Immune toxicity and flow cytometry of circulating blood cells in cancer patients receiving immune therapy.

Abstract

e15126 Background: The mechanisms by which immune checkpoint inhibitors (ICI) cause immune related adverse events (irAE) are not clear. No current blood tests predict which patients will suffer irAEs. Methods: 70 cancer patients had blood drawn prior to ICI and then 4-6 weeks later. 63 had lung cancer; median age was 69 yrs; 21 were female and median ICI treatment duration was 142 days. Patient blood samples were analyzed using flow cytometry for over 50 lymphoid and myeloid markers to determine levels of circulating white blood cells, as well as their changes over time. Medical records were reviewed for irAE’s grades in 8 groups: rash, colitis, hypothyroidism, hyperthyroidism, hypophysitis, hepatitis, adrenal insufficiency, and pneumonitis. To determine association between toxicity and cell counts we used logistic regression and random forest for the univariate and multivariate analysis. We included a random permutation of the baseline age (which is associated with higher toxicity rates) as the analytical control in the random forest model to help identify the threshold of association strength of a random noise and to identify only those cell subsets that were statistically significant (reported below). Results: Patients with elevated baseline levels of specific subsets of T-cells were more likely to develop irAE’s, and elderly patients were more likely to develop Grade 2 irAE’s. T-cell populations associated with any toxicity included CD3+CD8+PD1+; CD3+CD8+; CD25+CD127lowFoxP3+ and CD3+CD8+Ki-67+. Also, patients who developed any irAE’s demonstrated changes in the following cell types after 2 doses of immune therapy compared to their baseline readings. These included changes in the percentage of circulating cells that were CD3+CD8+PD1+; CD3+CD8+HLA-DR+; CD3+CD4+HLA-DR+; CD3+CD8+CCR7-; CD56+; CD3-CD19-; CD3+CD8+; CD16+CD56dim. Patients with grade ≥2 toxicities, or organ specific toxicities demonstrated statistically significant differing levels of T-cells subsets, though the number of patients in each group was limited. Patients with > median baseline expression of CD3+ CD8+ PD1+ cells had greater survival compared with those that had < median levels. Conclusions: Flow cytometry analysis of patients receiving immune therapy at baseline and 6 weeks after initiating therapy may predict which patients are at greater likelihood of any/organ specific irAEs, potentially leading to an improved understanding their mechanisms. Responding to leukocyte changes by substitution of drugs or shortening of treatment duration may reduce irAEs.