Low Utility of the H-Score and HLH-2004 Criteria to Identify Patients with Secondary Hemophagocytic Lymphohistiocytosis after CAR-T Cell Therapy for Relapsed/Refractory Diffuse Large B-Cell Lymphoma

Introduction: Secondary hemophagocytic lymphohistiocytosis (HLH) is an aggressive life-threatening activation of the immune system triggered by an underlying condition. Use of chimeric antigen receptor therapy (CAR-T) to treat relapsed/refractory diffuse large B-cell lymphoma (R/R DLBCL) is associated with cytokine release syndrome (CRS), Immune Effector Cell Associated Neurotoxicity Syndrome (ICANS) and development of secondary HLH. Application of HLH scoring systems such as the H-score or HLH-2004 criteria to identify CAR-T triggered HLH is not validated in this setting. Inability to promptly recognize the development of secondary HLH in CAR-T patients and to distinguish it from CRS may lead to delay in HLH specific therapy and increased mortality. We analyzed the utility of applying the H-score and the HLH-2004 criteria to identify possible patients with HLH post-CAR-T for R/R DLBCL.

Methods: A single center retrospective analysis of 75 patients with R/R DLBCL following CAR-T was performed. Using a peak ferritin of 500 mcg/L or higher within 30 days of CAR-T administration, 43 out of 75 patients were identified at risk for HLH. The H-score and HLH-2004 criteria were applied retrospectively. Measurement of NK activity was not available for any patients and soluble IL2 was collected intermittently. The mean H-score was calculated, and two sample t-test performed to evaluate for a difference in the mean H-score between low versus high grade CRS, low versus high grade ICANS, as well as presence versus absence of cytopenias at days 30, 90 and 180. Low grade CRS/ICANS was defined as 1, and high grade as 2 or higher. CRS was graded per Lee 2014 criteria and ICANS per CTCAEv.4. The 43 patients were then subdivided by H-score threshold of 169 into H-score Low (< 169) and H-score High (≥ 169). The threshold of 169 was used given a prior validation study showing optimal sensitivity and specificity for identifying acquired HLH in adult patients with an H-score equal to or higher than 169. Both Progression Free Survival (PFS) and Overall Survival (OS) were defined as time from CAR-T infusion until an event of last follow up. Kaplan-Meier curves were produced to describe the distribution of PFS and OS between two subpopulations of low and high H-Score.

Results: Median peak ferritin was 1571.8 mcg/L (range: 375 mcg/L - >50,000 mcg/L, table 1). Median H-score was 135 (range 61 -250). Four patients were treated with HLH directed therapy receiving steroids with either tocilizumab, siltuximab and/or anakinra. Of these 4 patients, only 2 met five HLH-2004 criteria. All 4 patients had H-scores above 169 (209, 211, 228 and 250). Of these 4 patients, one died from complications thought secondary to HLH. Fourteen patients (14/43, 32%) had an H-score >169. Ten patients did not require any HLH directed therapy and had no clinical evidence of HLH. The mean H-score was not different between patients with low vs high grade CRS (148.2 vs 141.8, p=0.7) or low vs high grade ICANS (145.9 vs 142.6, p=0.8). No statistical correlation was seen between H-score and cytopenias at any time point. There was no significant difference in PFS or OS between the H-score low vs high subgroups (p=0.7, p=0.1 respectively, figure 1 and 2). Patients with higher H-score tended to have longer length of hospitalization for CAR-T (Pearson correlation coefficient 0.289).

Conclusions: In patients with R/R DLBCL post CAR-T, the use of either the H-score or application of HLH-2004 criteria had low utility in identifying possible HLH in patients who screened in based on peak ferritin within 30 days of CAR-T administration. While apparent differences between the H-score high and low categories may not reach statistical significance due to the small number of patients, our exploratory analysis does not support the use of H-Score to evaluate for HLH post-CAR-T. The immediate post-CAR-T period is characterized by a high inflammatory state, which likely results in high H-scores. Results from our study suggest a need for further characterization of HLH following CAR-T and the role for a CAR-T specific HLH scoring system to distinguish secondary HLH from CAR-T related inflammation in this specific patient population.