Accuracy of T2 magnetic resonance assays as point-of-care methods in the intensive care unit

Abstract

Background: Novel molecular diagnostic methods are being evaluated in order to expedite pathogen identification in patients with bacteraemia.

Aims: To evaluate the feasibility and diagnostic accuracy of the T2 magnetic resonance (T2MR) assays - T2 Bacteria (T2B) and T2 Resistance (T2R) - as point-of-care tests in the intensive care unit compared with blood-culture-based tests.

Methods: Prospective cross-sectional study of consecutive patients with suspected bacteraemia. Diagnostic accuracy was evaluated using blood culture as the reference method.

Findings: In total, 208 cases were included in the study. The mean time from sampling to report was lower for the T2MR assays compared with blood-culture-based methods (P<0.001). The rate of invalid reports was 6.73% for the T2B assay and 9.9% for the T2R assay. For the T2B assay, overall positive percentage agreement (PPA) was 84.6% [95% confidence interval (CI) 71.9-93.1%], negative percentage agreement (NPA) was 64.3% (95% CI 55.4-72.6%), positive predictive value (PPV) was 48.9% (95% CI 42.5-55.3%) and negative predictive value (NPV) was 91.2% (95% CI 84.4-95.2%). Cohen's kappa coefficient was 0.402. For the T2R assay, overall PPA was 80% (95% CI 51.9-95.7%), NPA was 69.2% (95% CI 54.9-81.3%), PPV was 42.9% (95% CI 31.7-54.8%) and NPV was 92.3% (95% CI 81.1-97.1%). Cohen's kappa coefficient was 0.376.

Conclusion: T2MR assays have high NPV for rapid exclusion of bacteraemia, and could potentially assist with antimicrobial stewardship when applied as point-of-care diagnostic tests in the intensive care unit.

Keywords: Bacteraemia; Blood culture; Sepsis; T2 magnetic resonance.