The research was conducted at Beijing Shijitan Hospital, Beijing, China. The study protocols and informed consent forms were reviewed and approved by the Institutional Review Board (IRB) of Beijing Shijitan Hospital (2021(6), 2021(7), 2022(41)). The research adhered to ethical principles derived from international guidelines, including the Declaration of Helsinki and the International Council for Harmonisation Guidelines for Good Clinical Practice (ICH GCP), as well as all applicable legal and regulatory requirements. All participants provided written informed consent prior to the initiation of any procedures. The studies were registered at chinadrugtrials. org.cn (Study Numbers: CTR20210575, CTR20210576, CTR20223278; date: 25 March 2021, 26 March 2021, 27 December 2022).

The test formulation used was a regorafenib tablet manufactured by Beijing SL Pharmaceutical Co., Ltd., Beijing, China (40 mg/tablet, batch number: 20200102, expiry date: 9 July 2023). The reference formulation was Stivarga^® from Bayer AG (40 mg/tablet; for studies 1 and 2, batch number: BXJAZ32, expiry date: 13 August 2022; for study 3, batch number: BXJRNL4, expiry date: 5 October 2024).

Eligible participants were healthy Chinese males aged over 30 years with a Body Mass Index (BMI) within the range of 19.0–28.0 kg/m² at screening. Participants were deemed healthy based on a comprehensive clinical assessment, which included a detailed medical history review, thorough physical examination, vital signs monitoring, electrocardiogram (ECG) analysis, and standard laboratory evaluations. Inclusion criteria also required participants to agree to use an effective method of contraception throughout the study period. Participants were prohibited from using concurrent medications, consuming alcohol, smoking tobacco, or taking dietary supplements during the studies. Individuals who had taken drugs known to induce or inhibit drug-metabolizing enzymes (such as barbiturates) within 30 days prior to the administration were excluded. Additionally, those with a history of liver disease or elevated levels of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) above the upper limit of normal, or with a history of allergic disease were not included in the studies.

Three separated studies involving 216 healthy Chinese male participants under varying dietary conditions (fasting, post low-fat breakfast, post high-fat breakfast) were conducted. Each study was a randomized, single-center, open-label, two-treatment, two-period, two-sequence crossover study with a 12-day washout interval. Given the half-lives of regorafenib and its metabolite M-2 (approximately 25 and 28 h, respectively) [Stivarga^® (regorafenib) label. 2012; Stivarga^® summary. 2013], the washout period was set at 12 days to ensure complete elimination of the drug prior to the next dosing period. According to the requirements of the Chinese National Medical Products Administration (NMPA), the dosage should be the maximum specification of a single tablet. Since the only specification for regorafenib tablets is 40mg, the single-dose in the studies was set at 40 mg. SAS statistical software (v9.4) generated a random number table assigning participants to either sequence A (Test (T)/Reference (R)) or B (R/T) in a 1:1 ratio. Sixty-four participants were randomized into Study 1 under fasting conditions, while 76 participants each were randomized into Studies 2 and 3 under low-fat and high-fat fed conditions, respectively. Subjects were allocated random numbers sequentially based on their screening numbers. Due to preclinical findings of increased necrotic corpora lutea in female rats’ ovaries, only male adults were enrolled.

In Study 1, all participants received the regorafenib tablet (40 mg) or Stivarga^® (40 mg) under fasting conditions (after at least a 10-h fast) on day 1 of each period. In Studies 2 and 3, participants received the regorafenib tablet or Stivarga^® under fed conditions. Specifically, Study 2 utilized a low-fat breakfast consisting of a boiled egg, a slice of bread, and 250 mL of skimmed milk (totaling 411 kcal with 25, 380, and 6 kcal from protein, carbohydrates, and fat, respectively). Study 3 involved a high-fat breakfast comprising an egg, a piece of beef cooked with oil, a steamed bun, and 250 mL of whole milk (totaling 925 kcal with 150, 265, and 510 kcal from protein, carbohydrates, and fat, respectively). Breakfasts were designed to align with FDA recommendations regarding calorie and fat content (FDA, 2021; Parsad and Ratain, 2017). Participants were required to complete their breakfast within 30 min and abstain from food for at least 4 h post-administration. Standardized meals were provided at least 4 and 10 h post-administration for lunch and dinner, respectively.

Blood samples were collected at specified time points post-administration for all studies. In Study 1, venous blood (approximately 4 mL) was drawn at 18 predetermined intervals: 0 h (within 1 h pre-administration), 1 h, 2 h, 2.5 h, 3 h, 3.5 h, 4 h, 4.5 h, 5 h, 6 h, 8 h, 12 h, 24 h, 48 h, 72 h, 96 h, 120 h, and 168 h. The previous food effect study showed that T_max of regorafenib was delayed due to food effect, thus in Studies 2 and 3, a blood sampling points of 5.5 h was added and point two h was deleted. The blood samples were 0 h (within 1 h pre-administration), 1 h, 2.5 h, 3 h, 3.5 h, 4 h, 4.5 h, 5 h, 5.5 h, 6 h, 8 h, 12 h, 24 h, 48 h, 72 h, 96 h, 120 h, and 168 h.

Blood samples were centrifuged at 2°C–8°C and 1,700 g for 10 min. Plasma was separated and divided into two parts: one for testing and one for backup storage. Samples were stored at −20°C and later transferred to a long-term storage freezer at −60 to −90°C prior to pharmacokinetic analysis.

Plasma samples were analyzed at Wuhan Hongren Biopharmaceutical Inc. (Wuhan, China) using a validated liquid chromatography-tandem mass spectrometry (LC-MS/MS) method. Briefly, regorafenib and M-2 were quantified by mixing 50 μL of K₂EDTA human plasma with 50 μL of internal standard followed by protein precipitation with acetonitrile. The supernatant was then transferred and diluted before injection into the LC-MS/MS system. Data were acquired and processed using Analyst 1.6.3 software, and sample management and regression calculations were performed using Watson^® LIMS software (version 7.5). The linearity range for regorafenib and M-2 was 3.00–1,500 ng/mL and 1.00–500 ng/mL, respectively. The interday precision (CV% coefficient of variation) for regorafenib was <5.4%, and the accuracy ranged within −2.3%–2.7%. The interday precision (CV% coefficient of variation) for M-2 was <6.8%, and the accuracy ranged within −2.0%–0.4%.

Primary pharmacokinetic parameters included the peak plasma concentration (C_max), the area under the concentration-time curve from time zero to 168 h (AUC_0–168h), and the extrapolated area under the curve from time zero to infinity (AUC_0–∞) for regorafenib. Secondary parameters included time to reach C_max (T_max), half-life (t_1/2), elimination rate constant (λ_z), and the same primary parameters for M-2. Due to the low blood concentration and pharmacological activity of M-5, the metabolite M-5 was not detected, only M-2 was measured.

Participants were continuously monitored through vital sign assessments, physical examinations, laboratory tests (including hematology, biochemistry, and urinalysis), and standard 12-lead electrocardiograms (ECGs). Adverse events (AEs) were graded using the Common Terminology Criteria for Adverse Events (CTCAE) version 5.0 and categorized by System Organ Class or Preferred Term according to the Medical Dictionary for Regulatory Activities (MedDRA).

The sample size was determined based on statistical considerations. For Study 1, assuming the true ratio was 0.92 and the intra-individual coefficient of variation (Intra-CV) was 26% in a two-way crossover design, 52 participants were randomized to achieve at least 80% power for meeting the bioequivalence criteria. Considering a subject withdrawal rate of 20%, the number of participants was increased to 64. For Studies 2 and 3, assuming the true ratio was 0.9, Intra-CV was 28%, and other settings were consistent with those in the fasting study, the sample number was set at 76. Bioequivalence was assessed using a mixed-effects model with analysis of variance in SAS (version 9.4). Participants who had at least one evaluable pharmacokinetic parameter were included in the pharmacokinetic analysis set, while those who completed at least one treatment period were included in the bioequivalence analysis set. If there are cases where the extrapolated area percentage exceeds 20%, a sensitivity analysis should be conducted to further determine whether such data can be excluded. Primary endpoints (C_max, AUC_0–168h, and AUC_0–∞ of regorafenib) were log-transformed, and the 90% confidence intervals (CIs) for the geometric mean ratios (GMRs) of the primary parameters were calculated. Bioequivalence was assumed if the 90% CIs of GMRs for C_max, AUC_0–168h, and AUC_0–∞ of the prototype of regorafenib fell within the predefined range of 80.00%–125.00%. Descriptive statistics were provided for all other parameters.

Study 1 was conducted from 22 September 2021, to 8 November 2021. Initially, sixty-four subjects were enrolled. However, two participants withdrew before the first administration due to personal reasons, and six discontinued participation due to adverse events (AEs) after the first period. Consequently, fifty-six subjects completed both treatment periods.

Study 2 ran from 16 June 2021, to 2 August 2021, with seventy-six subjects being enrolled. One participant withdrew before the first administration for personal reasons, and two subjects discontinued due to AEs. In total, seventy-three subjects completed both treatment periods.

Study 3 took place from 2 February 2023, to 13 March 2023, involving seventy-six subjects. One subject withdrew before the second administration due to an AE, resulting in seventy-five subjects completing both treatment periods. Baseline demographics and participant characteristics are detailed in Table 1, while the participant disposition is illustrated in Figure 1.

The mean blood concentration-time profiles of regorafenib and M-2 for both the test and reference formulations under three dietary conditions are graphically represented in Figure 2. Summaries of the pharmacokinetic (PK) parameters for both formulations are provided in Table 2. The geometric mean ratios (GMRs) and corresponding 90% confidence intervals (CIs) for the primary PK parameters of regorafenib are presented in Table 3. The GMRs and 90% CIs for C_max, AUC_0–168h, and AUC_0–∞ of the prototype of regorafenib across studies 1, 2, and 3 all fell within the bioequivalence acceptance range of 80.00%–125.00%, thereby establishing bioequivalence. The BE assessment of M-2 for the two formulations was equivalent under fasting and low-fat conditions, but not equivalent under high-fat conditions. However, M-2 data were provided as Supplementary Material only and were not considered as the result of equivalence assessment. Details are presented in Supplementary Table 1.

In Study 1, thirty-three out of sixty-four subjects (53.2%) experienced a total of eighty-eight AEs, eighty-four of which were considered possibly related to the test or reference formulations. Fifty-two and thirty-two AEs were likely related to the test and reference formulations, respectively. All AEs were mild (Grade 1), and all participants fully recovered or showed improvement.

In Study 2, fifty-three out of seventy-five subjects (70.7%) experienced one hundred forty-seven AEs, with one hundred thirty-nine considered possibly related to the test or reference formulations. Seventy-four and sixty-five AEs were likely attributable to the test and reference products, respectively.

In Study 3, forty-four out of seventy-six subjects (57.9%) experienced ninety-one AEs, with seventy-one considered possibly related to the test or reference formulations. Thirty-seven and thirty-four AEs were likely associated with the test and reference products, respectively. Most AEs were asymptomatic abnormalities in laboratory results or vital signs, and no vomiting was reported. No other AEs were likely to have impacted the PK parameters. No hand-foot skin reactions were observed, though one subject experienced dry skin. Most AEs were Grade 1, and all participants fully recovered or improved.

No serious AEs were reported in any of the studies, and there was no statistically significant difference in the incidence of AEs between the test and reference formulations. Further details are available in Supplementary Table S2. The single 40 mg dose of regorafenib was generally safe and well tolerated.