Increased Cortical Glycolysis Following CD19 CART Therapy: A Radiographic Surrogate for an Altered Blood-Brain Barrier

Background: Cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS) are known complications of chimeric antigen receptor T-cell (CAR-T) therapy. These clinical syndromes develop as a result of CAR-T activation, proliferation, and tumor lysis with resultant cytokine secretion. In prior reports of CD19 CAR-T therapy patients, those who developed ICANS showed evidence of endothelial activation and disruption of the blood-brain barrier as a result of cytokine release while only approximately one-third demonstrated changes on Brain MRI (Gust et al. Cancer Discov 2017). As such, further predictive markers and studies are needed to identify patients at risk for ICANS to allow for expedited management and improved outcomes. Herein we report a single-center analysis exploring glycolytic activity on PET/CT and the association with clinical outcomes for patients with relapsed/refractory diffuse large B-cell lymphoma (R/R DLBCL) after CAR-T therapy.

Methods: An organ-based evaluation of uninvolved sites was conducted in R/R DLBCL patients (n=32) who underwent CD19 CAR-T therapy with evaluable PET/CT imaging at baseline immediately prior to CAR-T therapy and at 30 days post-infusion (D+30). All patients in this analysis were treated with axicabtagene ciloleucel as standard of care therapy after 2 or more lines of therapy. Tumor metabolic volume (TMV) and mean standard uptake value (SUV_mean) of various organs were quantified using ROVER [Region of interest (ROI) visualization, evolution, and image registration] software (ABX advanced biochemical compounds GmbH, Radeberg, Germany). Statistical analysis was completed using STATA 14 (StataCorp. 2015. Stata Statistical Software: Release 14. College Station, TX: StataCorp LP). All tests were performed after testing the normality distribution assumption. Temporal changes were assessed using paired t-tests, and between-group analyses were completed with two-sample t-tests.

Results: SUV_mean increased significantly after CAR-T therapy in the following organs (D+30 v baseline pre-CAR-T PET/CT): cerebral cortex (8.23 v 7.09, p=0.036), cerebellum (6.26 v 5.56, p=0.024), basal ganglia (9.22 v 7.61, p=0.005), parotid gland (1.61 v 1.42, p=0.004), liver (2.47 v 2.17, p=0.002), spleen (2.08 v 1.84, p=0.043), and pancreas (1.76 v 1.48, p<0.001). No differences in SUV_mean were seen in the lung, testes, retroperitoneal or subcutaneous fat, or paraspinal and psoas muscles. A significant increase in cortical activity was seen in patients with CRS grades ≥2 when compared to those with CRS grades 0-1 (Δ2.65 v Δ0.33, p=0.03). No changes in glycolytic activity were observed between patients stratified by CRS in the cerebellum (Δ1.31 v Δ0.36, p=0.12) or liver (Δ0.21 v Δ0.34, p=0.55). In contrast, changes in glycolytic activity were not significantly associated with development of ICANS or with treatment responses.

Conclusion: For patients with R/R DLBCL undergoing CD19 CAR-T therapy, significantly increased CNS glycolytic activity is seen on PET/CT at D+30 post-infusion when compared to baseline. Interestingly, these changes do not correlate with development of ICANS or lymphoma response; however, changes in cortical activity were associated with CRS grade ≥2. Overall, our findings illustrate a functional and radiographic link between cytokine release and subsequent disruption of the blood-brain barrier as quantified by increased cortical glycolysis 30 days post-CAR-T therapy. While findings are limited by small sample size, further validation in a larger data set is warranted.