Purpose:

Progression-free survival (PFS) was significantly improved with nivolumab 480 mg plus relatlimab 160 mg fixed-dose combination (FDC) every 4 weeks (Q4W) versus nivolumab alone in patients with previously untreated advanced melanoma in RELATIVITY-047. In addition, RELATIVITY-020 (Part D) demonstrated a manageable safety profile and potential for durable response with nivolumab plus relatlimab in previously treated patients. Here, we evaluate the clinical pharmacology profile (CPP) of nivolumab plus relatlimab to support the approved regimen for adult and adolescent patients with advanced melanoma and its continued clinical development in solid tumors.

Experimental Design:

The pharmacokinetics (PK) and immunogenicity of relatlimab and nivolumab were assessed using data from RELATIVITY-047 and RELATIVITY-020. Patients with advanced solid tumors received relatlimab alone or nivolumab plus relatlimab as single-agent vials (SAV) or FDC. PK was characterized using a population PK (popPK) model.

Results:

Relatlimab demonstrated nonlinear and time-varying PK. Nonlinearity in relatlimab PK represented approximately 31% of total CL of relatlimab 160 mg Q4W. Relatlimab PK was dose proportional at doses ≥160 mg Q4W. Geometric mean exposures were similar for SAV and FDC cohorts receiving equivalent dosing regimens. No dose adjustment was required for covariates. Incidence of relatlimab antidrug antibodies was <6% for nivolumab plus relatlimab and had no clinically meaningful impact. There was no PK-related drug interaction of nivolumab plus relatlimab.

Conclusions:

The CPP of relatlimab alone or in combination with nivolumab supports the approved dosing in advanced melanoma and the continued evaluation of nivolumab and relatlimab across other solid tumors.

See related commentary by Gopalakrishnan and Amaria, p. 2862

Translational Relevance

This analysis characterized the pharmacokinetics (PK) of a novel fixed-dose combination of nivolumab and relatlimab and assessed the effect of intrinsic and extrinsic factors, special populations, formulation, infusion duration, and combination effect on PK. There was no clinically relevant effect of the other intrinsic or extrinsic factors evaluated on relatlimab or nivolumab PK, including age, sex, race, albumin (relatlimab only), LDH, Eastern Cooperative Oncology Group status, renal function [estimated glomerular filtration rate (eGFR)], and hepatic function. On the basis of the evaluation of these factors, no dose adjustment is required in special populations, including the elderly, adolescents, or patients with renal or hepatic impairment. This analysis supports the approved dose in adult and adolescent patients with melanoma as well as the continued evaluation of the safety and efficacy of this combination across other solid tumors.

The introduction of immune checkpoint inhibitors (ICI) has dramatically improved clinical outcomes for patients with cancer (1, 2). However, despite clinical benefit, some patients will not respond or will relapse due to various mechanisms of tumor resistance. Therefore, there is an unmet need for the development of combination therapies to overcome resistance to ICI monotherapy (3, 4). Lymphocyte activation gene-3 (LAG-3 or CD223) is an immune checkpoint and is a cancer immunotherapeutic target due to its inhibitory biologic effects on the function of T cells (5).

Relatlimab is a human LAG-3–blocking antibody that binds to the LAG-3 receptor with high affinity, and blocks interactions with its ligands such as MHC class II and other emerging ligands, including fibrinogen-like protein 1 (5, 6). In preclinical studies, the combination of relatlimab plus nivolumab, a fully human IgG4 anti-programmed cell death 1 (PD-1) antibody (7), enhanced antitumor activity more than either agent alone, leading to T-cell activation, proliferation, and restored effector function (5).

Relatlimab 160 mg administered in combination with nivolumab 480 mg as a 1:3 protein-mass ratio, fixed-dose combination (FDC) over 60 minutes every 4 weeks (Q4W) demonstrated a statistically significant improvement in progression-free survival (PFS) by blinded independent central review (BICR) compared with nivolumab alone for the treatment of patients with previously untreated unresectable or metastatic melanoma in the randomized phase 2/3 trial RELATIVITY-047 (NCT03470922; ref. 8). Nivolumab plus relatlimab FDC demonstrated a clinically meaningful (although not statistically significant) improvement in overall survival (OS) and descriptively increased objective response rate by BICR versus nivolumab, with a manageable safety profile (9).

RELATIVITY-020 (NCT01968109) is an ongoing, phase 1/2a, dose-escalation and cohort-expansion trial evaluating relatlimab as monotherapy or in combination with nivolumab in advanced solid tumors (10). In RELATIVITY-020 part D, this Q4W nivolumab and relatlimab dosing regimen demonstrated a manageable safety profile and the potential for durable clinical activity in patients with advanced melanoma whose disease had progressed on prior anti–PD-1 or anti–programmed death ligand-1 (PD-L1) containing therapies (11).

In this study, we reported PK characterization of nivolumab plus relatlimab when administered in combination. This model-based characterization provided clinical pharmacology profiling (CPP) in patients with advanced solid tumors and sought to determine the magnitude of effect of intrinsic covariates on relatlimab pharmaco-kinetics (PK), focusing on covariates determined to have statistically significant effects in previous analyses of therapeutic mAbs (12, 13). The analysis also supported the posology of this novel immunooncology (IO) FDC in different patient populations (prior-IO–treated and IO-naive advanced melanoma); in special populations such as adolescents, patients with hepatic impairment or renal impairment, the elderly, and patients with heavy body weight; and in the assessment of drug–drug interactions and biocomparability between the single-agent vial (SAV) and FDC drug products. In addition, the shortened infusion time and PK-based extrapolation and dose recommendation in adolescent patients was solely modelinformed in the absence of clinical data. Finally, the model-informed CPP supported the bridging of safety and efficacy data from the phase 1/2a study in prior-IO melanoma using SAV and FDC drug products with that from patients with IO-naive melanoma using an FDC drug product in a pivotal phase-3 study. On the basis of these clinical studies and the clinical pharmacology evaluation, nivolumab plus relatlimab 480 mg/160 mg FDC Q4W was approved by the FDA and European Medicines Agency in adults and adolescents with unresectable or metastatic melanoma with a 30-minute administration time (see indication details in the label; refs. 6, 14). In addition, this approach supports continued evaluation of the safety and efficacy of this combination across other solid tumor types (6, 10).

Relatlimab population PK

The relatlimab analysis dataset included 10,015 relatlimab concentration values from 1,713 patients from two studies conducted in patients with unresectable or metastatic melanoma (RELATIVITY-047, n = 332 with ≥1 evaluable PK sample) or melanoma and other solid tumors (RELATIVITY-020, n = 1,381 with ≥1 evaluable PK sample) receiving nivolumab plus relatlimab. Patients with solid tumors received relatlimab monotherapy intravenously [20–800 mg every 2 weeks (Q2W)], nivolumab and relatlimab as SAVs administered sequentially (nivolumab followed by relatlimab) or coadministered, or FDC (nivolumab/relatlimab 80 mg/20 mg to 240 mg/240 mg Q2W and 480 mg/160 mg to 480 mg/1,440 mg Q4W). Relatlimab was infused over 60 minutes in both studies. The PK analyses evaluated both 30- and 60-minute infusion rates by comparing the model predicted exposures following both infusion rates. Both studies allowed the enrollment of pediatric patients ages >12 years; however, no adolescent patients were enrolled.

Patient demographics and baseline covariates in the relatlimab analysis dataset are provided in Table 1. The relatlimab population PK (popPK) model was developed in the following 3 stages: base, full, and final models. Base model development determined structural, inter-individual variability, and residual error components of the popPK model. The effect of covariates on PK model parameters was assessed by the full model. The potential clinical relevance of continuous and categorical covariates was assessed using a range of 80% to 125% as a reference (see Supplementary Information for detailed criteria). The final model was developed from the full model by elimination of the nonsignificant covariates. The detailed model development is described in the Supplementary Information (relatlimab model development).

Table 1.

Patient demographics and baseline characteristics in the relatlimab dataset.

First-line melanoma (n = 456)Prior immunotherapy melanoma (n = 791)Other solid tumors (n = 466)Total (n = 1,713)
Age, years 
 Mean (SD) 61.2 (14) 60.9 (13.4) 59.7 (11.5) 60.7 (13.1) 
 Median (min–max) 63 (20–90) 62 (17–92) 61 (18–88) 62 (17–92) 
Baseline body weight (kg) 
 Mean (SD) 80.9 (18.6) 79.1 (18.4) 73.3 (16.8) 78 (18.3) 
 Median (min–max) 80 (39–163) 77 (41.1–170) 71.5 (37–145) 76.5 (37–170) 
 Missing, n (%) 3 (0.379) 3 (0.175) 
Sex N (%) 
 Male 266 (58.3) 473 (59.8) 317 (68.0) 1,056 (61.6) 
 Female 190 (41.7) 318 (40.2) 149 (32.0) 657 (38.4) 
Race N (%) 
 White 443 (97.1) 750 (94.8) 431 (92.5) 1,624 (94.8) 
 Black/African American 7 (0.9) 10 (2.1) 17 (1.0) 
 Asian 1 (0.2) 29 (3.7) 11 (2.4) 41 (2.4) 
 Other 7 (1.5) 5 (0.6) 14 (3.0) 26 (1.5) 
 Missing 5 (1.1) 5 (0.3) 
Baseline serum albumin, g/dL 
 Mean (SD) 4.14 (0.457) 3.96 (0.473) 3.75 (0.478) 3.95 (0.492) 
 Median (min–max) 4.2 (2.1–5.2) 4 (2–6) 3.8 (2.2–4.8) 4 (2–6) 
 Missing, n (%) 23 (5.04) 69 (8.72) 37 (7.94) 129 (7.53) 
Baseline lactate dehydrogenase (U/L) 
 Mean (SD) 355 (417) 377 (482) 272 (227) 342 (411) 
 Median (min–max) 238 (60–4,641) 237 (85–7,464) 206 (90—2,367) 230 (60—7,464) 
 Missing, n (%) 2 (0.439) 11 (1.39) 1 (0.215) 14 (0.817) 
Baseline eGFR (mL/min/1.73 m2
 Mean (SD) 86.7 (18.6) 86.8 (18.5) 86.6 (20.9) 86.7 (19.2) 
 Median (min–max) 89.6 (25.7–132) 88 (34.2–144) 89.5 (27.3–159) 88.5 (25.7–159) 
 Missing, n (%) 6 (1.32) 2 (0.253) 8 (0.467) 
Liver dysfunction group N (%) 
 Group A: Normal 396 (86.8) 701 (88.6) 354 (76.0) 1,451 (84.7) 
 Group B: Mild 53 (11.6) 84 (10.6) 108 (23.2) 245 (14.3) 
 Group C: Moderate 5 (1.1) 3 (0.4) 4 (0.9) 12 (0.7) 
 Group D: Severe 1 (0.2) 1 (0.1) 
 Missing 1 (0.2) 3 (0.4) 4 (0.2) 
Baseline LAG-3 expression N (covariance; 1% cutoff) 
 Negative 100 (21.9) 246 (31.1) 346 (20.2) 
 Positive 281 (61.6) 379 (47.9) 660 (38.5) 
 Unknown 75 (16.4) 166 (21.0) 466 (100.0) 707 (41.3) 
Baseline ECOG performance status N (%) 
 0 322 (70.6) 540 (68.3) 224 (48.1) 1,086 (63.4) 
 1 134 (29.4) 240 (30.3) 227 (48.7) 601 (35.1) 
 2 11 (1.4) 3 (0.6) 14 (0.8) 
 Missing 12 (2.6) 12 (0.7) 
ADA N (%) 
 Negativea 422 (92.5) 726 (91.8) 374 (80.3) 1,522 (88.8) 
 Positiveb 34 (7.5) 65 (8.2) 92 (19.7) 191 (11.2) 
Therapy N (%) 
 Combination therapy 456 (100.0) 791 (100.0) 445 (95.5) 1,692 (98.8) 
 Monotherapy 21 (4.5) 21 (1.2) 
First-line melanoma (n = 456)Prior immunotherapy melanoma (n = 791)Other solid tumors (n = 466)Total (n = 1,713)
Age, years 
 Mean (SD) 61.2 (14) 60.9 (13.4) 59.7 (11.5) 60.7 (13.1) 
 Median (min–max) 63 (20–90) 62 (17–92) 61 (18–88) 62 (17–92) 
Baseline body weight (kg) 
 Mean (SD) 80.9 (18.6) 79.1 (18.4) 73.3 (16.8) 78 (18.3) 
 Median (min–max) 80 (39–163) 77 (41.1–170) 71.5 (37–145) 76.5 (37–170) 
 Missing, n (%) 3 (0.379) 3 (0.175) 
Sex N (%) 
 Male 266 (58.3) 473 (59.8) 317 (68.0) 1,056 (61.6) 
 Female 190 (41.7) 318 (40.2) 149 (32.0) 657 (38.4) 
Race N (%) 
 White 443 (97.1) 750 (94.8) 431 (92.5) 1,624 (94.8) 
 Black/African American 7 (0.9) 10 (2.1) 17 (1.0) 
 Asian 1 (0.2) 29 (3.7) 11 (2.4) 41 (2.4) 
 Other 7 (1.5) 5 (0.6) 14 (3.0) 26 (1.5) 
 Missing 5 (1.1) 5 (0.3) 
Baseline serum albumin, g/dL 
 Mean (SD) 4.14 (0.457) 3.96 (0.473) 3.75 (0.478) 3.95 (0.492) 
 Median (min–max) 4.2 (2.1–5.2) 4 (2–6) 3.8 (2.2–4.8) 4 (2–6) 
 Missing, n (%) 23 (5.04) 69 (8.72) 37 (7.94) 129 (7.53) 
Baseline lactate dehydrogenase (U/L) 
 Mean (SD) 355 (417) 377 (482) 272 (227) 342 (411) 
 Median (min–max) 238 (60–4,641) 237 (85–7,464) 206 (90—2,367) 230 (60—7,464) 
 Missing, n (%) 2 (0.439) 11 (1.39) 1 (0.215) 14 (0.817) 
Baseline eGFR (mL/min/1.73 m2
 Mean (SD) 86.7 (18.6) 86.8 (18.5) 86.6 (20.9) 86.7 (19.2) 
 Median (min–max) 89.6 (25.7–132) 88 (34.2–144) 89.5 (27.3–159) 88.5 (25.7–159) 
 Missing, n (%) 6 (1.32) 2 (0.253) 8 (0.467) 
Liver dysfunction group N (%) 
 Group A: Normal 396 (86.8) 701 (88.6) 354 (76.0) 1,451 (84.7) 
 Group B: Mild 53 (11.6) 84 (10.6) 108 (23.2) 245 (14.3) 
 Group C: Moderate 5 (1.1) 3 (0.4) 4 (0.9) 12 (0.7) 
 Group D: Severe 1 (0.2) 1 (0.1) 
 Missing 1 (0.2) 3 (0.4) 4 (0.2) 
Baseline LAG-3 expression N (covariance; 1% cutoff) 
 Negative 100 (21.9) 246 (31.1) 346 (20.2) 
 Positive 281 (61.6) 379 (47.9) 660 (38.5) 
 Unknown 75 (16.4) 166 (21.0) 466 (100.0) 707 (41.3) 
Baseline ECOG performance status N (%) 
 0 322 (70.6) 540 (68.3) 224 (48.1) 1,086 (63.4) 
 1 134 (29.4) 240 (30.3) 227 (48.7) 601 (35.1) 
 2 11 (1.4) 3 (0.6) 14 (0.8) 
 Missing 12 (2.6) 12 (0.7) 
ADA N (%) 
 Negativea 422 (92.5) 726 (91.8) 374 (80.3) 1,522 (88.8) 
 Positiveb 34 (7.5) 65 (8.2) 92 (19.7) 191 (11.2) 
Therapy N (%) 
 Combination therapy 456 (100.0) 791 (100.0) 445 (95.5) 1,692 (98.8) 
 Monotherapy 21 (4.5) 21 (1.2) 

Abbreviations: ADA, antidrug–antibody; ECOG, Eastern Cooperative Oncology Group; eGFR, estimated glomerular filtration rate; LAG-3, lymphocyte activation gene 3.

a

Patients without presence of ADAs at all sampling times.

b

Patients with presence of ADAs at ≥1 sampling time during the study.

Nivolumab popPK

The nivolumab popPK analysis dataset included 17,680 nivolumab concentration values from 2,974 patients from 10 clinical studies. Two studies were conducted in patients receiving nivolumab and relatlimab or nivolumab monotherapy with unresectable or metastatic melanoma (RELATIVITY-047; n = 658) or melanoma and other solid tumors (RELATIVITY-020; n = 1,341). Seven studies [MDX1106–01 (CA209001), MDX1106–03 (CA209003), ONO 4538–01 (CA209005), CA209017, ONO 4538–02 (CA209051), CA209057, and CA209063; n = 915] were conducted in patients receiving nivolumab monotherapy with solid tumors (MEL, non–small cell lung cancer, renal cell carcinoma, and others). Study ADVL1412 (CA209070; n = 60) was conducted in pediatric patients receiving nivolumab monotherapy with recurrent or refractory tumors (Parts A and B). Pediatric patients <18 years with non-hematological solid tumors were included in the nivolumab analysis dataset to conduct PK extrapolation to support dose selection for nivolumab plus relatlimab FDC in adolescent populations. Patient demographics and baseline covariates in the analysis dataset are provided in Supplementary Table S1. As nivolumab popPK has been extensively characterized in multiple tumor types, the nivolumab popPK model was developed in two steps: base and full models. Base model development consisted of re-estimating parameters of the previously developed final model with the current analysis dataset (12). The previously developed final model was a 2 compartment, zero-order intravenous infusion with time varying elimination according to a sigmoidal Emax function. A proportional residual error model was used, and the random effects include log-normally distributed random effects on clearance (CL), volume of central compartment (VC), and volume of peripheral compartment (VP), normally distributed random effect on Emax, and a correlation between the CL and VC random effects. The full model was developed from the base model by incorporating additional covariates. To support PK extrapolation in adolescent patients, the covariate effect of patient population on CL was assessed as the following: adolescent (ages, 12–17 years), pediatric (ages, <12 years), adult patients with melanoma previously treated with IO (prior-IO MEL) and adult patient with advanced melanoma receiving first-line treatment (1L MEL). Additional pediatric and adolescent effects were also evaluated on VC. The detailed model development was described in the Supplementary Information (nivolumab model development).

Both relatlimab and nivolumab popPK models were specified in terms of fixed- and random-effect parameters estimated by nonlinear regression using the first-order conditional estimation (FOCEI) method in the NONMEM software program (15).

Each study protocol was approved by the institutional review board at each participating study site, and the studies were conducted in accordance with the Declaration of Helsinki and with Good Clinical Practice guidelines as defined by the International Council for Harmonisation. All patients (or their legal representatives) provided written informed consent before enrollment.

Model evaluation and application

For both popPK models, model evaluations were performed by using standard goodness-of-fit plots, and prediction-corrected visual predictive check (pcVPC) to provide a graphical assessment of the agreement between the time course of observed and model-predicted concentrations. The pcVPC was performed with 1,000 simulated datasets, which were obtained by using parameter values from the full model. Model-based simulations were performed to evaluate relatlimab CL under various conditions, estimate effective half-life, compare exposure metrics for relatlimab 160 mg Q4W versus 80 mg Q2W dose regimens, and assess the clinical relevance of covariates of interest in the full popPK model, such as sex, hepatic function, renal function, Eastern Cooperative Oncology Group (ECOG) performance status (PS), patient population, race, LAG-3 expression, drug product (SAV vs. FDC), and shorter infusion time (30 vs. 60 minutes). Immunogenicity was also assessed using model-based simulations. A patient was considered antidrug antibody (ADA) positive if ADA was positive for any visit during the posttreatment period, and only patients who received the nivolumab and relatlimab combination therapy were included in the comparison. The exposure was generated using individual PK parameters obtained from the popPK model using the NONMEM computer program (Version 7.4, ICON Development Solutions), compiled using Intel Fortran. All exploratory data analyses and presentations of data were performed using SAS Version 9.2 and R Version 3.6.1 (16).

Simulations to inform adolescent pediatric dose recommendations

The popPK models were applied to simulate exposures of nivolumab and relatlimab in adolescent patients with advanced melanoma. The estimated adolescent population type effect on nivolumab CL was used to simulate nivolumab exposures. The same estimated adolescent population type effect on nivolumab CL and VC was applied to linear relatlimab CL and VC for the simulation. A virtual pediatric population (800 patients ages ≥12–17 years) was created by random sampling from data from the National Health and Nutrition Examination Survey (NHANES) database (2017–2018). Body weight, sex, age, and race were obtained from the NHANES database. The patients in the simulation dataset were evenly distributed with respect to sex, and the body weight ranged from 29.3 to 155 kg (data not shown). Exposures in adolescent patients were predicted following the same adult dosing regimen at nivolumab and relatlimab 480 and 160 mg Q4W and compared with the corresponding exposures for adult melanoma patients receiving 1L treatment.

Data availability

Qualified researchers may submit a proposal to access de-identified and anonymized datasets for this study to Bristol-Myers Squibb. The option to submit data requests as well as review criteria for data requests is available at https://vivli.org/ourmember/bristol-myers-squibb/. Additional information on Bristol-Myers Squibb’s policy on data sharing may be found at https://www.bms.com/researchers-and-partners/clinical-trialsand-research/disclosure-commitment.html.

Relatlimab PK

Relatlimab PK was well described by a 2-compartment, zero-order intravenous infusion PK model with parallel linear and nonlinear CL. The typical baseline CL of relatlimab was 7.00 mL/h for a reference male patient with advanced melanoma receiving 1L treatment, 60 years of age, and weighing 75 kg, with a baseline estimated glomerular filtration rate (eGFR) of 90 mL/min/1.73 m2, baseline lactate dehydrogenase [LDH; normalized to the upper limit of normal (ULN) of 0.8], normal hepatic function, and an ECOG PS of 0. The linear component of CL was time-varying that decreased with time. The typical maximum decrease in CL with time was 5%, with half-maximal reduction occurring after approximately 133 days.

A parallel nonlinear CL was described by Michaelis–Menten kinetics with a maximum elimination rate of 0.0751 mg/h. The nonlinear (Michaelis–Menten) process reached half-maximal saturation at 1.37 μg/mL [95% confidence interval (CI), 1.08–1.65 μg/mL]. Parameter estimates of the full model are presented in Supplementary Table S2. The fraction of relatlimab nonlinear CL of the total CL (linear and nonlinear) at steady-state decreased with increasing dose for the Q4W regimens (as shown in Fig. 1), although a limited number of patients were available for the assessment of doses ≥240 mg among the Q4W dose regimens. The total CL at steady-state was calculated using the model predicted steady-state concentrations. Nonlinearity in relatlimab PK represented approximately 31% of total relatlimab CL at the recommended dose of relatlimab 160 mg Q4W when given in combination with nivolumab 480 mg Q4W. A final model was developed to represent a parsimonious model, and parameter estimates for the final model are presented in Supplementary Table S3. The PK parameters were comparable between the full and final models.

Figure 1.

Model predicted the percentage of relatlimab nonlinear CL at steady-state by dose with Q4W dose regimens. %CV, percent coefficient of variation; CL, clearance; Q4W, every 4 weeks.

Figure 1.

Model predicted the percentage of relatlimab nonlinear CL at steady-state by dose with Q4W dose regimens. %CV, percent coefficient of variation; CL, clearance; Q4W, every 4 weeks.

Close modal

Dose-normalized time-averaged relatlimab concentration after the first dose (Cavg1) increased approximately dose proportionally at doses ≥160 mg Q4W (Table 2). Following nivolumab and relatlimab Q4W administration, steady-state concentrations of relatlimab were reached by approximately 16 weeks, with approximately a 2-fold systemic accumulation. Approximately 97% of the maximum concentration at steady-state achieved with the proposed dosing regimen was predicted to be eliminated in 68.6 days.

Table 2.

Summary of predicted relatlimab exposures and accumulation ratio by dose.

Dose (mg)NDose-normalized Cavg1 (μg/mL/mg dose)aDose-normalized Cavgss (μg/mL/mg dose)aRACb
Q2W 
 20 21 0.099 (31.4) 0.17 (60.4) 1.72 (30.8) 
 80 736 0.103 (24.7) 0.265 (46.9) 2.58 (29.1) 
 160 0.129 (25.1) 0.376 (39) 2.93 (34.7) 
 240 12 0.11 (24.2) 0.323 (38.2) 2.94 (23.8) 
 800 0.123 (25.9) 0.384 (31.9) 3.11 (11.2) 
Q4W 
 160 734 0.0921 (26.3) 0.172 (46.9) 1.87 (26.5) 
 240 0.0784 (45.3) 0.133 (66.9) 1.69 (25.4) 
 320 0.0915 (21.7) 0.158 (39.1) 1.73 (25.6) 
 480 132 0.0978 (27.2) 0.189 (41.4) 1.93 (21.6) 
 960 17 0.104 (19.1) 0.187 (27.7) 1.8 (14.4) 
 1,440 19 0.113 (24.4) 0.24 (42.6) 2.13 (26.7) 
Dose (mg)NDose-normalized Cavg1 (μg/mL/mg dose)aDose-normalized Cavgss (μg/mL/mg dose)aRACb
Q2W 
 20 21 0.099 (31.4) 0.17 (60.4) 1.72 (30.8) 
 80 736 0.103 (24.7) 0.265 (46.9) 2.58 (29.1) 
 160 0.129 (25.1) 0.376 (39) 2.93 (34.7) 
 240 12 0.11 (24.2) 0.323 (38.2) 2.94 (23.8) 
 800 0.123 (25.9) 0.384 (31.9) 3.11 (11.2) 
Q4W 
 160 734 0.0921 (26.3) 0.172 (46.9) 1.87 (26.5) 
 240 0.0784 (45.3) 0.133 (66.9) 1.69 (25.4) 
 320 0.0915 (21.7) 0.158 (39.1) 1.73 (25.6) 
 480 132 0.0978 (27.2) 0.189 (41.4) 1.93 (21.6) 
 960 17 0.104 (19.1) 0.187 (27.7) 1.8 (14.4) 
 1,440 19 0.113 (24.4) 0.24 (42.6) 2.13 (26.7) 

Abbreviations: Q2W, every 2 weeks; Q4W, every 4 weeks; RAC, accumulation ratio.

a

Dose normalized concentration calculated as mg/mL per mg of dose.

b

RAC, Cavgss/Cavg1.

Nivolumab PK

Nivolumab PK was well-described by a linear, 2-compartment model with zero-order intravenous infusion, first-order elimination, and time-varying CL, the results agreed with the previous analysis (12). Parameter estimates of the full model are presented in Supplementary Table S4.

Covariate evaluation

Graphical representations of the effect of categorical and continuous covariates on the typical value of the structural model parameters for relatlimab are presented in Fig. 2. The estimated covariate effects (and 95% CIs) shown in Fig. 2 are relative to CL, VC, VP, Q, and the ratio of steady-state CL to baseline CL (CLss/CL0) at the reference values of the covariates. Drug product (SAV vs. FDC), ECOG PS, sex, eGFR, baseline albumin, and baseline LDH were statistically significant covariates on relatlimab baseline CL (95% CI did not include the null value); however, the magnitude of the differences was not considered to be clinically relevant (≤25%). Effect of baseline body weight was statistically significant for baseline relatlimab CL and VC (95% CI did not include the null value) with the magnitude of impact on CL and VC approximately 35% relative to median body weight (75 kg) in 1L patients with melanoma. Patient population (participants with prior anti–PD-1–treated IO-naive advanced melanoma) was not a significant covariate on CL, with a slightly lower CL (∼7%) for prior anti–PD-1–treated patients. In addition, PS was a statistically significant covariate on CLss/CL0. The model predicted relatlimab CL decreased by approximately 18% for patients with a PS >0 compared with an approximately 5% decrease in patients with PS = 0 (Supplementary Fig. S1).

Figure 2.

Covariate effects on full relatlimab population pharmacokinetic model. REF patient is a 60-year-old white male weighing 75 kg, with 1L MEL, baseline eGFR of 90 mL/min/1.73 m2, baseline albumin of 4 g/dL, baseline LDH (normalized to upper limit of normal) of 0.8, and a normal PS (PS = 0), who is receiving nivolumab plus relatlimab FDC. CLss/CL0REF is a typical value of change in magnitude of CL in a REF patient. VCREF, QREF, and VPREF are typical values in a REF patient weighing 75 kg. Parameter estimate in a REF patient is considered as 100% (vertical solid line), and dashed vertical lines are at 80% and 125% of this value. 1L MEL, first-line treatment; BBWT, baseline body weight; CL, clearance; eGFR, estimated glomerular filtration rate; FDC, fixed-dose combination; LDH, lactate dehydrogenase; PS, performance status; REF, reference; SAV, single-agent vials; VC, volume of central compartment; VP, volume of peripheral compartment.

Figure 2.

Covariate effects on full relatlimab population pharmacokinetic model. REF patient is a 60-year-old white male weighing 75 kg, with 1L MEL, baseline eGFR of 90 mL/min/1.73 m2, baseline albumin of 4 g/dL, baseline LDH (normalized to upper limit of normal) of 0.8, and a normal PS (PS = 0), who is receiving nivolumab plus relatlimab FDC. CLss/CL0REF is a typical value of change in magnitude of CL in a REF patient. VCREF, QREF, and VPREF are typical values in a REF patient weighing 75 kg. Parameter estimate in a REF patient is considered as 100% (vertical solid line), and dashed vertical lines are at 80% and 125% of this value. 1L MEL, first-line treatment; BBWT, baseline body weight; CL, clearance; eGFR, estimated glomerular filtration rate; FDC, fixed-dose combination; LDH, lactate dehydrogenase; PS, performance status; REF, reference; SAV, single-agent vials; VC, volume of central compartment; VP, volume of peripheral compartment.

Close modal

Similarly, additional covariate effects on CL, Q, VC, VP, and Emax were assessed in the nivolumab full model (Supplementary Fig. S2). As shown in this figure, the baseline nivolumab CL in patients receiving nivolumab monotherapy or in combination with relatlimab SAV was similar to patients receiving nivolumab and relatlimab FDC (≤5% difference in the typical value). The magnitude of the reduction in CL over time was comparable between nivolumab monotherapy and nivolumab plus relatlimab combination therapy. Baseline nivolumab CL was 16% lower in adult patients with melanoma previously treated with IO than in adults receiving 1L treatment. However, the magnitude of the reduction in CL was 20% less in adults with prior-IO–treated melanoma [indicated by an approximately 20% higher CLss/CL0 ratio (statistically significant)]. The adolescent (ages, ≥12–17 years) and pediatric patients (ages, <12 years) showed 36% and 62% lower baseline CL than adults receiving 1L therapy, respectively. In addition, the adolescent (ages, ≥12–17 years) and pediatric patients (ages, <12 years) had 16% and 32% lower VC than adult patients.

Model evaluation

Overall, the observed 5th, 50th (median), and 95th percentiles generally fell within the 90% PI (the shaded band) up to the first 60 days after the previous dose and the first 600 days after the first dose (trough concentrations). These results indicate that the model adequately characterized the data and predicted relatlimab and nivolumab concentrations (Supplementary Figs. S3 and S4).

Model application

Relatlimab exposures at 160 mg Q4W were predicted for nivolumab and relatlimab FDC, SAV coadministration, SAV sequential administration, and relatlimab monotherapy for the popPK analysis population (Fig. 3). The difference of model predicted relatlimab exposures (Cavg1 and Cavgss) at 160 mg Q4W was less than 25% for patients with 1L MEL, prior-IO MEL, and other tumor types regardless of drug product or administration. The observed data were also analyzed using noncompartmental analysis. Results showed that observed relatlimab maximum concentration (Cmax) and area under the concentration–time curve in one dosing interval (AUC(TAU)) were similar between 1:1 randomized SAV coadministration and FDC cohorts as indicated by the geometric mean ratios of approximately 1.0 (Supplementary Table S5). Nivolumab concentration at the end of infusion (Ceoi) and trough concentration (Ctrough) were also similar between SAV coadministration and FDC cohorts.

Figure 3.

Distribution of model predicted relatlimab exposures (Cavg1 and Cavgss) at 160 mg Q4W by drug product (FDC, SAV coadministration, SAV sequential) and patient population. Cavg1 and Cavgss, difference of model predicted relatlimab exposures; FDC, fixed-dose combination; Q4W, every 4 weeks; RELA, relatlimab; SAV, single-agent vials.

Figure 3.

Distribution of model predicted relatlimab exposures (Cavg1 and Cavgss) at 160 mg Q4W by drug product (FDC, SAV coadministration, SAV sequential) and patient population. Cavg1 and Cavgss, difference of model predicted relatlimab exposures; FDC, fixed-dose combination; Q4W, every 4 weeks; RELA, relatlimab; SAV, single-agent vials.

Close modal

Nivolumab and relatlimab exposures at 480 and 160 mg FDC Q4W and PK parameters of CL0, VC, VP, VMAX, CLss/CL0, and RAC were consistent across 1L MEL, prior-IO MEL, and all patients with melanoma, shown in Table 3. Predicted nivolumab and relatlimab exposures (Cavg1 and Cavgss) at 480 and 160 mg Q4W were higher for patients with lower body weight compared with patients with higher body weights (Supplementary Figs. S5 and S6). Both Cavg1 and Cavgss were lower in patients with higher body weight mainly because of the flat dose being administered. Predicted relatlimab and nivolumab CL was also affected by patient body weight and is summarized in Supplementary Table S6. Cavg at steady-state was approximately 26% and 10% higher in patients with lower body weight (at 5th percentile); and was approximately 31% and 16% lower in patients with higher body weight (at 95th percentile) relative to the exposure in typical patients at a body weight of 75 kg for relatlimab and nivolumab, respectively.

Table 3.

Summary of predicted relatlimab exposures at 160 mg/480 mg FDC Q4W in patients with melanoma.

Exposure (μg/mL)1L Melanoma Geo. Mean (%CV)Prior-IO Melanoma Geo. Mean (%CV)Melanoma (1L + Prior-IO) Geo. Mean (%CV)
N = 333N = 80N = 413
Exposure (μg/mL) 
Cmin1 5.7 (49.5) 5.4 (55.7) 5.64 (50.8) 
Cmax1 44.4 (23.3) 43.2 (22) 44.1 (23) 
Cavg1 14.8 (25.3) 14.8 (28.4) 14.8 (25.9) 
Cminss 15.3 (64.3) 13.4 (68.8) 14.9 (65.1) 
Cmaxss 62.2 (30.1) 59.8 (32.6) 61.7 (30.6) 
Cavgss 28.8 (44.8) 27.3 (48.3) 28.5 (45.4) 
CL0 (mL/h) 6.06 (38.9) 5.99 (35.2) 6.05 (38.2) 
CLss (mL/h) 5.48 (41.3) 5.67 (35.6) 5.51 (40.2) 
VC (L) 3.59 (22) 3.68 (20.7) 3.61 (21.7) 
VP (L) 3 (26.9) 2.84 (28.3) 2.97 (27.2) 
VSS (L) 6.65 (19.8) 6.58 (19.9) 6.64 (19.8) 
VMAX (μg/h) 80.3 (32.3) 81.5 (42.9) 80.5 (34.7) 
CLss/CL0 (%) 90.3 (26.2) 94.6 (12.3) 91.2 (24.1) 
RAC 1.94 (26.3) 1.84 (24.2) 1.92 (26) 
Exposure (μg/mL)1L Melanoma Geo. Mean (%CV)Prior-IO Melanoma Geo. Mean (%CV)Melanoma (1L + Prior-IO) Geo. Mean (%CV)
N = 333N = 80N = 413
Exposure (μg/mL) 
Cmin1 5.7 (49.5) 5.4 (55.7) 5.64 (50.8) 
Cmax1 44.4 (23.3) 43.2 (22) 44.1 (23) 
Cavg1 14.8 (25.3) 14.8 (28.4) 14.8 (25.9) 
Cminss 15.3 (64.3) 13.4 (68.8) 14.9 (65.1) 
Cmaxss 62.2 (30.1) 59.8 (32.6) 61.7 (30.6) 
Cavgss 28.8 (44.8) 27.3 (48.3) 28.5 (45.4) 
CL0 (mL/h) 6.06 (38.9) 5.99 (35.2) 6.05 (38.2) 
CLss (mL/h) 5.48 (41.3) 5.67 (35.6) 5.51 (40.2) 
VC (L) 3.59 (22) 3.68 (20.7) 3.61 (21.7) 
VP (L) 3 (26.9) 2.84 (28.3) 2.97 (27.2) 
VSS (L) 6.65 (19.8) 6.58 (19.9) 6.64 (19.8) 
VMAX (μg/h) 80.3 (32.3) 81.5 (42.9) 80.5 (34.7) 
CLss/CL0 (%) 90.3 (26.2) 94.6 (12.3) 91.2 (24.1) 
RAC 1.94 (26.3) 1.84 (24.2) 1.92 (26) 

Note: RAC, Cavgss/Cavg1; Vss, VC + VP.

Abbreviations: %CV, percent coefficient of variation; IO, immuno-oncology; Vmax, maximum rate of elimination via the nonlinear pathway; RAC, accumulation ratio.

Baseline eGFR was a significant covariate on relatlimab CL (Fig. 2); patients with higher baseline eGFR had higher CL (<10% difference). Renal impairment did not appear to affect relatlimab CL based on comparisons of model predicted relatlimab exposure at 160 mg Q4W presented in Supplementary Fig. S7. At 160 mg Q4W, exposures (Cavg1 and Cavgss) among patients with mild and moderate renal impairment were similar compared with patients with normal renal function (with a difference <17%). Exposures of patients with severe renal impairment were slightly higher; however, the sample size was limited (n = 2). The effect of hepatic impairment was evaluated by comparing the model predicted relatlimab exposures at 160 mg Q4W across patients with normal, mild, moderate, and severe hepatic function as shown in Supplementary Fig. S8. There were 12 patients with moderate hepatic impairment and only 1 patient with severe hepatic impairment. At 160 mg Q4W, exposures among patients with mild and moderate hepatic impairment were similar compared with patients with normal hepatic function (with a difference of <15%). The exposure of the patient with severe hepatic impairment was higher; however, the sample size was also limited.

In study CA224047, the incidence of relatlimab and nivolumab treatment emergent ADA and NAb were low (<6%) in patients who received nivolumab plus relatlimab FDC and nivolumab monotherapy (6). Comparisons of model predicted nivolumab and relatlimab exposures between nivolumab or relatlimab ADA+ and ADA patients at 480 or 160 mg Q4W further supports that there is no dose adjustment needed for this special population (Supplementary Figs. S9 and S10). Both Cavg1 and Cavgss were lower in ADA+ patients compared with ADA patients (19% and 24% for nivolumab and approximately 16% and 23% for relatlimab, respectively); however, the magnitude of these differences was not considered to be clinically relevant.

In addition, ADA did not have an impact on safety, particularly hypersensitivity/infusion reactions (Supplementary Table S7). Only 2 ADA+ patients had hypersensitivity/infusion reactions, and no consistent trend in the difference of adverse event occurrence between ADA-positive and ADA negative patients was observed.

Relatlimab exposures from 30- to 60-minute infusion durations at 160 mg Q4W were predicted using modeling. The predicted exposures were the same between the two infusion durations evaluated (Supplementary Table S8).

Simulation for pediatric dose recommendation

Comparisons of the predicted relatlimab and nivolumab cycle 1 exposures in adolescent patients weighing at least 30 kg (stratified by weight in 10 kg increments from 30–100 kg, and >100 kg) with that of adult exposures with nivolumab 480 mg and relatlimab 160 mg Q4W are provided in Supplementary Tables S9 and S10, respectively. The predicted median Cavg1 for adolescent patients are generally within the range of the 5th and 95th percentile of adult exposures.

The popPK analyses were performed to comprehensively characterize the CPP of relatlimab in solid tumors and to justify the approved dosing of nivolumab and relatlimab FDC in adult and adolescent patients with advanced melanoma. The nivolumab PK has been previously characterized to support the approval of nivolumab in multiple indications and tumor types (12). This report is focused on the characterization of relatlimab PK, and assessment of effect of intrinsic and extrinsic factors, special populations, formulation, infusion duration, and combination with nivolumab on relatlimab PK.

Relatlimab CL was described by parallel nonlinear and time-varying CL processes in the popPK model. The observed nonlinearity in relatlimab CL could be attributed to target-mediated disposition at a lower dose (e.g., 20 mg and 80 mg Q2W). The CL of relatlimab 160 mg Q4W was not affected when given in combination with nivolumab, and relatlimab exposure increased dose proportionally at doses ≥160 mg Q4W. With the currently approved dosing regimen of nivolumab 480 mg and relatlimab 160 mg FDC Q4W, steady-state concentrations of relatlimab would occur at approximately 16 weeks with the majority of the maximum relatlimab concentration eliminated in 68.6 days.

There was no clinically relevant PK interaction observed between nivolumab plus relatlimab when administered in combination regardless of drug formulation (SAV or FDC) or frequency or administration (Q2W or Q4W). Relatlimab baseline CL in patients receiving relatlimab monotherapy and relatlimab SAV nivolumab (sequential or coadministered) was similar to that of patients receiving nivolumab plus relatlimab FDC (≤5% and ≤18% difference, respectively). Similarly, nivolumab baseline CL in patients receiving nivolumab monotherapy or relatlimab SAV nivolumab was similar (≤5% difference) to that of patients receiving nivolumab plus relatlimab FDC. The magnitude of the reductions in nivolumab CL over time were also similar between nivolumab plus relatlimab and nivolumab monotherapy, with a maximal reduction of 21.1% in adult 1L patients with melanoma. Thus, no PK interaction between nivolumab and relatlimab was observed when these agents were given in combination.

Baseline body weight was found to have a potential clinically relevant (>25%) effect on relatlimab CL. Relative to 75 kg (selected reference body weight value), the effect of body weight at the 95th percentile on CL was 34% (95% CI, 27–40). In addition, at steady-state, patients with lower body weight (at the 5th percentile) exhibited exposures that were <30% higher, whereas patients with higher body weight (at the 95th percentile) showed exposures that were <45% lower than the exposure observed in typical patients with a body weight of 75 kg. For nivolumab, the adult patients with 1L melanoma with higher body weight (at 95th percentile) also had lower overall nivolumab exposures compared with patients with lower body weight (at 5th percentile). These differences can be explained by a single flat dose administration of both nivolumab and relatlimab to all patients. The potential clinical relevance of the exposure differences across different body weight quartiles was evaluated during the exposure–response analysis and will be reported separately. Despite the statistically significant difference in CL, no dose reduction or modification is recommended for the approved patient population.

In addition, there were no clinically relevant effects (defined as a <25% effect) of the other intrinsic or extrinsic factors evaluated on relatlimab or nivolumab PK; these included age, sex, race, albumin (relatlimab only), LDH, ECOG status, renal function (measured eGFR), and hepatic function. On the basis of the evaluation of these factors, no dose adjustment is required in special populations, including the elderly, adolescents, or patients with renal or hepatic impairment. Similarly, there was no clinically relevant difference in the predicted exposures between either relatlimab or nivolumab ADA+ and ADA patients treated with nivolumab and relatlimab. Finally, there was no consistent trend in the difference of hypersensitivity/infusion reaction occurrence between ADA+ and ADA patients.

In RELATIVITY-047, nivolumab and relatlimab FDC was administered with a 60-minute infusion time. This PK analysis evaluated nivolumab and relatlimab therapy at both 30- and 60-minute infusion rates and demonstrated comparable exposures. Furthermore, no clinically meaningful difference in the frequency of hypersensitivity/infusion-related reactions was observed at the studied nivolumab plus relatlimab combination regimen up to 480 mg/1,440 mg with a 60-minute infusion time. On the basis of modeling, the shorter duration of the 30-minute infusion is not expected to have clinically meaningful differences in hypersensitivity and PK compared with the 60-minute infusion, while providing convenience for patients, caregivers, and healthcare providers.

A PK-based extrapolation approach was used to determine a recommended dosing regimen of nivolumab plus relatlimab FDC for the treatment of adolescents (ages, ≥12–17 years) with unresectable or metastatic melanoma. Melanoma in the pediatric and adolescent populations is rare. Although there is a demand for new therapeutic approaches for this disease, the low incidence in the pediatric population creates a challenge for enrollment in clinical studies. Primary melanoma tumor characteristics, such as histology, clinical presentation, and risk factors, are considered to be comparable between adolescent and adult patients with melanoma (17). In addition, current treatment strategies for pediatric and adolescent melanoma are based on clinical guidelines for adult patients, and data suggest that the therapeutic safety profiles and efficacy (such as tumor shrinkage or PD effects of immunotherapy) in pediatric patients are comparable with that of adult patients (18).

The assessment of similarity of the disease and outcome of melanoma in adult and adolescent patients supported the extrapolation approach (1921). The nivolumab popPK model was developed, including pediatric PK data from multiple tumor types, and pediatric effects were identified both on the baseline CL and VC in adolescents relative to adults. It is to be noted that these effects were identified after accounting for the differences in body size. In the absence of clinical data in adolescent patients for relatlimab, the PK of relatlimab was characterized by extrapolation from adults, and with the assumption that the linear relatlimab CL and VC would follow the same pediatric effect on nivolumab CL and VC, respectively. This assumption was considered reasonable because of the expected similarity in the PK of IgG4 mAbs, and similar effects of covariates on nivolumab and relatlimab PK parameters (particularly CL and VC) were observed. The recommended dose of nivolumab and relatlimab FDC was determined by using popPK models to identify nivolumab and relatlimab doses that are predicted to produce exposures in adolescent patients with melanoma that are similar to that of the corresponding exposures produced by nivolumab and relatlimab 480 mg/160 mg FDC in adults. The exposures of nivolumab and relatlimab in pediatric patients ages ≥12 years who weigh at least 30 kg are expected to be in the range of exposures in adult patients at the recommended dosage.

In summary, the pharmacology and clinical PK characterization of relatlimab as monotherapy or in combination with nivolumab supports the FDA-approved dosing of nivolumab 480 mg and relatlimab 160 mg FDC every 4 weeks administered over 30 minutes for adult and adolescent patients with advanced melanoma, and the ongoing evaluation of this combination in other solid tumors.

Y. Zhao reports other support from Bristol-Myers Squibb outside the submitted work. Z. Hu reports employment and stock ownership at Bristol-Myers Squibb. S.P. Bathena reports other support from Bristol-Myers Squibb outside the submitted work. S. Keidel reports other support from BMS during the conduct of the study as well as other support from BMS outside the submitted work; in addition, S. Keidel reports employment at Bristol-Myers Squibb. K. Miller-Moslin reports personal fees from Bristol-Myers Squibb during the conduct of the study as well as personal fees from Bristol-Myers Squibb outside the submitted work. P. Statkevich reports other support from Bristol-Myers Squibb outside the submitted work. A. Bello reports other support from Bristol-Myers Squibb outside the submitted work. A. Roy reports personal fees and other support from Bristol-Myers Squibb outside the submitted work; in addition, Bristol-Myers Squibb is a client of A. Roy’s current employer. S. Suryawanshi reports other support from Bristol-Myers Squibb during the conduct of the study as well as other support from Bristol-Myers Squibb outside the submitted work. No other disclosures were reported.

Y. Zhao: Conceptualization, formal analysis, supervision, writing–review and editing. Z. Hu: Formal analysis, writing–review and editing. S.P. Bathena: Conceptualization, data curation, writing–review and editing. S. Keidel: Investigation, writing–review and editing. K. Miller-Moslin: Investigation, writing–review and editing. P. Statkevich: Investigation, writing–review and editing. A. Bello: Conceptualization, investigation, writing–review and editing. A. Roy: Conceptualization, investigation, writing–review and editing, interpretation. S. Suryawanshi: Conceptualization, formal analysis, supervision, writing–review and editing.

The authors thank the patients and investigators who participated in all clinical trials included in this analysis. They acknowledge Ono Pharmaceutical Company, Ltd. (Osaka, Japan) for contributions to nivolumab development and Dako, an Agilent Technologies, Inc. company (Santa Clara, CA) for collaborative development of the PD-L1 IHC 28–8 pharmDx assay. They thank Baylea Boyle, Prema Sukumar, Weidong Chen, and Erin Dombrowsky for their support of analysis dataset preparation. Professional medical writing and editorial assistance were provided by Jessica R. Augello and Michele Salernitano at Ashfield MedComms, an Inizio Company, funded by Bristol-Myers Squibb.

Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/).

1.
Long
L
,
Zhang
X
,
Chen
F
,
Pan
Q
,
Phiphatwatchara
P
,
Zeng
Y
, et al
.
The promising immune checkpoint LAG-3: from tumor microenvironment to cancer immunotherapy
.
Genes Cancer
2018
;
9
:
176
89
.
2.
Huang
R-Y
,
Francois
A
,
McGray
AJ
,
Miliotto
A
,
Odunsi
K
.
Compensatory upregulation of PD-1, LAG-3, and CTLA-4 limits the efficacy of single-agent checkpoint blockade in metastatic ovarian cancer
.
Oncoimmunology
2017
;
6
:
e1249561
.
3.
Zaremba
A
,
Eggermont
AMM
,
Robert
C
,
Dummer
R
,
Ugurel
S
,
Livingstone
E
, et al
.
The concepts of rechallenge and retreatment with immune checkpoint blockade in melanoma patients
.
Eur J Cancer
2021
;
155
:
268
80
.
4.
Turnis
ME
,
Andrews
LP
,
Vignali
DA
.
Inhibitory receptors as targets for cancer immunotherapy
.
Eur J Immunol
2015
;
45
:
1892
905
.
5.
Woo
SR
,
Turnis
ME
,
Goldberg
V
,
Bankoti
J
,
Selby
M
,
Nirschl
CJ
, et al
.
Immune inhibitory molecules LAG-3 and PD-1 synergistically regulate T-cell function to promote tumoral immune escape
.
Cancer Res
2012
;
72
:
917
27
.
6.
Opdualag [package insert]
.
Princeton, NJ
:
Bristol-Myers Squibb Company
;
2022
.
7.
Opdivo [package insert]
.
Princeton, NJ
:
Bristol-Myers Squibb Company
;
2014
.
8.
Tawbi
HA
,
Schadendorf
D
,
Lipson
EJ
,
Ascierto
PA
,
Matamala
L
,
Castillo Guti'errez
E
, et al
.
Relatlimab and nivolumab versus nivolumab in untreated advanced melanoma
.
N Engl J Med
2022
;
386
:
24
34
.
9.
Long
G
,
Hodi
FS
,
Lipson
EJ
,
Schadendorf
D
,
Ascierto
PA
,
Matamata
L
, et al
.
Overall survival and response with nivolumab and relatlimab in advanced melanoma
.
NEJM Evidence
2023
;
2
.
10.
ClinicalTrials.gov
.
An investigational immuno-therapy study to assess the safety, tolerability and effectiveness of anti–LAG-3 with and without anti–PD-1 in the treatment of solid tumors. ClinicalTrials.gov Identifier: NCT01968109
.
Updated Marched 17, 2023. Available from:
https://clinicaltrials.gov/ct2/ show/NCT01968109.
Accessed March 23, 2023
.
11.
Ascierto
PA
,
Lipson
EJ
,
Dummer
R
,
Larkin
J
,
Long
GV
,
Sanborn
RA
, et al
.
Nivolumab and relatlimab in patients with advanced melanoma that had progressed on anti-programmed death-1/programmed death ligand 1 therapy: results from the phase I/IIa RELATIVITY-020 Trial
.
J Clin Oncol
2023
;
41
:
2724
35
.
12.
Bajaj
G
,
Wang
X
,
Agrawal
S
,
Gupta
M
,
Roy
A
,
Feng
Y
.
Model-based population pharmacokinetic analysis of nivolumab in patients with solid tumors
.
CPT Pharmacometrics Syst Pharmacol
2017
;
6
:
58
66
.
13.
Bajaj
G
,
Suryawanshi
S
,
Roy
A
,
Gupta
M
.
Evaluation of covariate effects on pharmacokinetics of monoclonal antibodies in oncology
.
Br J Clin Pharmacol
2019
;
85
:
2045
58
.
14.
Opdualag SmPC
.
European Medicines Agency (EMA)
.
Bristol-Myers Squibb
:
Dublin, Ireland
.
Available from
: https://www.ema.europa.eu/en/documents/product-information/opdualag-epar-product-information_en.pdf.
15.
Beal
SL
,
Sheiner
LB
,
Boeckmann
AJ
, et al
. eds.
NONMEM 7.3.0 Users Guides
.
Hanover, MD
:
ICON Development Solutions
;
1989–2013
.
16.
R Development Core Team
.
R: a language and environment for statistical computing
.
R Foundation for Statistical Computing, Vienna, Austria
;
2020
.
17.
Strouse
JJ
,
Fears
TR
,
Tucker
MA
,
Wayne
AS
.
Pediatric melanoma: risk factor and survival analysis of the surveillance, epidemiology, and end results database
.
J Clin Oncol
2005
;
23
:
4735
41
.
18.
Bagnoni
G
,
Fidanzi
C
,
D’Erme
AM
,
Viacava
P
,
Leoni
M
,
Strambi
S
, et al
.
Melanoma in children, adolescents and young adults: anatomo-clinical features and prognostic study on 426 cases
.
Pediatr Surg Int
2019
;
35
:
159
65
.
19.
Food and Drug Administration (FDA) website
.
Guidance for industry: exposure–response relationships—study design, data analysis, and regulatory applications
.
Center for Drug Evaluation and Research
,
May 2003
.
Available from:
https://www.fda.gov/regulatory-information/search-fda-guidance-documents/exposure-response-relationships-study-design-data-analysis-and-regulatory-applications.
Accessed September 10, 2023
.
20.
European Medicines Agency (EMA) website
.
Extrapolation of efficacy and safety in paediatric medicine development—scientific guideline
.
Available from:
https://www.ema.europa.eu/en/extrapolation-efficacy-safety-paediatric-medicine-development-scientific-guideline#current-version-section.
Accessed September 10, 2023
.
21.
Food and Drug Administration (FDA) website
.
Change in pediatric extrapolation of efficacy from adults
.
Available from:
https://www.fda.gov/science-research/fda-stem-outreach-education-and-engagement/change-pediatric-extrapolation-efficacy-adults.
Accessed September 10, 2023
.
This open access article is distributed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0) license.

Supplementary data