Incidence of hematological toxicity with use of immunotherapy and chemotherapy in advanced non-small cell lung cancer

To the Editor: Immunotherapy has overtaken chemotherapy (Chemo) as the frontline mode of treatment for advanced non-small cell lung cancer (NSCLC). The efficacy of immunotherapies such as immune checkpoint inhibitors (ICIs) can be further increased by combining them with Chemo and/or anti-angiogenic therapy. However, these combination treatments are associated with more adverse events, which need to be carefully managed. Although previous meta-analyses have shown that ICI use increased gastrointestinal and skin toxicities, the association between ICIs and hematological toxicity has not been adequately assessed. A study by Kramer et al[1] found that 0.6% of the 7626 ICI-treated patients experienced hematological adverse events; however, only 16% of this group of patients received ICIs + Chemo. Thus, the risks associated with combining ICIs and Chemo may not be fully appreciated. The present study reviewed data from phase III trials of ICIs used in combination with Chemo and/or anti-angiogenic agents to evaluate the hematological toxicity in advanced NSCLC. This study followed the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) extension for network meta-analysis and was registered in PROSPERO (CRD42022369808). PubMed, MEDLINE, Embase, and the Cochrane Library were retrieved for studies on “NSCLC,” “Immunotherapy,” “Antiangiogenic,” “Chemotherapy,” and “Randomized controlled trials (RCT),” published up to April, 2024. Studies were included if they were phase III trials of immunotherapies or anti-angiogenic therapies combined with Chemo involving patients with advanced (stages III–IV) NSCLC, confirmed by histology or cytology, and reported all-grade (including grade ≥3) hematological toxicities. Studies that compared unapproved treatments were excluded. Three investigators independently selected studies, extracted data, and assessed bias using the revised Cochrane risk-of-bias tool (RoB2). STATA MP (https://www.stata.com/statamp/) was used to create network plots and assess evidence connectivity. Random-effects models were employed for network meta-analysis using the gemtc R package (https://posit.co/download/rstudio-desktop/) and JAGS software (https://sourceforge.net/projects/mcmc-jags/files/). Random-effects models were also used to perform pairwise and single-arm meta-analyses of the incidence of hematologic toxicity. Of the 4657 articles identified, 38 eligible RCTs, which trialed 13 interventions and involved 24,951 patients, were included in our study [Supplementary Figure 1, https://links.lww.com/CM9/C148]. Detailed study information is presented in Supplementary Table 1, https://links.lww.com/CM9/C149, and risk of bias assessment is outlined in Supplementary Figure 2, https://links.lww.com/CM9/C148. Comparative network plots for the incidence of grade ≥3 hematologic toxicity are displayed in Supplementary Figure 3, https://links.lww.com/CM9/C148. Compared with standard Chemo, both cytotoxic T lymphocyte-associated antigen-4 inhibitor (CTLA-4) + Chemo (odds ratio [OR] = 1.68, 95% confidence interval [CI]: 0.42, 6.68) and programmed death-1 (PD-1) inhibitor + vascular endothelial growth factor inhibitor (VEGF) + Chemo increased the risk of grade ≥3 anemia (OR = 1.18, 95% CI: 0.49, 2.89) [Supplementary 4A, https://links.lww.com/CM9/C148]. The programmed death-ligand 1 inhibitor (PD-L1) + VEGF + Chemo (OR = 1.71, 95% CI: 0.83, 3.49) and PD-1 + CTLA-4 + Chemo (OR = 1.33, 95% CI: 0.50, 3.55) greatly increased the risk of grade ≥3 neutropenia. PD-1 + VEGF + Chemo (OR = 2.18, 95% CI: 1.11, 4.30), PD-L1 + VEGF + Chemo (OR = 2.33, 95% CI: 0.79, 6.73), and PD-1 + Chemo (OR = 1.34, 95% CI: 1.04, 1.81) regimens increased the risk of grade ≥3 thrombocytopenia by varying degrees [Supplementary 4B, https://links.lww.com/CM9/C148]. PD-1 + CTLA-4 + Chemo (OR = 2.72, 95% CI: 0.53, 24.08), and PD-1 + VEGF + Chemo (OR = 1.60, 95% CI: 0.92, 2.76) increased the risk of grade ≥3 leukopenia. The results of all-grade hematologic toxicities were similar with those of the grade ≥3 hematologic toxicity analysis [Supplementary Figures 3 and 5, https://links.lww.com/CM9/C148]. Next, the patients were stratified into first-line and post-line therapy groups [Supplementary Figure 6, https://links.lww.com/CM9/C148]. In the first-line therapy group [Supplementary Figure 7, https://links.lww.com/CM9/C148], CTLA-4/PD-1 + Chemo increased the risk of grade ≥3 anemia, while PD-L1/PD-1 + VEGF + Chemo increased the risk of grade ≥3 neutropenia, thrombocytopenia, and leukopenia. In the post-line therapy group [Supplementary Figure 8, https://links.lww.com/CM9/C148], PD-1 + VEGF + Chemo increased risks for all grade ≥3 hematologic toxicities. Patients were also stratified based on whether their chemotherapy regimens contain paclitaxel [Supplementary Figure 9, https://links.lww.com/CM9/C148]. In the non-paclitaxel subgroup, PD-1 + VEGF + Chemo and PD-1 + Chemo increased grade ≥3 anemia and leukopenia, while PD-1 + CTLA-4 + Chemo increased grade ≥3 neutropenia [Supplementary Figure 10, https://links.lww.com/CM9/C148]. In the paclitaxel subgroup, CTLA-4/PD-1 + Chemo increased grade ≥3 anemia, PD-L1 + VEGF + Chemo increased grade ≥3 neutropenia, and PD-1 + VEGF + Chemo increased grade ≥3 thrombocytopenia and leukopenia [Supplementary Figure 11, https://links.lww.com/CM9/C148]. The Bayesian ranking results of the various treatment regimens [Supplementary Table 2, https://links.lww.com/CM9/C149] were consistent with the direct and indirect summary results. A forest plot was used to evaluate the extent of heterogeneity in the pairwise meta-analysis [Supplementary Figure 12, https://links.lww.com/CM9/C148]. We found that the deviance information criterion (DIC) of the consistency model was superior to that of the inconsistency model [Supplementary Table 3, https://links.lww.com/CM9/C149]. Local consistency analysis using the node-splitting method revealed a significant degree of inconsistency between CTLA-4 and PD-L1 for a grade ≥3 decrease in white blood cell counts [Supplementary Table 4, https://links.lww.com/CM9/C149]; no inconsistencies were detected for the other therapeutic agents. The meta-regression analysis showed that publication year, patient age, sample size, or Chemo regimen did not significantly impact the results [Supplementary Table 5, https://links.lww.com/CM9/C149]. Single-arm meta-analysis [Figure 1A] and forest plots [Figure 1B, C] showed that combining ICIs (especially PD-1 inhibitors) with Chemo significantly increased the risk of developing hematological toxicities and severe adverse events compared with Chemo alone [Supplementary Figures 13 and 14, https://links.lww.com/CM9/C148].Figure 1: Incidence of hematological toxicity of the single-arm meta-analysis. (A) Summary of the incidence of hematological toxicity. (B) Forest plots of the single-arm meta-analysis for incidence of hematological toxicity of grade 3 or higher. (C) Forest plots of the single-arm meta-analysis for incidence of hematological toxicity of all grades. CIs: Confidence intervals; CTLA-4: Cytotoxic T lymphocyte-associated antigen-4 inhibitor; Chemo: Chemotherapy; NA: Not available; ORs: Odds ratios; PD-1: Programmed death-1 inhibitor; PD-L1: Programmed death-ligand 1 inhibitor; VEGF: Vascular endothelial growth factor inhibitor.This study included data from 38 RCTs (24,951 patients) and assessed the impact of combining PD-1, PD-L1, VEGF, and CTLA-4 inhibitors with Chemo on the incidence of hematologic toxicities. Our analysis revealed that combining ICIs with Chemo significantly increased the risk of hematologic adverse events compared with that observed with Chemo alone. These findings were consistent across all grades of hematologic toxicities. Subgroup analyses showed these risks were widespread across different Chemo regimens and treatment settings. The variability in the hematologic toxicity risks (indicated by wide CIs) is related to the heterogeneity of patient responses to therapy, which may be influenced by factors such as sex, age, health status, nutritional status, and genetic susceptibility. The mechanisms underlying the increased hematologic toxicities associated with combining ICIs with Chemo and anti-angiogenic agents are not fully understood; however, interactions between the immune system and the bone marrow microenvironment are likely implicated. By blocking checkpoint proteins (e.g., PD-1, PD-L1, CTLA-4), ICIs disrupt immune tolerance and enhance immune responses against tumors and other host tissues, including the bone marrow. This can induce inflammation and damage hematopoietic stem cells.[2] Chemo targets rapidly dividing cells, affecting both cancer and bone marrow cells, and ICIs may increase the susceptibility of hematopoietic cells to chemo-induced damage. Anti-angiogenic therapies inhibit blood vessel growth, disrupting the vascular support for hematopoiesis and impairing blood cell production.[3] These combined therapies synergize to increase hematologic toxicity, as immune activation by ICIs enhances the cytotoxic effects of Chemo, and combining Chemo with anti-angiogenic agents reduces hematopoietic progenitor cell[4] proliferation and impairs recovery after myelosuppression, exacerbating hematologic defects. Serious hematologic toxicities complicate the clinical management of patients. As immunotherapy has become central to cancer treatment, effective management strategies are urgently needed. Personalized medicine, routine hematologic monitoring, and preemptive strategies, such as using cyclin-dependent kinase (CDK) 4/6 inhibitors or adjusting dosages, are crucial. Future research should focus on targeted trials to understand the long-term effects of specific cancer treatment combinations, develop toxicity mitigation interventions, and identify predicting biomarkers of severe toxicities. Although our study included data from high-quality RCTs, it had some limitations. For instance, we did not have access to detailed clinical information. Moreover, we did not further evaluate adverse events associated with specific drug and dosage combinations. In summary, our analysis revealed that combining ICIs with Chemo significantly increased the risk of hematologic toxicities, regardless of whether anti-angiogenic therapy was used. The heightened hematologic toxicity profiles of these combination regimens highlight the need for personalized treatments, which could be tailored to individual risk profiles. Future research should explore the molecular mechanisms and predictive markers of synergistic hematologic toxicity, as well as develop protective strategies. Funding This study was funded by the China National Science Foundation (Nos. 82022048 and 82373121), the Science and Technology Planning Project of Guangzhou (No. 202206080013), and the National Key Research & Development Programme (No. 2022YFC2505100). Conflicts of interest None.