We aim to determine the minimally clinically important difference (MCID) of the Clinical Assessment Scale in Autoimmune Encephalitis (CASE) to benchmark clinical response in autoimmune encephalitis (AE).

The CASE score is used to track disease severity in AE, with an increasing number of clinical applications. However, no accepted threshold for significant change currently exists, which hampers its use in clinical trials for AE as the primary endpoint.

Using our AE cohort, receiver operating characteristic (ROC) curves were constructed to determine optimal CASE MCID thresholds, with ≥1-point improvement in mRS over 3-month time intervals as the anchor. Sensitivity and specificity of multiple candidate cutoffs were subsequently compared to derive a standardized threshold.

We analyzed 222 adult AE patients (77 N-methyl-D-aspartate receptor-antibody (NMDAR), 49 leucine-rich glioma-inactivated 1 (LGI1) and 113 seronegative AE) with longitudinal CASE scores. A 30% threshold demonstrated good sensitivity and specificity in the first 6 months, especially for the pooled AE cohort, NMDAR and seronegative subgroups. From baseline to 3 months, the sensitivities and specificities of a 30% change in CASE score were: pooled AE (AUC 0.95, sensitivity 0.90, specificity 0.88), NMDAR (AUC 1.00, sensitivity 1.00, specificity 0.95), LGI1 (AUC 0.84, sensitivity 0.67, specificity 0.82) and seronegative (AUC 0.96, sensitivity 0.92, specificity 0.88). Performance at 3 to 6 months was similar: pooled AE (AUC 0.93, sensitivity 0.88, specificity 0.85), NMDAR (AUC 0.92, sensitivity 1.00, specificity 0.74), LGI1 (AUC 0.94, sensitivity 1.00, specificity 0.83), and seronegative (AUC 0.90, sensitivity 0.62, specificity 0.93).

Adoption of a 30% change in CASE score as the MCID threshold may facilitate standardized definitions of clinical response in autoimmune encephalitis therapeutic trials.