A-001 External Clinical Verification of the Atellica CI High-Sensitivity Troponin I Assay

Abstract Background The purpose of this clinical study was to verify the performance characteristics of the Atellica IM High-Sensitivity Troponin I (TnIH) assay on the Atellica CI and Atellica IM Analyzers. Methods The reproducibility study consisted of three sites (two external and one internal), 5 days, two runs per day 2 hours apart, and three replicates/run per CLSI EP05-A3. Eight reproducibility panels (four Medical Decision Pools, one serum pool, three lithium heparin plasma pools), and three control levels were tested on one reagent lot at three sites. Clinical performance was determined by testing patient samples from a clinical study of emergency department patients presenting with chest pain. Samples were collected at baseline, =0.5-1.5, =1.5-4, =4-9, and =9-24 hr. Samples were split among three sites and assayed in singlicate on both Atellica CI and Atellica IM Analyzers. The Atellica IM TnIH assay clinical concordance study was performed on both Atellica CI and Atellica IM Analyzers using the 99th percentile upper reference limit (URL) determined on the Atellica IM Analyzer. Clinical concordance was calculated between the 99th percentile URL cutoff point and the adjudicated diagnosis of AMI at each time point. The following assessments were conducted for serum and lithium heparin plasma samples: assessment of all patients using a single overall 99th percentile URL cut-off value; assessment of each sex subgroup using sex-specific cut-offs and presented separately per sex; sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). Comparison of clinical performance was performed with available results tested on both systems (n=2455). Results Reproducibility coefficients of variation (CVs) for the Atellica CI Analyzer were 2.0–7.9% at 8.19–23361.95 pg/mL (ng/L) and Atellica IM Analyzer, 2.7–5.2% at 8.11–23531.55 pg/mL (ng/L). The Atellica CI vs IM Analyzer SD Ratio (95% Confidence Interval [CI]) ranged from 0.6 (0.3, 1.0) to 1.5 (0.9, 3.7) for reproducibility samples demonstrating Atellica CI and IM Analyzer statistical equivalence. Across both matrices for the overall 99th percentile cutoff, Atellica CI % (95%CI) for sensitivity was >=85.3% (80.9, 88.8); specificity, >=86.1% (83.7, 88.3); PPV, >=57.8% (52.9, 62.5); NPV, >=97.6% (96.8, 98.2). For the female cutoff, sensitivity was >=86.8% (79.0, 92.0); specificity, >=85.1% (81.0, 88.4); PPV, >=50.6% (42.9, 58.3); NPV, >=98.4% (97.3, 99.1). For the male cutoff, sensitivity was >=83.4% (77.6, 87.9); specificity, >=87.5% (84.8, 89.9); PPV, >=61.6% (55.3, 67.5); NPV, >=97.0% (95.7, 97.8). Clinical concordance testing comparing the Atellica CI and the Atellica IM Analyzers demonstrated that assay sensitivity, specificity, PPV, and NPV (overall, female, male) are nearly indistinguishable between the systems, as determined by the 95%CI of the percent differences. This was demonstrated using Newcombe’s approach in which the 95%CI encompassed a difference of 0. Conclusion The SD ratio analyses showed that the Atellica CI Analyzer was statistically equivalent to the Atellica IM Analyzer across all reproducibility sample levels tested. Both Atellica CI Analyzer and Atellica IM Analyzer demonstrated equivalent clinical performance in terms of sensitivity, specificity, NPV and PPV.