Three prospective BE studies with albumin-bound paclitaxel (injectable suspension) were conducted, with drugs provided by three different sponsors, i.e., QL (n = 25), HR (n = 25), or ZDTQ (n = 24), and the reference product, Abraxane®. These single-center, randomized two-period crossover, BE studies, were performed between March 2016 and March 2018, according to the US FDA guidance draft on Paclitaxel. The tolerability and PK of the test products (albumin-bound Paclitaxel, QL, HR, and ZDTQ) and Abraxane® (reference) were compared in patients with breast cancer in these three studies respectively. The inclusion criteria were: (1) age > 18 years; (2) histologic diagnosis of advanced breast cancer for which there is no curative therapy and treatment with single-agent paclitaxel has been considered appropriate by the treating physician; (3) Eastern Cooperative Oncology Group (ECOG) performance status of 0/1; (4) life expectancy of >12 weeks; and (5) complete recovery from acute toxicities of prior treatment. Subjects were excluded if they did not have adequate hematologic, kidney, and liver function (hemoglobin ≥ 90g/L [not having blood transfusion within 14 days], absolute neutrophil count ≥1.5 × 10⁹/L, blood platelet count ≥100 × 10⁹/L, total bilirubin < 1 upper limit normal [ULN], alanine aminotransferase [ALT], and aspartate aminotransferase [AST] < 2.5 × ULN [if liver metastasis, then ALT and AST ≤ 5 × ULN], creatinine ≤ 1.5 × ULN), or had received radiotherapy, chemotherapy, immunotherapy, or endocrine therapy within 4 weeks prior to the use of the study drug and residual effects were still present.

This study was carried out in accordance with the recommendations of the Good Clinical Practice and the Declaration of Helsinki. The protocol was approved by the Ethics Committee of the First Hospital of Jilin University, Changchun, Jilin, China. All subjects gave written informed consent in accordance with the Declaration of Helsinki.

A screening visit was scheduled within 14 days prior to administration of the study drug. Then the eligible subjects were admitted to the clinical research unit 1 day before dosing. Following an overnight fast of at least 8-h, subjects were randomized to receive a single intravenous dose of 260 mg/m² (infusion 30 ± 3 min) of albumin-bound paclitaxel (test product; QL, HR, or ZDTQ) or Abraxane® (reference product, from the US market) in a 1:1 ratio according to a computer-generated randomization schedule for each study in the first period (Figures 1, 2). Then the same dosing method for the reference or test formulation was followed in second period, or vice versa. Each drug had a unique batch number. The washout period was of 3 weeks. Subjects were administered the drug at the same time on first day of first period and day 22 of second period (Figure 1). Albumin-bound paclitaxel by sponsor 1 (QL) and sponsor 3 (ZDTQ) was administered after breakfast (light food), whereas the HR product (sponsor 2) was administered after 8 h of fasting. Patients were carefully monitored, particularly during the infusion. Subjects were discharged after 72 h of drug administration. Blood samples for the primary PK analysis were collected prior to treatment and at specified time points during the 72-h follow-up. Subjects were followed up for safety assessment at 7 ± 1 and 21 ± 1 days.

According to the current US FDA guidelines, to achieve an 80–90% power (1-β) in BE studies at the 5% nominal level (α = 5%), the geometric mean ratio (GMR) is usually set to be 95–105% (Karalis et al., 2012). The coefficient of variation (CV) is evaluated as the intra-subject variability (intra-cv). The intra-cv for paclitaxel is assumed to be 20–21.3%². According to the initial estimation by R software, sample size should be of 21 patients. Based on the above sample size estimation result and considering a loss to follow-up of 10% and the opinion of sponsor and investigator, the final needed sample size was considered to be 24 (Table 1) (Zhang et al., 2018a,b).

Blood samples (5 ml each time) for PK evaluation were collected into heparin anticoagulant-containing tubes within 0.2 h of initiation of albumin-bound paclitaxel infusion (pre-dose), 0.25 h after the start of infusion, immediately after stopping the infusion, and 0.75, 1, 1.5, 2, 4, 6, 8, 12, 24, 48, and 72 h after the start of infusion. Blood samples were centrifuged at 3,000 rpm for approximately 10 min at 2–8°C in a refrigerated centrifuge. The serum was stored at −70°C until analysis. The serum concentrations of total paclitaxel and unbound paclitaxel were analyzed using a validated, sensitive and specific liquid chromatography tandem mass spectrometry (LC-MS/MS) method at the Shanghai Drug Metabolism Research Center for QL and HR, and Covance for ZDTQ.

The lower limits of quantification (LLOQ) for un-bound paclitaxel and total paclitaxel were 0.2 ng/mL and 5.0 ng/mL, respectively, and the upper limits of quantification (ULOQ) for unbound paclitaxel and total paclitaxel were 2,000 ng/mL and 15,000 ng/mL, respectively for QL and HR.

The LLOQ for un-bound paclitaxel and total paclitaxel were 2 ng/mL and 10.0 ng/mL, respectively, and the ULOQ for unbound paclitaxel and total paclitaxel were 2,000 ng/mL and 10,000 ng/mL, respectively for ZDTQ. The validated concentration ranges were between the LLOQ and ULOQ.

The LC-MS/MS method and rapid equilibrium method have been validated for the determination of paclitaxel concentration in human raw serum and equilibrium serum samples (Ronghao and Jun, 2015). In each analysis batch, the quantity of quality control samples accounted for more than 5% of the unknown samples in the analysis batch. In all quality control samples, the ratio of relative deviation within 15% is more than at least 67% of all quality control samples, and the ratio of relative deviation within 15% is more than at least 50% of each concentration of quality control samples. All analysis batch quality control met the above criteria (Supplement Table 1).

Total levels of free and protein-bound paclitaxel were later quantified. Paclitaxel-D5 served as the internal standard. The analytical column was a Eclipse Plus C18 column (100 × 4.6 mm, 3.5 μm, Agilent), and the formula used for calculation of unbound paclitaxel was as follows: Cunbound Paclitaxel=(Creceiver unbound Paclitaxel/Cdonor Paclitaxel)×Ctotal Paclitaxel

The inter-run assay accuracy, expressed as percent relative error for quality control samples [BIAS(%)]. The assay precision, expressed as the inter-run CV of the measured concentrations of quality control samples [relative standard deviation, RSD(%)].

All analysis batch quality control met the above criteria (Supplement table 1).

Medical history, physical examination, electrocardiography (ECG), and laboratory test (hematology, biochemistry, and urinalysis, etc.) results were obtained at the time of screening (2–14 days before the first dose of study drug) and at 7 ± 1 and 21 ± 1 days for all subjects in each test period. Vital signs were recorded immediately before the first dose and after administration of the drug. Adverse events were recorded daily from the day of administration of the first dose through the end of the study. The National Cancer Institute Common Toxicity Criteria for Adverse Events version 4.03 was used to describe and grade all toxicities and adverse events (AEs).

The serum concentrations vs. time data were analyzed with non-compartmental methods using WinNonlin Professional, Version 6.4 (Pharsight Corporation, NC, USA). The PK analysis used actual sample collection times. Samples below the LLOQ were set to zero before T_max and not detectable after T_max for the PK analysis. The PK parameters for paclitaxel included C_max, area under the curve (AUC)_0−t, AUC_0−∞, T_max, and T_1/2. Descriptive statistics were calculated for PK parameters, demographics, and safety variables and these were analyzed by t-test or analysis of variance (ANOVA). ANOVA also was used to compare the AUC and C_max, with factors fitted for the effect of sequence, subject within sequence, period, and treatment. The comparisons are presented in terms of the geometric least square means and the 90% confidence interval (CI). BE was established if the 90% CI of the treatment ratio was within the equivalence range of 0.8–1.25. T_max and T_1/2 were analyzed with a Wilcoxon rank test. All statistical tests were performed using SAS 9.1 Statistical Package, and P < 0.05 was considered statistically significant.

A total of 120 Patients with breast cancer were initially screened for these studies. Of these, 74 patients with breast cancer (n = 25, 25, 24) were enrolled and received the assigned study drug from sponsor 1-3 (HR, QL, ZDTQ) respectively; these patients also constituted the safety analysis set for each sponsor and 71 patients constituted the BE analysis set (Figure 2). The demographics and baseline characteristics of patients treated by the three sponsors were comparable (Table 2). Most of the study subjects were Han Chinese. The mean age of the study subjects was 48.8–52 years old.

The total paclitaxel and unbound paclitaxel serum concentrations increased rapidly in all study subjects and reached C_max at 0.5 h after the start of infusion. The serum concentrations showed a decline in a biphasic manner, which initially decreased rapidly after the end of infusion and then demonstrated a slight decrease until the lower limit of quantification. The study drugs exhibited a similar mean serum concentration–time profile in the R and T formulations in studies 1, 2, and 3 (Figures 3, 4).

AUC_0−t accounted for >90% of the AUC_0−∞ in all subjects, which indicated that the plasma concentration vs. time profiles were well characterized. The mean t_1/2 were 20.61–27.31 and 20.3–26.74 h, and the intra-cv values ranged from 6.4 to 11% and 9.85 to 15.87% for total paclitaxel and unbound paclitaxel, respectively (except the t_1/2 of 2.53–3 h of unbound paclitaxel of ZDTQ). The Inter-cv values were small, and almost all of these were less than 30%. There were no differences in PK parameters of total paclitaxel among the Sponsor 1 (QL), Sponsor 2 (HR), and Sponsor 3 (ZDTQ) products, which indicate that food did not affect the PK of paclitaxel. However, the unbound paclitaxel exposure was higher and the elimination rate lower with the Sponsor 1 (QL) and Sponsor 3 (ZDTQ) product than with the Sponsor 2 (HR) product (Table 3, Figure 5). The t_1/2 of ZDTQ is obviously shorter than those of QL and HR (Table 3, Figure 5).

The relative bioavailability of the test products as compared with the reference formulation was 92.16–106.44% for unbound paclitaxel and 93.21–100.8% for total paclitaxel. Both the assessments met the 80–125% BE range recommended by the US FDA (Table 4).

We re-estimated the sample size for the three studies based on their BE analysis results (α = 0.05, power = 0.8, GMR, and intra-cv) and the original hypothesis. The re-estimated sample size was 6–20, which is less than our enrollment size (Table 4) (Zhang et al., 2018a,b).

For QL product, 25 subjects were included in the safety evaluation. The incidence of AEs was 100% with the QL product and reference product. The incidence of AEs of grade II or higher severity was 32.0% (5/25) vs. 41.6% (10/24) in the QL and reference groups, respectively.

For HR product, 25 subjects were included in the safety evaluation. The incidence of AEs was 100% with the HR product and 91.7% with the reference product. The incidence of AEs of grade II or higher severity was 32.0% (8/25) vs. 37.5% (9/24) in the HR and reference groups, respectively.

For ZDTQ product, 24 subjects were included in the safety evaluation. The incidence of AEs was 100% with the ZDTQ product and the reference product. The incidence of AEs of grade II or higher severity was 83.0% (20/24) vs. 87.5% (21/24) in the HR and reference groups, respectively.

There was one SAE (cataract) with ZDTQ; which was found to be unrelated to the drug. No fatal AEs or study drug injection site reactions of the drugs were observed. The test and reference groups had a similar incidence and pattern of AEs. There were no reports of unexpected AEs. The common hematologic adverse reactions included: neutropenia, leucopenia, thrombocytopenia, and anemia. The common non-hematologic adverse reactions included: increased ALT, AST, and fasting serum glucose levels; hyperesthesia; skin rashes; itching; fever; fatigue; nausea; diarrhea; and vomiting. In these studies, 24 (32.4%) and 23 (31%) subjects had concomitant medication use for co-existing diseases and administration of colony-stimulating factors was the most common group. There was no use of metabolic inducers or inhibitors, such as cyclosporine, phenobarbital, and ketoconazole, among the patients.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.