Assessing Commercial Tissue-Based Assays for Autoimmune Neurologic Disorders (II): Antibodies to Surface Antigens

Abstract

Background and objectives: Detecting neural surface antibodies (NSAbs) is essential for diagnosing autoimmune encephalitis. The recommended diagnostic strategy involves initial screening with tissue-based assays (TBAs), followed by confirmation with cell-based assays (CBAs). While specialized centers use in-house TBAs, many clinical laboratories depend on commercial TBAs, whose accuracy is yet to be fully assessed.

Methods: We selected 92 CSF and 99 serum samples from patients with autoimmune encephalitis and NSAbs confirmed by in-house TBAs and CBAs (20 samples each for AMPAR, GABA_AR, GABA_BR, IgLON5, LGI1, NMDAR, and CASPR2; 19 for mGluR5; 17 for DPPX; and 15 for mGluR1 antibodies), along with 50 CSF and 50 serum samples from negative controls. We assessed the performance of a commercial indirect immunofluorescence (IIF)-TBA (EUROIMMUN). Slides were evaluated as "positive" or "negative" by 2 experienced investigators and 2 less experienced raters. Discordant results were re-evaluated through interrater discussion and assessed using Cohen's kappa.

Results: The experienced raters agreed on 94% (133/142) of CSF and 88% (131/149) of serum classifications (Cohen's kappa = 0.87 and 0.75, respectively, p < 0.001). Among CSF samples, 75% (106/142) were correctly identified while 19% (27/142) were misclassified (13 false positives, 14 false negatives). Among serum samples, 66% (98/149) were correctly identified while 22% (33/149) were misclassified (11 false positives, 22 false negatives). The poorest performance was seen in detecting NMDAR, GABA_AR, and mGluR5 Abs, which were not identified in 5 of 10, 6 of 10, and 5 of 9 serum samples and in 4 of 10, 5 of 10, and 5 of 10 CSF samples, respectively. The overall sensitivity of the commercial IIF-TBA was 84% for CSF and 76% for serum while the specificity was 72% for CSF and 73% for serum. Less experienced raters correctly identified 69% (98/142) of CSF samples and 73% (109/149) of serum samples and misclassified 13% (18/142) of CSF samples and 11% (16/149) of serum samples, and 18% (26/142) of CSF samples and 16% (24/149) of serum samples remained discordant.

Discussion: The diagnostic performance of EUROIMMUN IIF-TBA in detecting NSAbs in autoimmune encephalitis is suboptimal. NMDAR antibodies, among the most common NSAbs, can be missed in 50% of cases. This commercial TBA should not be used alone as a screening method nor as a confirmatory technique for NSAbs.

Figure 1. Algorithm of the Study

CBA = cell-based assay; IIF = indirect immunofluorescence; NSAbs = neural surface antibodies; TBA = tissue-based assay.

Figure 2. Pie Charts Showing the Percentages of True/False-Positive/Negative Results in CSF (A) and Serum (B) Samples Using EUROIMMUN IIF-TBA

IIF = indirect immunofluorescence; NSAb+ = neural surface antibody positive; NC = negative control; TBA = tissue-based assay.

Figure 3. Examples of NMDAR and LGI1 Antibody Staining on In-House IHC and Commercial IIF-TBA

Staining patterns of samples positive for NMDAR (left column) and LGI1 (right column) antibodies on rat hippocampus with in-house IHC and IIF-TBA. Characteristic staining patterns of NMDAR (A) and LGI1 (B) antibodies with in-house IHC. In some cases, the same recognizable typical staining pattern was observed on IIF-TBA (C, D), whereas in others, the sample was labeled as “positive” although a typical staining pattern was not observed (E, F). Examples of NMDAR (G) and LGI1 (H) antibody–positive samples with negative staining with IIF-TBA. Scale bar = 500 μm. Abs = antibodies; IHC = immunohistochemistry; IIF = indirect immunofluorescence; LGI1 = leucine-rich glioma inactivated 1; NMDAR = N-methyl-d-aspartate receptor; TBA = tissue-based assay.

Figure 4. Heatmaps Showing the Detailed Results for Each Sample Tested With the Commercial IIF-TBA

CSF (A, upper plot) and serum (A, lower plot) positive samples are grouped into different columns based on the specific antibody. Each sample is represented with a different color according to the results after the interrater agreement between the 2 expert raters (in green: true positive; in purple: false negative; in gray: discordant samples). CSF (B, top row) and serum (B, bottom row) negative controls (unpaired samples) are represented with different colors according to the results after the interrater agreement between the 2 expert raters (in green: true negative; in purple: false positive; in gray: discordant samples). NC = negative control; NSAb+ = neural surface antibody positive.