Ramucirumab is an IgG1 monoclonal antibody specific for the vascular endothelial growth factor receptor-2. Ramucirumab, 8 mg/kg every 2 weeks, administered as monotherapy (REGARD) or in combination with paclitaxel (RAINBOW), was safe and effective in patients with previously treated advanced gastric or gastroesophageal junction (GEJ) cancer. We evaluated exposure–efficacy and exposure–safety relationships of ramucirumab from two randomized, placebo-controlled phase III trials. Sparse pharmacokinetic samples were collected, and a population pharmacokinetic analysis was conducted to predict ramucirumab minimum trough concentration at steady state (Cmin,ss). Kaplan–Meier methods and Cox proportional hazards models were used to evaluate the ramucirumab exposure (Cmin,ss)–efficacy relationship to overall survival (OS) and progression-free survival (PFS). Logistic regression analyses were used to evaluate exposure–safety relationships. Analyses included 321 ramucirumab + paclitaxel and 335 placebo + paclitaxel patients from RAINBOW and 72 ramucirumab and 35 placebo patients from REGARD. Exposure–efficacy analysis showed ramucirumab Cmin,ss was a significant predictor of OS and PFS in both trials. Higher ramucirumab exposure was associated with longer OS and PFS. In RAINBOW, grade ≥3 hypertension, leukopenia, and neutropenia, but not febrile neutropenia, significantly correlated with Cmin,ss, with increased exposure leading to increased incidence. Exploratory exposure–response analyses suggest a positive relationship between efficacy and ramucirumab exposure with manageable toxicities at exposures generated from a dose of 8 mg/kg ramucirumab given every 2 weeks for patients with advanced gastric/GEJ cancer. These findings suggest an opportunity to further optimize benefit versus risk profiles of ramucirumab treatment in patients with gastric/GEJ cancer. Mol Cancer Ther; 16(10); 2215–22. ©2017 AACR.

Ramucirumab is a recombinant human IgG1-neutralizing monoclonal antibody (mAb) with a 50 pmol/L Kd specific for the ectodomain of the vascular endothelial growth factor receptor-2 (VEGFR-2; ref. 1). The maximum tolerated dose of single-agent ramucirumab determined in a phase I trial was 13 mg/kg/week (1). Safety and efficacy of ramucirumab in patients with previously treated advanced gastric or gastroesophageal junction (GEJ) cancer were evaluated in two randomized, double-blind phase III trials (REGARD and RAINBOW; refs. 2, 3). In both the REGARD (ramucirumab 8 mg/kg versus placebo every 2 weeks) and RAINBOW (ramucirumab 8 mg/kg or placebo every 2 weeks in combination with weekly paclitaxel 80 mg/m2) studies, overall survival (OS) and progression-free survival (PFS) were significantly improved in the ramucirumab arm. The most common grade ≥3 adverse events (AE) in patients treated with ramucirumab included neutropenia, leukopenia, hypertension, fatigue, abdominal pain, and anemia (2, 3). Based on the outcomes of these trials, ramucirumab has been approved for use as monotherapy or in combination with paclitaxel for the treatment of patients with previously treated advanced gastric/GEJ cancer (4).

Exposure–response analyses are the foundation for determining safety and efficacy of mAbs used for cancer treatments. Assessing the pharmacokinetics (PK) of mAbs is challenging. Often the maximum tolerated dose is not determined in phase I trials, and measuring target occupancy in cancer tissues is not easily achieved; indeed many tumors show inhomogeneous antibody penetration (5, 6). Several studies on mAbs used to treat various types of cancer have demonstrated that the relationship of efficacy and AEs with a defined drug exposure is critical to determining if a treatment is considered safe and effective and if changes to dose or schedule could possibly improve the benefit versus risk profile (7–11).

Although the REGARD and RAINBOW studies showed improved OS and PFS for ramucirumab 8 mg/kg alone or in combination with paclitaxel versus placebo, exploratory exposure–response analyses were conducted to determine if increased exposure to ramucirumab could maximize the potential therapeutic benefit in patients with advanced gastric/GEJ cancer. We conducted separate exposure–response analyses of efficacy and safety from patients on ramucirumab in the REGARD and RAINBOW trials. The objectives of these analyses were to evaluate the relationships between predicted ramucirumab exposure and survival, as well as commonly reported AEs in previously treated patients with advanced gastric/GEJ cancer.

The details of the REGARD and RAINBOW trials, including informed consent, trial design, and clinical definitions of OS, PFS, and objective response rate (ORR), were previously described (2, 3). Each center's institutional review board or independent ethics committee approved the study. The trial followed the principles of the Declaration of Helsinki and the Good Clinical Practice Guidelines of the International Conference on Harmonization. All patients provided written informed consent.

Briefly, patients in the RAINBOW trial with advanced gastric/GEJ cancer who had progressed on or within 4 months of first-line chemotherapy (platinum plus fluoropyrimidine with or without an anthracycline) were randomized (1:1) to receive ramucirumab 8 mg/kg or placebo on days 1 and 15, plus paclitaxel 80 mg/m2 on days 1, 8, and 15 of a 28-day cycle. In the REGARD trial, patients with advanced gastric/GEJ cancer who had progressed on or within 4 months of the last dose of first-line platinum- or fluoropyrimidine-containing chemotherapy were randomized (2:1) to receive ramucirumab (8 mg/kg) plus best supportive care or placebo plus best supportive care every 2 weeks. The primary endpoint for both trials was OS, defined as the time from randomization to death from any cause. Key secondary endpoints included PFS (defined as time from randomization to progressive disease or death, whichever occurred first), ramucirumab PK parameters, and safety. Samples for PK analysis from patients in the RAINBOW trial were collected prior to and 1 hour after infusions 1, 4, and 7 and at the end of the trial; samples from patients in REGARD were collected prior to cycles 1, 4, and 7 infusions and at the 30-day follow-up.

Population PK (popPK) analyses were conducted as previously described, using a nonlinear mixed-effect modeling approach (12). Population PK-predicted minimum concentrations at steady-state [Cmin,ss] were then determined for ramucirumab-treated patients using the developed model (Supplementary Fig. S1). Analyses were conducted in accordance with the FDA Guidance for Industry on Population Pharmacokinetics (13).

Exposure–efficacy analyses

Univariate and multivariable Cox regression analyses were utilized to determine whether there was a significant relationship between Cmin,ss and survival for each of the phase III trials. Only patients with nonmissing ramucirumab concentration data were included in each of the exposure populations. For the multivariable analyses, hazard ratios (HR) were adjusted for significant baseline covariates. A stepwise Cox regression model was used to select baseline covariates at an entry P value of 0.05 and exit P value of 0.1. The list of factors evaluated for potential prognostic significance included geographical region, disease measurability, time to progression on first-line therapy, gender, age, race, Eastern Cooperative Oncology Group performance status (ECOG PS), weight loss in the 3 months prior to randomization, primary tumor location, prior first-line chemotherapy, histologic subtype, number of metastatic sites, peritoneal metastasis, liver metastasis, presence of ascites, tumor differentiation, number of previous treatment lines including neoadjuvant and adjuvant, and prior gastrectomy.

To evaluate the exposure–response relationship compared with the control groups for both studies, patients in the exposure populations were stratified into exposure groups by defined Cmin,ss quartiles (Q) (RAINBOW) or by the Cmin,ss median (REGARD). The quartiles for the patients from the RAINBOW study were defined as: Q1 = Cmin,ss 20.1 to ≤42.0 μg/mL (<25%); Q2 = Cmin,ss >42.0 to ≤53.8 μg/mL (25 to <50%); Q3 = Cmin,ss >53.8 to ≤72.2 μg/mL (50 to <75%); and Q4 = Cmin,ss >72.2 to 227 μg/mL (≥75%). The REGARD exposure population was divided by the median Cmin,ss of 65.35 μg/mL. The Kaplan–Meier method was used to evaluate OS and PFS for each of the individual exposure groups versus the control groups. The HR for each exposure group versus the control arm was estimated using a Cox proportional hazards model. All HRs for the multivariable analyses were adjusted for the significant baseline covariates.

For REGARD, the first post-baseline PK sample was collected before the cycle 4 infusion; thus, only the subset of placebo patients who had at least 3 cycles of treatment were included in the analyses to have an unbiased comparison.

All statistical analyses were undertaken using SAS version 9.2 (SAS Institute; ref. 2).

Exposure–safety analyses

The three most common grade ≥3 consolidated treatment-emergent AEs (TEAE) occurring in ≥10% of patients in the RAINBOW intent-to-treat (ITT) population and at a higher rate in the ramucirumab plus paclitaxel arm were selected as the safety endpoints for all exposure–safety analyses. These TEAEs were neutropenia (including febrile neutropenia), leukopenia, and hypertension and were graded per the National Cancer Institute Common Terminology Criteria for Adverse Events version 4.02 (14). A Cmin,ss value was calculated for each patient receiving ramucirumab using the popPK model, while patients in the control groups were assigned a value of zero. The relationship between TEAE incidence and ramucirumab Cmin,ss was evaluated using logistic regression.

Exposure–response populations

RAINBOW.

In the phase III RAINBOW trial that established the efficacy and safety of ramucirumab plus paclitaxel versus placebo plus paclitaxel for patients with advanced gastric/GEJ cancer, 330 patients were randomized to receive ramucirumab plus paclitaxel, 327 of whom were treated (3). Among those treated, ramucirumab concentration data were available for 321 patients (98.2%) who were included in the exposure–response population. Thus, the baseline characteristics from the ramucirumab exposure–response population were very similar to those of the RAINBOW ITT population (Table 1).

Table 1.

RAINBOW ITT and exposure–response population baseline demographics and disease characteristics

ITT population, n (%)Exposure–response population, n (%)
RAM+PACPBO+PACRAM+PAC overallRAM+PAC Cmin,ss Q1RAM+PAC Cmin,ss Q2RAM+PAC Cmin,ss Q3RAM+PAC Cmin,ss Q4
CharacteristicsN = 330N = 335N = 321N = 80N = 80N = 80N = 81
Gender, male  229 (69) 243 (73) 220 (69) 52 (65) 57 (71) 58 (73) 53 (65) 
Age <65 years  204 (62) 212 (63) 202 (63) 56 (70) 47 (59) 56 (70) 43 (53) 
Race White 208 (63) 199 (59) 201 (63) 47 (59) 54 (68) 54 (68) 46 (57) 
 Asian 110 (33) 121 (36) 109 (34) 26 (33) 23 (29) 25 (31) 35 (43) 
 Other 12 (4) 15 (4) 11 (3) 7 (9) 3 (4) 1 (1) 0 (0) 
Baseline body weight (kg) Mean (SD) 64 (15) 63 (15) 64 (15) 58 (13) 63 (13) 67 (18) 67 (16) 
 Median (range) 61 (32–134) 62 (30–135) 61 (32–134) 55 (40–96) 60 (32–96) 66 (38–134) 64 (44–131) 
ECOG PS 117 (35) 144 (43) 116 (36) 19 (24) 22 (28) 31 (39) 44 (54) 
 213 (65) 191 (57) 205 (64) 61 (76) 58 (73) 49 (61) 37 (46) 
Histological subtype Diffuse 115 (35) 133 (40) 114 (36) 34 (43) 33 (41) 29 (36) 18 (22) 
 Intestinal 145 (44) 135 (40) 140 (44) 33 (41) 34 (43) 31 (39) 42 (52) 
 Mixed/Missing/Unknown 70 (21) 67 (20) 67 (21) 13 (16) 13 (16) 20 (25) 21 (26) 
Primary tumor location GEJ 66 (20) 71 (21) 62 (19) 11 (14) 12 (15) 22 (28) 17 (21) 
 Gastric 264 (80) 264 (79) 259 (81) 69 (86) 68 (85) 58 (73) 64 (79) 
Primary tumor present  209 (63) 209 (62) 200 (62) 52 (65) 50 (63) 50 (63) 48 (59) 
Number of metastatic sites 0–2 209 (63) 232 (69) 203 (63) 46 (58) 50 (63) 53 (66) 54 (67) 
 ≥3 121 (37) 103 (31) 118 (37) 34 (43) 30 (38) 27 (34) 27 (33) 
Presence of ascites  130 (39) 107 (32) 128 (40) 38 (48) 33 (41) 27 (34) 30 (37) 
Most common metastases Lymph nodes 215 (65) 205 (61) 209 (65) 56 (70) 50 (63) 47 (59) 56 (69) 
 Peritoneal 163 (49) 152 (45) 158 (49) 48 (60) 42 (53) 32 (40) 36 (44) 
 Liver 150 (45) 138 (41) 144 (45) 28 (35) 32 (40) 40 (50) 44 (54) 
 Lung 77 (23) 70 (21) 76 (24) 19 (24) 18 (23) 20 (25) 19 (23) 
Tumor grade Well differentiated 28 (8) 22 (7) 26 (8) 9 (11) 4 (5) 4 (5) 9 (11) 
 Moderately differentiated 96 (29) 106 (32) 93 (29) 22 (28) 24 (30) 21 (26) 26 (32) 
 Poorly differentiated 186 (56) 186 (56) 183 (57) 46 (58) 44 (55) 50 (63) 43 (53) 
 Unknown 20 (6) 21 (6) 19 (6) 3 (4) 8 (10) 5 (6) 3 (4) 
Prior gastrectomy  132 (40) 124 (37) 132 (41) 27 (34) 36 (45) 31 (39) 38 (47) 
Weight loss, prior 3 monthsa <10% 277 (84) 286 (85) 269 (84) 65 (81) 64 (80) 64 (80) 76 (94) 
 ≥10% 53 (16) 47 (14) 52 (16) 15 (19) 16 (20) 16 (20) 5 (6) 
Disease progression During first-line therapy 227 (69) 217 (65) 222 (69) 63 (79) 53 (66) 51 (64) 55 (68) 
 ≤4 months after first-line therapy 94 (28) 108 (32) 90 (28) 17 (21) 24 (30) 25 (31) 24 (30) 
 Missing 9 (3) 10 (3) 9 (3) 0 (0) 3 (4) 4 (5) 2 (2) 
ITT population, n (%)Exposure–response population, n (%)
RAM+PACPBO+PACRAM+PAC overallRAM+PAC Cmin,ss Q1RAM+PAC Cmin,ss Q2RAM+PAC Cmin,ss Q3RAM+PAC Cmin,ss Q4
CharacteristicsN = 330N = 335N = 321N = 80N = 80N = 80N = 81
Gender, male  229 (69) 243 (73) 220 (69) 52 (65) 57 (71) 58 (73) 53 (65) 
Age <65 years  204 (62) 212 (63) 202 (63) 56 (70) 47 (59) 56 (70) 43 (53) 
Race White 208 (63) 199 (59) 201 (63) 47 (59) 54 (68) 54 (68) 46 (57) 
 Asian 110 (33) 121 (36) 109 (34) 26 (33) 23 (29) 25 (31) 35 (43) 
 Other 12 (4) 15 (4) 11 (3) 7 (9) 3 (4) 1 (1) 0 (0) 
Baseline body weight (kg) Mean (SD) 64 (15) 63 (15) 64 (15) 58 (13) 63 (13) 67 (18) 67 (16) 
 Median (range) 61 (32–134) 62 (30–135) 61 (32–134) 55 (40–96) 60 (32–96) 66 (38–134) 64 (44–131) 
ECOG PS 117 (35) 144 (43) 116 (36) 19 (24) 22 (28) 31 (39) 44 (54) 
 213 (65) 191 (57) 205 (64) 61 (76) 58 (73) 49 (61) 37 (46) 
Histological subtype Diffuse 115 (35) 133 (40) 114 (36) 34 (43) 33 (41) 29 (36) 18 (22) 
 Intestinal 145 (44) 135 (40) 140 (44) 33 (41) 34 (43) 31 (39) 42 (52) 
 Mixed/Missing/Unknown 70 (21) 67 (20) 67 (21) 13 (16) 13 (16) 20 (25) 21 (26) 
Primary tumor location GEJ 66 (20) 71 (21) 62 (19) 11 (14) 12 (15) 22 (28) 17 (21) 
 Gastric 264 (80) 264 (79) 259 (81) 69 (86) 68 (85) 58 (73) 64 (79) 
Primary tumor present  209 (63) 209 (62) 200 (62) 52 (65) 50 (63) 50 (63) 48 (59) 
Number of metastatic sites 0–2 209 (63) 232 (69) 203 (63) 46 (58) 50 (63) 53 (66) 54 (67) 
 ≥3 121 (37) 103 (31) 118 (37) 34 (43) 30 (38) 27 (34) 27 (33) 
Presence of ascites  130 (39) 107 (32) 128 (40) 38 (48) 33 (41) 27 (34) 30 (37) 
Most common metastases Lymph nodes 215 (65) 205 (61) 209 (65) 56 (70) 50 (63) 47 (59) 56 (69) 
 Peritoneal 163 (49) 152 (45) 158 (49) 48 (60) 42 (53) 32 (40) 36 (44) 
 Liver 150 (45) 138 (41) 144 (45) 28 (35) 32 (40) 40 (50) 44 (54) 
 Lung 77 (23) 70 (21) 76 (24) 19 (24) 18 (23) 20 (25) 19 (23) 
Tumor grade Well differentiated 28 (8) 22 (7) 26 (8) 9 (11) 4 (5) 4 (5) 9 (11) 
 Moderately differentiated 96 (29) 106 (32) 93 (29) 22 (28) 24 (30) 21 (26) 26 (32) 
 Poorly differentiated 186 (56) 186 (56) 183 (57) 46 (58) 44 (55) 50 (63) 43 (53) 
 Unknown 20 (6) 21 (6) 19 (6) 3 (4) 8 (10) 5 (6) 3 (4) 
Prior gastrectomy  132 (40) 124 (37) 132 (41) 27 (34) 36 (45) 31 (39) 38 (47) 
Weight loss, prior 3 monthsa <10% 277 (84) 286 (85) 269 (84) 65 (81) 64 (80) 64 (80) 76 (94) 
 ≥10% 53 (16) 47 (14) 52 (16) 15 (19) 16 (20) 16 (20) 5 (6) 
Disease progression During first-line therapy 227 (69) 217 (65) 222 (69) 63 (79) 53 (66) 51 (64) 55 (68) 
 ≤4 months after first-line therapy 94 (28) 108 (32) 90 (28) 17 (21) 24 (30) 25 (31) 24 (30) 
 Missing 9 (3) 10 (3) 9 (3) 0 (0) 3 (4) 4 (5) 2 (2) 

Abbreviations: N, total number of patients; n, number of patients in group; PAC, paclitaxel; PBO, placebo; Q, quartile; RAM, ramucirumab.

aTwo patients in the placebo plus paclitaxel arm were missing data.

REGARD.

In the phase III REGARD trial that established the efficacy of ramucirumab plus best supportive care versus placebo plus best supportive care for patients with advanced gastric/GEJ cancer, 238 patients were randomized to receive ramucirumab plus best supportive care, 236 of whom were treated (2). The first post-baseline PK sample was collected before the cycle 4 infusion; therefore, only the placebo patients who had at least 3 cycles of treatment (n = 35) were included in exposure–response analyses to facilitate an unbiased comparison. Of the ramucirumab patients who were treated, 72 (30.5%) had ramucirumab concentration data and were included in the exposure–response population. Although the REGARD exposure–response population was notably smaller than the ITT population, the baseline demographics and disease characteristics were mostly similar (Table 2).

Table 2.

REGARD ITT and exposure–response population baseline demographics and disease characteristics

ITT population, n (%)Exposure–response population, n (%)
RAM+BSCPBO+BSCRAM+BSCRAM+BSC, Cmin,ss <MedianRAM+BSC, Cmin,ss ≥MedianPBO+BSC
CharacteristicN = 238N = 117Overall, N = 72N = 36N = 36N = 35
Gender, male  169 (71) 79 (68) 49 (68) 27 (75) 22 (61) 22 (63) 
Age <65 years  156 (66) 71 (61) 43 (60) 25 (69) 18 (50) 22 (63) 
Race White 181 (76) 91 (78) 61 (85) 27 (75) 34 (94) 26 (74) 
 Asian 39 (16) 17 (15) 7 (10) 6 (17) 1 (3) 6 (17) 
 Other 18 (8) 9 (8) 4 (6) 3 (8) 1 (3) 3 (9) 
Baseline body weight (kg) Mean (SD) 65 (15) 66 (17) 67 (16) 63 (13) 72 (17) 62 (15) 
 Median (range) 63 (36–113) 64 (31–118) 66 (45–113) 62 (45–110) 70 (50–113) 59 (31–94) 
ECOG PSa 67 (28) 31 (26) 23 (32) 12 (33) 11 (31) 18 (51) 
 171 (72) 85 (73) 49 (68) 24 (67) 25 (69) 17 (49) 
Histological subtype Diffuse 96 (40) 44 (38) 34 (47) 18 (50) 16 (44) 16 (46) 
 Intestinal 52 (22) 35 (30) 13 (18) 10 (28) 3 (8) 7 (20) 
 Undetermined/Not available/Other 90 (38) 38 (32) 25 (35) 8 (22) 17 (47) 12 (34) 
Primary tumor location GEJ 60 (25) 30 (26) 20 (28) 7 (19) 13 (36) 8 (23) 
 Gastric 178 (75) 87 (74) 52 (72) 29 (81) 23 (64) 27 (77) 
Primary tumor present  174 (73) 86 (74) 50 (69) 27 (75) 23 (64) 23 (66) 
Most common metastases Lymph nodes 154 (65) 74 (63) 43 (60) 24 (67) 19 (53) 19 (54) 
 Peritoneal 64 (27) 45 (38) 13 (18) 8 (22) 5 (14) 13 (37) 
 Liver 104 (44) 56 (48) 29 (40) 13 (36) 16 (44) 19 (54) 
 Lung 56 (24) 28 (24) 19 (26) 9 (25) 10 (28) 10 (29) 
Number of metastatic sites 0–2 163 (68) 71 (61) 54 (75) 27 (75) 27 (75) 22 (63) 
 ≥3 75 (32) 46 (39) 18 (25) 9 (25) 9 (25) 13 (37) 
Tumor grade Well differentiated 18 (8) 3 (3) 4 (6) 2 (6) 2 (6) 
 Moderately differentiated 75 (32) 29 (25) 27 (38) 15 (42) 12 (33) 11 (31) 
 Poorly differentiated 90 (38) 66 (56) 23 (32) 11 (31) 12 (33) 19 (54) 
 Undifferentiated 8 (3) 4 (3) 3 (4) 1 (3) 2 (6) 1 (3) 
 Unknown 47 (20) 15 (13) 15 (21) 7 (19) 8 (22) 4 (11) 
Weight loss, prior 3 months <10% 197 (83) 97 (83) 63 (88) 32 (89) 31 (86) 30 (86) 
 ≥10% 41 (17) 20 (17) 9 (13) 4 (11) 5 (14) 5 (14) 
Disease progression <6 months 154 (65) 83 (71) 47 (65) 29 (81) 18 (50) 24 (69) 
 ≥6 monthsb 81 (34) 34 (29) 25 (35) 7 (19) 18 (50) 11 (31) 
ITT population, n (%)Exposure–response population, n (%)
RAM+BSCPBO+BSCRAM+BSCRAM+BSC, Cmin,ss <MedianRAM+BSC, Cmin,ss ≥MedianPBO+BSC
CharacteristicN = 238N = 117Overall, N = 72N = 36N = 36N = 35
Gender, male  169 (71) 79 (68) 49 (68) 27 (75) 22 (61) 22 (63) 
Age <65 years  156 (66) 71 (61) 43 (60) 25 (69) 18 (50) 22 (63) 
Race White 181 (76) 91 (78) 61 (85) 27 (75) 34 (94) 26 (74) 
 Asian 39 (16) 17 (15) 7 (10) 6 (17) 1 (3) 6 (17) 
 Other 18 (8) 9 (8) 4 (6) 3 (8) 1 (3) 3 (9) 
Baseline body weight (kg) Mean (SD) 65 (15) 66 (17) 67 (16) 63 (13) 72 (17) 62 (15) 
 Median (range) 63 (36–113) 64 (31–118) 66 (45–113) 62 (45–110) 70 (50–113) 59 (31–94) 
ECOG PSa 67 (28) 31 (26) 23 (32) 12 (33) 11 (31) 18 (51) 
 171 (72) 85 (73) 49 (68) 24 (67) 25 (69) 17 (49) 
Histological subtype Diffuse 96 (40) 44 (38) 34 (47) 18 (50) 16 (44) 16 (46) 
 Intestinal 52 (22) 35 (30) 13 (18) 10 (28) 3 (8) 7 (20) 
 Undetermined/Not available/Other 90 (38) 38 (32) 25 (35) 8 (22) 17 (47) 12 (34) 
Primary tumor location GEJ 60 (25) 30 (26) 20 (28) 7 (19) 13 (36) 8 (23) 
 Gastric 178 (75) 87 (74) 52 (72) 29 (81) 23 (64) 27 (77) 
Primary tumor present  174 (73) 86 (74) 50 (69) 27 (75) 23 (64) 23 (66) 
Most common metastases Lymph nodes 154 (65) 74 (63) 43 (60) 24 (67) 19 (53) 19 (54) 
 Peritoneal 64 (27) 45 (38) 13 (18) 8 (22) 5 (14) 13 (37) 
 Liver 104 (44) 56 (48) 29 (40) 13 (36) 16 (44) 19 (54) 
 Lung 56 (24) 28 (24) 19 (26) 9 (25) 10 (28) 10 (29) 
Number of metastatic sites 0–2 163 (68) 71 (61) 54 (75) 27 (75) 27 (75) 22 (63) 
 ≥3 75 (32) 46 (39) 18 (25) 9 (25) 9 (25) 13 (37) 
Tumor grade Well differentiated 18 (8) 3 (3) 4 (6) 2 (6) 2 (6) 
 Moderately differentiated 75 (32) 29 (25) 27 (38) 15 (42) 12 (33) 11 (31) 
 Poorly differentiated 90 (38) 66 (56) 23 (32) 11 (31) 12 (33) 19 (54) 
 Undifferentiated 8 (3) 4 (3) 3 (4) 1 (3) 2 (6) 1 (3) 
 Unknown 47 (20) 15 (13) 15 (21) 7 (19) 8 (22) 4 (11) 
Weight loss, prior 3 months <10% 197 (83) 97 (83) 63 (88) 32 (89) 31 (86) 30 (86) 
 ≥10% 41 (17) 20 (17) 9 (13) 4 (11) 5 (14) 5 (14) 
Disease progression <6 months 154 (65) 83 (71) 47 (65) 29 (81) 18 (50) 24 (69) 
 ≥6 monthsb 81 (34) 34 (29) 25 (35) 7 (19) 18 (50) 11 (31) 

Abbreviations: BSC, best supportive care; N, total number of patients; n, number in group; RAM, ramucirumab.

aOne patient in the placebo plus best supportive care arm had an ECOG PS of 2.

bThree patients missing in the ramucirumab plus best supportive care arm.

Exposure–efficacy analyses

RAINBOW.

An exploratory univariate analysis evaluating ramucirumab exposure as a continuous variable identified a significant association between ramucirumab Cmin,ss and OS (P < 0.0001) as well as PFS (P = 0.0003). Stepwise Cox regression identified prognostic factors that were significant independent predictors for improved OS, including geographical region 3 (Japan, South Korea, Hong Kong, Singapore, and Taiwan), ECOG PS 0, weight loss less than 10%, up to 2 metastatic sites, absence of ascites, well or moderately differentiated tumor, and prior gastrectomy. After accounting for these factors in a multivariable analysis, the association between OS and Cmin,ss remained statistically significant (P = 0.0008). Improved PFS was significantly associated with gender (male), weight loss <10%, up to 2 metastatic sites, and absence of liver metastasis. After adjusting for these factors, the association between PFS and Cmin,ss also remained statistically significant (P = 0.0004).

Ramucirumab exposure was also evaluated as a categorical variable in order to facilitate comparisons with the control group. Kaplan–Meier plots of OS and PFS stratified by ramucirumab Cmin,ss quartiles demonstrated clear separation (Fig. 1A and B). Median OS for each quartile and the placebo group was 6.6 months (Q1), 9.6 months (Q2), 11.1 months (Q3), 12.6 months (Q4), and 7.4 months (placebo). Median PFS for each quartile and the placebo group was 2.9 months (Q1), 4.3 months (Q2), 5.2 months (Q3), 5.7 months (Q4), and 2.9 months (placebo; Table 3). When each Cmin,ss quartile was compared with the placebo plus paclitaxel arm, the HRs for OS and PFS decreased with increasing ramucirumab exposure. This effect was consistent after adjusting HRs for significant prognostic factors (Table 3).

Figure 1.

Kaplan–Meier analysis of RAINBOW (A and B) and REGARD (C and D). Efficacy outcomes in exposure–response populations by RAM Cmin,ss Group. RAINBOW overall survival (OS) and progression-free survival (PFS) (A and B) are stratified by Cmin,ss quartile and each compared with paclitaxel plus placebo patients. REGARD OS and PFS (C and D) are dichotomized by Cmin,ss median and each compared with patients in the placebo arm who had at least 3 cycles of treatment. Abbreviations: PAC, paclitaxel; PBO, placebo; Q, quartile; RAM, ramucirumab. Quartiles: Q1 = <25%; Q2 = 25%–<50%; Q3 = 50%–<75%; Q4 = ≥75%; RAM, ramucirumab; RAM Cmin,ss low, ramucirumab patients with median Cmin,ss <65.35 μg/mL; RAM Cmin,ss high, ramucirumab patients with median Cmin,ss ≥65.35 μg/mL.

Figure 1.

Kaplan–Meier analysis of RAINBOW (A and B) and REGARD (C and D). Efficacy outcomes in exposure–response populations by RAM Cmin,ss Group. RAINBOW overall survival (OS) and progression-free survival (PFS) (A and B) are stratified by Cmin,ss quartile and each compared with paclitaxel plus placebo patients. REGARD OS and PFS (C and D) are dichotomized by Cmin,ss median and each compared with patients in the placebo arm who had at least 3 cycles of treatment. Abbreviations: PAC, paclitaxel; PBO, placebo; Q, quartile; RAM, ramucirumab. Quartiles: Q1 = <25%; Q2 = 25%–<50%; Q3 = 50%–<75%; Q4 = ≥75%; RAM, ramucirumab; RAM Cmin,ss low, ramucirumab patients with median Cmin,ss <65.35 μg/mL; RAM Cmin,ss high, ramucirumab patients with median Cmin,ss ≥65.35 μg/mL.

Close modal
Table 3.

RAINBOW exposure–efficacy outcomes by Cmin,ss quartile

Efficacy parametern/#eventsMedian (months) (95% CI)Unadjusted HR (95% CI)Adjusted HRa,b (95% CI)
OS 
 PAC+PBO 335/260 7.4 (6.3–8.4)   
 RAM+PAC Q1 80/71 6.6 (5.3–7.5) 1.32 (1.02–1.72) 1.08 (0.82–1.41) 
 RAM+PAC Q2 80/62 9.6 (6.9–11.4) 0.87 (0.66–1.15) 0.78 (0.58–1.03) 
 RAM+PAC Q3 80/59 11.1 (8.2–12.6) 0.74 (0.55–0.98) 0.65 (0.49–0.87) 
 RAM+PAC Q4 81/58 12.6 (11.0–15.5) 0.57 (0.43–0.75) 0.59 (0.44–0.78) 
PFS 
 PAC+PBO 335/296 2.9 (2.8–3.0)   
 RAM+PAC Q1 80/71 2.9 (2.8–4.2) 0.90 (0.70–1.17) 0.87 (0.67–1.14) 
 RAM+PAC Q2 80/70 4.3 (3.0–5.6) 0.72 (0.55–0.93) 0.68 (0.52–0.89) 
 RAM+PAC Q3 80/68 5.2 (4.2–5.7) 0.60 (0.46–0.78) 0.54 (0.41–0.71) 
 RAM+PAC Q4 81/65 5.7 (4.6–7.2) 0.49 (0.37–0.64) 0.43 (0.33–0.57) 
Efficacy parametern/#eventsMedian (months) (95% CI)Unadjusted HR (95% CI)Adjusted HRa,b (95% CI)
OS 
 PAC+PBO 335/260 7.4 (6.3–8.4)   
 RAM+PAC Q1 80/71 6.6 (5.3–7.5) 1.32 (1.02–1.72) 1.08 (0.82–1.41) 
 RAM+PAC Q2 80/62 9.6 (6.9–11.4) 0.87 (0.66–1.15) 0.78 (0.58–1.03) 
 RAM+PAC Q3 80/59 11.1 (8.2–12.6) 0.74 (0.55–0.98) 0.65 (0.49–0.87) 
 RAM+PAC Q4 81/58 12.6 (11.0–15.5) 0.57 (0.43–0.75) 0.59 (0.44–0.78) 
PFS 
 PAC+PBO 335/296 2.9 (2.8–3.0)   
 RAM+PAC Q1 80/71 2.9 (2.8–4.2) 0.90 (0.70–1.17) 0.87 (0.67–1.14) 
 RAM+PAC Q2 80/70 4.3 (3.0–5.6) 0.72 (0.55–0.93) 0.68 (0.52–0.89) 
 RAM+PAC Q3 80/68 5.2 (4.2–5.7) 0.60 (0.46–0.78) 0.54 (0.41–0.71) 
 RAM+PAC Q4 81/65 5.7 (4.6–7.2) 0.49 (0.37–0.64) 0.43 (0.33–0.57) 

Abbreviations: CI, confidence interval; n, number of patients in group; PAC, paclitaxel; PBO, placebo; Q, quartile; RAM, ramucirumab.

Quartiles: Q1= <25%; Q2 = 25%–<50%; Q3 = 50%–<75%; Q4 = ≥75%.

aOS: Hazard ratios were adjusted for Eastern Cooperative Oncology Group performance status, weight loss, number of metastatic sites, presence of ascites, tumor differentiation, prior gastrectomy, and geographic region.

bPFS: Hazard ratios were adjusted for gender, weight loss, number of metastatic sites, and liver metastasis.

REGARD.

In a univariate analysis utilizing ramucirumab exposure as a continuous variable, a significant positive association was observed between ramucirumab Cmin,ss and efficacy for the REGARD exposure–response population (OS: P = 0.0005 and PFS: P = 0.0134). Baseline factors that were significantly associated with reduced OS and PFS were the presence of peritoneal metastasis, ECOG PS of 1 or greater, and location of primary tumor (GEJ), which were used in all subsequent multivariable analyses. The association between survival and Cmin,ss remained significant after adjusting for these factors (OS: P = 0.0002 and PFS: P = 0.0030).

In addition, ramucirumab exposure was reevaluated categorically in order to facilitate comparison with the placebo arm. The Kaplan–Meier plots of OS and PFS dichotomized by median Cmin,ss in Fig. 1C and D demonstrate clear separation between the two exposure groups. Median OS was 6.2 months in the placebo group, 6.6 months in the ramucirumab group with Cmin,ss < median, and 10.7 months in the ramucirumab group with Cmin,ss ≥ median (Table 4). The OS and PFS are higher in the exposure–response population than in the REGARD ITT population because the first PK sample was taken prior to cycle 4 at 8 weeks; thus, the patients with early progression were not included, and the matched placebo patients had the same consideration for the analysis. Median PFS was 2.8 months in the placebo group and in the ramucirumab group with Cmin,ss < median and 5.2 months in the ramucirumab group with Cmin,ss ≥ median (Table 4).

Table 4.

REGARD exposure-efficacy outcomes by Cmin,ss group

Efficacy parametern/#eventsMedian (months)Adjusted HRa (95% CI)
OS 
 PBO 35/27 6.2  
 RAM Cmin,ss low vs. PBO 36/23 6.6 0.87 (0.49–1.58) 
 RAM Cmin,ss high vs. PBO 36/18 10.7 0.38 (0.19–0.73) 
PFS 
 PBO 35/30 2.8  
 RAM Cmin,ss low vs. PBO 36/25 2.8 0.70 (0.40–1.24) 
 RAM Cmin,ss high vs. PBO 36/27 5.2 0.38 (0.20–0.71) 
Efficacy parametern/#eventsMedian (months)Adjusted HRa (95% CI)
OS 
 PBO 35/27 6.2  
 RAM Cmin,ss low vs. PBO 36/23 6.6 0.87 (0.49–1.58) 
 RAM Cmin,ss high vs. PBO 36/18 10.7 0.38 (0.19–0.73) 
PFS 
 PBO 35/30 2.8  
 RAM Cmin,ss low vs. PBO 36/25 2.8 0.70 (0.40–1.24) 
 RAM Cmin,ss high vs. PBO 36/27 5.2 0.38 (0.20–0.71) 

Abbreviations: CI, confidence interval; N, total number of patients; PBO, patients in placebo arm who had at least 3 cycles of treatment; RAM, ramucirumab; RAM Cmin,ss low, ramucirumab patients with median Cmin,ss <65.35 μg/mL; RAM Cmin,ss high, ramucirumab patients with median Cmin,ss ≥65.35 μg/mL.

aMultivariable analysis of OS and PFS hazard ratios were adjusted for peritoneal metastasis, Eastern Cooperative Oncology Group performance status, and location of primary tumor.

In summary, evaluating ramucirumab exposure as a continuous variable or as a categorical variable showed the same trend that increased ramucirumab exposure is associated with longer survival and smaller HRs (Table 4).

Exposure–safety analyses

RAINBOW.

The observed incidences of each safety endpoint by Cmin,ss quartile are shown in Fig. 2. Logistic regression analyses were performed to evaluate the relationship between predicted Cmin,ss and the incidences of grade ≥3 hypertension, neutropenia, and leukopenia. A statistically significant relationship (P < 0.001) was identified between the incidence of grade ≥3 hypertension, leukopenia, and neutropenia and model-predicted ramucirumab Cmin,ss. The febrile neutropenia rate in the RAINBOW study was low, at 2.4% for the placebo plus paclitaxel group, and increased to only 3.1% with the addition of ramucirumab to paclitaxel; due to the low incidence rate, logistic regression was not performed for febrile neutropenia.

Figure 2.

RAINBOW exposure–safety. The observed incidence of selected grade ≥3 adverse events by ramucirumab plus paclitaxel Cmin,ss quartiles. Abbreviations: PAC, paclitaxel; PBO, placebo; Q, quartile; RAM, ramucirumab. Quartiles, concentration range (μg/mL): Q1 = 20.1–≤42.0; Q2 = >42.0–≤53.8; Q3 = >53.8–≤72.2; Q4 = >72.2–227.

Figure 2.

RAINBOW exposure–safety. The observed incidence of selected grade ≥3 adverse events by ramucirumab plus paclitaxel Cmin,ss quartiles. Abbreviations: PAC, paclitaxel; PBO, placebo; Q, quartile; RAM, ramucirumab. Quartiles, concentration range (μg/mL): Q1 = 20.1–≤42.0; Q2 = >42.0–≤53.8; Q3 = >53.8–≤72.2; Q4 = >72.2–227.

Close modal

Covariates from the RAINBOW trial were evaluated for significant (P < 0.001) relationships with the incidences of hypertension, neutropenia, and leukopenia. Regardless of treatment arm, patients with a previous history of hypertension may have an increased risk of hypertension, older patients may have an increased risk of neutropenia, and Asian patients may have an increased risk of both neutropenia and leukopenia. The relationships between ramucirumab Cmin,ss and the three safety endpoints remained significant (P < 0.001) after adjusting for the covariates.

REGARD.

The number of patients with hypertension in the REGARD exposure population was limited (n = 10 with grade ≥3 hypertension) among the 72 ramucirumab-treated patients who provided PK exposure data. The incidence of grade ≥3 neutropenia was also low (n = 5 in the ramucirumab group); therefore, exposure–safety analyses were not conducted on these toxicity outcomes. Although it appears that higher exposure may increase the incidence of hypertension, with an odds ratio of 1.60 (95% CI, 0.34–8.45; Supplementary Table S1), data were too limited to draw a firm conclusion regarding the relationship between exposure and hypertension incidence.

RAINBOW and REGARD were two phase III clinical trials that established the efficacy and safety of ramucirumab plus paclitaxel or ramucirumab alone in patients with previously treated advanced gastric/GEJ cancer (2, 3). Exploratory analyses identified associations between model-predicted exposure parameters and survival and safety outcomes.

These exposure–response analyses suggest a positive relationship between ramucirumab exposure and survival, with an increased risk of toxicities that were manageable in the original phase III trials (2, 3). When the ramucirumab exposure–efficacy populations were stratified by exposure quartiles or dichotomized by the exposure median, there was clear separation between the OS curves, suggesting that higher ramucirumab exposure ranges may result in longer survival over the range of exposures achieved by a dose of 8 mg/kg given once every 2 weeks as a single agent or in combination with paclitaxel.

The safety profile of ramucirumab alone or in combination with paclitaxel is largely consistent with the safety profiles of the individual treatment components, and the combination has revealed no unexpected safety findings (2, 3). In the RAINBOW trial, grade ≥3 TEAEs occurring in at least 10% of patients and at a higher rate in the ramucirumab plus paclitaxel arm were neutropenia, leukopenia, and hypertension. Independent of the treatment arm, neutropenia was the most frequently reported AE; however, the incidence of febrile neutropenia was low and similar between treatment arms (3). The exposure–safety results presented here demonstrate that the incidence of grade ≥3 hypertension, neutropenia, and leukopenia significantly correlated with ramucirumab exposure, with increased exposure leading to increased incidence. AE data from the REGARD exposure–safety population were sparse, precluding any firm conclusions from this population. However, it is noteworthy that in the RAINBOW ITT population, grade ≥3 neutropenia was deemed as a clinically manageable safety event (3). Similarly, grade ≥3 hypertension was controlled with antihypertensive medication; additionally, there were no reports of grade 4 or 5 hypertension in the RAINBOW or the REGARD study (2, 3). Although patients in the lowest ramucirumab exposure groups may benefit from an increase in dose normalized by body weight, the risk for significant AEs would potentially increase in the patient population as a whole.

The intersubject variability for mAb exposure is likely dependent on a complex interaction of multiple factors, including disease types, age, sex, race, ethnicity, body size, genetic polymorphisms, concomitant medication, comorbidities, immune status, duration of therapy, and dose. In a previous pooled popPK meta-analysis of 11 ramucirumab studies, the PK profile of ramucirumab was well characterized by a 2-compartment model with interpatient variability estimated on all parameters. However, the analysis did not identify any significant variables that accounted for the differences in ramucirumab exposure (12). In that same analysis, although body weight did not meet the prespecified significance criteria for covariate selection, results did indicate a relationship between body weight and ramucirumab PK. Thus, although we were not able to identify patients who would be more likely to have higher or lower ramucirumab exposure, the previous PK meta-analysis confirmed the appropriateness of weight-based dosing (12). Likewise, biomarker analyses from the RAINBOW study have not identified a predictive biomarker for ramucirumab treatment, but did reveal several pharmacodynamic trends and potential markers for prognostic effects (15). Levels of the proangiogenic cytokines VEGF-D and placental growth factor increased, and levels of angiopoietin 2, a key regulator of angiogenesis, decreased in the ramucirumab treatment arm following initiation of treatment. Additionally, low baseline levels of several potential prognostic markers corresponded with longer OS and PFS across treatment arms. Further exploration of these biomarkers is warranted.

Exposure–response relationships similar to those described here for ramucirumab have been reported for several other mAbs being used to treat various types of cancer. The ToGA trial established the efficacy of the addition of trastuzumab to chemotherapy for patients with HER2-positive metastatic gastric or GEJ cancer. The analysis of the PK of trastuzumab in this population showed that some patients with gastric cancer had a higher clearance of trastuzumab than that seen in metastatic breast cancer patients (7, 16). These data suggest that a higher dose of trastuzumab may be required in at least some patients with advanced gastric cancer (16). A similar phenomenon was seen with bevacizumab in patients from the AVAGAST study with advanced gastric cancer and with pertuzumab from the JOSHUA study (9, 10). These PK and safety studies in patients with advanced gastric cancer determined that a higher dose of treatment is more likely to achieve the target exposure in a higher percentage of patients and have been used to establish the dosing regimen for additional trials (7, 10). Additionally, patients with HER2-positive metastatic breast cancer treated with ado-trastuzumab emtansine (T-DM1) who had trough concentrations below the median value had OS and PFS times comparable to those in the control arm; higher T-DM1 exposures were associated with improved OS and PFS (11). The popPK analysis of T-DM1 determined body weight, tumor burden, HER2 extracellular domain concentrations, aspartate aminotransferase levels, serum albumin levels, and baseline trastuzumab concentrations as the most statistically significant covariates for T-DM1 exposure, but the magnitude of the effect of each of these covariates was less than 10% and did not result in dose modifications (11). Higher steady-state trough concentrations of ipilimumab used to treat patients with advanced melanoma were also associated with improved survival (8). Exposure–response relationships of ramucirumab in patients with metastatic non–small cell lung cancer and in patients with advanced second-line colorectal cancer also showed associations between higher exposure to ramucirumab and longer OS and PFS, with increased incidence of grade ≥3 AEs (17, 18).

The exposure data from RAINBOW were recently analyzed by the FDA and, using a matched case–control approach, the same conclusions were reached (19). The exposure–response relationship seen in REGARD and RAINBOW led the FDA to request a Post Marketing Commitment trial to test the hypothesis that higher doses of ramucirumab could be beneficial with a manageable toxicity profile (20). This phase II trial will treat patients with paclitaxel in combination with alternative dosing schedules of ramucirumab (ClinicalTrials.gov identifier NCT02514551). In Europe, the EMA requested a monotherapy study of ramucirumab in a four-arm trial testing different doses and schedules (ClinicalTrials.gov identifier NCT02443883). These trials have efficacy or PK and safety as primary endpoints, respectively. These exploratory exposure–response findings have also informed the decision to increase the ramucirumab dose in RAINFALL, a first-line trial in gastric/GEJ cancer investigating ramucirumab 8 mg/kg or placebo given on days 1 and 8 of a 21-day cycle in combination with chemotherapy (21).

In conclusion, exposure–response results presented here suggest a positive relationship between efficacy and ramucirumab exposures generated from a dose of 8 mg/kg administered every 2 weeks as monotherapy or in combination with weekly paclitaxel with manageable toxicities. The hypothesis generated from these results, that increasing exposure to ramucirumab may increase efficacy, is currently being tested in randomized clinical trials.

J. Tabernero is a consultant/advisory board member for Amgen, Bayer, Roche, Sanofi, Symphogen, Taiho, Takeda, Boehringer Ingelheim, Celgene, Chugai, Lilly, MSD, Merck Serono, Novartis, and Pfizer. A. Ohtsu reports receiving a commercial research grant from BMS. E. Van Cutsem reports receiving a commercial research grant from and is a consultant/advisory board member for Lilly. G. Bodoky is a consultant/advisory board member for Novartis, Bayer, Janssen, Servier, Roche, Pfizer, and Lilly. S. Hironaka has received honoraria from speakers bureau of Eli Lilly, Taiho Pharma Co. Ltd., and MSD. J.A. Ajani has received honoraria from speakers bureau of Lilly. J. Tomasek has received honoraria from speakers bureau as a regional speaker. Y. Hsu has ownership interest (including patents) in Eli Lilly & Co. M. Heathman is a senior research advisor for Eli Lilly & Co. L. Ni has ownership interest (including patents) from Eli Lilly Stock. C. Fuchs declares consulting fees from Sanofi, Genentech, Merck, Eli Lilly & Co., Bristol Myers Squibb, Medimmune, Dicerna, Entrinsic Health, Macrogenics, Amgen, Pfizer, and Gilead. No potential conflicts of interest were disclosed by the other authors.

Conception and design: J. Tabernero, A. Ohtsu, J.A. Ajani, A. Khan, A.S. Melemed, L. Gao, D. Ferry, C. Fuchs

Development of methodology: J. Tabernero, A. Khan, L. Ni, L. Gao, D. Ferry

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J. Tabernero, K. Muro, E. Van Cutsem, S.C. Oh, Y. Shimada, S. Hironaka, J. Tomasek, H. Safran, L. Gao, C. Fuchs

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J. Tabernero, E. Van Cutsem, Y. Shimada, J.A. Ajani, K. Chandrawansa, Y. Hsu, M. Heathman, A. Khan, L. Ni, A.S. Melemed, L. Gao, D. Ferry, C. Fuchs

Writing, review, and/or revision of the manuscript: J. Tabernero, A. Ohtsu, K. Muro, E. Van Cutsem, S.C. Oh, G. Bodoky, Y. Shimada, S. Hironaka, J.A. Ajani, J. Tomasek, K. Chandrawansa, Y. Hsu, M. Heathman, A. Khan, L. Ni, A.S. Melemed, L. Gao, D. Ferry, C. Fuchs

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y. Shimada

Study supervision: J. Tabernero, D. Ferry, C. Fuchs

We would like to thank the patients, investigators, and institutions involved in this study. Medical writing assistance and editorial support were provided by Andrea Humphries, PhD, and Casie Polanco of inVentiv Health Clinical and were funded by Eli Lilly and Company.

This study was funded by Eli Lilly and Company.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Spratlin
JL
,
Cohen
RB
,
Eadens
M
,
Gore
L
,
Camidge
DR
,
Diab
S
, et al
Phase I pharmacologic and biologic study of ramucirumab (IMC-1121B), a fully human immunoglobulin G1 monoclonal antibody targeting the vascular endothelial growth factor receptor-2
.
J Clin Oncol
2010
;
28
:
780
7
.
2.
Fuchs
CS
,
Tomasek
J
,
Yong
CJ
,
Dumitru
F
,
Passalacqua
R
,
Goswami
C
, et al
Ramucirumab monotherapy for previously treated advanced gastric or gastro-oesophageal junction adenocarcinoma (REGARD): an international, randomised, multicentre, placebo-controlled, phase 3 trial
.
Lancet
2014
;
383
:
31
9
.
3.
Wilke
H
,
Muro
K
,
Van Cutsem
E
,
Oh
SC
,
Bodoky
G
,
Shimada
Y
, et al
Ramucirumab plus paclitaxel versus placebo plus paclitaxel in patients with previously treated advanced gastric or gastro-oesophageal junction adenocarcinoma (RAINBOW): a double-blind, randomised phase 3 trial
.
Lancet Oncol
2014
;
15
:
1224
35
.
4.
National Comprehensive Cancer Network (NCCN)
. 
NCCN clinical practice guidelines in oncology: esophageal and esophagogastric junction cancers, version 2
. 
2016
. Available from: https://www.nccn.org/professionals/physician_gls/pdf/esophageal.pdf
5.
Thurber
GM
,
Schmidt
MM
,
Wittrup
KD
. 
Antibody tumor penetration: transport opposed by systemic and antigen-mediated clearance
.
Adv Drug Deliv Rev
2008
;
60
:
1421
34
.
6.
Thurber
GM
,
Schmidt
MM
,
Wittrup
KD
. 
Factors determining antibody distribution in tumors
.
Trends Pharmacol Sci
2008
;
29
:
57
61
.
7.
Cosson
VF
,
Ng
VW
,
Lehle
M
,
Lum
BL
. 
Population pharmacokinetics and exposure-response analyses of trastuzumab in patients with advanced gastric or gastroesophageal junction cancer
.
Cancer Chemother Pharmacol
2014
;
73
:
737
47
.
8.
Feng
Y
,
Roy
A
,
Masson
E
,
Chen
TT
,
Humphrey
R
,
Weber
JS
. 
Exposure-response relationships of the efficacy and safety of ipilimumab in patients with advanced melanoma
.
Clin Cancer Res
2013
;
19
:
3977
86
.
9.
Han
K
,
Jin
J
,
Maia
M
,
Lowe
J
,
Sersch
MA
,
Allison
DE
. 
Lower exposure and faster clearance of bevacizumab in gastric cancer and the impact of patient variables: analysis of individual data from AVAGAST phase III trial
.
AAPS J
2014
;
16
:
1056
63
.
10.
Kang
YK
,
Rha
SY
,
Tassone
P
,
Barriuso
J
,
Yu
R
,
Szado
T
, et al
A phase IIa dose-finding and safety study of first-line pertuzumab in combination with trastuzumab, capecitabine and cisplatin in patients with HER2-positive advanced gastric cancer
.
Br J Cancer
2014
;
111
:
660
6
.
11.
Wang
J
,
Song
P
,
Schrieber
S
,
Liu
Q
,
Xu
Q
,
Blumenthal
G
, et al
Exposure-response relationship of T-DM1: insight into dose optimization for patients with HER2-positive metastatic breast cancer
.
Clin Pharmacol Ther
2014
;
95
:
558
64
.
12.
O'Brien
L
,
Westwood
P
,
Gao
L
,
Heathman
M
. 
Population pharmacokinetic meta-analysis of ramucirumab in cancer patients
.
J Pharmacokinet Pharmacodyn
2015
:
42
:
abstract M-10
.
13.
U.S. Department of Health and Human Services
. 
Guidance for industry, exposure-response relationships–study design, data analysis, and regulatory applications
, 
April 2003
.
Available from:
http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/ucm072109.pdf.
Accessed
Mar 11, 2014
.
14.
Cancer Therapy Evaluation Program
. 
Common terminology criteria for adverse events v4.02
. 
2009
. Available from: http://ctep.cancer.gov/protocolDevelopment/electronic_applications/ctc.htm#ctc_40
Accessed
Feb 28, 2014
.
15.
Van Cutsem
E
,
Muro
K
,
Cunningham
D
,
Bodoky
G
,
Sobrero
A
,
Cascinu
S
, et al
Biomarker analyses of second-line ramucirumab in patients with advanced gastric cancer from RAINBOW, a global, randomized, double-blind, phase 3 study
.
Ann Oncol
2016
;
27
:
abstract O-007
.
16.
Bang
YJ
,
Van Cutsem
E
,
Feyereislova
A
,
Chung
HC
,
Shen
L
,
Sawaki
A
, et al
Trastuzumab in combination with chemotherapy versus chemotherapy alone for treatment of HER2-positive advanced gastric or gastro-oesophageal junction cancer (ToGA): a phase 3, open-label, randomised controlled trial
.
Lancet
2010
;
376
:
687
97
.
17.
Smit
EF
,
Perol
M
,
Reck
M
,
Cappuzzo
F
,
Bidoli
P
,
Cohen
RB
, et al
Exposure-response relationship for ramucirumab (RAM) from the randomized, double-blind, phase III REVEL trial (docetaxel [DOC] vs DOC plus RAM) in second-line treatment of metastatic non-small cell lung cancer (NSCLC)
.
J Clin Oncol
2015
;
33
:
abstract 8053
.
18.
Cohn
AL
,
Yoshino
T
,
Obermannova
R
,
Bodoky
G
,
Prausova
J
,
Garcia-Carbonero
R
, et al
Exposure-response (E-R) relationship of ramucirumab (RAM) from a global, randomized, double-blind, Phase 3 study of patients (Pts) with advancd 2nd line colorectal cancer
.
Eur J Cancer
2015
;
51
:
abstract 2123
.
19.
Jin
R
,
Li
H
,
Zhang
LH
,
Zhao
H
,
Fashoyin-Aje
L
,
Lemery
S
, et al
Exposure-response (E-R) and case-control analyses of ramucirumab leading to recommendation for dosing optimization in patients with gastric cancer
.
J Clin Oncol
2015
;
33
:
abstract 2578
.
20.
Casak
SJ
,
Fashoyin-Aje
I
,
Lemery
SJ
,
Zhang
L
,
Jin
R
,
Li
H
, et al
FDA approval summary: ramucirumab for gastric cancer
.
Clin Cancer Res
2015
;
21
:
3372
6
.
21.
Fuchs
CS
,
Tabernero
J
,
Al-Batran
S-E
,
Chau
I
,
Ilson
DH
,
Van Cutsem
E
, et al
A randomized, double-blind, placebo-controlled phase III study of cisplatin plus a fluoropyrimidine with or without ramucirumab as first-line therapy in patients with metastatic gastric or gastroesophageal junction (GEJ) adenocarcinoma (RAINFALL, NCT02314117)
.
J Clin Oncol
2016
;
34
:
abstract TPS178
.