in LVEF; this was reported in 3/19 (15.8%) tra-bectedin+PLD and 5/10 (50.0%) PLD monotherapy patients.

| Trabectedin monotherapy
Trabectedin-treated patients who experienced a cTEAE were generally older (18.4% aged ≥65 years vs. 9.6% aged <65 years). Results from multivariate analyses of cTEAEs when controlling for potential risk factors are presented in Figure 2. These showed that patients aged ≥65 years and those with cardiovascular medical history had an increased risk of cTEAEs. The effect of cumulative anthracycline dose of ≥300 versus <300 mg/m 2 and baseline LVEF <lower limit of normal (LLN) versus ≥LLN, however, could not be evaluated in the ten phase 2 studies due to differences in study designs.

| Trabectedin in combination with PLD
In the multivariate analyses, when controlling for potential risk factors, results showed that patients receiving trabectedin+PLD were at increased risk for experiencing a cTEAE compared with PLD monotherapy (risk ratio [RR] 2.70; 95% CI: 1.75-4.17; p < 0.0001). Furthermore, patients with a history of prior cardiac medication use who received trabectedin+PLD versus PLD were also at increased risk of experiencing cTEAEs (RR 1.88; 95% CI: 1.16-3.05; p = 0.010). Patients with a cumulative anthracycline dose of ≥300 mg/m 2 who received trabectedin+PLD in the OVC-3006 and OVA-301 trials were at increased risk for a significant decrease in LVEF compared with patients who received PLD monotherapy (RR 0.54; 95% CI: 0.30-0.99; p = 0.046) (Figure 3).

| DISCUSSION
Trabectedin was developed based on its novel chemical structure and promising preclinical activity in several types of human tumors. The development program focused on STS 3 and ROC 4,5 in which trabectedin was active at very low concentrations in both preclinical models and clinical trials. 1 Trabectedin binds to the N2 position of guanine in the minor groove of DNA and bends the helix toward the major groove, a unique property in the class of DNAbinding agents; it triggers a cascade of events affecting several transcription factors, DNA-binding proteins, and DNA-repair pathways (e.g., transcription-coupled NER), resulting in slowed progression through S and G2/M phases and p53-independent apoptosis. Trabectedin also prevents the binding of translocation-related oncogenic fusion proteins to DNA promoter regions, thereby interfering with the function of proteins that contribute to the malignant phenotype and tumor progression. [6][7][8][9] PLD is doxorubicin hydrochloride encapsulated in STEALTH ® liposomes for intravenous administration. PLD was granted approvals for advanced ovarian cancer in June 1999 and October 2000 in the United States and European Union, respectively. As with any anthracycline, PLD can cause myocardial damage, including congestive heart failure, as the total cumulative dose of doxorubicin hydrochloride approaches 550 mg/m 2 . In a clinical study of 250 patients with advanced cancer who were treated with PLD, the risk of cardiotoxicity was 11% when the cumulative anthracycline dose was 450 to 550 mg/m 2 . 10 This is the most comprehensive analysis of cardiac safety in the setting of trabectedin administration from clinical trial data including more than 1600 patients. Strengths include pooled analyses of one phase 3 trial and ten phase 2 trials of T in STS and two phase 3 trials of trabectedin in combination with PLD for ROC. Limitations include patient heterogeneity and varying dosing, scheduling, and infusion times for T (1.5 mg/m 2 q3wk; 24 h) compared with PLD combination therapy (trabectedin 1.1 mg/m 2 q3wk; 3 h). The authors recognize that cardiac adverse events with diverse etiologies make it difficult to ascribe the outcomes to trabectedin alone or identify specific causal mechanisms. Cardiotoxicity may be mediated by multiple mechanisms including damage from prior cardiotoxic therapies (anthracyclines), preexisting cardiovascular comorbidities, and the alkylating MOA among others. Last, the retrospective nature of data collection and other events, such as sepsis (that could contribute to cardiac events) are additional limitations.
In phase 3 SAR-3007 study 3 , no difference in the overall incidence rate of any-grade cTEAEs was observed between trabectedin-and dacarbazine-treated patients. Multivariate analysis of safety data (data not shown) indicated that cumulative anthracycline dose of ≥300 mg/m 2 and baseline LVEF <LLN were risk factors for the development of cTEAEs in STS. In pooled analyses, only age ≥65 years and cardiovascular medical history were associated with an increased risk of cTEAEs. This difference could be attributed to variability in patient populations, pretreatment history, and consistent baseline LVEF testing in the ten phase 2 studies compared with SAR-3007. In summation, cardiac safety signals observed with T may be, in part, due to the history of prior or concurrent therapy with an anthracycline, known for potential short-and long-term cardiotoxicity, and longer median duration of treatment for patients receiving trabectedin.
In OVA-301 and OVC-3006, patients in the trabected-in+PLD groups experienced cTEAEs at a higher incidence, regardless of toxicity grade, compared with PLD monotherapy patients. In OVA-301, multivariate analyses indicated an increased risk of cTEAEs among patients in the trabected-in+PLD group compared with PLD monotherapy (data not shown). In OVC-3006, however, a cumulative anthracycline dose of ≥300 mg/m 2 and prior cardiac medication use were also identified as independently associated with increased risk of cTEAEs (data not shown). Differences between the two studies may be attributed to enrollment criteria for each study. Inclusion criteria for OVC-3006 differed from OVA-301 in that patients were allowed to have received two prior lines versus one line of chemotherapy for ROC, and prior PLD combination therapy was also allowed.
Ultimately, these data suggest that some patients receiving T after prior therapy with anthracyclines are at risk for cTEAEs, which may be serious in a small number. The overall risk of fatal events is relatively low but appears to be higher for patients with existing myocardial dysfunction (abnormal LVEF) or prior cardiovascular medical history. In the setting of STS, the available data support recommendations to assess LVEF by echocardiogram or multigated acquisition radionuclide scan before the initiation of trabectedin and at two-to three-month intervals thereafter until trabectedin is discontinued, particularly for patients with prior cardiovascular disease. Additionally, when using trabectedin in combination with PLD for ROC, cumulative anthracycline dose ≥300 mg/m 2 and prior cardiac medication may increase the risk of cTEAEs if prior lines of therapy involved PLD.
In conclusion, as with any systemic cytotoxic therapy, benefit versus risk should be carefully considered when instituting treatment with trabectedin in patients with few other treatment options and with risk factors for developing cTE-AEs. In consultation with a cardiologist or cardio-oncology service, baseline cardiovascular risk should be comprehensively assessed before commencing treatments with cardiotoxic potential as noted above. 11,12 Cardiotoxicity risk can be minimized by primary prevention strategies; signs and symptoms of myocardial toxicity including decreases in LVEF should be assessed routinely as described above. Dose reductions or temporary or permanent discontinuation