Immunosuppression for immune-related adverse events during checkpoint inhibition: an intricate balance

Abstract

Immune checkpoint inhibitors (ICIs) have changed perspectives for patients with cancer, but come with severe immune-related adverse events (irAEs). To prevent fatality or chronicity, these irAEs are often promptly treated with high-dose immunosuppressants. Until recently, evidence on the effects of irAE management on ICI efficacy has been sparse. As a result, algorithms for irAE management are mostly expert-opinion based and barely consider possible detrimental effects of immunosuppressants on ICI efficacy. However, recent growing evidence suggests that vigorous immunosuppressive management of irAEs comes with unfavourable effects on ICI efficacy and survival. With expansion of the indications of ICIs, evidence-based treatment of irAEs without hampering tumour control becomes more and more important. In this review, we discuss novel evidence from pre-clinical and clinical studies on the effects of different irAE management regimens including corticosteroids, TNF inhibition and tocilizumab on cancer control and survival. We provide recommendations for pre-clinical research, cohort studies and clinical trials that can help clinicians in tailored irAE management, minimising patients' burden while maintaining ICI efficacy.

Fig. 1. Main findings and limitations of animal models and human studies investigating impact of immunosuppressants on immune checkpoint inhibitor efficacy and consequential clinical recommendations.

ICI immune checkpoint inhibition, irAE immune-related adverse event, TNF tumour necrosis factor, IL-6(R) interleukin-6(receptor), JAKi Janus kinase inhibitor, mTORi mammalian target of rapamycin inhibitor. Created with Biorender.com.

Fig. 2. Fictitious Kaplan-Meier curves of ICI treated patients with or without immune-related adverse events (irAEs), treated with immunosuppressants (IS) or not.

At baseline, it is unknown whether a patient will be treated with IS for irAEs. Comparing patients treated with IS for irAEs to all other patients (also those without irAEs) is likely incorrect. Under the assumption that patients with irAEs have a longer survival than patients without irAEs (lower left), survival of patients with irAEs treated with IS might appear similar to survival of all other patients (including those without irAEs; upper right), while it may actually be worse than survival of patients with irAEs which were not treated with IS (lower right). This is irrespective of whether the survival difference between patients with irAEs and those without is (partially) due to immortal-time bias.

Fig. 3. Survival in patients with immune-related adverse events (irAEs) with or without corticosteroid treatment.

Overall survival (a) and progression-free survival (b) in patients with irAEs with or without corticosteroid treatment. Corticosteroid dose cut-offs represent prednisone equivalent dose. For immune checkpoint inhibitor (ICI) type and, tumour heterogeneity type, green (+) indicates restriction to one ICI regimen/tumour type, yellow (−) indicates adjustment for ICI regimen/tumour type and red (X) indicates that heterogeneity was not accounted for. In studies by Li et al. and Faje et al., corticosteroids irrespective of their indication (irAE or non-irAE) were included, which may have induced bias. In well-intended efforts to account for immortal-time bias, Robert et al. may have introduced bias rather than corrected for time to irAE onset and Lafayolle et al. have potentially overestimated the negative effect of corticosteroids on survival. Other studies accounted for immortal-time by analysing post-irAE survival (green +), did not account for difference in time-to-irAE onset (yellow −) or did not clearly describe start-point of survival analyses (blue?). *Thompson et al. compared two dosages of corticosteroids against the lowest dose of <7.5 mg/d for at least 2 weeks prednisone equivalent. HR hazard ratio for death, CI confidence interval, n number of participants in analysis, IS immunosuppressant, N/A not applicable.

Fig. 4. Survival in patients with immune-related adverse events (irAEs) with or without tumour necrosis factor (TNF) inhibition.

Overall survival (a) and progression-free survival (b) in patients with irAEs with or without TNF inhibition. For immune checkpoint inhibitor (ICI) type and tumour, heterogeneity type, green (+) indicates restriction to one ICI regimen/tumour type, yellow (−) indicates adjustment for ICI regimen/tumour type and red (X) indicates that heterogeneity was not accounted for. Zou et al. analysed post-irAE survival which is theoretically less susceptible for difference in time-to-irAE (therefore green +), than survival from ICI initiation (yellow −, as in Fig. 2). Favara et al. did not clearly describe start-point of survival analyses (blue?). HR hazard ratio for death, CI confidence interval, n number of participants in analysis, IS immunosuppressant, GC glucocorticoid, VEDO vedolizumab, N/A not applicable.