Grading of Neurotoxicity in Patients with Relapsed or Refractory Diffuse Large B-Cell Lymphoma (r/r DLBCL) Receiving Tisagenlecleucel Treatment in the JULIET Study

Introduction: CAR-T cell therapy has demonstrated prompt and durable clinical responses in patients with r/r DLBCL, but is associated with unique toxicities such as cytokine-release syndrome (CRS) and neurotoxicity (NT). NT is the second most common unique toxicity frequently attributed to CAR-T therapy and is present in boxed warnings for all approved CD19 targeted therapies. Similar to other organ toxicities, NT is graded using the Common Terminology Criteria for Adverse Events (CTCAE). However, the CTCAE grading system does not adequately characterize the severity, timing and spectrum of CAR-T related NT. New grading tools are needed for this syndrome-specific AE. The CARTOX working group introduced a novel system for CAR-T Related Encephalopathy Syndrome (CRES), i.e. the CRES grading (Neelapu, Nat Rev Clin Oncol, 2017). To better understand CAR-T related NT and move towards harmonized toxicity reporting, this study retrospectively assessed concordance and variances between the CTCAE and a modified version of the CRES (mCRES) grading system among JULIET patients.

Methods: Patient level data from case report forms collected for JULIET, a single-arm, open-label, multicenter, global phase 2 trial of tisagenlecleucel in adult patients with r/r DLBCL (NCT02445248) were used. Four medical experts with experience treating DLBCL patients with different CAR-T therapy products independently reviewed the data and definitions of NT proposed by the FDA using CTCAE and mCRES system. Patients were graded using these two systems; however, only NT attributable to CAR-T therapy were considered. For example, headache without temporal association or evidence of cognitive impairment was graded 0. The CARTOX group's CRES grading criteria were modified in this study since the CARTOX-10 questionnaire, a new tool to assess overall cognitive function, was not prospectively utilized. Hence, mCRES grades 1 and 2, distinguished by CARTOX-10 score, could not be distinctly defined and were assigned based upon investigator report of cognitive or attention dysfunction by CTCAE. Results were discussed and reconciled among all medical experts in a live meeting. As per the research group charter, the highest grading by any of the four experts would determine the final grading for an individual event. Graded results were also compared with those in the FDA label of tisagenlecleucel, in which NT was broadly defined as the occurrence of any CTCAE graded neurological or psychiatric AE (e.g., anxiety, dizziness, headache, peripheral neuropathy, and sleep disorder).

Results: Among 111 patients infused with tisagenlecleucel (as of December 2017), 68 who had NT per FDA definition were graded. With the CTCAE grading system, the medical experts identified 50 (45%) patients as having experienced CAR-T related NT, including 34 with grade 1/2, 11 with grade 3, and 5 with grade 4; the mCRES system identified 19 (17%) patients, 5 of whom were grade 1/2, 6 were grade 3, and 8 were grade 4 (Figure 1). Among the subgroup of 64 patients who experienced CRS, the CTCAE and the mCRES systems identified 30 (47%) and 15 (23%) patients with any grade NT, respectively (grade ≥3: CTCAE vs. mCRES: 11 vs. 10). For 47 patients without CRS, the CTCAE and the mCRES systems identified 20 (43%) and 4 (9%) patients with NT, respectively (grade ≥3: 5 vs. 4; Table 1). These grades by medical experts also varied from those reported by FDA: among 106 patients receiving tisagenlecleucel (as of September 2017), 62 (58%) had NT including 19 (18%) with grade ≥3.

Conclusions: This exploratory study is the first to retrospectively apply a modified version of the new CARTOX-CRES grading system for CAR-T related NT. Using data from JULIET patients, medical experts were able to achieve consensus NT grading using both the CTCAE and the mCRES grading systems. Using the mCRES system, 19 (17%) patients had any grade NT (5 with grade 1/2, 6 with grade 3, and 8 with grade 4) versus the CTCAE system, which identified 50 (45%) patients as having NT (34 with grade 1/2, 11 with grade 3, and 5 with grade 4). The differences between the two grading systems and the NT grading the FDA reported highlight how the same patient data can be represented variably on different scales and highlight the divergent focus of each system, where encephalopathy is the principal focus of CARTOX-10. These results raise an urgent need for broader consensus on a specific grading scale for CAR-T related NT.

View largeDownload slide

View largeDownload slide

Close modal