Clinical dose rationale of tislelizumab in patients with solid or hematological advanced tumors

Abstract

Tislelizumab, an anti‐programmed cell death protein 1 monoclonal antibody, has demonstrated improved survival benefits over standard of care for multiple cancer indications. We present the clinical rationale and data supporting tislelizumab dose recommendation in patients with advanced tumors. The phase I, first‐in‐human, dose‐finding BGB‐A317‐001 study (data cutoff [DCO]: August 2017) examined the following tislelizumab dosing regimens: 0.5–10 mg/kg every 2 weeks (q2w), 2–5 mg/kg q2w or q3w, and 200 mg q3w. Similar objective response rates (ORRs) were reported in the 2 and 5 mg/kg q2w or q3w cohorts. Safety outcomes (grade ≥3 adverse events [AEs], AEs leading to dose modification/discontinuation, immune‐mediated AEs, and infusion‐related reactions) were generally comparable across the dosing range examined. These results, alongside the convenience of a fixed q3w dose, formed the basis of choosing 200 mg q3w as the recommended dosing regimen for further clinical use. Pooled exposure–response (E–R) analyses by logistic regression using data from study BGB‐A317‐001 (DCO: August 2020) and three additional phase I/II studies (DCOs: 2018–2020) showed no statistically significant correlation between tislelizumab pharmacokinetic exposure and ORR across multiple solid tumor types or classical Hodgkin's lymphoma, nor was exposure associated with any of the safety end points evaluated over the dose range tested. Hence, tislelizumab showed a relatively flat E–R relationship. Overall, the totality of data, including efficacy, safety, and E–R analyses, together with the relative convenience of a fixed q3w dose, provided clinical rationale for the recommended dosing regimen of tislelizumab 200 mg q3w for multiple cancer indications.

Study Highlights.

• WHAT IS THE CURRENT KNOWLEDGE ON THE TOPIC?

Tislelizumab, an anti‐programmed cell death protein 1 monoclonal antibody, has demonstrated improved survival benefits over standard of care in multiple cancer indications.

• WHAT QUESTION DID THIS STUDY ADDRESS?

This study aimed to provide the rationale for selecting the dose regimen of tislelizumab 200 mg once every 3 weeks (q3w) by evaluating the efficacy and safety findings of a previously reported early‐phase tislelizumab clinical trial. It also used exposure–response (E–R) analyses to assess whether data from four early‐phase clinical studies supported the use of this dosing regimen in patients with advanced solid tumors or classical Hodgkin's lymphoma.

• WHAT DOES THIS STUDY ADD TO OUR KNOWLEDGE?

This study highlights a lack of difference across efficacy and safety outcomes for examined dosing levels and frequency of administration. It also revealed that across four early‐phase clinical studies and in patients with multiple tumor types, there were no statistically significant E–R relationships between tislelizumab exposure and the selected efficacy or safety end points. Results from this analysis support the recommendation of intravenous tislelizumab 200 mg q3w, which is clinically more convenient to administer.

• HOW MIGHT THIS CHANGE CLINICAL PHARMACOLOGY OR TRANSLATIONAL SCIENCE?

This analysis provides clinically relevant data underpinning the dose recommendation for tislelizumab. This is of value when planning future studies evaluating tislelizumab, especially combination therapy regimens.

INTRODUCTION

Tislelizumab is a humanized IgG4 monoclonal antibody (mAb) with a specificity for programmed cell death protein 1 (PD‐1). ¹ , ² It is engineered to inhibit PD‐1 signaling by binding to PD‐1 with high affinity and specificity. ¹ , ² Across phase III clinical trials, tislelizumab 200 mg once every 3 weeks (q3w) has been associated with a significant improvement in overall survival or prolonged progression‐free survival versus comparator, and a manageable safety profile, as: a first‐ and second‐line treatment for advanced or metastatic esophageal squamous cell carcinoma (ESCC), ³ , ⁴ a first‐line treatment for advanced non‐squamous and squamous non‐small cell lung cancer (NSCLC), ⁵ , ⁶ a first‐line treatment for advanced gastric or gastroesophageal junction adenocarcinoma (interim analysis), ⁷ a first‐line treatment for recurrent or metastatic nasopharyngeal cancer, ⁸ a treatment for previously treated advanced NSCLC, ⁹ and a first‐line treatment for small cell lung cancer. ¹⁰

The pharmacokinetic (PK) profile of tislelizumab was first characterized along with its safety, tolerability, and preliminary antitumor activity in a phase Ia/Ib, open‐label, dose‐finding study (BGB‐A317‐001), which evaluated the following weight‐based doses: 0.5, 2, 5, and 10 mg/kg every 2 weeks (q2w), 2 and 5 mg/kg q3w, and a fixed dose of 200 mg q3w. ² In this study, tislelizumab exhibited dose‐proportional PKs in the dose ranges tested with no correlation between clearance and baseline body weight. Tislelizumab also demonstrated a wide therapeutic index with no maximum tolerated dose identified up to 10 mg/kg q2w. Safety and efficacy were generally comparable between all dose levels and frequencies examined. In light of these results and given the relative convenience of a fixed q3w dosing regimen, the fixed dose of intravenous (i.v.) tislelizumab 200 mg q3w was selected for further clinical evaluation. In addition, 200 mg q3w lies to the lower end of the 2–5 mg/kg q3w dose for patients with a body weight of ~70 kg. This dosing regimen of 200 mg q3w has since been used in multiple phase III studies ³ , ⁴ , ⁵ , ⁹ and is the approved clinical dose for multiple indications in China and as a second‐line treatment for ESCC in Europe and Korea. ¹¹ , ¹² , ¹³

In a subsequent population PK analysis based on pooled data from 2596 patients with cancer across 12 clinical studies, none of the covariates evaluated, including age, body weight, race, sex, hepatic function, renal function, tumor size or type, and line of treatment, had a clinically meaningful impact on the PKs of tislelizumab and no dose modifications were required based on patient baseline characteristics or clinical factors. ¹⁴ In this analysis, the time to reach 90% of the steady‐state level was ~84 days (12 weeks) and the elimination half‐life was estimated to be 23.8 days. ¹⁴

To support and justify the use of the 200 mg q3w dose in clinical trials, it is important to understand the relationship between dose‐response and exposure‐response. It is generally assumed that a positive exposure–response (E–R) correlation suggests that a higher exposure to therapy (or a higher dose) would lead to improved responses; however, data from recent immuno‐oncology clinical trials have demonstrated that the E–R relationship for mAbs can be confounded by various factors, such as patient baseline characteristics, drug clearance, prior exposure, and disease status. ¹⁵ This has resulted in increased regulatory scrutiny to justify the appropriateness of the chosen clinical dose, and whether a higher dose would lead to better efficacy outcomes. Thus, it is crucial to utilize early clinical data, where multiple dose levels are tested, to uncover E–R relationships for dose optimization. ¹⁶

The primary purpose of the current analysis was to provide the clinical rationale for selecting the dose regimen of tislelizumab 200 mg q3w for the treatment of various advanced tumor types by further evaluating the efficacy and safety findings of the previously reported phase Ia BGB‐A317‐001 study. ² The secondary purpose of the current analysis aimed to provide additional support for this clinical rationale by performing E–R analyses on pooled data from four early‐phase clinical trials in patients with advanced tumors.

METHODS

Clinical studies and patient populations

This analysis included four early‐phase tislelizumab clinical trials: BGB‐A317‐001 (NCT02407990), ² BGB‐A317‐102 (NCT04068519), ¹⁷ BGB‐A317‐203 (NCT03209973), ¹⁸ and BGB‐A317‐204 (NCT04004221). ¹⁹ All of these studies enrolled adult patients with histologically or cytologically confirmed cancer with measurable disease. Study BGB‐A317‐001 investigated multiple dose regimens, as described below. ² Patients from the other three studies all received tislelizumab 200 mg q3w. ¹⁷ , ¹⁸ , ¹⁹

The first part of the present analysis evaluated data from phase Ia of study BGB‐A317‐001. Study BGB‐A317‐001 was a multicenter, open‐label, phase Ia/Ib, dose escalation and expansion study to investigate the safety, PKs, and antitumor activities of i.v. tislelizumab in patients with advanced or metastatic tumors, including NSCLC, ovarian cancer, gastric cancer, hepatocellular carcinoma, head and neck squamous cell carcinoma, ESCC, triple‐negative breast cancer, cholangiocarcinoma, renal cell cancer, bladder cancer, melanoma, Merkel‐cell carcinoma, sarcoma, gastrointestinal stromal tumor, cutaneous squamous cell carcinoma, or tumors with known microsatellite instability‐high (MSI‐H). ² Phase Ia of the study included: part (1) a dose escalation phase that tested four weight‐based dose levels (0.5, 2, 5, and 10 mg/kg q2w); part (2) a schedule expansion that evaluated tislelizumab at 2 and 5 mg/kg q2w and q3w; and part (3) a fixed‐dose expansion that evaluated tislelizumab 200 mg q3w. Phase Ib of the study was an indication expansion phase that evaluated tislelizumab 5 mg/kg q3w in nine disease‐specific cohorts.

The second part of the present analysis was to perform E–R analyses on pooled data from all four clinical trials (BGB‐A317‐001, BGB‐A317‐102, BGB‐A317‐203, and BGB‐A317‐204). Study BGB‐A317‐102 was a multicenter, open‐label, noncomparative, phase I/II study investigating the safety, tolerability, PKs, and antitumor activity of i.v. tislelizumab in Chinese patients with advanced or metastatic, unresectable solid tumors. ¹⁷ Study BGB‐A317‐203 was a multicenter, open‐label, single‐arm, phase II study of tislelizumab in relapsed or refractory classical Hodgkin's lymphoma (cHL). ¹⁸ Study BGB‐A317‐204 was a multicenter, single‐arm, phase II study of tislelizumab in patients with previously treated, PD‐1‐positive, locally advanced or metastatic, urothelial bladder cancer. ¹⁹

Efficacy and safety end points

This analysis evaluated efficacy and safety end points as binary responses by incidence (yes/no) from the four studies.

Efficacy (antitumor activity) was assessed by computed tomography imaging or magnetic resonance imaging at screening, every 8 or 9 weeks (for the q2w or q3w groups, respectively) for the first 12 months, and every 12 weeks thereafter. Objective response rate (ORR) was used as the efficacy end point in this analysis. The ORR was defined as the proportion of patients who had a complete or partial response according to the investigator per the Response Evaluation Criteria in Solid Tumors version 1.1 in patients with solid tumors, ²⁰ or according to the Revised International Working Group Criteria for Malignant Lymphomas (the Lugano Classification) in patients with cHL. ²¹

Safety and tolerability assessments included monitoring of adverse events (AEs), as well as vital signs, physical examinations, electrocardiograms, and laboratory investigations at prespecified study visits. AEs were graded according to National Cancer Institute Common Terminology Criteria for AEs, version 4.03. Safety end points assessed included the proportion of patients with grade 3 or higher AEs, AEs leading to treatment discontinuation, AEs leading to dose modification, immune‐mediated AEs, and infusion‐related reactions.

Efficacy and safety by dose stratification in study BGB‐A317‐001

Efficacy and safety data were analyzed in patients who received at least one dose of tislelizumab during phase Ia of study BGB‐A317‐001 to determine which of the evaluated doses (0.5, 2, 5, and 10 mg/kg q2w; 2 and 5 mg/kg q3w; and 200 mg q3w) were associated with the best ORR while minimizing AEs. Results from patients who received tislelizumab 2 mg/kg q2w or 5 mg/kg q2w were pooled across the phase Ia study.

E–R analyses based on pooled data

Empirical Bayesian estimates of individual PK parameters for all of the tislelizumab‐treated patients were derived from the final population PK model previously described in Budha et al., 2023. ¹⁴ The model was used to predict exposure metrics (steady‐state maximal serum concentration [C _max,ss] and first dose average serum concentration [C _avg,dose1]) for 932 patients across the four studies. The C _avg,dose1 was calculated based on the area under the concentration–time curve for the first dose over time (AUC_dose1/τ), where τ was 21 days for the q3w groups and 14 days for the q2w groups. C _max,ss was the maximum concentration of tislelizumab at a steady state.

The probabilities of objective response and safety events were calculated and plotted against exposure using a quantile binning approach. If an E–R trend was observed, further analysis with linear logistic regression models was performed with PK metrics and other baseline covariates as potential predictors of the probability of events. Model building was conducted with a forward‐addition and backward‐elimination method. Models were compared using a log‐likelihood criterion and a significance level of p < 0.05 for the forward‐addition and p < 0.01 for the backward‐elimination steps.

Two separate exposure–efficacy analyses were performed due to the differences in the definition of objective response used in solid tumors versus blood cancers. The E–R logistic regression models for exposure and ORR in patients with solid tumors treated with tislelizumab (C _avg,dose1) were developed using combined data from all solid tumor types in three studies (BGB‐A317‐001, BGB‐A317‐102, and BGB‐A317‐204), ² , ¹⁷ , ¹⁹ and a separate exposure–efficacy analysis was conducted in patients with cHL (BGB‐A317‐203). ¹⁸ The relationship between exposure and the efficacy end point was further explored by solid tumor type, that is, NSCLC, ESCC, MSI‐H tumors, or urothelial carcinoma (UC). Exposure–safety analyses were conducted using the combined data from all four studies (across solid tumors and cHL).

RESULTS

Patient disposition and baseline characteristics

Patients included in phase Ia of study BGB‐A317‐001

A total of 116 patients from Australia, South Korea, New Zealand, and Taiwan were enrolled in phase Ia of study BGB‐A317‐001 and included in the analyses of efficacy and safety by dose stratification. Patients had a mean ± standard deviation (SD) age of 59.9 ± 12.14 years, and a balanced sex distribution (men, 47.4% [n = 55]; and women, 52.6% [n = 61]). The full patient demographics and baseline disease characteristics for patients enrolled in phase Ia of study BGB‐A317‐001 have been previously described. ²

Patients included in the pooled E–R analyses

A total of 932 patients were included in the pooled E–R analyses from the BGB‐A317‐001 (n = 450), BGB‐A317‐102 (n = 300), BGB‐A317‐203 (n = 70), and BGB‐A317‐204 (n = 112) studies. The mean ± SD age of patients was 56.8 ± 13.0 years and the mean ± SD baseline body weight was 67.3 ± 14.4 kg. The majority of the patients enrolled in these studies were Asian (from either China or South Korea; 65.7% [n = 612/932]). Baseline patient characteristics are summarized in Table 1.

TABLE 1.

Baseline patient characteristics overall and by study.

Mean ± SD, unless otherwise stated	All (N = 932)	BGB‐A317‐001 2 (N = 450)	BGB‐A317‐102 17 (N = 300)	BGB‐A317‐203 18 (N = 70)	BGB‐A317‐204 19 (N = 112)
Age (years)	56.8 ± 13.0	59.5 ± 12.0	55.5 ± 11.5	36.2 ± 12.7	62.0 ± 7.86
Sex (M/F), n	576/356	246/204	207/93	40/30	83/29
Weight (kg)	67.3 ± 14.4	70.6 ± 16.2	62.9 ± 11.6	66.9 ± 13.2	65.8 ± 11.2
Bilirubin (μmol/L)	10.1 ± 6.26	8.79 ± 6.34	12.1 ± 5.11	11.6 ± 9.62	9.15 ± 4.24
ALT (IU/L)	23.0 ± 22.1	24.3 ± 27.7	23.3 ± 15.9	20.1 ± 16.4	18.8 ± 11.5
AST (IU/L)	29.0 ± 20.8	31.7 ± 25.1	29.2 ± 16.9	22.0 ± 11.9	22.2 ± 11.2
Albumin (g/L)	40.3 ± 19.8	36.7 ± 5.62	41.6 ± 4.59	41.9 ± 5.85	50.6 ± 53.9
Creatine (μmol/L)	76.4 ± 23.6	75.3 ± 21.9	73.0 ± 20.1	62.7 ± 20.0	98.8 ± 27.4
EGFR (mL/min/m2)	89.5 ± 22.2	87.7 ± 20.0	93.5 ± 19.1	116 ± 20.8	69.5 ± 18.4
LDH (U/L)	297 ± 301	312 ± 367	304 ± 256	253 ± 135	244 ± 160
Race (White/Asian/Other), n	289/612/31	289/130/31	0/300/0	0/70/0	0/112/0
ECOG PS (0/1), n	350/582	169/281	80/220	48/22	53/59

Abbreviations: ALT, alanine transaminase; AST, aspartate transaminase; ECOG PS, Eastern Cooperative Oncology Group performance status;

EGFR, estimated glomerular filtration rate; F, female; LDH, lactate dehydrogenase; M, male; SD, standard deviation.

A small number of patients were treated with the lowest dose of 0.5 mg/kg (n = 3) and the highest dose of 10 mg/kg (n = 7). A larger number of patients were treated in the 2 mg/kg q2w (n = 26), 5 mg/kg q2w (n = 26), 2 mg/kg q3w (n = 21), and 5 mg/kg q3w (n = 354) cohorts and in the fixed dose expansion cohort of 200 mg q3w (n = 495; Table S1).

Efficacy and safety by dose stratification in phase Ia of study BGB‐A317‐001

The confirmed ORRs in patients treated with tislelizumab 2 mg/kg and 5 mg/kg q2w in phase Ia of this study were both 11.5% (3 responders out of 26 treated patients in both treatment groups); whereas ORRs for patients treated with 2 and 5 mg/kg q3w were 38.1% (n = 8/21) and 15.0% (n = 3/20), respectively (Table 2). In the 200 mg q3w group, the ORR was 23.1% (n = 3/13). No patients achieved a clinical response in the 0.5 or 10 mg/kg q2w groups; these groups included less than 10 patients each.

TABLE 2.

Response stratified by dose in phase Ia of study BGB‐A317‐001. ^{a , 2}

	Weight‐based doses									Fixed dose
	q2w					q3w			q2w and q3w	q3w
	0.5 mg/kg (N = 3)	2.0 mg/kg (N = 26)	5.0 mg/kg (N = 26)	10.0 mg/kg (N = 7)	All q2w (N = 62)	2.0 mg/kg (N = 21)	5.0 mg/kg (N = 20)	All q3w (N = 41)	All (N = 103)	200 mg (N = 13)
ORR (CR, PR)
n (%)	0 (0)	3 (11.5)	3 (11.5)	0 (0)	6 (9.7)	8 (38.1)	3 (15.0)	11 (26.8)	17 (16.5)	3 (23.1)
Best overall response – confirmed, n (%)
CR	0 (0)	1 (3.8)	0 (0)	0 (0)	1 (1.6)	1 (4.8)	0 (0)	1 (2.4)	2 (1.9)	0 (0)
PR	0 (0)	2 (7.7)	3 (11.5)	0 (0)	5 (8.1)	7 (33.3)	3 (15.0)	10 (24.4)	15 (14.6)	3 (23.1)
SD	3 (1.0)	11 (42.3)	9 (34.6)	1 (14.3)	24 (38.7)	6 (28.6)	8 (40.0)	14 (34.1)	38 (36.9)	4 (30.8)
PD	0 (0)	11 (42.3)	13 (50.0)	6 (85.7)	30 (48.4)	7 (33.3)	7 (35.0)	14 (34.1)	44 (42.7)	6 (46.2)
Missing	0 (0)	1 (3.8)	1 (3.8)	0 (0)	2 (3.2)	0 (0)	2 (10.0)	2 (4.9)	4 (3.9)	0 (0)

Abbreviations: CR, complete response; ORR, objective response rate; PD, progressive disease; PR, partial response; q2w, every 2 weeks; q3w, every 3 weeks; SD, stable disease.

^a Data cutoff: August 28, 2017.

The safety profiles observed in patients receiving 2 mg/kg and 5 mg/kg q2w and q3w, including the incidence of grade 3 or higher AEs, were generally comparable (Table 3), and did not indicate any clear dose‐dependence across these regimens. In addition, the incidences of AEs were generally similar between the fixed dose of 200 mg q3w and the weight‐based doses of 2 or 5 mg/kg q2w and q3w.

TABLE 3.

Summary of adverse events stratified by dose in phase Ia of study BGB‐A317‐001. ^{a , 2}

n (%)	Weight‐based doses									Fixed dose
	q2w					q3w			q2w and q3w	q3w
	0.5 mg/kg (N = 3)	2.0 mg/kg (N = 26)	5.0 mg/kg (N = 26)	10.0 mg/kg (N = 7)	All q2w (N = 62)	2.0 mg/kg (N = 21)	5.0 mg/kg (N = 20)	All q3w (N = 41)	All q2w and q3w (N = 103)	200 mg (N = 13)
Any grade ≥3 AE	0 (0)	12 (46.2)	13 (50.0)	3 (42.9)	28 (45.2)	5 (23.8)	10 (50.0)	15 (36.6)	43 (41.7)	4 (30.8)
AE leading to treatment discontinuation	0 (0)	2 (7.7)	3 (11.5)	0 (0)	5 (8.1)	1 (4.8)	2 (10.0)	3 (7.3)	8 (7.8)	0 (0)
AE leading to dose modification b	0 (0)	5 (19.2)	7 (26.9)	0 (0)	12 (19.4)	4 (19.0)	6 (30.0)	10 (24.4)	22 (21.4)	3 (23.1)
Immune‐related AE	0 (0)	9 (34.6)	9 (34.6)	0 (0.0)	18 (29.0)	7 (33.3)	7 (35.0)	14 (34.1)	32 (31.1)	4 (30.8)
Infusion‐related reaction	0 (0)	3 (11.5)	8 (30.8)	1 (14.3)	12 (19.4)	3 (14.3)	4 (20.0)	7 (17.1)	19 (18.4)	0 (0)

Abbreviations: AE, adverse event; q2w, every 2 weeks; q3w, every 3 weeks.

^a Data cutoff: August 28, 2017. AEs were graded according to National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.03.

^b Dose modification comprised treatment interruption or dose delay.

Exposure–efficacy analyses in patients with solid tumors and cHL

The mean C _avg,dose1 values were similar between responders (37.4 μg/mL, n = 136) and nonresponders (39.3 μg/mL, n = 684) in patients with solid tumors. Consistent with these observations, the results from the E–R logistic regression model suggested that tislelizumab C _avg,dose1 was not associated with the probability of objective response in patients with solid tumors in the pooled analysis (Figure 1a; n = 820, p = 0.230). There was no statistically significant relationship between tislelizumab C _avg,dose1 and objective response in patients with the different subcategories of solid tumors evaluated in this analysis: NSCLC (n = 99, p = 0.844), ESCC (n = 71, p = 0.053), MSI‐H tumors (n = 26, p = 0.739), or UC (n = 148, p = 0.975; Figures 1b–e). A positive trend was observed between the probability of objective response and increasing tislelizumab exposure in patients with ESCC; however, this was not statistically significant (p = 0.053). Among patients with cHL treated with tislelizumab 200 mg q3w, 87.1% achieved objective response (n = 61). The E–R logistic regression model showed that tislelizumab C _avg,dose1 was not associated with the probability of objective response in patients with cHL (n = 112, p = 0.749; Figure 1f).

FIGURE 1.

Logistic regression of the probability of achieving objective response versus tislelizumab exposure in patients by tumor type. Logistic modeling showed that there was no statistically significant relationship between tislelizumab C _avg,dose1 and objective response for (a) all solid tumors (n = 820, p = 0.230), (b) NSCLC (n = 99, p = 0.844), (c) ESCC (n = 71, p = 0.053), (d) MSI‐H tumors (n = 26, p = 0.739), (e) UC (n = 148, p = 0.975), or (f) cHL (n = 112, p = 0.749). The open blue circles reflect the observed events (objective response). The filled black symbols are the observed probability of achieving objective response, and the error bars are SE for quantiles (at 100×(1/4)^th percentiles) of exposures plotted at the median value within each quantile, where p is the probability of objective response and n is the number of patients in each quantile. The blue line and light blue shaded area are the median and 95% prediction interval, respectively, based on the 1000 bootstrap samples of the model. The horizon boxplots represent the observed exposure range of 200 mg q3w and 5 mg/kg q3w. C _avg,dose1, first dose average plasma concentration; cHL, classical Hodgkin lymphoma; ESCC, esophageal squamous cell carcinoma; MSI‐H, microsatellite instability‐high; NSCLC, non‐small cell lung cancer; q3w, every 3 weeks; SE, standard error; UC, urothelial carcinoma.

Pooled exposure–safety analyses across patients with solid tumors and cHL

Table 4 provides a summary of the five selected safety outcomes in each of the individual studies and pooled across the four studies. The probabilities of the occurrence of selected AEs by quantiles of tislelizumab exposure (C _max,ss or C _avg,dose1) are shown in Figure 2. Similar ranges of tislelizumab exposure were observed in patients regardless of whether they experienced one of the selected AEs.

TABLE 4.

Summary of selected safety outcomes overall and by study. ^a

	All (N = 932)	BGB‐A317‐001 2 (N = 450)	BGB‐A317‐102 17 (N = 300)	BGB‐A317‐203 18 (N = 70)	BGB‐A317‐204 19 (N = 112)
Data cutoff dates		August 26, 2020 (pooled analysis)	May 31, 2020	November 26, 2018	September 18, 2020
Any grade ≥3 AE, n (%)	442 (47.4)	220 (48.9)	140 (46.7)	22 (31.4)	60 (53.6)
AE leading to treatment discontinuation, n (%)	104 (11.2)	40 (8.9)	38 (12.7)	4 (5.7)	22 (19.6)
AE leading to dose modification, n (%)	269 (28.9)	129 (28.7)	94 (31.3)	12 (17.1)	34 (30.4)
Immune‐mediated AE, n (%)	135 (14.5)	62 (13.8)	29 (9.7)	21 (30.0)	23 (20.5)
Infusion‐related reaction, n (%)	66 (7.1)	29 (6.4)	6 (2.0)	27 (38.6)	4 (3.6)

Abbreviations: AE, adverse event; q2w, every 2 weeks; q3w, every 3 weeks.

^a AEs were graded according to National Cancer Institute Common Terminology Criteria for Adverse Events, version 5.0. The following doses were evaluated: BGB‐A317‐001: 0.5, 2, 5, and 10 mg/kg q2w, 2 and 5 mg/kg q3w, 200 mg q3w; BGB‐A317‐102, BGB‐A317‐203 and BGB‐A317‐204: 200 mg q3w.

FIGURE 2.

E–R analysis for safety end points: probability of selected AEs versus tislelizumab exposure in patients with advanced tumors. This figure presents the probability of selected AEs versus tislelizumab exposure using pooled data (n = 932 patients) from four studies (BGB‐A317‐001, BGB‐A317‐102, BGB‐A317‐203, and BGB‐A317‐204). There were no apparent relationships between tislelizumab exposure and any of the safety end points evaluated (any grade ≥3 AEs, AEs leading to treatment discontinuation, AEs leading to dose modification, immune‐mediated AEs, and infusion‐related reactions) among tislelizumab‐treated patients. The open blue circles reflect the observed AEs (yes or no). The filled black symbols are the observed probability of AEs and the error bars are SE for quantiles (at 100×(1/6)^th percentiles, vertical dotted lines) of exposures plotted at the median value within each quantile. The red lines are smooth curves to show the relationship between two variables. AE, adverse event; C _avg,dose1, first dose average plasma concentration; C _max,ss, steady‐state maximal plasma concentration;E–R, exposure–response; SE, standard error.

There were no apparent relationships between tislelizumab exposure and any of the safety end points evaluated (including grade 3 or higher AEs, AEs leading to treatment discontinuation, AEs leading to dose modification, immune‐mediated AEs, and infusion‐related reactions) among tislelizumab‐treated patients. Hence, no logistic regression models were developed for these end points.

DISCUSSION

The present analysis summarized the efficacy and safety data from study BGB‐A317‐001 that led to the recommendation of the 200 mg q3w fixed dose for use in clinical trials evaluating tislelizumab. The pooled E–R results of this analysis further justify the use of the 200 mg q3w dose by demonstrating that exposure to tislelizumab across the dose regimens evaluated was associated with similar clinical efficacy and safety outcomes in patients with solid tumors or cHL.

To identify the clinical dose of tislelizumab, phase Ia of study BGB‐A317‐001 comprised three parts: dose escalation, schedule expansion, and fixed‐dose expansion. ² Prior reports from this study briefly outlined the comparable safety, PKs, and efficacy profiles between the doses evaluated. The current analysis presents a more detailed breakdown of these findings by dosing level and dosing frequency.

The dose escalation and schedule expansion parts of the study evaluated various weight‐based doses of tislelizumab. This analysis demonstrated that there was a lack of difference across the efficacy and safety outcomes for both the dosing levels (2 mg/kg vs. 5 mg/kg) and frequencies (q2w vs. q3w) evaluated. We also observed that the higher and more frequent dosing regimen of 10 mg/kg q2w did not appear to provide any additional therapeutic benefit, suggesting that further dose escalation is not necessary to achieve optimal clinical outcomes; however, these findings require cautious interpretation owing to the small sample size (n = 7) for the 10 mg/kg q2w dose cohort.

The fixed‐dose expansion part of study BGB‐A317‐001 aimed to assess the potential use of a fixed tislelizumab dose, which would be more convenient than weight‐based dosing. As reported previously, based on results from the dose escalation and schedule expansion, the fixed dose of 200 mg q3w was selected for evaluation as it lies toward the lower end of the range of the 2–5 mg/kg q3w doses for a typical body weight of 70 kg (i.e., approximately the mean body weight of patients enrolled in study BGB‐A317‐001). ¹⁴ This would mean that the majority of patients receiving the fixed dose would have a PK exposure covered by 2–5 mg/kg q3w to avoid PK exposures beyond the established PK exposure range. The current analysis of study BGB‐A317‐001 results showed that efficacy and safety outcomes between tislelizumab 200 mg q3w and the weight‐based dosing regimens were generally comparable. This is consistent with previous reports, which have shown that body weight does not have a clinically meaningful impact on PK exposure, and that the dose of tislelizumab 200 mg q3w provides considerable overlap with the 2 and 5 mg/kg weight‐based doses. ² , ¹⁴ Thus, the 200 mg q3w fixed dose was associated with similar efficacy and safety but had the added benefit of being the most convenient, when compared with the other dose regimens evaluated. For this reason, the 200 mg q3w regimen was chosen for further clinical evaluation. ² , ¹⁴

Although a small number of patients were treated with the lowest dose of 0.5 mg/kg (n = 3) and the highest dose of 10 mg/kg (n = 7), more than 20 patients were treated in each of the 2 and 5 mg/kg cohorts at the two different dose schedules (q2w and q3w; Table S1). In addition to 354 patients treated with 5 mg/kg q3w in BGB‐A317‐001, almost 500 patients were treated with a fixed dose of 200 mg q3w across the four studies. As such, a sufficiently robust dataset was used to conduct the E–R analyses to justify the tislelizumab dose recommendation. The efficacious dose range (2–5 mg/kg) and the highest tested dose with no dose‐limiting toxicities (10 mg/kg) for tislelizumab is similar to other anti‐PD‐1 agents, ²² , ²³ further supporting the recommended dose.

The E–R logistic regression models evaluated in this study did not show a correlation between tislelizumab exposure and ORR in patients with either solid tumors or cHL, further justifying the use of the 200 mg q3w fixed dose. These findings were also generally true across the different types of solid tumors evaluated, including NSCLC, MSI‐H tumors, and UC, although there appeared to be a positive correlation for ESCC that was not statistically significant. This suggests that, within the evaluated dose range, greater tislelizumab exposure, or a higher dose, is generally not expected to confer additional therapeutic benefits in terms of antitumor activity. The pooled data from the four early‐phase clinical studies also showed that there was no observable clinically relevant E–R relationship between tislelizumab exposure and any of the safety end points evaluated. Therefore, a greater exposure to, or higher dose of, tislelizumab among the evaluated doses did not appear to be associated with a higher risk of these safety events.

The 200 mg q3w dosing regimen has been evaluated in subsequent phase II/III clinical studies, and has demonstrated efficacy and manageable safety across patients with a wide range of tumors. ³ , ⁴ , ⁵ , ⁹ , ¹⁸ , ¹⁹ These results support the clinical efficacy and safety of the 200 mg q3w dose.

This work has several limitations. First, the dose‐ranging analyses of tislelizumab were limited to the dose escalation/dose expansion portion of the first‐in‐human study, meaning that the number of patients treated with the lowest (i.e., 0.5, 2 mg/kg) and highest doses (10 mg/kg) was relatively small. Second, for certain disease indications, such as MSI‐H cancers or cHL, the number of enrolled patients was small (n = 26 for MSI‐H) or only one dose level was tested (200 mg q3w for cHL), thus limiting the ability to interpret E–R relations in these indications. Third, the effect of baseline PD‐L1 expression was not assessed in the current analysis, an imbalance in which across different dose groups could potentially obscure inherent E–R relations and the interpretation thereof. Last, a randomized dose optimization study was not conducted prior to the phase III studies. Nevertheless, this analysis included data from a relatively wide range of doses and schedules from a sufficiently large number of patients and utilized pooled data to provide a comprehensive evaluation of tislelizumab clinical dosing schedule.

This analysis demonstrated that the more convenient dosing schedule of i.v. tislelizumab 200 mg q3w has comparable efficacy and safety outcomes with the weight‐based doses and dosing frequencies evaluated in an early‐phase clinical study in patients with advanced cancer. Furthermore, no statistically significant E–R relationships were observed between tislelizumab exposure and the efficacy or safety end points in four early‐phase clinical studies in patients with advanced cancer. Hence, this analysis provides the rationale underpinning the recommendation of i.v. tislelizumab 200 mg q3w for clinical use.

AUTHOR CONTRIBUTIONS

T.Y., X.L., C.‐Y.W., Z.T., H.W., P.S., Y.W., K.W., L.L., Y.G., S.S., and N.B. wrote the manuscript. T.Y., C.‐Y.W., S.S., and N.B. designed and performed the research. T.Y., X.L., C.‐Y.W., Z.T., Y.W., K.W., L.L., Y.G., S.S., and N.B. analyzed data.

FUNDING INFORMATION

This work was supported by BeiGene, Ltd.

CONFLICT OF INTEREST STATEMENT

T.Y., X.L., Z.T., H.W., P.S., and N.B. are employees of BeiGene USA, Inc. and own stock in BeiGene. C.‐Y.W. and S.S. are former employees of BeiGene USA, Inc. Y.W. is an employee of BeiGene (Shanghai) Co., Ltd. and owns stock in BeiGene. K.W., L.L., and Y.G. are employees of Shanghai Qiangshi Information Technology Co., Ltd.

Supporting information

Table S1.

Yu T, Liu X, Wu C‐Y, et al. Clinical dose rationale of tislelizumab in patients with solid or hematological advanced tumors. Clin Transl Sci. 2024;17:e13769. doi: 10.1111/cts.13769

DATA AVAILABILITY STATEMENT

On request, and subject to specific criteria, conditions, and exceptions, BeiGene Ltd., will provide access to individual de‐identified participant data from BeiGene‐sponsored global interventional clinical studies conducted for medicines for (1) indications that have been approved or (2) in programs that have been terminated. BeiGene will also consider requests for the protocol, data dictionary, and statistical analysis plan. Data requests may be submitted to datadisclosure@beigene.com.