• 2019-07
  • 2019-08
  • 2019-09
  • 2019-10
  • 2019-11
  • 2020-03
  • 2020-07
  • 2020-08
  • The proportion of patients developing


    The proportion of patients developing a severe irAE in this series was low (8.4%) and there were no treatment related deaths. Furthermore, only 8 patients required overnight hospitalization (of whom 1 required admission to an intensive care unit) to treat irAE. The pattern of irAE was, on the whole, consistent with that observed in the literature. As a notable exception, 2 patients in the present study developed grade 3 pericarditis; pembrolizumab induced cardiotoxicity (typically myocarditis) is a recognized but uncommon complication [32]. The proportion of patients requiring treatment discontinuation due to irAE is inconsistently reported in clinical trials; in KEYNOTE- 024, 7.1% of pembrolizumab treated patients had therapy discontinued due to an adverse event [2]. In a retrospective review of 482 patients treated at Memorial Sloane Kettering Cancer Center both on and outside of a clinical trial with PD-L1/PD-1 BQ-788 sodium salt (10% received combination therapy with a CTLA-4 inhibitor), 6.9% needed treatment stopped permanently due to an irAE [33]. In the present series, 12.1% of patients had treatment discontinued due to irAE as documented specifically in clinic notes. While BC Cancer healthcare practitioners are supposed to follow irAE treatment algorithms, it is possible that pembrolizumab was discontinued prematurely with low grade irAE due to concern about potential future worsening. In multivariate logistic regression analysis, ECOG PS of 2 or 3 significantly increased the odds (p = 0.05) of severe irAE within 3 months of initiation of pembrolizumab while a weaker association existed if CCI score <3 (vs. ≥ 3, p = 0.06). It is important to note that in our series patients with a poor performance status demonstrated a median OS 10.9 months lower and odds of severe irAE 6.3 fold higher than those with a good performance status at initiation of pembrolizumab. This highlights the need for prospective clinic trials to determine the safety and efficacy of immunotherapy in this patient. The trend to association of severe irAE in patients with a CCI score < 3 (versus ≥ 3) is surprising. CCI has been identified as a poor prognostic factor in retrospective studies for patients with aNSCLC that are EGFR/ALK wildtype [34]. It is possible that individuals with multiple medical conditions have more frequent contact with a variety of healthcare practitioners and, as such, irAE would be detected earlier. Multiple theories have been postulated to explain the physiologic mechanism of irAE. PD-1 inhibitors might modulate the humoral immune system to unmask latent auto-immunity [10]. For example, anti-thyroid antibodies (which can cause hyperthyroid conditions such Grave’s disease) are present in 11% of the population, even though only 4% demonstrate abnormal thyroid function tests. In a retrospective review of aNSCLC patients enrolled onto KEYNOTE-001 at a single institution, Osario et al. observed that hypothyroidism developed in 80% of patients who had thyroid antibodies present at baseline, compared to 8% who did not [17]. The authors believe this finding consistent with auto-immune destruction of the thyroid gland. Cross reactivity of tumor and self neoantigens (so-called molecular mimicry) has been postulated to explain skin toxicity due to nivolumab [35]. Hasan Ali et al. performed immunohistochemical staining (CD3, CD4+, and CD8+) from punch biopsy specimens of patients with aNSCLC who developed nivolumab induced dermatitis; 2 punch biopsies were obtained from patients with squamous tumors and 2 from patients with adenocarcinoma histology. The distribution of CD3 and CD8 + T cells was dependent on cancer subtype, with lymphocytes being more prominent above the basal cell membrane in patients with squamous histology, while patients with adenocarcinoma histology displayed more lymphocytes in the dermis. The authors hypothesize that molecular mimicry (i.e., similarities between antigens expressed on keratinocytes near the basal cell membrane and those present on squamous tumors) could account for skin toxicity in nivolumab treated aNSCLC patients.