B-110 Validation of Digital and Automated anti-dsDNA IFA Interpretations Using the dIFine Automated Fluorescent Microscope System

Abstract

Abstract Background Detection of anti-double stranded DNA antibodies (anti-dsDNA) by the indirect fluorescent antibody (IFA) assay is a well-established tool for aiding in the diagnosis of systemic lupus erythematosus (SLE). ZEUS Scientific has manufactured an FDA-cleared anti-dsDNA IFA test system for decades, which relies upon manual interpretation of qualitative results using a standard fluorescent microscope. This same anti-dsDNA IFA test system was recently utilized to train software, in conjunction with a specialized fluorescent microscope (collectively termed dIFine), towards automating qualitative result interpretation. The objective of this study was to validate the reproducibility, sensitivity, and specificity of the ZEUS anti-dsDNA IFA test system, in conjunction with the dIFine automated microscope. Methods Multi-site reproducibility study–2 low-positive, 2 mid-positive, 2 high-positive, and 2 negative serum samples were assembled to create a reproducibility panel. The sample identities of the panel members were scrambled prior to testing. The panel was assayed via the ZEUS anti-dsDNA-IFA test system at a 1:10 screening dilution, in triplicate, twice per day, on five days, at three different laboratories (including ZEUS Scientific). At each laboratory, the processed slides were interpreted manually by two operators, then scanned by dIFine to generate digital images of the wells, as well as an automated qualitative result interpretation. The digital images of each slide well were also interpreted by two operators, resulting in 3 independent interpretations per sample, for each operator and instrument (Manual, Digital, Automated). The number of slide wells processed for interpretation at each laboratory was: 8 samples × 30 replicates each = 240 total. Clinical study–300 serum samples derived from donors previously diagnosed with SLE were assembled. A control cohort consisting of 360 samples from donors previously diagnosed with other rheumatic and non-rheumatic diseases was also assembled. The sample identities were scrambled for both cohorts prior to testing. The 660 samples were assayed at a 1:10 screening dilution using the ZEUS anti-dsDNA IFA test system at ZEUS Scientific. Manual, digital, and automated interpretations were carried out at ZEUS Scientific as described above. Results Reproducibility (Actual/Expected): [Site 1 - Operator A - Manual (240/240), Digital (240/240); Site 1 - Operator B - Manual (240/240), Digital (240/240); Site 1 - Automated - (231/240)]. [Site 2 - Operator A - Manual (240/240), Digital (240/240); [Site 2 - Operator B - Manual (240/240), Digital (240/240); Site 2 - Automated - (236/240)]. [Site 3 - Operator A - Manual (240/240), Digital (238/240); Site 3 - Operator B - Manual (240/240), Digital (238/240); Site 3 - Automated - (231/240)]. Sensitivity for SLE cohort (Positive/Total, %): [Operator A - Manual, (79/300, 26.3%), Digital (79/300, 26.3%)]; [Operator B - Manual, (79/300, 26.3%), Digital (79/300, 26.3%), Automated (81/300, 27.0%)]. Specificity for control cohort (Negative/Total): [Operator A - Manual, (358/360, 99.4%), Digital (358/360, 99.4%)]; [Operator B - Manual, (358/360, 99.4%), Digital (358/360, 99.4%), Automated (358/360, 99.4%)]. Conclusions These validation studies demonstrate that the ZEUS anti-dsDNA-IFA test system, in conjunction with the automated dIFine microscope, yields highly reproducible and accurate results derived from digital and automated interpretations.