The Use of APE2 Score to Diagnose Autoimmune Encephalitis
Bijoya Basu1, Hesham Abboud2
1Case Western Reserve University, 2University Hospitals Cleveland Medical Center
Objective:

To assess the APE2 score as a diagnostic tool for autoimmune encephalitis (AE).

Background:

The antibody prevalence in epilepsy and encephalopathy score (APE2) was developed to help select patients for neuronal antibody testing. However, the score can also be used to diagnose AE and may be less restrictive than the 2016 Graus criteria especially in seronegative patients. 

Design/Methods:

We conducted a retrospective study on patients referred to a tertiary neuroimmunology clinic for evaluation of AE from January 2017 to May 2023. We applied the APE2 score to patients who were and were not eventually diagnosed with AE and assessed sensitivity, specificity, and positive/negative predictive values.

Results:

Seventy-four patients were evaluated (59.5% female, average age 52.28±19.75, 17 seronegative). Forty-seven patients (63.5%) were diagnosed with autoimmune encephalitis per expert opinion and twenty-seven (36.5%) were found to have alternative diagnoses. The average APE2 score in the AE patients was 4.91±2.11 while it was 2.11±1.82 in the patients without AE. Of the AE patients, according to the APE2 score, 2% had possible AE, 34% had probable AE, and 44.6 % had seropositive AE. To diagnose AE, an APE2 score of ≥4 with good response to immunotherapy had a 78% sensitivity, an 81.4% specificity, an 88.1% positive predictive value, and a 68.8% negative predictive value relative to expert opinion.  An APE2 score ≥ 7 improved specificity to 96.3% and the PPV to 92.3%. However, the sensitivity dropped to 37.5% and the NPV to 56.5%.

Conclusions:

The APE2 score criteria could have diagnostic utility for autoimmune encephalitis patients, particularly those that are seronegative and can differentiate AE from mimickers. We are currently comparing the utility of the APE2 score to the Graus 2016 criteria in seronegative patients.

10.1212/WNL.0000000000205896