Deeper insights into combinations of immune checkpoint inhibitors with other drug classes

Pooled data on chemo-IO vs. IO in PD-L1 ≥ 50 %

Various regimens consisting of anti-PD-(L)1 antibodies with or without chemotherapy have been approved for the first-line treatment of patients with advanced NSCLC that does not harbor genomic alterations. The analysis reported at ASCO 2022 by Akinboro et al. used pooled data from 12 pivotal studies to compare overall survival (OS) obtained with chemoimmunotherapy (n = 455) vs. immunotherapy (n = 1,298) in patients with ALK– and EGFR-negative tumors that showed ≥ 50 % PD-L1 expression [1]. Six randomized controlled trials each had evaluated chemoimmunotherapy and immunotherapy alone. All comparator regimens consisted of platinum-based chemotherapy.

Despite a slight numerical advantage favoring chemoimmunotherapy, the data did not suggest an OS difference compared to the immunotherapy-only treatment. Median OS was 25.0 vs. 20.9 months, which translated into a 18 % risk reduction (HR, 0.82). The Kaplan-Meier estimation suggested a marginal separation of the curves within the first 18 months. For progression-free survival (PFS), the exploratory analysis revealed superiority of chemoimmunotherapy, with a median of 9.6 vs. 7.1 months (HR, 0.69). Here, the Kaplan-Meier curves separated clearly within the first year of treatment, which might represent a potentially greater cytoreductive effect; this warrants further exploration. Also, the overall response rates (ORRs) were higher with the combined approach (61 % vs. 43 %; OR, 1.2). According to the subgroup analyses, however, patients aged ≥ 75 years appeared to derive greater benefit from immunotherapy alone with respect to both OS and PFS.

As the authors pointed out, limitations arise from the retrospective and exploratory nature of the analyses, and the results are only hypothesis-generating. No factors were examined that might explain the lack of concordance between the OS results on one hand and the PFS/ORR findings on the other. Also, potential heterogeneity across the trials including differences in PD-L1 assays need to be taken into account. Nevertheless, these findings emphasize the importance of shared decision making in selecting a therapeutic approach.

Meta-analysis according to KRAS status

Another pooled analysis presented by Nakajima et al. investigated the outcomes of first-line chemoimmunotherapy vs. immunotherapy according to KRAS mutation status and PD-L1 expression [2]. The researchers assessed the question of whether patients with KRAS-mutated advanced NSCLC respond differently to immunotherapy with or without chemotherapy than those with KRAS wildtype. To this end, data from a total of 1,430 ­patients included in 12 randomized trials were analyzed. Sixty-one percent of these (n = 875) had KRAS wildtype, while 39 % had KRAS mutations (n = 555). The KRASG12C mutation was present in 11 % (n = 157). Roughly equal proportions within each of these groups were PD-L1–negative (TPS, < 1 %), PD-L1–low (1-49 %), and PD-L1–high (≥ 50 %). Chemo­immunotherapy had been administered in 35 % to 39 % across the groups, and immunotherapy alone had been used in 24 % to 29 % of the population.

Among patients with both KRAS wildtype and KRAS-mutant disease, ORRs were higher in the groups treated with checkpoint inhibitors plus chemotherapy than in those treated with immunotherapy alone (Table 1). ORRs did not differ across patients with KRAS wildtype and KRAS mutations; this was true for chemoimmunotherapy, immunotherapy only, and chemotherapy. Likewise, OS was highest in patients treated with chemoimmunotherapy, and no notable differences emerged according to KRAS status for either treatment regimen. The hazard ratios were exploratory and did not have prespecified alpha.

In patients with both PD-L1–high and PD-L1–low disease, those treated with chemoimmunotherapy appeared to obtain the greatest survival benefit independent of KRAS status. Also, in the group with PD-L1–negative tumors, median OS was similar among patients with KRAS wildtype and KRAS-mutated disease. The numbers of patients who received immunotherapy alone were low here, which limits conclusions on the survival benefit in this group.

This analysis represents the most comprehensive assessment of patients with KRAS-mutated NSCLC in response to first-line therapy, although it has limitations as a retrospective, exploratory analysis. Overall, all patients appeared to benefit from the addition of chemotherapy to immunotherapy regardless of KRAS mutation status and PD-L1 expression. Collectively, these data suggest that the optimal comparator for studies conducted in the first-line setting in patients with KRAS-mutant disease might consist of immune checkpoint inhibition plus chemotherapy. Additional data are needed to determine whether there is a subset of patients with KRAS-mutant NSCLC who can forgo first-line immunotherapy in favor of targeted therapy that has become available for patients with KRASG12C-mutant disease.

Table 1 Response rates and overall survival with different treatment regimens according to KRAS mutation status

Eftilagimod alpha plus pembrolizumab: TACTI-002

The antitumor activity of PD-1 antagonists is synergistically enhanced in combination with the soluble LAG-3 protein eftilagimod alpha, which targets a subset of MHC class II molecules, thus activating antigen-presenting cells and leading to an increase in activated T cells. Felip et al. reported initial results from Part A of the multinational, open-label, phase II TACTI-002 trial at ASCO 2022 [3]. In this group that comprised 114 patients recruited across 6 countries, eftilagimod alpha Q2W was administered together with pembrolizumab Q3W for 8 cycles followed by eftilagimod alpha plus pembrolizumab Q3W for 9 cycles. The combined phase lasted for up to 1 year; subsequently, pembrolizumab monotherapy Q3W was administered for another year. These patients had untreated, advanced or metastatic NSCLC not amenable to targeted therapy and were unselected for PD-L1 expression. Approximately 70 % had a TPS < 50 %. ORR by iRECIST was defined as the primary endpoint.

After a median follow-up of 11.2 months, the ORR by iRECIST was 38.6 % in the ITT population, with the analysis according to RECIST 1.1 revealing a similar ORR of 37.7 %. In the evaluable population that had ≥ 1 post-baseline radiological assessment (n = 103), 42.7 % and 41.8 % of patients responded according to iRECIST and RECIST 1.1, respectively. The ORR analysis by PD-L1 status showed that those with PD-L1 expression ≥ 50 % experienced the highest ORR (52.6 % according to iRECIST), which markedly exceeded the ORR of 28.1 % in the PD-L1–negative population. Disease control rates ranged from 68.8 % to 78.9 % across all PD-L1 subgroups. No ORR difference resulted between patients with squamous and non-squamous tumors (35.0 % and 38.9 %, respectively).

Responses were deep and durable. Two thirds of patients with a post-baseline assessment had decreases in target lesions, and only 8.6 % of those with confirmed response progressed within 6 months until data cutoff. Complete remissions occurred in 2 patients. Median duration of response had not been reached yet. The median PFS of 6.9 months was deemed promising in this PD-L1–unselected population. Again, PFS was more favorable in the PD-L1–high subgroup, with a median of 11.8 months in those with PD-L1 ≥ 50 % vs. 4.2 months in the PD-L1–negative cohort.

Eftilagimod alpha was demonstrated to be safe and well tolerated. Treatment-emergent AEs (TEAEs) included dyspnea, asthenia, decreased appetite, cough, anemia, and fatigue. Local injection site reactions were seen in 20.3 % of patients and rated as grade 1 in almost all cases; no grade ≥ 3 events were reported. Treatment-related grade ≥ 3 TEAEs occurred in 10.5 % and led to discontinuation in 9.6 %. The most common AEs with possible immune etiology were diarrhea (any grade, 15.8 %), hypothyroidism (8.8 %), hyperthyroidism (5.3 %), and pneumonitis (3.5 %). No cytokine release syndrome was observed. In their conclusion, the authors noted that eftilagimod alpha plus pembrolizumab showed encouraging efficacy in first-line, PD-L1–unselected patients and warrants late-stage clinical investigation.

COSMIC-021: cabozantinib alone and plus atezolizumab

The combination of the multikinase inhibitor cabozantinib and the anti-PD-L1 antibody atezolizumab has demonstrated encouraging clinical activity in immunotherapy-pretreated patients included in the phase IB COSMIC-021 study [4]. Outcomes for the combination in the expanded Cohort 7 and for cabozantinib alone in Cohort 20 were reported by Neal et al. [5]. COSMIC-021 enrolled a population with stage IV, non-squamous NSCLC and radiographic progression on or after one immune checkpoint inhibitor administered for metastatic disease. The patients had been treated with ≤ 2 prior lines of systemic anticancer therapy. In Cohort 7, cabozantinib 40 mg QD was administered together with atezolizumab 1,200 mg Q3W (n = 81), while Cohort 20 received cabozantinib 60 mg QD (n = 31). The PD-L1 status was not available in all patients; in those in whom it was available, approximately 70 % showed PD-L1 positivity. Platinum-based chemotherapy had been administered in > 80 % of patients in both cohorts. ORR was defined as the primary endpoint.

The combination demonstrated encouraging clinical activity, with an ORR of 19 %, a disease control rate of 80 %, median PFS of 4.5 months, and median OS of 13.8 months (Table 2). Responses were observed with cabozantinib plus atezolizumab irrespective of known PD-L1 expression. There was a trend towards improved PFS and OS in the groups with PD-L1–positive tumors and unknown PD-L1 status compared to the PD-L1–negative group. Seventy-six percent of patients experienced tumor reductions. For cabozantinib alone, on the other hand, the analysis revealed only modest activity. Six percent of patients responded, with 65 % achieving disease control. Median PFS and OS amounted to 3.4 and 9.4 months, respectively.

The safety profiles of both the combination and cabozantinib monotherapy were consistent with those previously reported. Diarrhea, decreased appetite, nausea, and fatigue occurred most commonly in both cohorts. Among adverse events of special interest, rash, liver function test abnormalities, laboratory pancreatitis and thyroid abnormalities were observed. Cabozantinib dose reductions due to AEs were necessary in 40 % and 58 % in Cohorts 7 and 20, respectively, although treatment-related AEs (TRAEs) led to discontinuation of cabozantinib treatment only in 14 % and 10 %, respectively. Cabozantinib is currently tested together with nivolumab in the phase II EA5191 trial (NCT04310007) and combined with atezolizumab in the phase III CONTACT-01 study (NCT04471428).

Table 2 COSMIC-021: clinical outcomes observed with cabozantinib/atezolizumab and cabozantinib monotherapy

Lung-MAP sub-study S1800A

Lung-MAP is a master protocol to evaluate biomarker-driven agents and immunotherapies in previously treated patients with stage IV or recurrent NSCLC. Those not eligible for biomarker-matched sub-studies enroll in unmatched sub-studies all of which operate independently. The unmatched sub-study S1800A was designed as a randomized phase II trial to compare pembrolizumab plus the anti-VEGFR2 antibody ramucirumab (n = 69) with standard treatment according to the investigator’s choice (i.e., docetaxel plus ramucirumab; docetaxel; gemcitabine; pemetrexed; n = 67). Pembrolizumab plus ramucirumab was chosen to overcome acquired resistance to immunotherapy, which is a major area of unmet need for patients with NSCLC. Direct and indirect effects of angiogenesis-modulating factors on the tumor microenvironment have been observed [6], and pembrolizumab combined with ramucirumab has shown preliminary activity and safety in a phase I trial in patients with advanced NSCLC [7, 8]. All patients included in the sub-study S1800A had previously received both PD-(L)1 inhibitor therapy and platinum-based doublet chemotherapy either sequentially or combined and had experienced disease progression at least 84 days after initiation of this treatment. The majority of those enrolled in the control arm received docetaxel plus ramucirumab (n = 45), followed by gemcitabine (n = 12). Reckamp et al. presented OS data and other outcomes from S1800A at ASCO 2022 [9].

Survival improvement with PFS/ORR discordance

Indeed, pembrolizumab plus ramucirumab prolonged OS compared to the standard of care (14.5 vs. 11.6 months; HR, 0.69; standard log-rank p = 0.05; Figure). All subgroups favored the combination. Similar reductions in mortality risk were observed independent of PD-L1 expression, and co-mutations did not affect the OS improvement. Patients with squamous/mixed histology showed a greater survival benefit (HR, 0.43) than those with non-squamous histology (HR, 0.95).

For PFS, the analysis did not reveal any superiority of the pembrolizumab-based regimen (4.5 vs. 5.2 months; HR, 0.86), which also applied to ORR (22 % vs. 28 %), although median duration of response was longer in the experimental arm (12.9 vs. 5.6 months). In both arms, > 70 % of patients experienced disease control. The authors noted that this discordance of PFS and ORR from OS has been reported in prior trials evaluating ­immune checkpoint inhibitors and is described as post-progression prolongation of survival.

With respect to grade 3-5 TRAEs, patients treated with pembrolizumab/ramucirumab fared better than those receiving standard therapy (42 % vs. 60 %). Nine grade 3-5 immune-related events were seen in the experimental arm (31 %). Overall, this is the first trial in the checkpoint-inhibitor–refractory setting without a chemotherapy backbone to demonstrate a survival benefit compared to standard-of-care regimens including docetaxel plus ramucirumab. Further evaluation of this approach is warranted.

Figure 1: A: Overall survival advantage for ramucirumab plus pembrolizumab vs. standard-of-care treatment in patients after immunotherapy and platinum-based chemotherapy

Figure 1: A: Overall survival advantage for ramucirumab plus pembrolizumab vs. standard-of-care treatment in patients after immunotherapy and platinum-based chemotherapy


  1. Akinboro O et al., Outcomes of anti-PD-(L)1 therapy with or without chemotherapy for first-line treatment of advanced non-small cell lung cancer with PD-L1 score ≥ 50 %: FDA pooled analysis. J Clin Oncol 40, 2022 (suppl 16; abstr 9000)
  2. Nakajima E et al., Outcomes of 1L therapy in patients with advanced NSCLC according to KRAS mutation status & PD-L1 expression: FDA pooled analysis. J Clin Oncol 40, 2022 (suppl 16; abstr 9001)
  3. Felip E et al., A phase II study (TACTI-002) in 1st line metastatic non-small cell lung cancer ­investigating eftilagimod alpha (soluble LAG-3 protein) and pembrolizumab: updated results from a PD-L1 unselected population. J Clin ­Oncol 40, 2022 (suppl 16; abstr 9003)
  4. Neal JW et al., Cabozantinib in combination with atezolizumab in non-small cell lung cancer patients previously treated with an immune checkpoint inhibitor: results from cohort 7 of the COSMIC-021 study. J Clin Oncol 2020; 38(15_suppl): 9610
  5. Neal JW et al., Cabozantinib plus atezolizumab or cabozantinib alone in patients with ­advanced non-small cell lung cancer previously treated with an immune checkpoint inhibitor: COSMIC-021 study cohorts 7 and 20. J Clin ­Oncol 40, 2022 (suppl 16; abstr 9005)
  6. Khan KA & Kerbel RS, Improving immunotherapy outcomes with anti-angiogenic treatments and vice versa. Nat Rev Clin Oncol 2018; 15(5): 310-324
  7. Herbst RS et al., Ramucirumab plus pembrolizumab in patients with previously treated advanced non-small-cell lung cancer, gastro-­oesophageal cancer, or urothelial carcinomas (JVDF): a multicohort, non-randomised, open-­label, phase 1a/b trial. Lancet Oncol 2019; 20(8): 1109-1123
  8. Herbst RS et al., Phase 1 expansion cohort of ramucirumab plus pembrolizumab in advanced treatment-naive NSCLC. J Thorac Oncol 2021; 16(2): 289-298
  9. Reckamp KL et al., Overall survival from a phase II randomized study of ramucirumab plus pembrolizumab versus standard of care for advanced non-small cell lung cancer previously treated with immunotherapy – Lung-MAP non-matched sub-study S1800A. J Clin Oncol 40, 2022 (suppl 16; abstr 9004)

© 2022 Springer-Verlag GmbH, Impressum