Efficacy of ramucirumab combination chemotherapy as second‐line treatment in patients with advanced adenocarcinoma of the stomach or gastroesophageal junction after exposure to checkpoint inhibitors and chemotherapy as first‐line therapy

FOLFOX plus nivolumab represents a standard of care for first‐line therapy of advanced gastroesophageal cancer (aGEC) with positive PD‐L1 expression. The efficacy of second‐line VEGFR‐2 inhibition with ramucirumab (RAM) plus chemotherapy after progression to immunochemotherapy remains unclear. Medical records of patients with aGEC enrolled in the randomized phase II AIO‐STO‐0417 trial after treatment failure to first‐line FOLFOX plus nivolumab and ipilimumab were retrospectively analyzed. Patients were divided into two groups based on second‐line therapy: RAM plus chemotherapy (RAM group) or treatment without RAM (control group). Eighty three patients were included. In the overall population, progression‐free survival (PFS) in the RAM group was superior to the control (4.5 vs 2.9 months). Responders (CR/PR) to first‐line immunochemotherapy receiving RAM containing second‐line therapy had prolonged OS from start of first‐line therapy (28.9 vs 16.5 months), as well as second‐line OS (9.6 vs 7.5 months), PFS (5.6 vs 2.9 months) and DCR (53% vs 29%) compared to the control. PD‐L1 CPS ≥1 was 42% and 44% for the RAM and the control, respectively. Patients with CPS ≥1 in the RAM group showed better tumor control (ORR 25% vs 10%) and improved survival (total OS 11.5 vs 8.0 months; second‐line OS 6.5 vs 3.9 months; PFS 4.5 vs 1.6 months) compared to the control. Prior exposure to first‐line FOLFOX plus dual checkpoint inhibition followed by RAM plus chemotherapy shows favorable response and survival rates especially in patients with initial response and positive PD‐L1 expression and has the potential to advance the treatment paradigm in aGEC.

therapy: RAM plus chemotherapy (RAM group) or treatment without RAM (control group).Eighty three patients were included.In the overall population, progressionfree survival (PFS) in the RAM group was superior to the control (4.5 vs 2.9 months).
Responders (CR/PR) to first-line immunochemotherapy receiving RAM containing second-line therapy had prolonged OS from start of first-line therapy (28.9 vs 16.5 months), as well as second-line OS (9.6 vs 7.5 months), PFS (5.6 vs 2.9 months) and DCR (53% vs 29%) compared to the control.PD-L1 CPS ≥1 was 42% and 44% for the RAM and the control, respectively.Patients with CPS ≥1 in the RAM group showed better tumor control (ORR 25% vs 10%) and improved survival (total OS 11.5 vs 8.0 months; second-line OS 6.5 vs 3.9 months; PFS 4.5 vs 1.6 months) compared to the control.Prior exposure to first-line FOLFOX plus dual checkpoint inhibition followed by RAM plus chemotherapy shows favorable response and survival rates especially in patients with initial response and positive PD-L1 expression and has the potential to advance the treatment paradigm in aGEC.
antiangiogenic treatment, anti-VEGFR-2 therapy, gastroesophageal cancer, immunotherapy, ramucirumab What's new?Evidence suggests that immunochemotherapy followed by second-line treatment with ramucirumab, a vascular endothelial growth factor receptor (VEGF) inhibitor, potentially boosts treatment response in patients with advanced gastroesophageal cancer (aGEC).Here, to more thoroughly explore this possibility, the authors analyzed data from a phase II trial involving aGEC patients treated with ramucirumab after failing first-line immunochemotherapy.Improved tumor control and lengthened survival were observed in patients following second-line ramucirumabcontaining therapy.The benefit was greatest among patients who initially responded to first-line therapy and whose tumors expressed programmed-death ligand-1.The findings indicate that integration of VEGF-targeted agents into immunochemotherapy-based regimens improves aGEC survival.

| INTRODUCTION
Immune checkpoint inhibition (ICI) targeting programmed-death (PD)-1 has demonstrated benefit in advanced gastric and gastroesophageal cancer (aGEC) patients with positive PD-L1 expression as well as in distinct subpopulations such as mismatch-repair deficient/ microsatellite instability high (dMMR/MSI-H) tumors 1,2 and is recommended by current treatment guidelines. 3,4Combination regimens of ICI and chemotherapy showed durable responses and prolonged survival in phase II and phase III trials leading to the approval of two anti-PD-1/PD-L1 monoclonal antibodies in combination with chemotherapy for PD-L1 positive tumors in the first-line setting.First-line nivolumab plus chemotherapy leads to an overall survival (OS) improvement of 14.4 months compared to 11.1 months with chemotherapy alone in patients with a PD-L1 combine positive score (CPS) of five or more. 5Moreover, the addition of pembrolizumab to chemotherapy improves OS in aGEC patients with PD-L1 CPS ≥1 compared to chemotherapy alone (13.0 vs 11.4 months). 6However, response rates are still limited and a significant subset of patients relapses after immunochemotherapy.
The antiangiogenic antibody ramucirumab (RAM), directed against the vascular endothelial growth factor receptor 2 (VEGFR-2), in combination with paclitaxel is effective after progression of first-line chemotherapy and improves OS compared to chemotherapy alone. 7Due to recent implementation of immunotherapy into first-line treatment, the evaluation of the efficacy of antiangiogenic therapy with RAM after progression to ICI is considered relevant.
Prior exposure to ICI is potentially able to enhance the efficacy of subsequent angiogenesis inhibition with RAM.Improved response rates of RAM in combination with docetaxel were reported in NSCLC patients after progression to PD-1 inhibition. 8,9Patients with aGEC enrolled in the KEYNOTE-059 trial showed durable responses to RAM/paclitaxel after progression on first-line pembrolizumab. 10Furthermore, overall response rates (ORR) of 60.5% were demonstrated in patients receiving RAM/paclitaxel after prior exposure to anti-PD-1 therapy. 11[18][19] In pretreated patients with aGEC, angiogenesis inhibition with RAM in combination with ICI showed encouraging activity as second-line treatment. 20,21Of note, the combination of tyrosine kinase inhibitors (TKIs) with activity against VEGFR (regorafenib, lenvatinib) and ICI demonstrated antitumor efficacy in aGEC. 22,23Regarding the promising therapeutic synergism of concurrent anti-VEGF and anti-PD-L1 combinatory therapies, further insight on the efficacy of sequential treatment of second-line ramucirumab after exposure to ICI is needed.
So far, no studies have prospectively evaluated the efficacy of second-line ramucirumab containing treatment after progression to first-line immunochemotherapy in aGEC.Here we report the outcomes of patients with aGEC who received second-line ramucirumab-containing therapy after exposure to checkpoint inhibitors and chemotherapy as first-line therapy within the AIO-STO-0417 trial (NCT03647969).

| Design and participating centers
The efficacy of second-line RAM-containing treatment in patients with aGEC after first-line ICI was evaluated retrospectively.Medical records of 83 consecutive patients receiving first-line immunochemotherapy within the randomized phase II AIO-STO-0417 trial followed by second-line treatment were reviewed.Within the AIO-STO-0417 trial patients received first-line therapy with mFOLFOX (oxaliplatin 85 mg/m 2 , leucovorin 400 mg/m 2 , 5-FU 400 mg/m 2 on day 1 followed by 5-FU 2400 mg/m 2 over 48 h, Q2W) combined with nivolumab (240 mg, Q2W) and ipilimumab (1 mg/kg, Q6W), sequential immunochemotherapy starting with 3 cycles of mFOLFOX, followed by dual ICI with nivolumab (Â4 240 mg Q2W) and ipilimumab (Â2 1 mg/kg Q6W) or chemotherapy alone (mFOLFOX).Molecular characteristic such as HER2-status was assessed via immunohistochemistry using formalin-fixed, paraffin-embedded tissue specimens.PD-L1 expression was defined by CPS which evaluates the number of PD-L1-staining cells (tumor cells, lymphocytes, macrophages) relative to all viable tumor cells.Laboratorydeveloped immunohistochemical PD-L1 staining was performed centrally according to current guidelines using the primary antibody clone 28-8 (Ab 205921).

| Statistical analysis
The baseline demographic data, tumor characteristics, first and second-line therapy and best overall response were reported as frequencies (percentages) for categorical variables and median (minimum-maximum) for continuous variables and the results were displayed separately between the RAM group and the control group.Time-to-event endpoints including PFS, and OS were summarized using Kaplan-Meier methods as well as log-rank tests and displayed both graphically and in tabular form with the number of events and median survival time with corresponding 95% confidence intervals (CIs) to compare the survival time between groups.Subgroup analyses were performed based on response to first-line immunochemotherapy and PD-L1 status.All parameters were analyzed in an exploratory, descriptive manner.All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., NC).

| Patient disposition, characteristics and treatment
In total 83 patients were included in our study.Baseline patients' demographics and tumor characteristics are displayed in Table 1

| Overall population
In the overall population, the median total OS (from randomization until death) was 12.5 months in the RAM group and 14.2 months in the control group.The median second-line OS (from start of the second-line therapy until death) was 6.5 months in the RAM group and 6.6 months in the control group (Figure 1).The median secondline PFS was 4.5 months in the RAM group and 2.9 months in the control group (Figure 1 3).

| Subgroup analysis according to response to first-line immunochemotherapy
Within the RAM group 15 patients (39%) and within the control group 24 patients (53%) achieved CR or PR to first-line therapy (defined as responders).In this subgroup of responders to first-line, median total OS after second-line treatment was clearly prolonged for patients in the RAM group compared to the control group (28.9 vs 16.5 months, P = .553;Figure 1).In contrast, for patients who did not respond to the first-line therapy (ie, with SD or PD) no relevant differences in OS were observed (Figure S1).Furthermore, median second-line PFS was increased for first-line responder in the RAM group compared to the control group (5.6 vs 2.9 months).Of note, DCR was achieved in 53% of first-line responders after RAM-containing second-line treatment vs 29% in the control group (Table 4).

| Subgroup analysis according to PD-L1 status
Within the RAM group 16 patients (42%) and within the control group 20 patients (44%) had a PD-L1 CPS ≥1.In this subgroup, median total OS and second-line OS was prolonged for patients who received RAM-containing second-line treatment compared to the control group (total OS: 11.5 vs 8.0 months; second-line OS: 6.5 vs 3.9 months).
second-line PFS was clearly increased in the RAM group compared to the control group (4.5 vs 1.6 months; Figure 1).Furthermore, ORR and DCR was increased for patients with PD-L1 CPS ≥1 after RAMcontaining second-line treatment compared to the control group (ORR: 25% vs 10%; DCR: 44% vs 30%; Table 5).In contrast, patients with a PD-L1 CPS <1 did not show improved OS, PFS, ORR and DCR to second-line RAM-containing treatment compared to control group (Figure S2).

| DISCUSSION
This retrospective study reports response and survival data of patients with aGEC enrolled in the randomized phase II AIO-STO-0417 trial who received second-line treatment with or without RAM after treatment failure on first-line mFOLFOX plus nivolumab and ipilimumab.
Our analysis showed that RAM-containing second-line treatment is effective after prior immunochemotherapy, especially in patients with initial response on immunochemotherapy and any PD-L1 expression (CPS ≥1).To our knowledge, this is the first report on the integration of ICI with VEGF targeted agents in the sequential treatment algorithm in a well-defined prospectively collected trial population.
The most frequent second-line therapy used in our study was the combination of RAM and paclitaxel which is the current standard of care after progression to first-line treatment, independent of PD-L1 expression. 3,4The observed improvement of PFS in patients who received second-line RAM-containing therapy compared to chemotherapy alone is consistent with previously reported findings. 11Of note, indirect comparison of the results of the retrospective analysis by Sasaki et al showed higher tumor response rates compared to this analysis, which might be attributed to a more homogenously selected patient population receiving exclusively taxanes in combination with RAM.Besides, cross-trial comparison should be carefully interpreted based on different patient characteristics.
Patients with immunotherapy-sensitive tumors seem to benefit in particular from sequential RAM compared to non-responders suggesting the ability of second-line anti-VEGF targeted therapy to reverse resistance to ICI.This can be explained by the immunomodulatory effects of antiangiogenic therapies like RAM, representing a distinct antitumor activity of these therapies beyond vessel growth inhibition and vessel normalization.5][26] Blocking VEGFR-2 with RAM can reverse these immunosuppressive mechanisms and can reinstall a more immunocompetent TME. 19Hence, anti-VEGF targeted therapy has the potential to enhance the efficacy of PD-1/PD-L1 inhibition.
Interestingly, ICI with nivolumab results in a PD-1 receptor occupation of >70% on circulating T cells that lasts more than 2 months after the last infusion of the anti-PD1 antibody. 27Subsequent antiangiogenic therapy with RAM would act on a tumor that progressed despite a sustained PD-1 checkpoint blockade.In patients with advanced NSCLC sequential RAM plus docetaxel led to improved PFS and response rates after progression on second-and third-line immunochemotherapy. 8,28Similarly, the KEYNOTE-059 trial reported durable responses to RAM/paclitaxel after progression on first-line pembrolizumab in patients with aGEC. 10 Since the immunogenic responsiveness to second-line RAM treatment.Thus, besides initial response to first-line immunochemotherapy, positive PD-L1 expression seems to be predictive of the responsiveness to RAM-containing second-line therapy.PD-L1 expression, as well as high tumor mutational burden (TMB) and MSI-H status correlate with response and improved survival under anti-PD-1/PD-L1 monotherapy and have been validated as predictors to the response of immunochemotherapy in aGEC. 5,6Thus, PD-L1 expression is not only prerequisite for response to immunochemotherapy but might also lead to better tumor T-cell responses, respectively. 29,30The anti-CTLA4 antibodies, ipilimumab and tremelimumab, have been studied in combination with anti-PD-1/PD-L1 antibodies in randomized trials for locally advanced and metastatic GEC. 31,32However, the efficacy of combinatory treatment with anti-CTLA-4 and anti-VEGF therapies is currently unclear but should be addressed in future prospective trials.
The main limitation of our study is the limited sample size, thus warranting further evaluation in a larger cohort.Of note, this trial exclusively evaluated data on the clinical efficacy of second-line treatment without assessing toxicity and safety of subsequent therapies.
Although no severe safety issues were reported for RAM-containing second-line treatment in previously published studies, 11 caution must be held when translating the therapeutic benefits observed in our study into clinical practice.Furthermore, the variety of different chemotherapy regimens used in combination with RAM as second-line treatment and the small sample size in this trial inevitably lead to the limitation that the chemotherapeutic regimens might influence the prognosis to a certain extend independently of RAM.Moreover, the presented data were retrospectively analyzed as it is within the nature of a secondary explorative analysis.However, since the results were not initially generated in a randomized trial, the patient Patients who met the following criteria were included: (a) Presence of advanced, histologically proven Her2-negative gastric or gastroesophageal adenocarcinoma.(b) Patients underwent at least one administration of both chemotherapy plus anti-PD-1 and anti-CTLA-4 therapy followed by second-line systemic treatment (chemotherapy, ICI or antibody-based therapy), radiotherapy or surgical treatment.(c) An Eastern Cooperative Oncology Group performance status (ECOG PS) of 0 to 2. Patients were divided into two groups based on subsequent second-line therapy: RAM group including all patients who received RAM-containing secondline therapy and the control group including all patients who did not receive RAM but any other treatment.The used second-line therapies were retrieved from the eCRF of the AIO-STO-0417 trial.Data were collected from 27 study sites participating in the AIO-STO-0417 trial and recorded in a standardized manner.Our study was conducted in accordance with the Declaration of Helsinki (as revised in 2013).

2. 2 |
Assessment of tumor response and efficacy of treatmentThe current analysis primarily aimed to investigate the efficacy of second-line VEGF-2 inhibition with RAM plus chemotherapy after treatment failure to immunochemotherapy.Clinical outcomes after second-line treatment were compared between RAM-containing therapies and treatment without RAM in the overall population.Furthermore, subgroup analyses were conducted according to responsiveness to first-line treatment (responders [complete response, CR; partial response, PR] vs non-responders [stable disease, SD; progressive disease PD]) and to PD-L1 expression (combined positive score (CPS) <1 vs CPS ≥1).We assessed the overall survival (OS), progression free survival (PFS), overall response rate (ORR) and disease control rate (DCR).Tumor response was evaluated in patients with measurable lesions according to the guidelines of the Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1.
effect of preceding PD-1 blockade seems to last beyond tumor progression, sequential VEGF inhibition with RAM activates this therapeutic synergism and might therefore be able to overcome resistance to ICI.The observation within our study of improved OS and PFS in responders to first-line immunochemotherapy indicates the ability of second-line RAM-containing therapy to reverse primary resistance to ICI.Moreover, patients with high PD-L1 tumor expression had better tumor control and prolonged survival under VEGFR2 inhibition indicating a correlation between PD-L1 tumor expression and F I G U R E 1 Kaplan-Meier curves for 2 nd -line OS (from start 2 nd -line treatment) and PFS.(A, B) Overall population, (C, D) subgroup of responders to 1st-line treatment, (E, F) subgroup of patients with tumor PD-L1 CPS ≥1.
response to second-line RAM-containing treatment.Ongoing and future trials investigating the synergism of VEGF plus PD-1/PD-L1 inhibition should incorporate the evaluation of the predictive value of these biomarkers.Moreover, identification of additional biomarkers aiming to predict the synergy between VEGF inhibition and ICI is clearly warranted.Combinatory ICI with PD-1/PD-L1 and CTLA-4 inhibition has distinct but complementary mechanisms of action that contribute to the restoration of antitumor T-cell function and induction of de novo anti-

a
Response assessment under first-line was missing for one patient in the control group.population was clearly biased for homogeneity.Given the limitation of missing statistical significance of the reported results, the current study only generates a hypothesis.The generalizability of our findings requires confirmation in a larger prospective randomized trial.Ongoing phase III trials such as the LEAP15 (NCT04662710) or INTEGRATE IIb (NCT0487936) evaluate the therapeutic synergism of ICI and VEGF inhibition in advanced or refractory GEC.If survival advantage is confirmed in these trials, the next step should be the evaluation of this treatment approach in multimodal therapy of earlystage GEC.Within the limitations of the exploratory nature of the data this analysis is of significant interest, showing that prior exposure to firstline FOLFOX plus dual ICI followed by chemotherapy plus RAM has promising potential to advance the treatment paradigm in gastroesophageal cancer, with most favorable tumor responses and survival outcomes in patients with initial response and positive PD-L1 expression.AUTHOR CONTRIBUTIONS Michael Masetti: Designed the model and the computational framework and analyzed the data; organized the data collection, performed and verified the statistical analysis; wrote the article with input from all authors.Salah-Eddin Al-Batran: Conceived the original idea of the study.Disorn Sookthai: Organized the data collection, performed and verified the statistical analysis.Sabine Junge: Organized the data collection, performed and verified the statistical analysis.Maria Loose: Organized the data collection, performed and verified the statistical analysis.Claudia Pauligk: Organized the data collection, performed and verified the statistical analysis.Sylvie Lorenzen: Conceived the original idea of the study; designed the model and the computational framework and analyzed the data.The work reported in the article has been performed by the authors, unless clearly specified in the text.All authors discussed the results, commented on the article and contributed to the final article.
First-and second-line treatment.
. All patients received second-line treatment after exposure to first line RAM (control group) (Table 2).The most common RAM-containing second-line therapy was ramucirumab plus paclitaxel (n = 32 [85%]).Most patients in both groups discontinued first-line treatment due to disease progression (RAM group: n = 35 [92%]; control group: n = 38 [84%]).The duration of second-line treatment was 2.1 months in the RAM group and 1.6 months in the control group.PD-L1 status was available in 77% of patients (n = 64).Of those, 16 patients (42%) in the RAM group and 20 patients (44%) in the control group had a PD-L1 CPS ≥1.Distribution of prognostic factors like ECOG-PS and tumors with signet-ring cells numerically slightly favored the control group.T A B L E 2 Abbreviations: CR, complete response; FLO, 5-FU + leucovorin + oxaliplatin; FLOT, 5-FU + leucovorin + oxaliplatin + docetaxel; FOLFIRI, 5-FU + leucovorin + irinotecan; FOLFOX, 5-FU + leucovorin + oxaliplatin; PR, partial response.a T A B L E 3 Best overall response under second-line therapy.DCR, disease control rate; OS, overall survival; ORR, objective response rate; PFS, progression-free survival.
Abbreviations: CPS, combined positive score; DCR, disease control rate; ORR, objective response rate; OS, overall survival; PFS, progression-free survival.T A B L E 4 Subgroup analysis according to best response under first-line therapy.